[www-releases] r262945 - Commit the 3.8.0 release
Hans Wennborg via llvm-commits
llvm-commits at lists.llvm.org
Tue Mar 8 10:28:25 PST 2016
Added: www-releases/trunk/3.8.0/tools/clang/docs/ClangFormatStyleOptions.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/ClangFormatStyleOptions.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/ClangFormatStyleOptions.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/ClangFormatStyleOptions.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,724 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Clang-Format Style Options — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="âClangâ CFE Internals Manual" href="InternalsManual.html" />
+ <link rel="prev" title="ClangFormat" href="ClangFormat.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Clang-Format Style Options</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="ClangFormat.html">ClangFormat</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="InternalsManual.html">“Clang” CFE Internals Manual</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="clang-format-style-options">
+<h1>Clang-Format Style Options<a class="headerlink" href="#clang-format-style-options" title="Permalink to this headline">¶</a></h1>
+<p><a class="reference internal" href=""><em>Clang-Format Style Options</em></a> describes configurable formatting style options
+supported by <a class="reference internal" href="LibFormat.html"><em>LibFormat</em></a> and <a class="reference internal" href="ClangFormat.html"><em>ClangFormat</em></a>.</p>
+<p>When using <strong class="program">clang-format</strong> command line utility or
+<tt class="docutils literal"><span class="pre">clang::format::reformat(...)</span></tt> functions from code, one can either use one of
+the predefined styles (LLVM, Google, Chromium, Mozilla, WebKit) or create a
+custom style by configuring specific style options.</p>
+<div class="section" id="configuring-style-with-clang-format">
+<h2>Configuring Style with clang-format<a class="headerlink" href="#configuring-style-with-clang-format" title="Permalink to this headline">¶</a></h2>
+<p><strong class="program">clang-format</strong> supports two ways to provide custom style options:
+directly specify style configuration in the <tt class="docutils literal"><span class="pre">-style=</span></tt> command line option or
+use <tt class="docutils literal"><span class="pre">-style=file</span></tt> and put style configuration in the <tt class="docutils literal"><span class="pre">.clang-format</span></tt> or
+<tt class="docutils literal"><span class="pre">_clang-format</span></tt> file in the project directory.</p>
+<p>When using <tt class="docutils literal"><span class="pre">-style=file</span></tt>, <strong class="program">clang-format</strong> for each input file will
+try to find the <tt class="docutils literal"><span class="pre">.clang-format</span></tt> file located in the closest parent directory
+of the input file. When the standard input is used, the search is started from
+the current directory.</p>
+<p>The <tt class="docutils literal"><span class="pre">.clang-format</span></tt> file uses YAML format:</p>
+<div class="highlight-yaml"><div class="highlight"><pre><span class="l-Scalar-Plain">key1</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">value1</span>
+<span class="l-Scalar-Plain">key2</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">value2</span>
+<span class="c1"># A comment.</span>
+<span class="nn">...</span>
+</pre></div>
+</div>
+<p>The configuration file can consist of several sections each having different
+<tt class="docutils literal"><span class="pre">Language:</span></tt> parameter denoting the programming language this section of the
+configuration is targeted at. See the description of the <strong>Language</strong> option
+below for the list of supported languages. The first section may have no
+language set, it will set the default style options for all lanugages.
+Configuration sections for specific language will override options set in the
+default section.</p>
+<p>When <strong class="program">clang-format</strong> formats a file, it auto-detects the language using
+the file name. When formatting standard input or a file that doesn’t have the
+extension corresponding to its language, <tt class="docutils literal"><span class="pre">-assume-filename=</span></tt> option can be
+used to override the file name <strong class="program">clang-format</strong> uses to detect the
+language.</p>
+<p>An example of a configuration file for multiple languages:</p>
+<div class="highlight-yaml"><div class="highlight"><pre><span class="nn">---</span>
+<span class="c1"># We'll use defaults from the LLVM style, but with 4 columns indentation.</span>
+<span class="l-Scalar-Plain">BasedOnStyle</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">LLVM</span>
+<span class="l-Scalar-Plain">IndentWidth</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">4</span>
+<span class="nn">---</span>
+<span class="l-Scalar-Plain">Language</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">Cpp</span>
+<span class="c1"># Force pointers to the type for C++.</span>
+<span class="l-Scalar-Plain">DerivePointerAlignment</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">false</span>
+<span class="l-Scalar-Plain">PointerAlignment</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">Left</span>
+<span class="nn">---</span>
+<span class="l-Scalar-Plain">Language</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">JavaScript</span>
+<span class="c1"># Use 100 columns for JS.</span>
+<span class="l-Scalar-Plain">ColumnLimit</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">100</span>
+<span class="nn">---</span>
+<span class="l-Scalar-Plain">Language</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">Proto</span>
+<span class="c1"># Don't format .proto files.</span>
+<span class="l-Scalar-Plain">DisableFormat</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">true</span>
+<span class="nn">...</span>
+</pre></div>
+</div>
+<p>An easy way to get a valid <tt class="docutils literal"><span class="pre">.clang-format</span></tt> file containing all configuration
+options of a certain predefined style is:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="go">clang-format -style=llvm -dump-config > .clang-format</span>
+</pre></div>
+</div>
+<p>When specifying configuration in the <tt class="docutils literal"><span class="pre">-style=</span></tt> option, the same configuration
+is applied for all input files. The format of the configuration is:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="go">-style='{key1: value1, key2: value2, ...}'</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="disabling-formatting-on-a-piece-of-code">
+<h2>Disabling Formatting on a Piece of Code<a class="headerlink" href="#disabling-formatting-on-a-piece-of-code" title="Permalink to this headline">¶</a></h2>
+<p>Clang-format understands also special comments that switch formatting in a
+delimited range. The code between a comment <tt class="docutils literal"><span class="pre">//</span> <span class="pre">clang-format</span> <span class="pre">off</span></tt> or
+<tt class="docutils literal"><span class="pre">/*</span> <span class="pre">clang-format</span> <span class="pre">off</span> <span class="pre">*/</span></tt> up to a comment <tt class="docutils literal"><span class="pre">//</span> <span class="pre">clang-format</span> <span class="pre">on</span></tt> or
+<tt class="docutils literal"><span class="pre">/*</span> <span class="pre">clang-format</span> <span class="pre">on</span> <span class="pre">*/</span></tt> will not be formatted. The comments themselves
+will be formatted (aligned) normally.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">int</span> <span class="n">formatted_code</span><span class="p">;</span>
+<span class="c1">// clang-format off</span>
+ <span class="kt">void</span> <span class="n">unformatted_code</span> <span class="p">;</span>
+<span class="c1">// clang-format on</span>
+<span class="kt">void</span> <span class="n">formatted_code_again</span><span class="p">;</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="configuring-style-in-code">
+<h2>Configuring Style in Code<a class="headerlink" href="#configuring-style-in-code" title="Permalink to this headline">¶</a></h2>
+<p>When using <tt class="docutils literal"><span class="pre">clang::format::reformat(...)</span></tt> functions, the format is specified
+by supplying the <a class="reference external" href="http://clang.llvm.org/doxygen/structclang_1_1format_1_1FormatStyle.html">clang::format::FormatStyle</a>
+structure.</p>
+</div>
+<div class="section" id="configurable-format-style-options">
+<h2>Configurable Format Style Options<a class="headerlink" href="#configurable-format-style-options" title="Permalink to this headline">¶</a></h2>
+<p>This section lists the supported style options. Value type is specified for
+each option. For enumeration types possible values are specified both as a C++
+enumeration member (with a prefix, e.g. <tt class="docutils literal"><span class="pre">LS_Auto</span></tt>), and as a value usable in
+the configuration (without a prefix: <tt class="docutils literal"><span class="pre">Auto</span></tt>).</p>
+<dl class="docutils">
+<dt><strong>BasedOnStyle</strong> (<tt class="docutils literal"><span class="pre">string</span></tt>)</dt>
+<dd><p class="first">The style used for all options not specifically set in the configuration.</p>
+<p>This option is supported only in the <strong class="program">clang-format</strong> configuration
+(both within <tt class="docutils literal"><span class="pre">-style='{...}'</span></tt> and the <tt class="docutils literal"><span class="pre">.clang-format</span></tt> file).</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">LLVM</span></tt>
+A style complying with the <a class="reference external" href="http://llvm.org/docs/CodingStandards.html">LLVM coding standards</a></li>
+<li><tt class="docutils literal"><span class="pre">Google</span></tt>
+A style complying with <a class="reference external" href="http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml">Google’s C++ style guide</a></li>
+<li><tt class="docutils literal"><span class="pre">Chromium</span></tt>
+A style complying with <a class="reference external" href="http://www.chromium.org/developers/coding-style">Chromium’s style guide</a></li>
+<li><tt class="docutils literal"><span class="pre">Mozilla</span></tt>
+A style complying with <a class="reference external" href="https://developer.mozilla.org/en-US/docs/Developer_Guide/Coding_Style">Mozilla’s style guide</a></li>
+<li><tt class="docutils literal"><span class="pre">WebKit</span></tt>
+A style complying with <a class="reference external" href="http://www.webkit.org/coding/coding-style.html">WebKit’s style guide</a></li>
+</ul>
+</dd>
+</dl>
+<dl class="docutils">
+<dt><strong>AccessModifierOffset</strong> (<tt class="docutils literal"><span class="pre">int</span></tt>)</dt>
+<dd>The extra indent or outdent of access modifiers, e.g. <tt class="docutils literal"><span class="pre">public:</span></tt>.</dd>
+<dt><strong>AlignAfterOpenBracket</strong> (<tt class="docutils literal"><span class="pre">BracketAlignmentStyle</span></tt>)</dt>
+<dd><p class="first">If <tt class="docutils literal"><span class="pre">true</span></tt>, horizontally aligns arguments after an open bracket.</p>
+<p>This applies to round brackets (parentheses), angle brackets and square
+brackets. This will result in formattings like</p>
+<p>Possible values:</p>
+<ul class="last">
+<li><p class="first"><tt class="docutils literal"><span class="pre">BAS_Align</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Align</span></tt>)
+Align parameters on the open bracket, e.g.:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">someLongFunction</span><span class="p">(</span><span class="n">argument1</span><span class="p">,</span>
+ <span class="n">argument2</span><span class="p">);</span>
+</pre></div>
+</div>
+</li>
+<li><p class="first"><tt class="docutils literal"><span class="pre">BAS_DontAlign</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">DontAlign</span></tt>)
+Don’t align, instead use <tt class="docutils literal"><span class="pre">ContinuationIndentWidth</span></tt>, e.g.:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">someLongFunction</span><span class="p">(</span><span class="n">argument1</span><span class="p">,</span>
+ <span class="n">argument2</span><span class="p">);</span>
+</pre></div>
+</div>
+</li>
+<li><p class="first"><tt class="docutils literal"><span class="pre">BAS_AlwaysBreak</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">AlwaysBreak</span></tt>)
+Always break after an open bracket, if the parameters don’t fit
+on a single line, e.g.:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">someLongFunction</span><span class="p">(</span>
+ <span class="n">argument1</span><span class="p">,</span> <span class="n">argument2</span><span class="p">);</span>
+</pre></div>
+</div>
+</li>
+</ul>
+</dd>
+<dt><strong>AlignConsecutiveAssignments</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd><p class="first">If <tt class="docutils literal"><span class="pre">true</span></tt>, aligns consecutive assignments.</p>
+<p>This will align the assignment operators of consecutive lines. This
+will result in formattings like</p>
+<div class="last highlight-c++"><div class="highlight"><pre><span class="kt">int</span> <span class="n">aaaa</span> <span class="o">=</span> <span class="mi">12</span><span class="p">;</span>
+<span class="kt">int</span> <span class="n">b</span> <span class="o">=</span> <span class="mi">23</span><span class="p">;</span>
+<span class="kt">int</span> <span class="n">ccc</span> <span class="o">=</span> <span class="mi">23</span><span class="p">;</span>
+</pre></div>
+</div>
+</dd>
+<dt><strong>AlignConsecutiveDeclarations</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd><p class="first">If <tt class="docutils literal"><span class="pre">true</span></tt>, aligns consecutive declarations.</p>
+<p>This will align the declaration names of consecutive lines. This
+will result in formattings like</p>
+<div class="last highlight-c++"><div class="highlight"><pre><span class="kt">int</span> <span class="n">aaaa</span> <span class="o">=</span> <span class="mi">12</span><span class="p">;</span>
+<span class="kt">float</span> <span class="n">b</span> <span class="o">=</span> <span class="mi">23</span><span class="p">;</span>
+<span class="n">std</span><span class="o">::</span><span class="n">string</span> <span class="n">ccc</span> <span class="o">=</span> <span class="mi">23</span><span class="p">;</span>
+</pre></div>
+</div>
+</dd>
+<dt><strong>AlignEscapedNewlinesLeft</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, aligns escaped newlines as far left as possible.
+Otherwise puts them into the right-most column.</dd>
+<dt><strong>AlignOperands</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, horizontally align operands of binary and ternary
+expressions.</dd>
+<dt><strong>AlignTrailingComments</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, aligns trailing comments.</dd>
+<dt><strong>AllowAllParametersOfDeclarationOnNextLine</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>Allow putting all parameters of a function declaration onto
+the next line even if <tt class="docutils literal"><span class="pre">BinPackParameters</span></tt> is <tt class="docutils literal"><span class="pre">false</span></tt>.</dd>
+<dt><strong>AllowShortBlocksOnASingleLine</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd><p class="first">Allows contracting simple braced statements to a single line.</p>
+<p class="last">E.g., this allows <tt class="docutils literal"><span class="pre">if</span> <span class="pre">(a)</span> <span class="pre">{</span> <span class="pre">return;</span> <span class="pre">}</span></tt> to be put on a single line.</p>
+</dd>
+<dt><strong>AllowShortCaseLabelsOnASingleLine</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, short case labels will be contracted to a single line.</dd>
+<dt><strong>AllowShortFunctionsOnASingleLine</strong> (<tt class="docutils literal"><span class="pre">ShortFunctionStyle</span></tt>)</dt>
+<dd><p class="first">Dependent on the value, <tt class="docutils literal"><span class="pre">int</span> <span class="pre">f()</span> <span class="pre">{</span> <span class="pre">return</span> <span class="pre">0;</span> <span class="pre">}</span></tt> can be put
+on a single line.</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">SFS_None</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">None</span></tt>)
+Never merge functions into a single line.</li>
+<li><tt class="docutils literal"><span class="pre">SFS_Empty</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Empty</span></tt>)
+Only merge empty functions.</li>
+<li><tt class="docutils literal"><span class="pre">SFS_Inline</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Inline</span></tt>)
+Only merge functions defined inside a class. Implies “empty”.</li>
+<li><tt class="docutils literal"><span class="pre">SFS_All</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">All</span></tt>)
+Merge all functions fitting on a single line.</li>
+</ul>
+</dd>
+<dt><strong>AllowShortIfStatementsOnASingleLine</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, <tt class="docutils literal"><span class="pre">if</span> <span class="pre">(a)</span> <span class="pre">return;</span></tt> can be put on a single
+line.</dd>
+<dt><strong>AllowShortLoopsOnASingleLine</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, <tt class="docutils literal"><span class="pre">while</span> <span class="pre">(true)</span> <span class="pre">continue;</span></tt> can be put on a
+single line.</dd>
+<dt><strong>AlwaysBreakAfterDefinitionReturnType</strong> (<tt class="docutils literal"><span class="pre">DefinitionReturnTypeBreakingStyle</span></tt>)</dt>
+<dd><p class="first">The function definition return type breaking style to use. This
+option is deprecated and is retained for backwards compatibility.</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">DRTBS_None</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">None</span></tt>)
+Break after return type automatically.
+<tt class="docutils literal"><span class="pre">PenaltyReturnTypeOnItsOwnLine</span></tt> is taken into account.</li>
+<li><tt class="docutils literal"><span class="pre">DRTBS_All</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">All</span></tt>)
+Always break after the return type.</li>
+<li><tt class="docutils literal"><span class="pre">DRTBS_TopLevel</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">TopLevel</span></tt>)
+Always break after the return types of top-level functions.</li>
+</ul>
+</dd>
+<dt><strong>AlwaysBreakAfterReturnType</strong> (<tt class="docutils literal"><span class="pre">ReturnTypeBreakingStyle</span></tt>)</dt>
+<dd><p class="first">The function declaration return type breaking style to use.</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">RTBS_None</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">None</span></tt>)
+Break after return type automatically.
+<tt class="docutils literal"><span class="pre">PenaltyReturnTypeOnItsOwnLine</span></tt> is taken into account.</li>
+<li><tt class="docutils literal"><span class="pre">RTBS_All</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">All</span></tt>)
+Always break after the return type.</li>
+<li><tt class="docutils literal"><span class="pre">RTBS_TopLevel</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">TopLevel</span></tt>)
+Always break after the return types of top-level functions.</li>
+<li><tt class="docutils literal"><span class="pre">RTBS_AllDefinitions</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">AllDefinitions</span></tt>)
+Always break after the return type of function definitions.</li>
+<li><tt class="docutils literal"><span class="pre">RTBS_TopLevelDefinitions</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">TopLevelDefinitions</span></tt>)
+Always break after the return type of top-level definitions.</li>
+</ul>
+</dd>
+<dt><strong>AlwaysBreakBeforeMultilineStrings</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd><p class="first">If <tt class="docutils literal"><span class="pre">true</span></tt>, always break before multiline string literals.</p>
+<p class="last">This flag is mean to make cases where there are multiple multiline strings
+in a file look more consistent. Thus, it will only take effect if wrapping
+the string at that point leads to it being indented
+<tt class="docutils literal"><span class="pre">ContinuationIndentWidth</span></tt> spaces from the start of the line.</p>
+</dd>
+<dt><strong>AlwaysBreakTemplateDeclarations</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, always break after the <tt class="docutils literal"><span class="pre">template<...></span></tt> of a
+template declaration.</dd>
+<dt><strong>BinPackArguments</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">false</span></tt>, a function call’s arguments will either be all on the
+same line or will have one line each.</dd>
+<dt><strong>BinPackParameters</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">false</span></tt>, a function declaration’s or function definition’s
+parameters will either all be on the same line or will have one line each.</dd>
+<dt><strong>BraceWrapping</strong> (<tt class="docutils literal"><span class="pre">BraceWrappingFlags</span></tt>)</dt>
+<dd><p class="first">Control of individual brace wrapping cases.</p>
+<p>If <tt class="docutils literal"><span class="pre">BreakBeforeBraces</span></tt> is set to <tt class="docutils literal"><span class="pre">custom</span></tt>, use this to specify how each
+individual brace case should be handled. Otherwise, this is ignored.</p>
+<p>Nested configuration flags:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">bool</span> <span class="pre">AfterClass</span></tt> Wrap class definitions.</li>
+<li><tt class="docutils literal"><span class="pre">bool</span> <span class="pre">AfterControlStatement</span></tt> Wrap control statements (if/for/while/switch/..).</li>
+<li><tt class="docutils literal"><span class="pre">bool</span> <span class="pre">AfterEnum</span></tt> Wrap enum definitions.</li>
+<li><tt class="docutils literal"><span class="pre">bool</span> <span class="pre">AfterFunction</span></tt> Wrap function definitions.</li>
+<li><tt class="docutils literal"><span class="pre">bool</span> <span class="pre">AfterNamespace</span></tt> Wrap namespace definitions.</li>
+<li><tt class="docutils literal"><span class="pre">bool</span> <span class="pre">AfterObjCDeclaration</span></tt> Wrap ObjC definitions (@autoreleasepool, interfaces, ..).</li>
+<li><tt class="docutils literal"><span class="pre">bool</span> <span class="pre">AfterStruct</span></tt> Wrap struct definitions.</li>
+<li><tt class="docutils literal"><span class="pre">bool</span> <span class="pre">AfterUnion</span></tt> Wrap union definitions.</li>
+<li><tt class="docutils literal"><span class="pre">bool</span> <span class="pre">BeforeCatch</span></tt> Wrap before <tt class="docutils literal"><span class="pre">catch</span></tt>.</li>
+<li><tt class="docutils literal"><span class="pre">bool</span> <span class="pre">BeforeElse</span></tt> Wrap before <tt class="docutils literal"><span class="pre">else</span></tt>.</li>
+<li><tt class="docutils literal"><span class="pre">bool</span> <span class="pre">IndentBraces</span></tt> Indent the wrapped braces themselves.</li>
+</ul>
+</dd>
+<dt><strong>BreakAfterJavaFieldAnnotations</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>Break after each annotation on a field in Java files.</dd>
+<dt><strong>BreakBeforeBinaryOperators</strong> (<tt class="docutils literal"><span class="pre">BinaryOperatorStyle</span></tt>)</dt>
+<dd><p class="first">The way to wrap binary operators.</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">BOS_None</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">None</span></tt>)
+Break after operators.</li>
+<li><tt class="docutils literal"><span class="pre">BOS_NonAssignment</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">NonAssignment</span></tt>)
+Break before operators that aren’t assignments.</li>
+<li><tt class="docutils literal"><span class="pre">BOS_All</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">All</span></tt>)
+Break before operators.</li>
+</ul>
+</dd>
+<dt><strong>BreakBeforeBraces</strong> (<tt class="docutils literal"><span class="pre">BraceBreakingStyle</span></tt>)</dt>
+<dd><p class="first">The brace breaking style to use.</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">BS_Attach</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Attach</span></tt>)
+Always attach braces to surrounding context.</li>
+<li><tt class="docutils literal"><span class="pre">BS_Linux</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Linux</span></tt>)
+Like <tt class="docutils literal"><span class="pre">Attach</span></tt>, but break before braces on function, namespace and
+class definitions.</li>
+<li><tt class="docutils literal"><span class="pre">BS_Mozilla</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Mozilla</span></tt>)
+Like <tt class="docutils literal"><span class="pre">Attach</span></tt>, but break before braces on enum, function, and record
+definitions.</li>
+<li><tt class="docutils literal"><span class="pre">BS_Stroustrup</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Stroustrup</span></tt>)
+Like <tt class="docutils literal"><span class="pre">Attach</span></tt>, but break before function definitions, ‘catch’, and ‘else’.</li>
+<li><tt class="docutils literal"><span class="pre">BS_Allman</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Allman</span></tt>)
+Always break before braces.</li>
+<li><tt class="docutils literal"><span class="pre">BS_GNU</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">GNU</span></tt>)
+Always break before braces and add an extra level of indentation to
+braces of control statements, not to those of class, function
+or other definitions.</li>
+<li><tt class="docutils literal"><span class="pre">BS_WebKit</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">WebKit</span></tt>)
+Like <tt class="docutils literal"><span class="pre">Attach</span></tt>, but break before functions.</li>
+<li><tt class="docutils literal"><span class="pre">BS_Custom</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Custom</span></tt>)
+Configure each individual brace in <tt class="docutils literal"><span class="pre">BraceWrapping</span></tt>.</li>
+</ul>
+</dd>
+<dt><strong>BreakBeforeTernaryOperators</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, ternary operators will be placed after line breaks.</dd>
+<dt><strong>BreakConstructorInitializersBeforeComma</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>Always break constructor initializers before commas and align
+the commas with the colon.</dd>
+<dt><strong>ColumnLimit</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd><p class="first">The column limit.</p>
+<p class="last">A column limit of <tt class="docutils literal"><span class="pre">0</span></tt> means that there is no column limit. In this case,
+clang-format will respect the input’s line breaking decisions within
+statements unless they contradict other rules.</p>
+</dd>
+<dt><strong>CommentPragmas</strong> (<tt class="docutils literal"><span class="pre">std::string</span></tt>)</dt>
+<dd>A regular expression that describes comments with special meaning,
+which should not be split into lines or otherwise changed.</dd>
+<dt><strong>ConstructorInitializerAllOnOneLineOrOnePerLine</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If the constructor initializers don’t fit on a line, put each
+initializer on its own line.</dd>
+<dt><strong>ConstructorInitializerIndentWidth</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>The number of characters to use for indentation of constructor
+initializer lists.</dd>
+<dt><strong>ContinuationIndentWidth</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>Indent width for line continuations.</dd>
+<dt><strong>Cpp11BracedListStyle</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd><p class="first">If <tt class="docutils literal"><span class="pre">true</span></tt>, format braced lists as best suited for C++11 braced
+lists.</p>
+<p>Important differences:
+- No spaces inside the braced list.
+- No line break before the closing brace.
+- Indentation with the continuation indent, not with the block indent.</p>
+<p class="last">Fundamentally, C++11 braced lists are formatted exactly like function
+calls would be formatted in their place. If the braced list follows a name
+(e.g. a type or variable name), clang-format formats as if the <tt class="docutils literal"><span class="pre">{}</span></tt> were
+the parentheses of a function call with that name. If there is no name,
+a zero-length name is assumed.</p>
+</dd>
+<dt><strong>DerivePointerAlignment</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, analyze the formatted file for the most common
+alignment of & and *. <tt class="docutils literal"><span class="pre">PointerAlignment</span></tt> is then used only as fallback.</dd>
+<dt><strong>DisableFormat</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>Disables formatting completely.</dd>
+<dt><strong>ExperimentalAutoDetectBinPacking</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd><p class="first">If <tt class="docutils literal"><span class="pre">true</span></tt>, clang-format detects whether function calls and
+definitions are formatted with one parameter per line.</p>
+<p>Each call can be bin-packed, one-per-line or inconclusive. If it is
+inconclusive, e.g. completely on one line, but a decision needs to be
+made, clang-format analyzes whether there are other bin-packed cases in
+the input file and act accordingly.</p>
+<p class="last">NOTE: This is an experimental flag, that might go away or be renamed. Do
+not use this in config files, etc. Use at your own risk.</p>
+</dd>
+<dt><strong>ForEachMacros</strong> (<tt class="docutils literal"><span class="pre">std::vector<std::string></span></tt>)</dt>
+<dd><p class="first">A vector of macros that should be interpreted as foreach loops
+instead of as function calls.</p>
+<p>These are expected to be macros of the form:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">FOREACH</span><span class="p">(</span><span class="o"><</span><span class="n">variable</span><span class="o">-</span><span class="n">declaration</span><span class="o">></span><span class="p">,</span> <span class="p">...)</span>
+ <span class="o"><</span><span class="n">loop</span><span class="o">-</span><span class="n">body</span><span class="o">></span>
+</pre></div>
+</div>
+<p>In the .clang-format configuration file, this can be configured like:</p>
+<div class="highlight-c++"><div class="highlight"><pre>ForEachMacros: ['RANGES_FOR', 'FOREACH']
+</pre></div>
+</div>
+<p class="last">For example: BOOST_FOREACH.</p>
+</dd>
+<dt><strong>IncludeCategories</strong> (<tt class="docutils literal"><span class="pre">std::vector<IncludeCategory></span></tt>)</dt>
+<dd><p class="first">Regular expressions denoting the different #include categories used
+for ordering #includes.</p>
+<p>These regular expressions are matched against the filename of an include
+(including the <> or “”) in order. The value belonging to the first
+matching regular expression is assigned and #includes are sorted first
+according to increasing category number and then alphabetically within
+each category.</p>
+<p>If none of the regular expressions match, UINT_MAX is assigned as
+category. The main header for a source file automatically gets category 0,
+so that it is kept at the beginning of the #includes
+(<a class="reference external" href="http://llvm.org/docs/CodingStandards.html#include-style">http://llvm.org/docs/CodingStandards.html#include-style</a>).</p>
+<p>To configure this in the .clang-format file, use:</p>
+<div class="last highlight-c++"><div class="highlight"><pre>IncludeCategories:
+ - Regex: '^"(llvm|llvm-c|clang|clang-c)/'
+ Priority: 2
+ - Regex: '^(<|"(gtest|isl|json)/)'
+ Priority: 3
+ - Regex: '.\*'
+ Priority: 1
+</pre></div>
+</div>
+</dd>
+<dt><strong>IndentCaseLabels</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd><p class="first">Indent case labels one level from the switch statement.</p>
+<p class="last">When <tt class="docutils literal"><span class="pre">false</span></tt>, use the same indentation level as for the switch statement.
+Switch statement body is always indented one level more than case labels.</p>
+</dd>
+<dt><strong>IndentWidth</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>The number of columns to use for indentation.</dd>
+<dt><strong>IndentWrappedFunctionNames</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>Indent if a function definition or declaration is wrapped after the
+type.</dd>
+<dt><strong>KeepEmptyLinesAtTheStartOfBlocks</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If true, empty lines at the start of blocks are kept.</dd>
+<dt><strong>Language</strong> (<tt class="docutils literal"><span class="pre">LanguageKind</span></tt>)</dt>
+<dd><p class="first">Language, this format style is targeted at.</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">LK_None</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">None</span></tt>)
+Do not use.</li>
+<li><tt class="docutils literal"><span class="pre">LK_Cpp</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Cpp</span></tt>)
+Should be used for C, C++, ObjectiveC, ObjectiveC++.</li>
+<li><tt class="docutils literal"><span class="pre">LK_Java</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Java</span></tt>)
+Should be used for Java.</li>
+<li><tt class="docutils literal"><span class="pre">LK_JavaScript</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">JavaScript</span></tt>)
+Should be used for JavaScript.</li>
+<li><tt class="docutils literal"><span class="pre">LK_Proto</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Proto</span></tt>)
+Should be used for Protocol Buffers
+(<a class="reference external" href="https://developers.google.com/protocol-buffers/">https://developers.google.com/protocol-buffers/</a>).</li>
+</ul>
+</dd>
+<dt><strong>MacroBlockBegin</strong> (<tt class="docutils literal"><span class="pre">std::string</span></tt>)</dt>
+<dd>A regular expression matching macros that start a block.</dd>
+<dt><strong>MacroBlockEnd</strong> (<tt class="docutils literal"><span class="pre">std::string</span></tt>)</dt>
+<dd>A regular expression matching macros that end a block.</dd>
+<dt><strong>MaxEmptyLinesToKeep</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>The maximum number of consecutive empty lines to keep.</dd>
+<dt><strong>NamespaceIndentation</strong> (<tt class="docutils literal"><span class="pre">NamespaceIndentationKind</span></tt>)</dt>
+<dd><p class="first">The indentation used for namespaces.</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">NI_None</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">None</span></tt>)
+Don’t indent in namespaces.</li>
+<li><tt class="docutils literal"><span class="pre">NI_Inner</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Inner</span></tt>)
+Indent only in inner namespaces (nested in other namespaces).</li>
+<li><tt class="docutils literal"><span class="pre">NI_All</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">All</span></tt>)
+Indent in all namespaces.</li>
+</ul>
+</dd>
+<dt><strong>ObjCBlockIndentWidth</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>The number of characters to use for indentation of ObjC blocks.</dd>
+<dt><strong>ObjCSpaceAfterProperty</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>Add a space after <tt class="docutils literal"><span class="pre">@property</span></tt> in Objective-C, i.e. use
+<tt class="docutils literal"><span class="pre">\@property</span> <span class="pre">(readonly)</span></tt> instead of <tt class="docutils literal"><span class="pre">\@property(readonly)</span></tt>.</dd>
+<dt><strong>ObjCSpaceBeforeProtocolList</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>Add a space in front of an Objective-C protocol list, i.e. use
+<tt class="docutils literal"><span class="pre">Foo</span> <span class="pre"><Protocol></span></tt> instead of <tt class="docutils literal"><span class="pre">Foo<Protocol></span></tt>.</dd>
+<dt><strong>PenaltyBreakBeforeFirstCallParameter</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>The penalty for breaking a function call after “call(”.</dd>
+<dt><strong>PenaltyBreakComment</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>The penalty for each line break introduced inside a comment.</dd>
+<dt><strong>PenaltyBreakFirstLessLess</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>The penalty for breaking before the first <tt class="docutils literal"><span class="pre"><<</span></tt>.</dd>
+<dt><strong>PenaltyBreakString</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>The penalty for each line break introduced inside a string literal.</dd>
+<dt><strong>PenaltyExcessCharacter</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>The penalty for each character outside of the column limit.</dd>
+<dt><strong>PenaltyReturnTypeOnItsOwnLine</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>Penalty for putting the return type of a function onto its own
+line.</dd>
+<dt><strong>PointerAlignment</strong> (<tt class="docutils literal"><span class="pre">PointerAlignmentStyle</span></tt>)</dt>
+<dd><p class="first">Pointer and reference alignment style.</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">PAS_Left</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Left</span></tt>)
+Align pointer to the left.</li>
+<li><tt class="docutils literal"><span class="pre">PAS_Right</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Right</span></tt>)
+Align pointer to the right.</li>
+<li><tt class="docutils literal"><span class="pre">PAS_Middle</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Middle</span></tt>)
+Align pointer in the middle.</li>
+</ul>
+</dd>
+<dt><strong>SpaceAfterCStyleCast</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, a space may be inserted after C style casts.</dd>
+<dt><strong>SpaceBeforeAssignmentOperators</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">false</span></tt>, spaces will be removed before assignment operators.</dd>
+<dt><strong>SpaceBeforeParens</strong> (<tt class="docutils literal"><span class="pre">SpaceBeforeParensOptions</span></tt>)</dt>
+<dd><p class="first">Defines in which cases to put a space before opening parentheses.</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">SBPO_Never</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Never</span></tt>)
+Never put a space before opening parentheses.</li>
+<li><tt class="docutils literal"><span class="pre">SBPO_ControlStatements</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">ControlStatements</span></tt>)
+Put a space before opening parentheses only after control statement
+keywords (<tt class="docutils literal"><span class="pre">for/if/while...</span></tt>).</li>
+<li><tt class="docutils literal"><span class="pre">SBPO_Always</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Always</span></tt>)
+Always put a space before opening parentheses, except when it’s
+prohibited by the syntax rules (in function-like macro definitions) or
+when determined by other style rules (after unary operators, opening
+parentheses, etc.)</li>
+</ul>
+</dd>
+<dt><strong>SpaceInEmptyParentheses</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, spaces may be inserted into ‘()’.</dd>
+<dt><strong>SpacesBeforeTrailingComments</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd><p class="first">The number of spaces before trailing line comments
+(<tt class="docutils literal"><span class="pre">//</span></tt> - comments).</p>
+<p class="last">This does not affect trailing block comments (<tt class="docutils literal"><span class="pre">/**/</span></tt> - comments) as those
+commonly have different usage patterns and a number of special cases.</p>
+</dd>
+<dt><strong>SpacesInAngles</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, spaces will be inserted after ‘<’ and before ‘>’ in
+template argument lists</dd>
+<dt><strong>SpacesInCStyleCastParentheses</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, spaces may be inserted into C style casts.</dd>
+<dt><strong>SpacesInContainerLiterals</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, spaces are inserted inside container literals (e.g.
+ObjC and Javascript array and dict literals).</dd>
+<dt><strong>SpacesInParentheses</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, spaces will be inserted after ‘(‘ and before ‘)’.</dd>
+<dt><strong>SpacesInSquareBrackets</strong> (<tt class="docutils literal"><span class="pre">bool</span></tt>)</dt>
+<dd>If <tt class="docutils literal"><span class="pre">true</span></tt>, spaces will be inserted after ‘[‘ and before ‘]’.</dd>
+<dt><strong>Standard</strong> (<tt class="docutils literal"><span class="pre">LanguageStandard</span></tt>)</dt>
+<dd><p class="first">Format compatible with this standard, e.g. use
+<tt class="docutils literal"><span class="pre">A<A<int></span> <span class="pre">></span></tt> instead of <tt class="docutils literal"><span class="pre">A<A<int>></span></tt> for LS_Cpp03.</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">LS_Cpp03</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Cpp03</span></tt>)
+Use C++03-compatible syntax.</li>
+<li><tt class="docutils literal"><span class="pre">LS_Cpp11</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Cpp11</span></tt>)
+Use features of C++11 (e.g. <tt class="docutils literal"><span class="pre">A<A<int>></span></tt> instead of
+<tt class="docutils literal"><span class="pre">A<A<int></span> <span class="pre">></span></tt>).</li>
+<li><tt class="docutils literal"><span class="pre">LS_Auto</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Auto</span></tt>)
+Automatic detection based on the input.</li>
+</ul>
+</dd>
+<dt><strong>TabWidth</strong> (<tt class="docutils literal"><span class="pre">unsigned</span></tt>)</dt>
+<dd>The number of columns used for tab stops.</dd>
+<dt><strong>UseTab</strong> (<tt class="docutils literal"><span class="pre">UseTabStyle</span></tt>)</dt>
+<dd><p class="first">The way to use tab characters in the resulting file.</p>
+<p>Possible values:</p>
+<ul class="last simple">
+<li><tt class="docutils literal"><span class="pre">UT_Never</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Never</span></tt>)
+Never use tab.</li>
+<li><tt class="docutils literal"><span class="pre">UT_ForIndentation</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">ForIndentation</span></tt>)
+Use tabs only for indentation.</li>
+<li><tt class="docutils literal"><span class="pre">UT_Always</span></tt> (in configuration: <tt class="docutils literal"><span class="pre">Always</span></tt>)
+Use tabs whenever we need to fill whitespace that spans at least from
+one tab stop to the next one.</li>
+</ul>
+</dd>
+</dl>
+</div>
+<div class="section" id="adding-additional-style-options">
+<h2>Adding additional style options<a class="headerlink" href="#adding-additional-style-options" title="Permalink to this headline">¶</a></h2>
+<p>Each additional style option adds costs to the clang-format project. Some of
+these costs affect the clang-format developement itself, as we need to make
+sure that any given combination of options work and that new features don’t
+break any of the existing options in any way. There are also costs for end users
+as options become less discoverable and people have to think about and make a
+decision on options they don’t really care about.</p>
+<p>The goal of the clang-format project is more on the side of supporting a
+limited set of styles really well as opposed to supporting every single style
+used by a codebase somewhere in the wild. Of course, we do want to support all
+major projects and thus have established the following bar for adding style
+options. Each new style option must ..</p>
+<blockquote>
+<div><ul class="simple">
+<li>be used in a project of significant size (have dozens of contributors)</li>
+<li>have a publicly accessible style guide</li>
+<li>have a person willing to contribute and maintain patches</li>
+</ul>
+</div></blockquote>
+</div>
+<div class="section" id="examples">
+<h2>Examples<a class="headerlink" href="#examples" title="Permalink to this headline">¶</a></h2>
+<p>A style similar to the <a class="reference external" href="https://www.kernel.org/doc/Documentation/CodingStyle">Linux Kernel style</a>:</p>
+<div class="highlight-yaml"><div class="highlight"><pre><span class="l-Scalar-Plain">BasedOnStyle</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">LLVM</span>
+<span class="l-Scalar-Plain">IndentWidth</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">8</span>
+<span class="l-Scalar-Plain">UseTab</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">Always</span>
+<span class="l-Scalar-Plain">BreakBeforeBraces</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">Linux</span>
+<span class="l-Scalar-Plain">AllowShortIfStatementsOnASingleLine</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">false</span>
+<span class="l-Scalar-Plain">IndentCaseLabels</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">false</span>
+</pre></div>
+</div>
+<p>The result is (imagine that tabs are used for indentation here):</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">void</span> <span class="nf">test</span><span class="p">()</span>
+<span class="p">{</span>
+ <span class="k">switch</span> <span class="p">(</span><span class="n">x</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">case</span> <span class="mi">0</span>:
+ <span class="k">case</span> <span class="mi">1</span>:
+ <span class="n">do_something</span><span class="p">();</span>
+ <span class="k">break</span><span class="p">;</span>
+ <span class="k">case</span> <span class="mi">2</span>:
+ <span class="n">do_something_else</span><span class="p">();</span>
+ <span class="k">break</span><span class="p">;</span>
+ <span class="nl">default:</span>
+ <span class="k">break</span><span class="p">;</span>
+ <span class="p">}</span>
+ <span class="k">if</span> <span class="p">(</span><span class="n">condition</span><span class="p">)</span>
+ <span class="n">do_something_completely_different</span><span class="p">();</span>
+
+ <span class="k">if</span> <span class="p">(</span><span class="n">x</span> <span class="o">==</span> <span class="n">y</span><span class="p">)</span> <span class="p">{</span>
+ <span class="n">q</span><span class="p">();</span>
+ <span class="p">}</span> <span class="k">else</span> <span class="k">if</span> <span class="p">(</span><span class="n">x</span> <span class="o">></span> <span class="n">y</span><span class="p">)</span> <span class="p">{</span>
+ <span class="n">w</span><span class="p">();</span>
+ <span class="p">}</span> <span class="k">else</span> <span class="p">{</span>
+ <span class="n">r</span><span class="p">();</span>
+ <span class="p">}</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>A style similar to the default Visual Studio formatting style:</p>
+<div class="highlight-yaml"><div class="highlight"><pre><span class="l-Scalar-Plain">UseTab</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">Never</span>
+<span class="l-Scalar-Plain">IndentWidth</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">4</span>
+<span class="l-Scalar-Plain">BreakBeforeBraces</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">Allman</span>
+<span class="l-Scalar-Plain">AllowShortIfStatementsOnASingleLine</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">false</span>
+<span class="l-Scalar-Plain">IndentCaseLabels</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">false</span>
+<span class="l-Scalar-Plain">ColumnLimit</span><span class="p-Indicator">:</span> <span class="l-Scalar-Plain">0</span>
+</pre></div>
+</div>
+<p>The result is:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">void</span> <span class="nf">test</span><span class="p">()</span>
+<span class="p">{</span>
+ <span class="k">switch</span> <span class="p">(</span><span class="n">suffix</span><span class="p">)</span>
+ <span class="p">{</span>
+ <span class="k">case</span> <span class="mi">0</span>:
+ <span class="k">case</span> <span class="mi">1</span>:
+ <span class="n">do_something</span><span class="p">();</span>
+ <span class="k">break</span><span class="p">;</span>
+ <span class="k">case</span> <span class="mi">2</span>:
+ <span class="n">do_something_else</span><span class="p">();</span>
+ <span class="k">break</span><span class="p">;</span>
+ <span class="nl">default:</span>
+ <span class="k">break</span><span class="p">;</span>
+ <span class="p">}</span>
+ <span class="k">if</span> <span class="p">(</span><span class="n">condition</span><span class="p">)</span>
+ <span class="n">do_somthing_completely_different</span><span class="p">();</span>
+
+ <span class="k">if</span> <span class="p">(</span><span class="n">x</span> <span class="o">==</span> <span class="n">y</span><span class="p">)</span>
+ <span class="p">{</span>
+ <span class="n">q</span><span class="p">();</span>
+ <span class="p">}</span>
+ <span class="k">else</span> <span class="k">if</span> <span class="p">(</span><span class="n">x</span> <span class="o">></span> <span class="n">y</span><span class="p">)</span>
+ <span class="p">{</span>
+ <span class="n">w</span><span class="p">();</span>
+ <span class="p">}</span>
+ <span class="k">else</span>
+ <span class="p">{</span>
+ <span class="n">r</span><span class="p">();</span>
+ <span class="p">}</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="ClangFormat.html">ClangFormat</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="InternalsManual.html">“Clang” CFE Internals Manual</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/ClangPlugins.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/ClangPlugins.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/ClangPlugins.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/ClangPlugins.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,147 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Clang Plugins — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="How to write RecursiveASTVisitor based ASTFrontendActions." href="RAVFrontendAction.html" />
+ <link rel="prev" title="LibFormat" href="LibFormat.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Clang Plugins</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="LibFormat.html">LibFormat</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="RAVFrontendAction.html">How to write RecursiveASTVisitor based ASTFrontendActions.</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="clang-plugins">
+<h1>Clang Plugins<a class="headerlink" href="#clang-plugins" title="Permalink to this headline">¶</a></h1>
+<p>Clang Plugins make it possible to run extra user defined actions during a
+compilation. This document will provide a basic walkthrough of how to write and
+run a Clang Plugin.</p>
+<div class="section" id="introduction">
+<h2>Introduction<a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>Clang Plugins run FrontendActions over code. See the <a class="reference internal" href="RAVFrontendAction.html"><em>FrontendAction
+tutorial</em></a> on how to write a <tt class="docutils literal"><span class="pre">FrontendAction</span></tt> using the
+<tt class="docutils literal"><span class="pre">RecursiveASTVisitor</span></tt>. In this tutorial, we’ll demonstrate how to write a
+simple clang plugin.</p>
+</div>
+<div class="section" id="writing-a-pluginastaction">
+<h2>Writing a <tt class="docutils literal"><span class="pre">PluginASTAction</span></tt><a class="headerlink" href="#writing-a-pluginastaction" title="Permalink to this headline">¶</a></h2>
+<p>The main difference from writing normal <tt class="docutils literal"><span class="pre">FrontendActions</span></tt> is that you can
+handle plugin command line options. The <tt class="docutils literal"><span class="pre">PluginASTAction</span></tt> base class declares
+a <tt class="docutils literal"><span class="pre">ParseArgs</span></tt> method which you have to implement in your plugin.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">bool</span> <span class="nf">ParseArgs</span><span class="p">(</span><span class="k">const</span> <span class="n">CompilerInstance</span> <span class="o">&</span><span class="n">CI</span><span class="p">,</span>
+ <span class="k">const</span> <span class="n">std</span><span class="o">::</span><span class="n">vector</span><span class="o"><</span><span class="n">std</span><span class="o">::</span><span class="n">string</span><span class="o">>&</span> <span class="n">args</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">for</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="n">i</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="n">e</span> <span class="o">=</span> <span class="n">args</span><span class="p">.</span><span class="n">size</span><span class="p">();</span> <span class="n">i</span> <span class="o">!=</span> <span class="n">e</span><span class="p">;</span> <span class="o">++</span><span class="n">i</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">if</span> <span class="p">(</span><span class="n">args</span><span class="p">[</span><span class="n">i</span><span class="p">]</span> <span class="o">==</span> <span class="s">"-some-arg"</span><span class="p">)</span> <span class="p">{</span>
+ <span class="c1">// Handle the command line argument.</span>
+ <span class="p">}</span>
+ <span class="p">}</span>
+ <span class="k">return</span> <span class="nb">true</span><span class="p">;</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="registering-a-plugin">
+<h2>Registering a plugin<a class="headerlink" href="#registering-a-plugin" title="Permalink to this headline">¶</a></h2>
+<p>A plugin is loaded from a dynamic library at runtime by the compiler. To
+register a plugin in a library, use <tt class="docutils literal"><span class="pre">FrontendPluginRegistry::Add<></span></tt>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">static</span> <span class="n">FrontendPluginRegistry</span><span class="o">::</span><span class="n">Add</span><span class="o"><</span><span class="n">MyPlugin</span><span class="o">></span> <span class="n">X</span><span class="p">(</span><span class="s">"my-plugin-name"</span><span class="p">,</span> <span class="s">"my plugin description"</span><span class="p">);</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="putting-it-all-together">
+<h2>Putting it all together<a class="headerlink" href="#putting-it-all-together" title="Permalink to this headline">¶</a></h2>
+<p>Let’s look at an example plugin that prints top-level function names. This
+example is checked into the clang repository; please take a look at
+the <a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/examples/PrintFunctionNames/PrintFunctionNames.cpp?view=markup">latest version of PrintFunctionNames.cpp</a>.</p>
+</div>
+<div class="section" id="running-the-plugin">
+<h2>Running the plugin<a class="headerlink" href="#running-the-plugin" title="Permalink to this headline">¶</a></h2>
+<p>To run a plugin, the dynamic library containing the plugin registry must be
+loaded via the <em class="xref std std-option">-load</em> command line option. This will load all plugins
+that are registered, and you can select the plugins to run by specifying the
+<em class="xref std std-option">-plugin</em> option. Additional parameters for the plugins can be passed with
+<em class="xref std std-option">-plugin-arg-</em>.</p>
+<p>Note that those options must reach clang’s cc1 process. There are two
+ways to do so:</p>
+<ul class="simple">
+<li>Directly call the parsing process by using the <em class="xref std std-option">-cc1</em> option; this
+has the downside of not configuring the default header search paths, so
+you’ll need to specify the full system path configuration on the command
+line.</li>
+<li>Use clang as usual, but prefix all arguments to the cc1 process with
+<em class="xref std std-option">-Xclang</em>.</li>
+</ul>
+<p>For example, to run the <tt class="docutils literal"><span class="pre">print-function-names</span></tt> plugin over a source file in
+clang, first build the plugin, and then call clang with the plugin from the
+source tree:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> <span class="nb">export </span><span class="nv">BD</span><span class="o">=</span>/path/to/build/directory
+<span class="gp">$</span> <span class="o">(</span><span class="nb">cd</span> <span class="nv">$BD</span> <span class="o">&&</span> make PrintFunctionNames <span class="o">)</span>
+<span class="gp">$</span> clang++ -D_GNU_SOURCE -D_DEBUG -D__STDC_CONSTANT_MACROS <span class="se">\</span>
+<span class="go"> -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -D_GNU_SOURCE \</span>
+<span class="go"> -I$BD/tools/clang/include -Itools/clang/include -I$BD/include -Iinclude \</span>
+<span class="go"> tools/clang/tools/clang-check/ClangCheck.cpp -fsyntax-only \</span>
+<span class="go"> -Xclang -load -Xclang $BD/lib/PrintFunctionNames.so -Xclang \</span>
+<span class="go"> -plugin -Xclang print-fns</span>
+</pre></div>
+</div>
+<p>Also see the print-function-name plugin example’s
+<a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/examples/PrintFunctionNames/README.txt?view=markup">README</a></p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="LibFormat.html">LibFormat</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="RAVFrontendAction.html">How to write RecursiveASTVisitor based ASTFrontendActions.</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/ClangTools.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/ClangTools.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/ClangTools.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/ClangTools.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,220 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Overview — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="ClangCheck" href="ClangCheck.html" />
+ <link rel="prev" title="JSON Compilation Database Format Specification" href="JSONCompilationDatabase.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Overview</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="JSONCompilationDatabase.html">JSON Compilation Database Format Specification</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ClangCheck.html">ClangCheck</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="overview">
+<h1>Overview<a class="headerlink" href="#overview" title="Permalink to this headline">¶</a></h1>
+<p>Clang Tools are standalone command line (and potentially GUI) tools
+designed for use by C++ developers who are already using and enjoying
+Clang as their compiler. These tools provide developer-oriented
+functionality such as fast syntax checking, automatic formatting,
+refactoring, etc.</p>
+<p>Only a couple of the most basic and fundamental tools are kept in the
+primary Clang Subversion project. The rest of the tools are kept in a
+side-project so that developers who don’t want or need to build them
+don’t. If you want to get access to the extra Clang Tools repository,
+simply check it out into the tools tree of your Clang checkout and
+follow the usual process for building and working with a combined
+LLVM/Clang checkout:</p>
+<ul class="simple">
+<li>With Subversion:<ul>
+<li><tt class="docutils literal"><span class="pre">cd</span> <span class="pre">llvm/tools/clang/tools</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">svn</span> <span class="pre">co</span> <span class="pre">http://llvm.org/svn/llvm-project/clang-tools-extra/trunk</span> <span class="pre">extra</span></tt></li>
+</ul>
+</li>
+<li>Or with Git:<ul>
+<li><tt class="docutils literal"><span class="pre">cd</span> <span class="pre">llvm/tools/clang/tools</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">git</span> <span class="pre">clone</span> <span class="pre">http://llvm.org/git/clang-tools-extra.git</span> <span class="pre">extra</span></tt></li>
+</ul>
+</li>
+</ul>
+<p>This document describes a high-level overview of the organization of
+Clang Tools within the project as well as giving an introduction to some
+of the more important tools. However, it should be noted that this
+document is currently focused on Clang and Clang Tool developers, not on
+end users of these tools.</p>
+<div class="section" id="clang-tools-organization">
+<h2>Clang Tools Organization<a class="headerlink" href="#clang-tools-organization" title="Permalink to this headline">¶</a></h2>
+<p>Clang Tools are CLI or GUI programs that are intended to be directly
+used by C++ developers. That is they are <em>not</em> primarily for use by
+Clang developers, although they are hopefully useful to C++ developers
+who happen to work on Clang, and we try to actively dogfood their
+functionality. They are developed in three components: the underlying
+infrastructure for building a standalone tool based on Clang, core
+shared logic used by many different tools in the form of refactoring and
+rewriting libraries, and the tools themselves.</p>
+<p>The underlying infrastructure for Clang Tools is the
+<a class="reference internal" href="LibTooling.html"><em>LibTooling</em></a> platform. See its documentation for much
+more detailed information about how this infrastructure works. The
+common refactoring and rewriting toolkit-style library is also part of
+LibTooling organizationally.</p>
+<p>A few Clang Tools are developed along side the core Clang libraries as
+examples and test cases of fundamental functionality. However, most of
+the tools are developed in a side repository to provide easy separation
+from the core libraries. We intentionally do not support public
+libraries in the side repository, as we want to carefully review and
+find good APIs for libraries as they are lifted out of a few tools and
+into the core Clang library set.</p>
+<p>Regardless of which repository Clang Tools’ code resides in, the
+development process and practices for all Clang Tools are exactly those
+of Clang itself. They are entirely within the Clang <em>project</em>,
+regardless of the version control scheme.</p>
+</div>
+<div class="section" id="core-clang-tools">
+<h2>Core Clang Tools<a class="headerlink" href="#core-clang-tools" title="Permalink to this headline">¶</a></h2>
+<p>The core set of Clang tools that are within the main repository are
+tools that very specifically complement, and allow use and testing of
+<em>Clang</em> specific functionality.</p>
+<div class="section" id="clang-check">
+<h3><tt class="docutils literal"><span class="pre">clang-check</span></tt><a class="headerlink" href="#clang-check" title="Permalink to this headline">¶</a></h3>
+<p><a class="reference internal" href="ClangCheck.html"><em>ClangCheck</em></a> combines the LibTooling framework for running a
+Clang tool with the basic Clang diagnostics by syntax checking specific files
+in a fast, command line interface. It can also accept flags to re-display the
+diagnostics in different formats with different flags, suitable for use driving
+an IDE or editor. Furthermore, it can be used in fixit-mode to directly apply
+fixit-hints offered by clang. See <a class="reference internal" href="HowToSetupToolingForLLVM.html"><em>How To Setup Clang Tooling For LLVM</em></a> for
+instructions on how to setup and used <cite>clang-check</cite>.</p>
+</div>
+<div class="section" id="clang-format">
+<h3><tt class="docutils literal"><span class="pre">clang-format</span></tt><a class="headerlink" href="#clang-format" title="Permalink to this headline">¶</a></h3>
+<p>Clang-format is both a <a class="reference internal" href="LibFormat.html"><em>library</em></a> and a <a class="reference internal" href="ClangFormat.html"><em>stand-alone tool</em></a> with the goal of automatically reformatting C++ sources files
+according to configurable style guides. To do so, clang-format uses Clang’s
+<tt class="docutils literal"><span class="pre">Lexer</span></tt> to transform an input file into a token stream and then changes all
+the whitespace around those tokens. The goal is for clang-format to serve both
+as a user tool (ideally with powerful IDE integrations) and as part of other
+refactoring tools, e.g. to do a reformatting of all the lines changed during a
+renaming.</p>
+</div>
+</div>
+<div class="section" id="extra-clang-tools">
+<h2>Extra Clang Tools<a class="headerlink" href="#extra-clang-tools" title="Permalink to this headline">¶</a></h2>
+<p>As various categories of Clang Tools are added to the extra repository,
+they’ll be tracked here. The focus of this documentation is on the scope
+and features of the tools for other tool developers; each tool should
+provide its own user-focused documentation.</p>
+<div class="section" id="clang-tidy">
+<h3><tt class="docutils literal"><span class="pre">clang-tidy</span></tt><a class="headerlink" href="#clang-tidy" title="Permalink to this headline">¶</a></h3>
+<p><a class="reference external" href="http://clang.llvm.org/extra/clang-tidy/">clang-tidy</a> is a clang-based C++
+linter tool. It provides an extensible framework for building compiler-based
+static analyses detecting and fixing bug-prone patterns, performance,
+portability and maintainability issues.</p>
+</div>
+</div>
+<div class="section" id="ideas-for-new-tools">
+<h2>Ideas for new Tools<a class="headerlink" href="#ideas-for-new-tools" title="Permalink to this headline">¶</a></h2>
+<ul>
+<li><p class="first">C++ cast conversion tool. Will convert C-style casts (<tt class="docutils literal"><span class="pre">(type)</span> <span class="pre">value</span></tt>) to
+appropriate C++ cast (<tt class="docutils literal"><span class="pre">static_cast</span></tt>, <tt class="docutils literal"><span class="pre">const_cast</span></tt> or
+<tt class="docutils literal"><span class="pre">reinterpret_cast</span></tt>).</p>
+</li>
+<li><p class="first">Non-member <tt class="docutils literal"><span class="pre">begin()</span></tt> and <tt class="docutils literal"><span class="pre">end()</span></tt> conversion tool. Will convert
+<tt class="docutils literal"><span class="pre">foo.begin()</span></tt> into <tt class="docutils literal"><span class="pre">begin(foo)</span></tt> and similarly for <tt class="docutils literal"><span class="pre">end()</span></tt>, where
+<tt class="docutils literal"><span class="pre">foo</span></tt> is a standard container. We could also detect similar patterns for
+arrays.</p>
+</li>
+<li><p class="first"><tt class="docutils literal"><span class="pre">tr1</span></tt> removal tool. Will migrate source code from using TR1 library
+features to C++11 library. For example:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#include <tr1/unordered_map></span>
+<span class="kt">int</span> <span class="nf">main</span><span class="p">()</span>
+<span class="p">{</span>
+ <span class="n">std</span><span class="o">::</span><span class="n">tr1</span><span class="o">::</span><span class="n">unordered_map</span> <span class="o"><</span><span class="kt">int</span><span class="p">,</span> <span class="kt">int</span><span class="o">></span> <span class="n">ma</span><span class="p">;</span>
+ <span class="n">std</span><span class="o">::</span><span class="n">cout</span> <span class="o"><<</span> <span class="n">ma</span><span class="p">.</span><span class="n">size</span> <span class="p">()</span> <span class="o"><<</span> <span class="n">std</span><span class="o">::</span><span class="n">endl</span><span class="p">;</span>
+ <span class="k">return</span> <span class="mi">0</span><span class="p">;</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>should be rewritten to:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#include <unordered_map></span>
+<span class="kt">int</span> <span class="nf">main</span><span class="p">()</span>
+<span class="p">{</span>
+ <span class="n">std</span><span class="o">::</span><span class="n">unordered_map</span> <span class="o"><</span><span class="kt">int</span><span class="p">,</span> <span class="kt">int</span><span class="o">></span> <span class="n">ma</span><span class="p">;</span>
+ <span class="n">std</span><span class="o">::</span><span class="n">cout</span> <span class="o"><<</span> <span class="n">ma</span><span class="p">.</span><span class="n">size</span> <span class="p">()</span> <span class="o"><<</span> <span class="n">std</span><span class="o">::</span><span class="n">endl</span><span class="p">;</span>
+ <span class="k">return</span> <span class="mi">0</span><span class="p">;</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+</li>
+<li><p class="first">A tool to remove <tt class="docutils literal"><span class="pre">auto</span></tt>. Will convert <tt class="docutils literal"><span class="pre">auto</span></tt> to an explicit type or add
+comments with deduced types. The motivation is that there are developers
+that don’t want to use <tt class="docutils literal"><span class="pre">auto</span></tt> because they are afraid that they might lose
+control over their code.</p>
+</li>
+<li><p class="first">C++14: less verbose operator function objects (<a class="reference external" href="http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2012/n3421.htm">N3421</a>).
+For example:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">sort</span><span class="p">(</span><span class="n">v</span><span class="p">.</span><span class="n">begin</span><span class="p">(),</span> <span class="n">v</span><span class="p">.</span><span class="n">end</span><span class="p">(),</span> <span class="n">greater</span><span class="o"><</span><span class="n">ValueType</span><span class="o">></span><span class="p">());</span>
+</pre></div>
+</div>
+<p>should be rewritten to:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">sort</span><span class="p">(</span><span class="n">v</span><span class="p">.</span><span class="n">begin</span><span class="p">(),</span> <span class="n">v</span><span class="p">.</span><span class="n">end</span><span class="p">(),</span> <span class="n">greater</span><span class="o"><></span><span class="p">());</span>
+</pre></div>
+</div>
+</li>
+</ul>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="JSONCompilationDatabase.html">JSON Compilation Database Format Specification</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ClangCheck.html">ClangCheck</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/CommandGuide/clang.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/CommandGuide/clang.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/CommandGuide/clang.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/CommandGuide/clang.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,689 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>clang - the Clang C, C++, and Objective-C compiler — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="../_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: '../',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="../_static/jquery.js"></script>
+ <script type="text/javascript" src="../_static/underscore.js"></script>
+ <script type="text/javascript" src="../_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="../index.html" />
+ <link rel="up" title="Clang âmanâ pages" href="index.html" />
+ <link rel="next" title="Frequently Asked Questions (FAQ)" href="../FAQ.html" />
+ <link rel="prev" title="Clang âmanâ pages" href="index.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="../index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>clang - the Clang C, C++, and Objective-C compiler</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="index.html">Clang “man” pages</a>
+ ::
+ <a class="uplink" href="../index.html">Contents</a>
+ ::
+ <a href="../FAQ.html">Frequently Asked Questions (FAQ)</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="clang-the-clang-c-c-and-objective-c-compiler">
+<h1>clang - the Clang C, C++, and Objective-C compiler<a class="headerlink" href="#clang-the-clang-c-c-and-objective-c-compiler" title="Permalink to this headline">¶</a></h1>
+<div class="section" id="synopsis">
+<h2>SYNOPSIS<a class="headerlink" href="#synopsis" title="Permalink to this headline">¶</a></h2>
+<p><strong class="program">clang</strong> [<em>options</em>] <em>filename ...</em></p>
+</div>
+<div class="section" id="description">
+<h2>DESCRIPTION<a class="headerlink" href="#description" title="Permalink to this headline">¶</a></h2>
+<p><strong class="program">clang</strong> is a C, C++, and Objective-C compiler which encompasses
+preprocessing, parsing, optimization, code generation, assembly, and linking.
+Depending on which high-level mode setting is passed, Clang will stop before
+doing a full link. While Clang is highly integrated, it is important to
+understand the stages of compilation, to understand how to invoke it. These
+stages are:</p>
+<dl class="docutils">
+<dt>Driver</dt>
+<dd>The clang executable is actually a small driver which controls the overall
+execution of other tools such as the compiler, assembler and linker.
+Typically you do not need to interact with the driver, but you
+transparently use it to run the other tools.</dd>
+<dt>Preprocessing</dt>
+<dd>This stage handles tokenization of the input source file, macro expansion,
+#include expansion and handling of other preprocessor directives. The
+output of this stage is typically called a ”.i” (for C), ”.ii” (for C++),
+”.mi” (for Objective-C), or ”.mii” (for Objective-C++) file.</dd>
+<dt>Parsing and Semantic Analysis</dt>
+<dd>This stage parses the input file, translating preprocessor tokens into a
+parse tree. Once in the form of a parse tree, it applies semantic
+analysis to compute types for expressions as well and determine whether
+the code is well formed. This stage is responsible for generating most of
+the compiler warnings as well as parse errors. The output of this stage is
+an “Abstract Syntax Tree” (AST).</dd>
+<dt>Code Generation and Optimization</dt>
+<dd><p class="first">This stage translates an AST into low-level intermediate code (known as
+“LLVM IR”) and ultimately to machine code. This phase is responsible for
+optimizing the generated code and handling target-specific code generation.
+The output of this stage is typically called a ”.s” file or “assembly” file.</p>
+<p class="last">Clang also supports the use of an integrated assembler, in which the code
+generator produces object files directly. This avoids the overhead of
+generating the ”.s” file and of calling the target assembler.</p>
+</dd>
+<dt>Assembler</dt>
+<dd>This stage runs the target assembler to translate the output of the
+compiler into a target object file. The output of this stage is typically
+called a ”.o” file or “object” file.</dd>
+<dt>Linker</dt>
+<dd>This stage runs the target linker to merge multiple object files into an
+executable or dynamic library. The output of this stage is typically called
+an “a.out”, ”.dylib” or ”.so” file.</dd>
+</dl>
+<p><strong class="program">Clang Static Analyzer</strong></p>
+<p>The Clang Static Analyzer is a tool that scans source code to try to find bugs
+through code analysis. This tool uses many parts of Clang and is built into
+the same driver. Please see <<a class="reference external" href="http://clang-analyzer.llvm.org">http://clang-analyzer.llvm.org</a>> for more details
+on how to use the static analyzer.</p>
+</div>
+<div class="section" id="options">
+<h2>OPTIONS<a class="headerlink" href="#options" title="Permalink to this headline">¶</a></h2>
+<div class="section" id="stage-selection-options">
+<h3>Stage Selection Options<a class="headerlink" href="#stage-selection-options" title="Permalink to this headline">¶</a></h3>
+<dl class="option">
+<dt id="cmdoption-E">
+<tt class="descname">-E</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-E" title="Permalink to this definition">¶</a></dt>
+<dd><p>Run the preprocessor stage.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fsyntax-only">
+<tt class="descname">-fsyntax-only</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fsyntax-only" title="Permalink to this definition">¶</a></dt>
+<dd><p>Run the preprocessor, parser and type checking stages.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-S">
+<tt class="descname">-S</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-S" title="Permalink to this definition">¶</a></dt>
+<dd><p>Run the previous stages as well as LLVM generation and optimization stages
+and target-specific code generation, producing an assembly file.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-c">
+<tt class="descname">-c</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-c" title="Permalink to this definition">¶</a></dt>
+<dd><p>Run all of the above, plus the assembler, generating a target ”.o” object file.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-arg-no">
+<tt class="descname">no</tt><tt class="descclassname"> stage selection option</tt><a class="headerlink" href="#cmdoption-arg-no" title="Permalink to this definition">¶</a></dt>
+<dd><p>If no stage selection option is specified, all stages above are run, and the
+linker is run to combine the results into an executable or shared library.</p>
+</dd></dl>
+
+</div>
+<div class="section" id="language-selection-and-mode-options">
+<h3>Language Selection and Mode Options<a class="headerlink" href="#language-selection-and-mode-options" title="Permalink to this headline">¶</a></h3>
+<dl class="option">
+<dt id="cmdoption-x">
+<tt class="descname">-x</tt><tt class="descclassname"> <language></tt><a class="headerlink" href="#cmdoption-x" title="Permalink to this definition">¶</a></dt>
+<dd><p>Treat subsequent input files as having type language.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-std">
+<tt class="descname">-std</tt><tt class="descclassname">=<language></tt><a class="headerlink" href="#cmdoption-std" title="Permalink to this definition">¶</a></dt>
+<dd><p>Specify the language standard to compile for.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-stdlib">
+<tt class="descname">-stdlib</tt><tt class="descclassname">=<library></tt><a class="headerlink" href="#cmdoption-stdlib" title="Permalink to this definition">¶</a></dt>
+<dd><p>Specify the C++ standard library to use; supported options are libstdc++ and
+libc++.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-ansi">
+<tt class="descname">-ansi</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-ansi" title="Permalink to this definition">¶</a></dt>
+<dd><p>Same as -std=c89.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-ObjC">
+<span id="cmdoption-ObjC"></span><tt class="descname">-ObjC</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-ObjC</tt><tt class="descclassname">++</tt><a class="headerlink" href="#cmdoption-ObjC" title="Permalink to this definition">¶</a></dt>
+<dd><p>Treat source input files as Objective-C and Object-C++ inputs respectively.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-trigraphs">
+<tt class="descname">-trigraphs</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-trigraphs" title="Permalink to this definition">¶</a></dt>
+<dd><p>Enable trigraphs.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-ffreestanding">
+<tt class="descname">-ffreestanding</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-ffreestanding" title="Permalink to this definition">¶</a></dt>
+<dd><p>Indicate that the file should be compiled for a freestanding, not a hosted,
+environment.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fno-builtin">
+<tt class="descname">-fno-builtin</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fno-builtin" title="Permalink to this definition">¶</a></dt>
+<dd><p>Disable special handling and optimizations of builtin functions like
+<tt class="xref c c-func docutils literal"><span class="pre">strlen()</span></tt> and <tt class="xref c c-func docutils literal"><span class="pre">malloc()</span></tt>.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fmath-errno">
+<tt class="descname">-fmath-errno</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fmath-errno" title="Permalink to this definition">¶</a></dt>
+<dd><p>Indicate that math functions should be treated as updating <tt class="xref c c-data docutils literal"><span class="pre">errno</span></tt>.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fpascal-strings">
+<tt class="descname">-fpascal-strings</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fpascal-strings" title="Permalink to this definition">¶</a></dt>
+<dd><p>Enable support for Pascal-style strings with “\pfoo”.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fms-extensions">
+<tt class="descname">-fms-extensions</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fms-extensions" title="Permalink to this definition">¶</a></dt>
+<dd><p>Enable support for Microsoft extensions.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fmsc-version">
+<tt class="descname">-fmsc-version</tt><tt class="descclassname">=</tt><a class="headerlink" href="#cmdoption-fmsc-version" title="Permalink to this definition">¶</a></dt>
+<dd><p>Set _MSC_VER. Defaults to 1300 on Windows. Not set otherwise.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fborland-extensions">
+<tt class="descname">-fborland-extensions</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fborland-extensions" title="Permalink to this definition">¶</a></dt>
+<dd><p>Enable support for Borland extensions.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fwritable-strings">
+<tt class="descname">-fwritable-strings</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fwritable-strings" title="Permalink to this definition">¶</a></dt>
+<dd><p>Make all string literals default to writable. This disables uniquing of
+strings and other optimizations.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-flax-vector-conversions">
+<tt class="descname">-flax-vector-conversions</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-flax-vector-conversions" title="Permalink to this definition">¶</a></dt>
+<dd><p>Allow loose type checking rules for implicit vector conversions.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fblocks">
+<tt class="descname">-fblocks</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fblocks" title="Permalink to this definition">¶</a></dt>
+<dd><p>Enable the “Blocks” language feature.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fobjc-gc-only">
+<tt class="descname">-fobjc-gc-only</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fobjc-gc-only" title="Permalink to this definition">¶</a></dt>
+<dd><p>Indicate that Objective-C code should be compiled in GC-only mode, which only
+works when Objective-C Garbage Collection is enabled.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fobjc-gc">
+<tt class="descname">-fobjc-gc</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fobjc-gc" title="Permalink to this definition">¶</a></dt>
+<dd><p>Indicate that Objective-C code should be compiled in hybrid-GC mode, which
+works with both GC and non-GC mode.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fobjc-abi-version">
+<tt class="descname">-fobjc-abi-version</tt><tt class="descclassname">=version</tt><a class="headerlink" href="#cmdoption-fobjc-abi-version" title="Permalink to this definition">¶</a></dt>
+<dd><p>Select the Objective-C ABI version to use. Available versions are 1 (legacy
+“fragile” ABI), 2 (non-fragile ABI 1), and 3 (non-fragile ABI 2).</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fobjc-nonfragile-abi-version">
+<tt class="descname">-fobjc-nonfragile-abi-version</tt><tt class="descclassname">=<version></tt><a class="headerlink" href="#cmdoption-fobjc-nonfragile-abi-version" title="Permalink to this definition">¶</a></dt>
+<dd><p>Select the Objective-C non-fragile ABI version to use by default. This will
+only be used as the Objective-C ABI when the non-fragile ABI is enabled
+(either via <a class="reference internal" href="#cmdoption-fobjc-nonfragile-abi"><em class="xref std std-option">-fobjc-nonfragile-abi</em></a>, or because it is the platform
+default).</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fobjc-nonfragile-abi">
+<tt class="descname">-fobjc-nonfragile-abi</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fobjc-nonfragile-abi" title="Permalink to this definition">¶</a></dt>
+<dd><p>Enable use of the Objective-C non-fragile ABI. On platforms for which this is
+the default ABI, it can be disabled with <em class="xref std std-option">-fno-objc-nonfragile-abi</em>.</p>
+</dd></dl>
+
+</div>
+<div class="section" id="target-selection-options">
+<h3>Target Selection Options<a class="headerlink" href="#target-selection-options" title="Permalink to this headline">¶</a></h3>
+<p>Clang fully supports cross compilation as an inherent part of its design.
+Depending on how your version of Clang is configured, it may have support for a
+number of cross compilers, or may only support a native target.</p>
+<dl class="option">
+<dt id="cmdoption-arch">
+<tt class="descname">-arch</tt><tt class="descclassname"> <architecture></tt><a class="headerlink" href="#cmdoption-arch" title="Permalink to this definition">¶</a></dt>
+<dd><p>Specify the architecture to build for.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-mmacosx-version-min">
+<tt class="descname">-mmacosx-version-min</tt><tt class="descclassname">=<version></tt><a class="headerlink" href="#cmdoption-mmacosx-version-min" title="Permalink to this definition">¶</a></dt>
+<dd><p>When building for Mac OS X, specify the minimum version supported by your
+application.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-miphoneos-version-min">
+<tt class="descname">-miphoneos-version-min</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-miphoneos-version-min" title="Permalink to this definition">¶</a></dt>
+<dd><p>When building for iPhone OS, specify the minimum version supported by your
+application.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-march">
+<tt class="descname">-march</tt><tt class="descclassname">=<cpu></tt><a class="headerlink" href="#cmdoption-march" title="Permalink to this definition">¶</a></dt>
+<dd><p>Specify that Clang should generate code for a specific processor family
+member and later. For example, if you specify -march=i486, the compiler is
+allowed to generate instructions that are valid on i486 and later processors,
+but which may not exist on earlier ones.</p>
+</dd></dl>
+
+</div>
+<div class="section" id="code-generation-options">
+<h3>Code Generation Options<a class="headerlink" href="#code-generation-options" title="Permalink to this headline">¶</a></h3>
+<dl class="option">
+<dt id="cmdoption-O0">
+<span id="cmdoption-O1"></span><span id="cmdoption-O2"></span><span id="cmdoption-O3"></span><span id="cmdoption-Ofast"></span><span id="cmdoption-Os"></span><span id="cmdoption-Oz"></span><span id="cmdoption-O"></span><span id="cmdoption-O4"></span><tt class="descname">-O0</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-O1</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-O2</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-O3</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-Ofast</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-Os</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-Oz</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-O</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><
tt class="descname">-O4</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-O0" title="Permalink to this definition">¶</a></dt>
+<dd><p>Specify which optimization level to use:</p>
+<blockquote>
+<div><p><a class="reference internal" href="#cmdoption-O0"><em class="xref std std-option">-O0</em></a> Means “no optimization”: this level compiles the fastest and
+generates the most debuggable code.</p>
+<p><a class="reference internal" href="#cmdoption-O1"><em class="xref std std-option">-O1</em></a> Somewhere between <a class="reference internal" href="#cmdoption-O0"><em class="xref std std-option">-O0</em></a> and <a class="reference internal" href="#cmdoption-O2"><em class="xref std std-option">-O2</em></a>.</p>
+<p><a class="reference internal" href="#cmdoption-O2"><em class="xref std std-option">-O2</em></a> Moderate level of optimization which enables most
+optimizations.</p>
+<p><a class="reference internal" href="#cmdoption-O3"><em class="xref std std-option">-O3</em></a> Like <a class="reference internal" href="#cmdoption-O2"><em class="xref std std-option">-O2</em></a>, except that it enables optimizations that
+take longer to perform or that may generate larger code (in an attempt to
+make the program run faster).</p>
+<p><a class="reference internal" href="#cmdoption-Ofast"><em class="xref std std-option">-Ofast</em></a> Enables all the optimizations from <a class="reference internal" href="#cmdoption-O3"><em class="xref std std-option">-O3</em></a> along
+with other aggressive optimizations that may violate strict compliance with
+language standards.</p>
+<p><a class="reference internal" href="#cmdoption-Os"><em class="xref std std-option">-Os</em></a> Like <a class="reference internal" href="#cmdoption-O2"><em class="xref std std-option">-O2</em></a> with extra optimizations to reduce code
+size.</p>
+<p><a class="reference internal" href="#cmdoption-Oz"><em class="xref std std-option">-Oz</em></a> Like <a class="reference internal" href="#cmdoption-Os"><em class="xref std std-option">-Os</em></a> (and thus <a class="reference internal" href="#cmdoption-O2"><em class="xref std std-option">-O2</em></a>), but reduces code
+size further.</p>
+<p><a class="reference internal" href="#cmdoption-O"><em class="xref std std-option">-O</em></a> Equivalent to <a class="reference internal" href="#cmdoption-O2"><em class="xref std std-option">-O2</em></a>.</p>
+<p><a class="reference internal" href="#cmdoption-O4"><em class="xref std std-option">-O4</em></a> and higher</p>
+<blockquote>
+<div>Currently equivalent to <a class="reference internal" href="#cmdoption-O3"><em class="xref std std-option">-O3</em></a></div></blockquote>
+</div></blockquote>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-g">
+<tt class="descname">-g</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-g" title="Permalink to this definition">¶</a></dt>
+<dd><p>Generate debug information. Note that Clang debug information works best at -O0.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-gmodules">
+<tt class="descname">-gmodules</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-gmodules" title="Permalink to this definition">¶</a></dt>
+<dd><p>Generate debug information that contains external references to
+types defined in clang modules or precompiled headers instead of
+emitting redundant debug type information into every object file.
+This option implies <em class="xref std std-option">-fmodule-format=obj</em>.</p>
+<p>This option should not be used when building static libraries for
+distribution to other machines because the debug info will contain
+references to the module cache on the machine the object files in
+the library were built on.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fstandalone-debug">
+<tt class="descname">-fstandalone-debug</tt><tt class="descclassname"> -fno-standalone-debug</tt><a class="headerlink" href="#cmdoption-fstandalone-debug" title="Permalink to this definition">¶</a></dt>
+<dd><p>Clang supports a number of optimizations to reduce the size of debug
+information in the binary. They work based on the assumption that the
+debug type information can be spread out over multiple compilation units.
+For instance, Clang will not emit type definitions for types that are not
+needed by a module and could be replaced with a forward declaration.
+Further, Clang will only emit type info for a dynamic C++ class in the
+module that contains the vtable for the class.</p>
+<p>The <a class="reference internal" href="../UsersManual.html#cmdoption-fstandalone-debug"><em class="xref std std-option">-fstandalone-debug</em></a> option turns off these optimizations.
+This is useful when working with 3rd-party libraries that don’t come with
+debug information. This is the default on Darwin. Note that Clang will
+never emit type information for types that are not referenced at all by the
+program.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fexceptions">
+<tt class="descname">-fexceptions</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fexceptions" title="Permalink to this definition">¶</a></dt>
+<dd><p>Enable generation of unwind information. This allows exceptions to be thrown
+through Clang compiled stack frames. This is on by default in x86-64.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-ftrapv">
+<tt class="descname">-ftrapv</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-ftrapv" title="Permalink to this definition">¶</a></dt>
+<dd><p>Generate code to catch integer overflow errors. Signed integer overflow is
+undefined in C. With this flag, extra code is generated to detect this and
+abort when it happens.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fvisibility">
+<tt class="descname">-fvisibility</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fvisibility" title="Permalink to this definition">¶</a></dt>
+<dd><p>This flag sets the default visibility level.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-fcommon">
+<tt class="descname">-fcommon</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fcommon" title="Permalink to this definition">¶</a></dt>
+<dd><p>This flag specifies that variables without initializers get common linkage.
+It can be disabled with <em class="xref std std-option">-fno-common</em>.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-ftls-model">
+<tt class="descname">-ftls-model</tt><tt class="descclassname">=<model></tt><a class="headerlink" href="#cmdoption-ftls-model" title="Permalink to this definition">¶</a></dt>
+<dd><p>Set the default thread-local storage (TLS) model to use for thread-local
+variables. Valid values are: “global-dynamic”, “local-dynamic”,
+“initial-exec” and “local-exec”. The default is “global-dynamic”. The default
+model can be overridden with the tls_model attribute. The compiler will try
+to choose a more efficient model if possible.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-flto">
+<span id="cmdoption-emit-llvm"></span><tt class="descname">-flto</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-emit-llvm</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-flto" title="Permalink to this definition">¶</a></dt>
+<dd><p>Generate output files in LLVM formats, suitable for link time optimization.
+When used with <a class="reference internal" href="#cmdoption-S"><em class="xref std std-option">-S</em></a> this generates LLVM intermediate language
+assembly files, otherwise this generates LLVM bitcode format object files
+(which may be passed to the linker depending on the stage selection options).</p>
+</dd></dl>
+
+</div>
+<div class="section" id="driver-options">
+<h3>Driver Options<a class="headerlink" href="#driver-options" title="Permalink to this headline">¶</a></h3>
+<dl class="option">
+<dt id="cmdoption-">
+<tt class="descname">-</tt><tt class="descclassname">###</tt><a class="headerlink" href="#cmdoption-" title="Permalink to this definition">¶</a></dt>
+<dd><p>Print (but do not run) the commands to run for this compilation.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption--help">
+<tt class="descname">--help</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption--help" title="Permalink to this definition">¶</a></dt>
+<dd><p>Display available options.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-Qunused-arguments">
+<tt class="descname">-Qunused-arguments</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-Qunused-arguments" title="Permalink to this definition">¶</a></dt>
+<dd><p>Do not emit any warnings for unused driver arguments.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-Wa">
+<tt class="descname">-Wa</tt><tt class="descclassname">,<args></tt><a class="headerlink" href="#cmdoption-Wa" title="Permalink to this definition">¶</a></dt>
+<dd><p>Pass the comma separated arguments in args to the assembler.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-Wl">
+<tt class="descname">-Wl</tt><tt class="descclassname">,<args></tt><a class="headerlink" href="#cmdoption-Wl" title="Permalink to this definition">¶</a></dt>
+<dd><p>Pass the comma separated arguments in args to the linker.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-Wp">
+<tt class="descname">-Wp</tt><tt class="descclassname">,<args></tt><a class="headerlink" href="#cmdoption-Wp" title="Permalink to this definition">¶</a></dt>
+<dd><p>Pass the comma separated arguments in args to the preprocessor.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-Xanalyzer">
+<tt class="descname">-Xanalyzer</tt><tt class="descclassname"> <arg></tt><a class="headerlink" href="#cmdoption-Xanalyzer" title="Permalink to this definition">¶</a></dt>
+<dd><p>Pass arg to the static analyzer.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-Xassembler">
+<tt class="descname">-Xassembler</tt><tt class="descclassname"> <arg></tt><a class="headerlink" href="#cmdoption-Xassembler" title="Permalink to this definition">¶</a></dt>
+<dd><p>Pass arg to the assembler.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-Xlinker">
+<tt class="descname">-Xlinker</tt><tt class="descclassname"> <arg></tt><a class="headerlink" href="#cmdoption-Xlinker" title="Permalink to this definition">¶</a></dt>
+<dd><p>Pass arg to the linker.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-Xpreprocessor">
+<tt class="descname">-Xpreprocessor</tt><tt class="descclassname"> <arg></tt><a class="headerlink" href="#cmdoption-Xpreprocessor" title="Permalink to this definition">¶</a></dt>
+<dd><p>Pass arg to the preprocessor.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-o">
+<tt class="descname">-o</tt><tt class="descclassname"> <file></tt><a class="headerlink" href="#cmdoption-o" title="Permalink to this definition">¶</a></dt>
+<dd><p>Write output to file.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-print-file-name">
+<tt class="descname">-print-file-name</tt><tt class="descclassname">=<file></tt><a class="headerlink" href="#cmdoption-print-file-name" title="Permalink to this definition">¶</a></dt>
+<dd><p>Print the full library path of file.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-print-libgcc-file-name">
+<tt class="descname">-print-libgcc-file-name</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-print-libgcc-file-name" title="Permalink to this definition">¶</a></dt>
+<dd><p>Print the library path for “libgcc.a”.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-print-prog-name">
+<tt class="descname">-print-prog-name</tt><tt class="descclassname">=<name></tt><a class="headerlink" href="#cmdoption-print-prog-name" title="Permalink to this definition">¶</a></dt>
+<dd><p>Print the full program path of name.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-print-search-dirs">
+<tt class="descname">-print-search-dirs</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-print-search-dirs" title="Permalink to this definition">¶</a></dt>
+<dd><p>Print the paths used for finding libraries and programs.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-save-temps">
+<tt class="descname">-save-temps</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-save-temps" title="Permalink to this definition">¶</a></dt>
+<dd><p>Save intermediate compilation results.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-integrated-as">
+<span id="cmdoption-no-integrated-as"></span><tt class="descname">-integrated-as</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-no-integrated-as</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-integrated-as" title="Permalink to this definition">¶</a></dt>
+<dd><p>Used to enable and disable, respectively, the use of the integrated
+assembler. Whether the integrated assembler is on by default is target
+dependent.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-time">
+<tt class="descname">-time</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-time" title="Permalink to this definition">¶</a></dt>
+<dd><p>Time individual commands.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-ftime-report">
+<tt class="descname">-ftime-report</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-ftime-report" title="Permalink to this definition">¶</a></dt>
+<dd><p>Print timing summary of each stage of compilation.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-v">
+<tt class="descname">-v</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-v" title="Permalink to this definition">¶</a></dt>
+<dd><p>Show commands to run and use verbose output.</p>
+</dd></dl>
+
+</div>
+<div class="section" id="diagnostics-options">
+<h3>Diagnostics Options<a class="headerlink" href="#diagnostics-options" title="Permalink to this headline">¶</a></h3>
+<dl class="option">
+<dt id="cmdoption-fshow-column">
+<span id="cmdoption-fshow-source-location"></span><span id="cmdoption-fcaret-diagnostics"></span><span id="cmdoption-fdiagnostics-fixit-info"></span><span id="cmdoption-fdiagnostics-parseable-fixits"></span><span id="cmdoption-fdiagnostics-print-source-range-info"></span><span id="cmdoption-fprint-source-range-info"></span><span id="cmdoption-fdiagnostics-show-option"></span><span id="cmdoption-fmessage-length"></span><tt class="descname">-fshow-column</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-fshow-source-location</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-fcaret-diagnostics</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-fdiagnostics-fixit-info</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-fdiagnostics-parseable-fixits</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descna
me">-fdiagnostics-print-source-range-info</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-fprint-source-range-info</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-fdiagnostics-show-option</tt><tt class="descclassname"></tt><tt class="descclassname">, </tt><tt class="descname">-fmessage-length</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-fshow-column" title="Permalink to this definition">¶</a></dt>
+<dd><p>These options control how Clang prints out information about diagnostics
+(errors and warnings). Please see the Clang User’s Manual for more information.</p>
+</dd></dl>
+
+</div>
+<div class="section" id="preprocessor-options">
+<h3>Preprocessor Options<a class="headerlink" href="#preprocessor-options" title="Permalink to this headline">¶</a></h3>
+<dl class="option">
+<dt id="cmdoption-D">
+<tt class="descname">-D</tt><tt class="descclassname"><macroname>=<value></tt><a class="headerlink" href="#cmdoption-D" title="Permalink to this definition">¶</a></dt>
+<dd><p>Adds an implicit #define into the predefines buffer which is read before the
+source file is preprocessed.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-U">
+<tt class="descname">-U</tt><tt class="descclassname"><macroname></tt><a class="headerlink" href="#cmdoption-U" title="Permalink to this definition">¶</a></dt>
+<dd><p>Adds an implicit #undef into the predefines buffer which is read before the
+source file is preprocessed.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-include">
+<tt class="descname">-include</tt><tt class="descclassname"> <filename></tt><a class="headerlink" href="#cmdoption-include" title="Permalink to this definition">¶</a></dt>
+<dd><p>Adds an implicit #include into the predefines buffer which is read before the
+source file is preprocessed.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-I">
+<tt class="descname">-I</tt><tt class="descclassname"><directory></tt><a class="headerlink" href="#cmdoption-I" title="Permalink to this definition">¶</a></dt>
+<dd><p>Add the specified directory to the search path for include files.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-F">
+<tt class="descname">-F</tt><tt class="descclassname"><directory></tt><a class="headerlink" href="#cmdoption-F" title="Permalink to this definition">¶</a></dt>
+<dd><p>Add the specified directory to the search path for framework include files.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-nostdinc">
+<tt class="descname">-nostdinc</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-nostdinc" title="Permalink to this definition">¶</a></dt>
+<dd><p>Do not search the standard system directories or compiler builtin directories
+for include files.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-nostdlibinc">
+<tt class="descname">-nostdlibinc</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-nostdlibinc" title="Permalink to this definition">¶</a></dt>
+<dd><p>Do not search the standard system directories for include files, but do
+search compiler builtin include directories.</p>
+</dd></dl>
+
+<dl class="option">
+<dt id="cmdoption-nobuiltininc">
+<tt class="descname">-nobuiltininc</tt><tt class="descclassname"></tt><a class="headerlink" href="#cmdoption-nobuiltininc" title="Permalink to this definition">¶</a></dt>
+<dd><p>Do not search clang’s builtin directory for include files.</p>
+</dd></dl>
+
+</div>
+</div>
+<div class="section" id="environment">
+<h2>ENVIRONMENT<a class="headerlink" href="#environment" title="Permalink to this headline">¶</a></h2>
+<dl class="envvar">
+<dt id="envvar-TMPDIR,TEMP,TMP">
+<tt class="descname">TMPDIR, TEMP, TMP</tt><a class="headerlink" href="#envvar-TMPDIR,TEMP,TMP" title="Permalink to this definition">¶</a></dt>
+<dd><p>These environment variables are checked, in order, for the location to write
+temporary files used during the compilation process.</p>
+</dd></dl>
+
+<dl class="envvar">
+<dt id="envvar-CPATH">
+<tt class="descname">CPATH</tt><a class="headerlink" href="#envvar-CPATH" title="Permalink to this definition">¶</a></dt>
+<dd><p>If this environment variable is present, it is treated as a delimited list of
+paths to be added to the default system include path list. The delimiter is
+the platform dependent delimiter, as used in the PATH environment variable.</p>
+<p>Empty components in the environment variable are ignored.</p>
+</dd></dl>
+
+<dl class="envvar">
+<dt id="envvar-C_INCLUDE_PATH,OBJC_INCLUDE_PATH,CPLUS_INCLUDE_PATH,OBJCPLUS_INCLUDE_PATH">
+<tt class="descname">C_INCLUDE_PATH, OBJC_INCLUDE_PATH, CPLUS_INCLUDE_PATH, OBJCPLUS_INCLUDE_PATH</tt><a class="headerlink" href="#envvar-C_INCLUDE_PATH,OBJC_INCLUDE_PATH,CPLUS_INCLUDE_PATH,OBJCPLUS_INCLUDE_PATH" title="Permalink to this definition">¶</a></dt>
+<dd><p>These environment variables specify additional paths, as for <span class="target" id="index-0"></span><a class="reference internal" href="#envvar-CPATH"><tt class="xref std std-envvar docutils literal"><span class="pre">CPATH</span></tt></a>, which are
+only used when processing the appropriate language.</p>
+</dd></dl>
+
+<dl class="envvar">
+<dt id="envvar-MACOSX_DEPLOYMENT_TARGET">
+<tt class="descname">MACOSX_DEPLOYMENT_TARGET</tt><a class="headerlink" href="#envvar-MACOSX_DEPLOYMENT_TARGET" title="Permalink to this definition">¶</a></dt>
+<dd><p>If <a class="reference internal" href="#cmdoption-mmacosx-version-min"><em class="xref std std-option">-mmacosx-version-min</em></a> is unspecified, the default deployment
+target is read from this environment variable. This option only affects
+Darwin targets.</p>
+</dd></dl>
+
+</div>
+<div class="section" id="bugs">
+<h2>BUGS<a class="headerlink" href="#bugs" title="Permalink to this headline">¶</a></h2>
+<p>To report bugs, please visit <<a class="reference external" href="http://llvm.org/bugs/">http://llvm.org/bugs/</a>>. Most bug reports should
+include preprocessed source files (use the <a class="reference internal" href="#cmdoption-E"><em class="xref std std-option">-E</em></a> option) and the full
+output of the compiler, along with information to reproduce.</p>
+</div>
+<div class="section" id="see-also">
+<h2>SEE ALSO<a class="headerlink" href="#see-also" title="Permalink to this headline">¶</a></h2>
+<p><em class="manpage">as(1)</em>, <em class="manpage">ld(1)</em></p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="index.html">Clang “man” pages</a>
+ ::
+ <a class="uplink" href="../index.html">Contents</a>
+ ::
+ <a href="../FAQ.html">Frequently Asked Questions (FAQ)</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/CommandGuide/index.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/CommandGuide/index.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/CommandGuide/index.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/CommandGuide/index.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,87 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Clang âmanâ pages — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="../_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: '../',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="../_static/jquery.js"></script>
+ <script type="text/javascript" src="../_static/underscore.js"></script>
+ <script type="text/javascript" src="../_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="../index.html" />
+ <link rel="next" title="clang - the Clang C, C++, and Objective-C compiler" href="clang.html" />
+ <link rel="prev" title="MSVC compatibility" href="../MSVCCompatibility.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="../index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Clang âmanâ pages</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="../MSVCCompatibility.html">MSVC compatibility</a>
+ ::
+ <a class="uplink" href="../index.html">Contents</a>
+ ::
+ <a href="clang.html">clang - the Clang C, C++, and Objective-C compiler</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="clang-man-pages">
+<h1>Clang “man” pages<a class="headerlink" href="#clang-man-pages" title="Permalink to this headline">¶</a></h1>
+<p>The following documents are command descriptions for all of the Clang tools.
+These pages describe how to use the Clang commands and what their options are.
+Note that these pages do not describe all of the options available for all
+tools. To get a complete listing, pass the <tt class="docutils literal"><span class="pre">--help</span></tt> (general options) or
+<tt class="docutils literal"><span class="pre">--help-hidden</span></tt> (general and debugging options) arguments to the tool you are
+interested in.</p>
+<div class="section" id="basic-commands">
+<h2>Basic Commands<a class="headerlink" href="#basic-commands" title="Permalink to this headline">¶</a></h2>
+<div class="toctree-wrapper compound">
+<ul>
+<li class="toctree-l1"><a class="reference internal" href="clang.html">clang - the Clang C, C++, and Objective-C compiler</a></li>
+</ul>
+</div>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="../MSVCCompatibility.html">MSVC compatibility</a>
+ ::
+ <a class="uplink" href="../index.html">Contents</a>
+ ::
+ <a href="clang.html">clang - the Clang C, C++, and Objective-C compiler</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/ControlFlowIntegrity.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/ControlFlowIntegrity.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/ControlFlowIntegrity.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/ControlFlowIntegrity.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,315 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Control Flow Integrity — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Control Flow Integrity Design Documentation" href="ControlFlowIntegrityDesign.html" />
+ <link rel="prev" title="Sanitizer special case list" href="SanitizerSpecialCaseList.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Control Flow Integrity</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="SanitizerSpecialCaseList.html">Sanitizer special case list</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ControlFlowIntegrityDesign.html">Control Flow Integrity Design Documentation</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="control-flow-integrity">
+<h1>Control Flow Integrity<a class="headerlink" href="#control-flow-integrity" title="Permalink to this headline">¶</a></h1>
+<div class="toctree-wrapper compound">
+</div>
+<div class="contents local topic" id="contents">
+<ul class="simple">
+<li><a class="reference internal" href="#introduction" id="id1">Introduction</a></li>
+<li><a class="reference internal" href="#available-schemes" id="id2">Available schemes</a></li>
+<li><a class="reference internal" href="#trapping-and-diagnostics" id="id3">Trapping and Diagnostics</a></li>
+<li><a class="reference internal" href="#forward-edge-cfi-for-virtual-calls" id="id4">Forward-Edge CFI for Virtual Calls</a><ul>
+<li><a class="reference internal" href="#performance" id="id5">Performance</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#bad-cast-checking" id="id6">Bad Cast Checking</a></li>
+<li><a class="reference internal" href="#non-virtual-member-function-call-checking" id="id7">Non-Virtual Member Function Call Checking</a><ul>
+<li><a class="reference internal" href="#strictness" id="id8">Strictness</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#indirect-function-call-checking" id="id9">Indirect Function Call Checking</a><ul>
+<li><a class="reference internal" href="#fsanitize-cfi-icall-and-fsanitize-function" id="id10"><tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt> and <tt class="docutils literal"><span class="pre">-fsanitize=function</span></tt></a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#blacklist" id="id11">Blacklist</a></li>
+<li><a class="reference internal" href="#shared-library-support" id="id12">Shared library support</a></li>
+<li><a class="reference internal" href="#design" id="id13">Design</a></li>
+<li><a class="reference internal" href="#publications" id="id14">Publications</a></li>
+</ul>
+</div>
+<div class="section" id="introduction">
+<h2><a class="toc-backref" href="#id1">Introduction</a><a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>Clang includes an implementation of a number of control flow integrity (CFI)
+schemes, which are designed to abort the program upon detecting certain forms
+of undefined behavior that can potentially allow attackers to subvert the
+program’s control flow. These schemes have been optimized for performance,
+allowing developers to enable them in release builds.</p>
+<p>To enable Clang’s available CFI schemes, use the flag <tt class="docutils literal"><span class="pre">-fsanitize=cfi</span></tt>.
+You can also enable a subset of available <a class="reference internal" href="#cfi-schemes"><em>schemes</em></a>.
+As currently implemented, all schemes rely on link-time optimization (LTO);
+so it is required to specify <tt class="docutils literal"><span class="pre">-flto</span></tt>, and the linker used must support LTO,
+for example via the <a class="reference external" href="http://llvm.org/docs/GoldPlugin.html">gold plugin</a>.</p>
+<p>To allow the checks to be implemented efficiently, the program must be
+structured such that certain object files are compiled with CFI
+enabled, and are statically linked into the program. This may preclude
+the use of shared libraries in some cases. Experimental support for
+<a class="reference internal" href="#cfi-cross-dso"><em>cross-DSO control flow integrity</em></a> exists that
+does not have these requirements. This cross-DSO support has unstable
+ABI at this time.</p>
+</div>
+<div class="section" id="available-schemes">
+<span id="cfi-schemes"></span><h2><a class="toc-backref" href="#id2">Available schemes</a><a class="headerlink" href="#available-schemes" title="Permalink to this headline">¶</a></h2>
+<p>Available schemes are:</p>
+<blockquote>
+<div><ul class="simple">
+<li><tt class="docutils literal"><span class="pre">-fsanitize=cfi-cast-strict</span></tt>: Enables <a class="reference internal" href="#cfi-strictness"><em>strict cast checks</em></a>.</li>
+<li><tt class="docutils literal"><span class="pre">-fsanitize=cfi-derived-cast</span></tt>: Base-to-derived cast to the wrong
+dynamic type.</li>
+<li><tt class="docutils literal"><span class="pre">-fsanitize=cfi-unrelated-cast</span></tt>: Cast from <tt class="docutils literal"><span class="pre">void*</span></tt> or another
+unrelated type to the wrong dynamic type.</li>
+<li><tt class="docutils literal"><span class="pre">-fsanitize=cfi-nvcall</span></tt>: Non-virtual call via an object whose vptr is of
+the wrong dynamic type.</li>
+<li><tt class="docutils literal"><span class="pre">-fsanitize=cfi-vcall</span></tt>: Virtual call via an object whose vptr is of the
+wrong dynamic type.</li>
+<li><tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt>: Indirect call of a function with wrong dynamic
+type.</li>
+</ul>
+</div></blockquote>
+<p>You can use <tt class="docutils literal"><span class="pre">-fsanitize=cfi</span></tt> to enable all the schemes and use
+<tt class="docutils literal"><span class="pre">-fno-sanitize</span></tt> flag to narrow down the set of schemes as desired.
+For example, you can build your program with
+<tt class="docutils literal"><span class="pre">-fsanitize=cfi</span> <span class="pre">-fno-sanitize=cfi-nvcall,cfi-icall</span></tt>
+to use all schemes except for non-virtual member function call and indirect call
+checking.</p>
+<p>Remember that you have to provide <tt class="docutils literal"><span class="pre">-flto</span></tt> if at least one CFI scheme is
+enabled.</p>
+</div>
+<div class="section" id="trapping-and-diagnostics">
+<h2><a class="toc-backref" href="#id3">Trapping and Diagnostics</a><a class="headerlink" href="#trapping-and-diagnostics" title="Permalink to this headline">¶</a></h2>
+<p>By default, CFI will abort the program immediately upon detecting a control
+flow integrity violation. You can use the <a class="reference internal" href="UsersManual.html#controlling-code-generation"><em>-fno-sanitize-trap=</em></a> flag to cause CFI to print a diagnostic
+similar to the one below before the program aborts.</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="go">bad-cast.cpp:109:7: runtime error: control flow integrity check for type 'B' failed during base-to-derived cast (vtable address 0x000000425a50)</span>
+<span class="go">0x000000425a50: note: vtable is of type 'A'</span>
+<span class="go"> 00 00 00 00 f0 f1 41 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 20 5a 42 00</span>
+<span class="go"> ^</span>
+</pre></div>
+</div>
+<p>If diagnostics are enabled, you can also configure CFI to continue program
+execution instead of aborting by using the <a class="reference internal" href="UsersManual.html#controlling-code-generation"><em>-fsanitize-recover=</em></a> flag.</p>
+</div>
+<div class="section" id="forward-edge-cfi-for-virtual-calls">
+<h2><a class="toc-backref" href="#id4">Forward-Edge CFI for Virtual Calls</a><a class="headerlink" href="#forward-edge-cfi-for-virtual-calls" title="Permalink to this headline">¶</a></h2>
+<p>This scheme checks that virtual calls take place using a vptr of the correct
+dynamic type; that is, the dynamic type of the called object must be a
+derived class of the static type of the object used to make the call.
+This CFI scheme can be enabled on its own using <tt class="docutils literal"><span class="pre">-fsanitize=cfi-vcall</span></tt>.</p>
+<p>For this scheme to work, all translation units containing the definition
+of a virtual member function (whether inline or not), other than members
+of <a class="reference internal" href="#cfi-blacklist"><em>blacklisted</em></a> types, must be compiled with
+<tt class="docutils literal"><span class="pre">-fsanitize=cfi-vcall</span></tt> enabled and be statically linked into the program.</p>
+<div class="section" id="performance">
+<h3><a class="toc-backref" href="#id5">Performance</a><a class="headerlink" href="#performance" title="Permalink to this headline">¶</a></h3>
+<p>A performance overhead of less than 1% has been measured by running the
+Dromaeo benchmark suite against an instrumented version of the Chromium
+web browser. Another good performance benchmark for this mechanism is the
+virtual-call-heavy SPEC 2006 xalancbmk.</p>
+<p>Note that this scheme has not yet been optimized for binary size; an increase
+of up to 15% has been observed for Chromium.</p>
+</div>
+</div>
+<div class="section" id="bad-cast-checking">
+<h2><a class="toc-backref" href="#id6">Bad Cast Checking</a><a class="headerlink" href="#bad-cast-checking" title="Permalink to this headline">¶</a></h2>
+<p>This scheme checks that pointer casts are made to an object of the correct
+dynamic type; that is, the dynamic type of the object must be a derived class
+of the pointee type of the cast. The checks are currently only introduced
+where the class being casted to is a polymorphic class.</p>
+<p>Bad casts are not in themselves control flow integrity violations, but they
+can also create security vulnerabilities, and the implementation uses many
+of the same mechanisms.</p>
+<p>There are two types of bad cast that may be forbidden: bad casts
+from a base class to a derived class (which can be checked with
+<tt class="docutils literal"><span class="pre">-fsanitize=cfi-derived-cast</span></tt>), and bad casts from a pointer of
+type <tt class="docutils literal"><span class="pre">void*</span></tt> or another unrelated type (which can be checked with
+<tt class="docutils literal"><span class="pre">-fsanitize=cfi-unrelated-cast</span></tt>).</p>
+<p>The difference between these two types of casts is that the first is defined
+by the C++ standard to produce an undefined value, while the second is not
+in itself undefined behavior (it is well defined to cast the pointer back
+to its original type).</p>
+<p>If a program as a matter of policy forbids the second type of cast, that
+restriction can normally be enforced. However it may in some cases be necessary
+for a function to perform a forbidden cast to conform with an external API
+(e.g. the <tt class="docutils literal"><span class="pre">allocate</span></tt> member function of a standard library allocator). Such
+functions may be <a class="reference internal" href="#cfi-blacklist"><em>blacklisted</em></a>.</p>
+<p>For this scheme to work, all translation units containing the definition
+of a virtual member function (whether inline or not), other than members
+of <a class="reference internal" href="#cfi-blacklist"><em>blacklisted</em></a> types, must be compiled with
+<tt class="docutils literal"><span class="pre">-fsanitize=cfi-derived-cast</span></tt> or <tt class="docutils literal"><span class="pre">-fsanitize=cfi-unrelated-cast</span></tt> enabled
+and be statically linked into the program.</p>
+</div>
+<div class="section" id="non-virtual-member-function-call-checking">
+<h2><a class="toc-backref" href="#id7">Non-Virtual Member Function Call Checking</a><a class="headerlink" href="#non-virtual-member-function-call-checking" title="Permalink to this headline">¶</a></h2>
+<p>This scheme checks that non-virtual calls take place using an object of
+the correct dynamic type; that is, the dynamic type of the called object
+must be a derived class of the static type of the object used to make the
+call. The checks are currently only introduced where the object is of a
+polymorphic class type. This CFI scheme can be enabled on its own using
+<tt class="docutils literal"><span class="pre">-fsanitize=cfi-nvcall</span></tt>.</p>
+<p>For this scheme to work, all translation units containing the definition
+of a virtual member function (whether inline or not), other than members
+of <a class="reference internal" href="#cfi-blacklist"><em>blacklisted</em></a> types, must be compiled with
+<tt class="docutils literal"><span class="pre">-fsanitize=cfi-nvcall</span></tt> enabled and be statically linked into the program.</p>
+<div class="section" id="strictness">
+<span id="cfi-strictness"></span><h3><a class="toc-backref" href="#id8">Strictness</a><a class="headerlink" href="#strictness" title="Permalink to this headline">¶</a></h3>
+<p>If a class has a single non-virtual base and does not introduce or override
+virtual member functions or fields other than an implicitly defined virtual
+destructor, it will have the same layout and virtual function semantics as
+its base. By default, casts to such classes are checked as if they were made
+to the least derived such class.</p>
+<p>Casting an instance of a base class to such a derived class is technically
+undefined behavior, but it is a relatively common hack for introducing
+member functions on class instances with specific properties that works under
+most compilers and should not have security implications, so we allow it by
+default. It can be disabled with <tt class="docutils literal"><span class="pre">-fsanitize=cfi-cast-strict</span></tt>.</p>
+</div>
+</div>
+<div class="section" id="indirect-function-call-checking">
+<h2><a class="toc-backref" href="#id9">Indirect Function Call Checking</a><a class="headerlink" href="#indirect-function-call-checking" title="Permalink to this headline">¶</a></h2>
+<p>This scheme checks that function calls take place using a function of the
+correct dynamic type; that is, the dynamic type of the function must match
+the static type used at the call. This CFI scheme can be enabled on its own
+using <tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt>.</p>
+<p>For this scheme to work, each indirect function call in the program, other
+than calls in <a class="reference internal" href="#cfi-blacklist"><em>blacklisted</em></a> functions, must call a
+function which was either compiled with <tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt> enabled,
+or whose address was taken by a function in a translation unit compiled with
+<tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt>.</p>
+<p>If a function in a translation unit compiled with <tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt>
+takes the address of a function not compiled with <tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt>,
+that address may differ from the address taken by a function in a translation
+unit not compiled with <tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt>. This is technically a
+violation of the C and C++ standards, but it should not affect most programs.</p>
+<p>Each translation unit compiled with <tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt> must be
+statically linked into the program or shared library, and calls across
+shared library boundaries are handled as if the callee was not compiled with
+<tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt>.</p>
+<p>This scheme is currently only supported on the x86 and x86_64 architectures.</p>
+<div class="section" id="fsanitize-cfi-icall-and-fsanitize-function">
+<h3><a class="toc-backref" href="#id10"><tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt> and <tt class="docutils literal"><span class="pre">-fsanitize=function</span></tt></a><a class="headerlink" href="#fsanitize-cfi-icall-and-fsanitize-function" title="Permalink to this headline">¶</a></h3>
+<p>This tool is similar to <tt class="docutils literal"><span class="pre">-fsanitize=function</span></tt> in that both tools check
+the types of function calls. However, the two tools occupy different points
+on the design space; <tt class="docutils literal"><span class="pre">-fsanitize=function</span></tt> is a developer tool designed
+to find bugs in local development builds, whereas <tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt>
+is a security hardening mechanism designed to be deployed in release builds.</p>
+<p><tt class="docutils literal"><span class="pre">-fsanitize=function</span></tt> has a higher space and time overhead due to a more
+complex type check at indirect call sites, as well as a need for run-time
+type information (RTTI), which may make it unsuitable for deployment. Because
+of the need for RTTI, <tt class="docutils literal"><span class="pre">-fsanitize=function</span></tt> can only be used with C++
+programs, whereas <tt class="docutils literal"><span class="pre">-fsanitize=cfi-icall</span></tt> can protect both C and C++ programs.</p>
+<p>On the other hand, <tt class="docutils literal"><span class="pre">-fsanitize=function</span></tt> conforms more closely with the C++
+standard and user expectations around interaction with shared libraries;
+the identity of function pointers is maintained, and calls across shared
+library boundaries are no different from calls within a single program or
+shared library.</p>
+</div>
+</div>
+<div class="section" id="blacklist">
+<span id="cfi-blacklist"></span><h2><a class="toc-backref" href="#id11">Blacklist</a><a class="headerlink" href="#blacklist" title="Permalink to this headline">¶</a></h2>
+<p>A <a class="reference internal" href="SanitizerSpecialCaseList.html"><em>Sanitizer special case list</em></a> can be used to relax CFI checks for certain
+source files, functions and types using the <tt class="docutils literal"><span class="pre">src</span></tt>, <tt class="docutils literal"><span class="pre">fun</span></tt> and <tt class="docutils literal"><span class="pre">type</span></tt>
+entity types.</p>
+<p>In addition, if a type has a <tt class="docutils literal"><span class="pre">uuid</span></tt> attribute and the blacklist contains
+the type entry <tt class="docutils literal"><span class="pre">attr:uuid</span></tt>, CFI checks are suppressed for that type. This
+allows all COM types to be easily blacklisted, which is useful as COM types
+are typically defined outside of the linked program.</p>
+<div class="highlight-bash"><div class="highlight"><pre><span class="c"># Suppress checking for code in a file.</span>
+src:bad_file.cpp
+src:bad_header.h
+<span class="c"># Ignore all functions with names containing MyFooBar.</span>
+fun:*MyFooBar*
+<span class="c"># Ignore all types in the standard library.</span>
+<span class="nb">type</span>:std::*
+<span class="c"># Ignore all types with a uuid attribute.</span>
+<span class="nb">type</span>:attr:uuid
+</pre></div>
+</div>
+</div>
+<div class="section" id="shared-library-support">
+<span id="cfi-cross-dso"></span><h2><a class="toc-backref" href="#id12">Shared library support</a><a class="headerlink" href="#shared-library-support" title="Permalink to this headline">¶</a></h2>
+<p>Use <strong>-f[no-]sanitize-cfi-cross-dso</strong> to enable the cross-DSO control
+flow integrity mode, which allows all CFI schemes listed above to
+apply across DSO boundaries. As in the regular CFI, each DSO must be
+built with <tt class="docutils literal"><span class="pre">-flto</span></tt>.</p>
+</div>
+<div class="section" id="design">
+<h2><a class="toc-backref" href="#id13">Design</a><a class="headerlink" href="#design" title="Permalink to this headline">¶</a></h2>
+<p>Please refer to the <a class="reference internal" href="ControlFlowIntegrityDesign.html"><em>design document</em></a>.</p>
+</div>
+<div class="section" id="publications">
+<h2><a class="toc-backref" href="#id14">Publications</a><a class="headerlink" href="#publications" title="Permalink to this headline">¶</a></h2>
+<p><a class="reference external" href="http://research.microsoft.com/pubs/64250/ccs05.pdf">Control-Flow Integrity: Principles, Implementations, and Applications</a>.
+Martin Abadi, Mihai Budiu, Ãlfar Erlingsson, Jay Ligatti.</p>
+<p><a class="reference external" href="http://www.pcc.me.uk/~peter/acad/usenix14.pdf">Enforcing Forward-Edge Control-Flow Integrity in GCC & LLVM</a>.
+Caroline Tice, Tom Roeder, Peter Collingbourne, Stephen Checkoway,
+Ãlfar Erlingsson, Luis Lozano, Geoff Pike.</p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="SanitizerSpecialCaseList.html">Sanitizer special case list</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ControlFlowIntegrityDesign.html">Control Flow Integrity Design Documentation</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/ControlFlowIntegrityDesign.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/ControlFlowIntegrityDesign.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/ControlFlowIntegrityDesign.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/ControlFlowIntegrityDesign.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,819 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Control Flow Integrity Design Documentation — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="up" title="Control Flow Integrity" href="ControlFlowIntegrity.html" />
+ <link rel="next" title="SafeStack" href="SafeStack.html" />
+ <link rel="prev" title="Control Flow Integrity" href="ControlFlowIntegrity.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Control Flow Integrity Design Documentation</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="ControlFlowIntegrity.html">Control Flow Integrity</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="SafeStack.html">SafeStack</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="control-flow-integrity-design-documentation">
+<h1>Control Flow Integrity Design Documentation<a class="headerlink" href="#control-flow-integrity-design-documentation" title="Permalink to this headline">¶</a></h1>
+<p>This page documents the design of the <a class="reference internal" href="ControlFlowIntegrity.html"><em>Control Flow Integrity</em></a> schemes
+supported by Clang.</p>
+<div class="section" id="forward-edge-cfi-for-virtual-calls">
+<h2>Forward-Edge CFI for Virtual Calls<a class="headerlink" href="#forward-edge-cfi-for-virtual-calls" title="Permalink to this headline">¶</a></h2>
+<p>This scheme works by allocating, for each static type used to make a virtual
+call, a region of read-only storage in the object file holding a bit vector
+that maps onto to the region of storage used for those virtual tables. Each
+set bit in the bit vector corresponds to the <a class="reference external" href="https://mentorembedded.github.io/cxx-abi/abi.html#vtable-general">address point</a> for a virtual
+table compatible with the static type for which the bit vector is being built.</p>
+<p>For example, consider the following three C++ classes:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">struct</span> <span class="n">A</span> <span class="p">{</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="n">f1</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f2</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f3</span><span class="p">();</span>
+<span class="p">};</span>
+
+<span class="k">struct</span> <span class="n">B</span> <span class="o">:</span> <span class="n">A</span> <span class="p">{</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="n">f1</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f2</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f3</span><span class="p">();</span>
+<span class="p">};</span>
+
+<span class="k">struct</span> <span class="n">C</span> <span class="o">:</span> <span class="n">A</span> <span class="p">{</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="n">f1</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f2</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f3</span><span class="p">();</span>
+<span class="p">};</span>
+</pre></div>
+</div>
+<p>The scheme will cause the virtual tables for A, B and C to be laid out
+consecutively:</p>
+<table border="1" class="docutils">
+<caption>Virtual Table Layout for A, B, C</caption>
+<colgroup>
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+<col width="7%" />
+</colgroup>
+<thead valign="bottom">
+<tr class="row-odd"><th class="head">0</th>
+<th class="head">1</th>
+<th class="head">2</th>
+<th class="head">3</th>
+<th class="head">4</th>
+<th class="head">5</th>
+<th class="head">6</th>
+<th class="head">7</th>
+<th class="head">8</th>
+<th class="head">9</th>
+<th class="head">10</th>
+<th class="head">11</th>
+<th class="head">12</th>
+<th class="head">13</th>
+<th class="head">14</th>
+</tr>
+</thead>
+<tbody valign="top">
+<tr class="row-even"><td>A::offset-to-top</td>
+<td>&A::rtti</td>
+<td>&A::f1</td>
+<td>&A::f2</td>
+<td>&A::f3</td>
+<td>B::offset-to-top</td>
+<td>&B::rtti</td>
+<td>&B::f1</td>
+<td>&B::f2</td>
+<td>&B::f3</td>
+<td>C::offset-to-top</td>
+<td>&C::rtti</td>
+<td>&C::f1</td>
+<td>&C::f2</td>
+<td>&C::f3</td>
+</tr>
+</tbody>
+</table>
+<p>The bit vector for static types A, B and C will look like this:</p>
+<table border="1" class="docutils">
+<caption>Bit Vectors for A, B, C</caption>
+<colgroup>
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+</colgroup>
+<thead valign="bottom">
+<tr class="row-odd"><th class="head">Class</th>
+<th class="head">0</th>
+<th class="head">1</th>
+<th class="head">2</th>
+<th class="head">3</th>
+<th class="head">4</th>
+<th class="head">5</th>
+<th class="head">6</th>
+<th class="head">7</th>
+<th class="head">8</th>
+<th class="head">9</th>
+<th class="head">10</th>
+<th class="head">11</th>
+<th class="head">12</th>
+<th class="head">13</th>
+<th class="head">14</th>
+</tr>
+</thead>
+<tbody valign="top">
+<tr class="row-even"><td>A</td>
+<td>0</td>
+<td>0</td>
+<td>1</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>1</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>1</td>
+<td>0</td>
+<td>0</td>
+</tr>
+<tr class="row-odd"><td>B</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>1</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+</tr>
+<tr class="row-even"><td>C</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>1</td>
+<td>0</td>
+<td>0</td>
+</tr>
+</tbody>
+</table>
+<p>Bit vectors are represented in the object file as byte arrays. By loading
+from indexed offsets into the byte array and applying a mask, a program can
+test bits from the bit set with a relatively short instruction sequence. Bit
+vectors may overlap so long as they use different bits. For the full details,
+see the <a class="reference external" href="http://llvm.org/docs/doxygen/html/structllvm_1_1ByteArrayBuilder.html">ByteArrayBuilder</a> class.</p>
+<p>In this case, assuming A is laid out at offset 0 in bit 0, B at offset 0 in
+bit 1 and C at offset 0 in bit 2, the byte array would look like this:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">char</span> <span class="n">bits</span><span class="p">[]</span> <span class="o">=</span> <span class="p">{</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">5</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span> <span class="p">};</span>
+</pre></div>
+</div>
+<p>To emit a virtual call, the compiler will assemble code that checks that
+the object’s virtual table pointer is in-bounds and aligned and that the
+relevant bit is set in the bit vector.</p>
+<p>For example on x86 a typical virtual call may look like this:</p>
+<div class="highlight-none"><div class="highlight"><pre>ca7fbb: 48 8b 0f mov (%rdi),%rcx
+ca7fbe: 48 8d 15 c3 42 fb 07 lea 0x7fb42c3(%rip),%rdx
+ca7fc5: 48 89 c8 mov %rcx,%rax
+ca7fc8: 48 29 d0 sub %rdx,%rax
+ca7fcb: 48 c1 c0 3d rol $0x3d,%rax
+ca7fcf: 48 3d 7f 01 00 00 cmp $0x17f,%rax
+ca7fd5: 0f 87 36 05 00 00 ja ca8511
+ca7fdb: 48 8d 15 c0 0b f7 06 lea 0x6f70bc0(%rip),%rdx
+ca7fe2: f6 04 10 10 testb $0x10,(%rax,%rdx,1)
+ca7fe6: 0f 84 25 05 00 00 je ca8511
+ca7fec: ff 91 98 00 00 00 callq *0x98(%rcx)
+ [...]
+ca8511: 0f 0b ud2
+</pre></div>
+</div>
+<p>The compiler relies on co-operation from the linker in order to assemble
+the bit vectors for the whole program. It currently does this using LLVM’s
+<a class="reference external" href="http://llvm.org/docs/BitSets.html">bit sets</a> mechanism together with link-time optimization.</p>
+<div class="section" id="optimizations">
+<h3>Optimizations<a class="headerlink" href="#optimizations" title="Permalink to this headline">¶</a></h3>
+<p>The scheme as described above is the fully general variant of the scheme.
+Most of the time we are able to apply one or more of the following
+optimizations to improve binary size or performance.</p>
+<p>In fact, if you try the above example with the current version of the
+compiler, you will probably find that it will not use the described virtual
+table layout or machine instructions. Some of the optimizations we are about
+to introduce cause the compiler to use a different layout or a different
+sequence of machine instructions.</p>
+<div class="section" id="stripping-leading-trailing-zeros-in-bit-vectors">
+<h4>Stripping Leading/Trailing Zeros in Bit Vectors<a class="headerlink" href="#stripping-leading-trailing-zeros-in-bit-vectors" title="Permalink to this headline">¶</a></h4>
+<p>If a bit vector contains leading or trailing zeros, we can strip them from
+the vector. The compiler will emit code to check if the pointer is in range
+of the region covered by ones, and perform the bit vector check using a
+truncated version of the bit vector. For example, the bit vectors for our
+example class hierarchy will be emitted like this:</p>
+<table border="1" class="docutils">
+<caption>Bit Vectors for A, B, C</caption>
+<colgroup>
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+</colgroup>
+<thead valign="bottom">
+<tr class="row-odd"><th class="head">Class</th>
+<th class="head">0</th>
+<th class="head">1</th>
+<th class="head">2</th>
+<th class="head">3</th>
+<th class="head">4</th>
+<th class="head">5</th>
+<th class="head">6</th>
+<th class="head">7</th>
+<th class="head">8</th>
+<th class="head">9</th>
+<th class="head">10</th>
+<th class="head">11</th>
+<th class="head">12</th>
+<th class="head">13</th>
+<th class="head">14</th>
+</tr>
+</thead>
+<tbody valign="top">
+<tr class="row-even"><td>A</td>
+<td> </td>
+<td> </td>
+<td>1</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>1</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>0</td>
+<td>1</td>
+<td> </td>
+<td> </td>
+</tr>
+<tr class="row-odd"><td>B</td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td>1</td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+</tr>
+<tr class="row-even"><td>C</td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td> </td>
+<td>1</td>
+<td> </td>
+<td> </td>
+</tr>
+</tbody>
+</table>
+</div>
+<div class="section" id="short-inline-bit-vectors">
+<h4>Short Inline Bit Vectors<a class="headerlink" href="#short-inline-bit-vectors" title="Permalink to this headline">¶</a></h4>
+<p>If the vector is sufficiently short, we can represent it as an inline constant
+on x86. This saves us a few instructions when reading the correct element
+of the bit vector.</p>
+<p>If the bit vector fits in 32 bits, the code looks like this:</p>
+<div class="highlight-none"><div class="highlight"><pre> dc2: 48 8b 03 mov (%rbx),%rax
+ dc5: 48 8d 15 14 1e 00 00 lea 0x1e14(%rip),%rdx
+ dcc: 48 89 c1 mov %rax,%rcx
+ dcf: 48 29 d1 sub %rdx,%rcx
+ dd2: 48 c1 c1 3d rol $0x3d,%rcx
+ dd6: 48 83 f9 03 cmp $0x3,%rcx
+ dda: 77 2f ja e0b <main+0x9b>
+ ddc: ba 09 00 00 00 mov $0x9,%edx
+ de1: 0f a3 ca bt %ecx,%edx
+ de4: 73 25 jae e0b <main+0x9b>
+ de6: 48 89 df mov %rbx,%rdi
+ de9: ff 10 callq *(%rax)
+[...]
+ e0b: 0f 0b ud2
+</pre></div>
+</div>
+<p>Or if the bit vector fits in 64 bits:</p>
+<div class="highlight-none"><div class="highlight"><pre>11a6: 48 8b 03 mov (%rbx),%rax
+11a9: 48 8d 15 d0 28 00 00 lea 0x28d0(%rip),%rdx
+11b0: 48 89 c1 mov %rax,%rcx
+11b3: 48 29 d1 sub %rdx,%rcx
+11b6: 48 c1 c1 3d rol $0x3d,%rcx
+11ba: 48 83 f9 2a cmp $0x2a,%rcx
+11be: 77 35 ja 11f5 <main+0xb5>
+11c0: 48 ba 09 00 00 00 00 movabs $0x40000000009,%rdx
+11c7: 04 00 00
+11ca: 48 0f a3 ca bt %rcx,%rdx
+11ce: 73 25 jae 11f5 <main+0xb5>
+11d0: 48 89 df mov %rbx,%rdi
+11d3: ff 10 callq *(%rax)
+[...]
+11f5: 0f 0b ud2
+</pre></div>
+</div>
+<p>If the bit vector consists of a single bit, there is only one possible
+virtual table, and the check can consist of a single equality comparison:</p>
+<div class="highlight-none"><div class="highlight"><pre>9a2: 48 8b 03 mov (%rbx),%rax
+9a5: 48 8d 0d a4 13 00 00 lea 0x13a4(%rip),%rcx
+9ac: 48 39 c8 cmp %rcx,%rax
+9af: 75 25 jne 9d6 <main+0x86>
+9b1: 48 89 df mov %rbx,%rdi
+9b4: ff 10 callq *(%rax)
+[...]
+9d6: 0f 0b ud2
+</pre></div>
+</div>
+</div>
+<div class="section" id="virtual-table-layout">
+<h4>Virtual Table Layout<a class="headerlink" href="#virtual-table-layout" title="Permalink to this headline">¶</a></h4>
+<p>The compiler lays out classes of disjoint hierarchies in separate regions
+of the object file. At worst, bit vectors in disjoint hierarchies only
+need to cover their disjoint hierarchy. But the closer that classes in
+sub-hierarchies are laid out to each other, the smaller the bit vectors for
+those sub-hierarchies need to be (see “Stripping Leading/Trailing Zeros in Bit
+Vectors” above). The <a class="reference external" href="http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/Transforms/IPO/LowerBitSets.h?view=markup">GlobalLayoutBuilder</a> class is responsible for laying
+out the globals efficiently to minimize the sizes of the underlying bitsets.</p>
+</div>
+<div class="section" id="alignment">
+<h4>Alignment<a class="headerlink" href="#alignment" title="Permalink to this headline">¶</a></h4>
+<p>If all gaps between address points in a particular bit vector are multiples
+of powers of 2, the compiler can compress the bit vector by strengthening
+the alignment requirements of the virtual table pointer. For example, given
+this class hierarchy:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">struct</span> <span class="n">A</span> <span class="p">{</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="n">f1</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f2</span><span class="p">();</span>
+<span class="p">};</span>
+
+<span class="k">struct</span> <span class="n">B</span> <span class="o">:</span> <span class="n">A</span> <span class="p">{</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="n">f1</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f2</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f3</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f4</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f5</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f6</span><span class="p">();</span>
+<span class="p">};</span>
+
+<span class="k">struct</span> <span class="n">C</span> <span class="o">:</span> <span class="n">A</span> <span class="p">{</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="n">f1</span><span class="p">();</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">f2</span><span class="p">();</span>
+<span class="p">};</span>
+</pre></div>
+</div>
+<p>The virtual tables will be laid out like this:</p>
+<table border="1" class="docutils">
+<caption>Virtual Table Layout for A, B, C</caption>
+<colgroup>
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+<col width="6%" />
+</colgroup>
+<thead valign="bottom">
+<tr class="row-odd"><th class="head">0</th>
+<th class="head">1</th>
+<th class="head">2</th>
+<th class="head">3</th>
+<th class="head">4</th>
+<th class="head">5</th>
+<th class="head">6</th>
+<th class="head">7</th>
+<th class="head">8</th>
+<th class="head">9</th>
+<th class="head">10</th>
+<th class="head">11</th>
+<th class="head">12</th>
+<th class="head">13</th>
+<th class="head">14</th>
+<th class="head">15</th>
+</tr>
+</thead>
+<tbody valign="top">
+<tr class="row-even"><td>A::offset-to-top</td>
+<td>&A::rtti</td>
+<td>&A::f1</td>
+<td>&A::f2</td>
+<td>B::offset-to-top</td>
+<td>&B::rtti</td>
+<td>&B::f1</td>
+<td>&B::f2</td>
+<td>&B::f3</td>
+<td>&B::f4</td>
+<td>&B::f5</td>
+<td>&B::f6</td>
+<td>C::offset-to-top</td>
+<td>&C::rtti</td>
+<td>&C::f1</td>
+<td>&C::f2</td>
+</tr>
+</tbody>
+</table>
+<p>Notice that each address point for A is separated by 4 words. This lets us
+emit a compressed bit vector for A that looks like this:</p>
+<table border="1" class="docutils">
+<colgroup>
+<col width="25%" />
+<col width="25%" />
+<col width="25%" />
+<col width="25%" />
+</colgroup>
+<thead valign="bottom">
+<tr class="row-odd"><th class="head">2</th>
+<th class="head">6</th>
+<th class="head">10</th>
+<th class="head">14</th>
+</tr>
+</thead>
+<tbody valign="top">
+<tr class="row-even"><td>1</td>
+<td>1</td>
+<td>0</td>
+<td>1</td>
+</tr>
+</tbody>
+</table>
+<p>At call sites, the compiler will strengthen the alignment requirements by
+using a different rotate count. For example, on a 64-bit machine where the
+address points are 4-word aligned (as in A from our example), the <tt class="docutils literal"><span class="pre">rol</span></tt>
+instruction may look like this:</p>
+<div class="highlight-none"><div class="highlight"><pre>dd2: 48 c1 c1 3b rol $0x3b,%rcx
+</pre></div>
+</div>
+</div>
+<div class="section" id="padding-to-powers-of-2">
+<h4>Padding to Powers of 2<a class="headerlink" href="#padding-to-powers-of-2" title="Permalink to this headline">¶</a></h4>
+<p>Of course, this alignment scheme works best if the address points are
+in fact aligned correctly. To make this more likely to happen, we insert
+padding between virtual tables that in many cases aligns address points to
+a power of 2. Specifically, our padding aligns virtual tables to the next
+highest power of 2 bytes; because address points for specific base classes
+normally appear at fixed offsets within the virtual table, this normally
+has the effect of aligning the address points as well.</p>
+<p>This scheme introduces tradeoffs between decreased space overhead for
+instructions and bit vectors and increased overhead in the form of padding. We
+therefore limit the amount of padding so that we align to no more than 128
+bytes. This number was found experimentally to provide a good tradeoff.</p>
+</div>
+<div class="section" id="eliminating-bit-vector-checks-for-all-ones-bit-vectors">
+<h4>Eliminating Bit Vector Checks for All-Ones Bit Vectors<a class="headerlink" href="#eliminating-bit-vector-checks-for-all-ones-bit-vectors" title="Permalink to this headline">¶</a></h4>
+<p>If the bit vector is all ones, the bit vector check is redundant; we simply
+need to check that the address is in range and well aligned. This is more
+likely to occur if the virtual tables are padded.</p>
+</div>
+</div>
+</div>
+<div class="section" id="forward-edge-cfi-for-indirect-function-calls">
+<h2>Forward-Edge CFI for Indirect Function Calls<a class="headerlink" href="#forward-edge-cfi-for-indirect-function-calls" title="Permalink to this headline">¶</a></h2>
+<p>Under forward-edge CFI for indirect function calls, each unique function
+type has its own bit vector, and at each call site we need to check that the
+function pointer is a member of the function type’s bit vector. This scheme
+works in a similar way to forward-edge CFI for virtual calls, the distinction
+being that we need to build bit vectors of function entry points rather than
+of virtual tables.</p>
+<p>Unlike when re-arranging global variables, we cannot re-arrange functions
+in a particular order and base our calculations on the layout of the
+functions’ entry points, as we have no idea how large a particular function
+will end up being (the function sizes could even depend on how we arrange
+the functions). Instead, we build a jump table, which is a block of code
+consisting of one branch instruction for each of the functions in the bit
+set that branches to the target function, and redirect any taken function
+addresses to the corresponding jump table entry. In this way, the distance
+between function entry points is predictable and controllable. In the object
+file’s symbol table, the symbols for the target functions also refer to the
+jump table entries, so that addresses taken outside the module will pass
+any verification done inside the module.</p>
+<p>In more concrete terms, suppose we have three functions <tt class="docutils literal"><span class="pre">f</span></tt>, <tt class="docutils literal"><span class="pre">g</span></tt>, <tt class="docutils literal"><span class="pre">h</span></tt>
+which are members of a single bitset, and a function foo that returns their
+addresses:</p>
+<div class="highlight-none"><div class="highlight"><pre>f:
+mov 0, %eax
+ret
+
+g:
+mov 1, %eax
+ret
+
+h:
+mov 2, %eax
+ret
+
+foo:
+mov f, %eax
+mov g, %edx
+mov h, %ecx
+ret
+</pre></div>
+</div>
+<p>Our jump table will (conceptually) look like this:</p>
+<div class="highlight-none"><div class="highlight"><pre>f:
+jmp .Ltmp0 ; 5 bytes
+int3 ; 1 byte
+int3 ; 1 byte
+int3 ; 1 byte
+
+g:
+jmp .Ltmp1 ; 5 bytes
+int3 ; 1 byte
+int3 ; 1 byte
+int3 ; 1 byte
+
+h:
+jmp .Ltmp2 ; 5 bytes
+int3 ; 1 byte
+int3 ; 1 byte
+int3 ; 1 byte
+
+.Ltmp0:
+mov 0, %eax
+ret
+
+.Ltmp1:
+mov 1, %eax
+ret
+
+.Ltmp2:
+mov 2, %eax
+ret
+
+foo:
+mov f, %eax
+mov g, %edx
+mov h, %ecx
+ret
+</pre></div>
+</div>
+<p>Because the addresses of <tt class="docutils literal"><span class="pre">f</span></tt>, <tt class="docutils literal"><span class="pre">g</span></tt>, <tt class="docutils literal"><span class="pre">h</span></tt> are evenly spaced at a power of
+2, and function types do not overlap (unlike class types with base classes),
+we can normally apply the <a class="reference internal" href="#alignment">Alignment</a> and <a class="reference internal" href="#eliminating-bit-vector-checks-for-all-ones-bit-vectors">Eliminating Bit Vector Checks
+for All-Ones Bit Vectors</a> optimizations thus simplifying the check at each
+call site to a range and alignment check.</p>
+</div>
+<div class="section" id="shared-library-support">
+<h2>Shared library support<a class="headerlink" href="#shared-library-support" title="Permalink to this headline">¶</a></h2>
+<p><strong>EXPERIMENTAL</strong></p>
+<p>The basic CFI mode described above assumes that the application is a
+monolithic binary; at least that all possible virtual/indirect call
+targets and the entire class hierarchy are known at link time. The
+cross-DSO mode, enabled with <strong>-f[no-]sanitize-cfi-cross-dso</strong> relaxes
+this requirement by allowing virtual and indirect calls to cross the
+DSO boundary.</p>
+<p>Assuming the following setup: the binary consists of several
+instrumented and several uninstrumented DSOs. Some of them may be
+dlopen-ed/dlclose-d periodically, even frequently.</p>
+<blockquote>
+<div><ul>
+<li><p class="first">Calls made from uninstrumented DSOs are not checked and just work.</p>
+</li>
+<li><p class="first">Calls inside any instrumented DSO are fully protected.</p>
+</li>
+<li><dl class="first docutils">
+<dt>Calls between different instrumented DSOs are also protected, with</dt>
+<dd><p class="first last">a performance penalty (in addition to the monolithic CFI
+overhead).</p>
+</dd>
+</dl>
+</li>
+<li><dl class="first docutils">
+<dt>Calls from an instrumented DSO to an uninstrumented one are</dt>
+<dd><p class="first last">unchecked and just work, with performance penalty.</p>
+</dd>
+</dl>
+</li>
+<li><dl class="first docutils">
+<dt>Calls from an instrumented DSO outside of any known DSO are</dt>
+<dd><p class="first last">detected as CFI violations.</p>
+</dd>
+</dl>
+</li>
+</ul>
+</div></blockquote>
+<p>In the monolithic scheme a call site is instrumented as</p>
+<div class="highlight-none"><div class="highlight"><pre>if (!InlinedFastCheck(f))
+ abort();
+call *f
+</pre></div>
+</div>
+<p>In the cross-DSO scheme it becomes</p>
+<div class="highlight-none"><div class="highlight"><pre>if (!InlinedFastCheck(f))
+ __cfi_slowpath(CallSiteTypeId, f);
+call *f
+</pre></div>
+</div>
+<div class="section" id="callsitetypeid">
+<h3>CallSiteTypeId<a class="headerlink" href="#callsitetypeid" title="Permalink to this headline">¶</a></h3>
+<p><tt class="docutils literal"><span class="pre">CallSiteTypeId</span></tt> is a stable process-wide identifier of the
+call-site type. For a virtual call site, the type in question is the class
+type; for an indirect function call it is the function signature. The
+mapping from a type to an identifier is an ABI detail. In the current,
+experimental, implementation the identifier of type T is calculated as
+follows:</p>
+<blockquote>
+<div><ul class="simple">
+<li>Obtain the mangled name for “typeinfo name for T”.</li>
+<li>Calculate MD5 hash of the name as a string.</li>
+<li>Reinterpret the first 8 bytes of the hash as a little-endian
+64-bit integer.</li>
+</ul>
+</div></blockquote>
+<p>It is possible, but unlikely, that collisions in the
+<tt class="docutils literal"><span class="pre">CallSiteTypeId</span></tt> hashing will result in weaker CFI checks that would
+still be conservatively correct.</p>
+</div>
+<div class="section" id="cfi-check">
+<h3>CFI_Check<a class="headerlink" href="#cfi-check" title="Permalink to this headline">¶</a></h3>
+<p>In the general case, only the target DSO knows whether the call to
+function <tt class="docutils literal"><span class="pre">f</span></tt> with type <tt class="docutils literal"><span class="pre">CallSiteTypeId</span></tt> is valid or not. To
+export this information, every DSO implements</p>
+<div class="highlight-none"><div class="highlight"><pre>void __cfi_check(uint64 CallSiteTypeId, void *TargetAddr)
+</pre></div>
+</div>
+<p>This function provides external modules with access to CFI checks for
+the targets inside this DSO. For each known <tt class="docutils literal"><span class="pre">CallSiteTypeId</span></tt>, this
+functions performs an <tt class="docutils literal"><span class="pre">llvm.bitset.test</span></tt> with the corresponding bit
+set. It aborts if the type is unknown, or if the check fails.</p>
+<p>The basic implementation is a large switch statement over all values
+of CallSiteTypeId supported by this DSO, and each case is similar to
+the InlinedFastCheck() in the basic CFI mode.</p>
+</div>
+<div class="section" id="cfi-shadow">
+<h3>CFI Shadow<a class="headerlink" href="#cfi-shadow" title="Permalink to this headline">¶</a></h3>
+<p>To route CFI checks to the target DSO’s __cfi_check function, a
+mapping from possible virtual / indirect call targets to
+the corresponding __cfi_check functions is maintained. This mapping is
+implemented as a sparse array of 2 bytes for every possible page (4096
+bytes) of memory. The table is kept readonly (FIXME: not yet) most of
+the time.</p>
+<p>There are 3 types of shadow values:</p>
+<blockquote>
+<div><ul class="simple">
+<li>Address in a CFI-instrumented DSO.</li>
+<li>Unchecked address (a âtrustedâ non-instrumented DSO). Encoded as
+value 0xFFFF.</li>
+<li>Invalid address (everything else). Encoded as value 0.</li>
+</ul>
+</div></blockquote>
+<p>For a CFI-instrumented DSO, a shadow value encodes the address of the
+__cfi_check function for all call targets in the corresponding memory
+page. If Addr is the target address, and V is the shadow value, then
+the address of __cfi_check is calculated as</p>
+<div class="highlight-none"><div class="highlight"><pre>__cfi_check = AlignUpTo(Addr, 4096) - (V + 1) * 4096
+</pre></div>
+</div>
+<p>This works as long as __cfi_check is aligned by 4096 bytes and located
+below any call targets in its DSO, but not more than 256MB apart from
+them.</p>
+</div>
+<div class="section" id="cfi-slowpath">
+<h3>CFI_SlowPath<a class="headerlink" href="#cfi-slowpath" title="Permalink to this headline">¶</a></h3>
+<p>The slow path check is implemented in compiler-rt library as</p>
+<div class="highlight-none"><div class="highlight"><pre>void __cfi_slowpath(uint64 CallSiteTypeId, void *TargetAddr)
+</pre></div>
+</div>
+<p>This functions loads a shadow value for <tt class="docutils literal"><span class="pre">TargetAddr</span></tt>, finds the
+address of __cfi_check as described above and calls that.</p>
+</div>
+<div class="section" id="position-independent-executable-requirement">
+<h3>Position-independent executable requirement<a class="headerlink" href="#position-independent-executable-requirement" title="Permalink to this headline">¶</a></h3>
+<p>Cross-DSO CFI mode requires that the main executable is built as PIE.
+In non-PIE executables the address of an external function (taken from
+the main executable) is the address of that functionâs PLT record in
+the main executable. This would break the CFI checks.</p>
+</div>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="ControlFlowIntegrity.html">Control Flow Integrity</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="SafeStack.html">SafeStack</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/CrossCompilation.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/CrossCompilation.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/CrossCompilation.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/CrossCompilation.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,259 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Cross-compilation using Clang — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Thread Safety Analysis" href="ThreadSafetyAnalysis.html" />
+ <link rel="prev" title="Attributes in Clang" href="AttributeReference.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Cross-compilation using Clang</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="AttributeReference.html">Attributes in Clang</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ThreadSafetyAnalysis.html">Thread Safety Analysis</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="cross-compilation-using-clang">
+<h1>Cross-compilation using Clang<a class="headerlink" href="#cross-compilation-using-clang" title="Permalink to this headline">¶</a></h1>
+<div class="section" id="introduction">
+<h2>Introduction<a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>This document will guide you in choosing the right Clang options
+for cross-compiling your code to a different architecture. It assumes you
+already know how to compile the code in question for the host architecture,
+and that you know how to choose additional include and library paths.</p>
+<p>However, this document is <em>not</em> a “how to” and won’t help you setting your
+build system or Makefiles, nor choosing the right CMake options, etc.
+Also, it does not cover all the possible options, nor does it contain
+specific examples for specific architectures. For a concrete example, the
+<a class="reference external" href="http://llvm.org/docs/HowToCrossCompileLLVM.html">instructions for cross-compiling LLVM itself</a> may be of interest.</p>
+<p>After reading this document, you should be familiar with the main issues
+related to cross-compilation, and what main compiler options Clang provides
+for performing cross-compilation.</p>
+</div>
+<div class="section" id="cross-compilation-issues">
+<h2>Cross compilation issues<a class="headerlink" href="#cross-compilation-issues" title="Permalink to this headline">¶</a></h2>
+<p>In GCC world, every host/target combination has its own set of binaries,
+headers, libraries, etc. So, it’s usually simple to download a package
+with all files in, unzip to a directory and point the build system to
+that compiler, that will know about its location and find all it needs to
+when compiling your code.</p>
+<p>On the other hand, Clang/LLVM is natively a cross-compiler, meaning that
+one set of programs can compile to all targets by setting the <tt class="docutils literal"><span class="pre">-target</span></tt>
+option. That makes it a lot easier for programers wishing to compile to
+different platforms and architectures, and for compiler developers that
+only have to maintain one build system, and for OS distributions, that
+need only one set of main packages.</p>
+<p>But, as is true to any cross-compiler, and given the complexity of
+different architectures, OS’s and options, it’s not always easy finding
+the headers, libraries or binutils to generate target specific code.
+So you’ll need special options to help Clang understand what target
+you’re compiling to, where your tools are, etc.</p>
+<p>Another problem is that compilers come with standard libraries only (like
+<tt class="docutils literal"><span class="pre">compiler-rt</span></tt>, <tt class="docutils literal"><span class="pre">libcxx</span></tt>, <tt class="docutils literal"><span class="pre">libgcc</span></tt>, <tt class="docutils literal"><span class="pre">libm</span></tt>, etc), so you’ll have to
+find and make available to the build system, every other library required
+to build your software, that is specific to your target. It’s not enough to
+have your host’s libraries installed.</p>
+<p>Finally, not all toolchains are the same, and consequently, not every Clang
+option will work magically. Some options, like <tt class="docutils literal"><span class="pre">--sysroot</span></tt> (which
+effectively changes the logical root for headers and libraries), assume
+all your binaries and libraries are in the same directory, which may not
+true when your cross-compiler was installed by the distribution’s package
+management. So, for each specific case, you may use more than one
+option, and in most cases, you’ll end up setting include paths (<tt class="docutils literal"><span class="pre">-I</span></tt>) and
+library paths (<tt class="docutils literal"><span class="pre">-L</span></tt>) manually.</p>
+<dl class="docutils">
+<dt>To sum up, different toolchains can:</dt>
+<dd><ul class="first last simple">
+<li>be host/target specific or more flexible</li>
+<li>be in a single directory, or spread out across your system</li>
+<li>have different sets of libraries and headers by default</li>
+<li>need special options, which your build system won’t be able to figure
+out by itself</li>
+</ul>
+</dd>
+</dl>
+</div>
+<div class="section" id="general-cross-compilation-options-in-clang">
+<h2>General Cross-Compilation Options in Clang<a class="headerlink" href="#general-cross-compilation-options-in-clang" title="Permalink to this headline">¶</a></h2>
+<div class="section" id="target-triple">
+<h3>Target Triple<a class="headerlink" href="#target-triple" title="Permalink to this headline">¶</a></h3>
+<p>The basic option is to define the target architecture. For that, use
+<tt class="docutils literal"><span class="pre">-target</span> <span class="pre"><triple></span></tt>. If you don’t specify the target, CPU names won’t
+match (since Clang assumes the host triple), and the compilation will
+go ahead, creating code for the host platform, which will break later
+on when assembling or linking.</p>
+<dl class="docutils">
+<dt>The triple has the general format <tt class="docutils literal"><span class="pre"><arch><sub>-<vendor>-<sys>-<abi></span></tt>, where:</dt>
+<dd><ul class="first last simple">
+<li><tt class="docutils literal"><span class="pre">arch</span></tt> = <tt class="docutils literal"><span class="pre">x86</span></tt>, <tt class="docutils literal"><span class="pre">arm</span></tt>, <tt class="docutils literal"><span class="pre">thumb</span></tt>, <tt class="docutils literal"><span class="pre">mips</span></tt>, etc.</li>
+<li><tt class="docutils literal"><span class="pre">sub</span></tt> = for ex. on ARM: <tt class="docutils literal"><span class="pre">v5</span></tt>, <tt class="docutils literal"><span class="pre">v6m</span></tt>, <tt class="docutils literal"><span class="pre">v7a</span></tt>, <tt class="docutils literal"><span class="pre">v7m</span></tt>, etc.</li>
+<li><tt class="docutils literal"><span class="pre">vendor</span></tt> = <tt class="docutils literal"><span class="pre">pc</span></tt>, <tt class="docutils literal"><span class="pre">apple</span></tt>, <tt class="docutils literal"><span class="pre">nvidia</span></tt>, <tt class="docutils literal"><span class="pre">ibm</span></tt>, etc.</li>
+<li><tt class="docutils literal"><span class="pre">sys</span></tt> = <tt class="docutils literal"><span class="pre">none</span></tt>, <tt class="docutils literal"><span class="pre">linux</span></tt>, <tt class="docutils literal"><span class="pre">win32</span></tt>, <tt class="docutils literal"><span class="pre">darwin</span></tt>, <tt class="docutils literal"><span class="pre">cuda</span></tt>, etc.</li>
+<li><tt class="docutils literal"><span class="pre">abi</span></tt> = <tt class="docutils literal"><span class="pre">eabi</span></tt>, <tt class="docutils literal"><span class="pre">gnu</span></tt>, <tt class="docutils literal"><span class="pre">android</span></tt>, <tt class="docutils literal"><span class="pre">macho</span></tt>, <tt class="docutils literal"><span class="pre">elf</span></tt>, etc.</li>
+</ul>
+</dd>
+</dl>
+<p>The sub-architecture options are available for their own architectures,
+of course, so “x86v7a” doesn’t make sense. The vendor needs to be
+specified only if there’s a relevant change, for instance between PC
+and Apple. Most of the time it can be omitted (and Unknown)
+will be assumed, which sets the defaults for the specified architecture.
+The system name is generally the OS (linux, darwin), but could be special
+like the bare-metal “none”.</p>
+<p>When a parameter is not important, it can be omitted, or you can
+choose <tt class="docutils literal"><span class="pre">unknown</span></tt> and the defaults will be used. If you choose a parameter
+that Clang doesn’t know, like <tt class="docutils literal"><span class="pre">blerg</span></tt>, it’ll ignore and assume
+<tt class="docutils literal"><span class="pre">unknown</span></tt>, which is not always desired, so be careful.</p>
+<p>Finally, the ABI option is something that will pick default CPU/FPU,
+define the specific behaviour of your code (PCS, extensions),
+and also choose the correct library calls, etc.</p>
+</div>
+<div class="section" id="cpu-fpu-abi">
+<h3>CPU, FPU, ABI<a class="headerlink" href="#cpu-fpu-abi" title="Permalink to this headline">¶</a></h3>
+<p>Once your target is specified, it’s time to pick the hardware you’ll
+be compiling to. For every architecture, a default set of CPU/FPU/ABI
+will be chosen, so you’ll almost always have to change it via flags.</p>
+<dl class="docutils">
+<dt>Typical flags include:</dt>
+<dd><ul class="first last simple">
+<li><tt class="docutils literal"><span class="pre">-mcpu=<cpu-name></span></tt>, like x86-64, swift, cortex-a15</li>
+<li><tt class="docutils literal"><span class="pre">-mfpu=<fpu-name></span></tt>, like SSE3, NEON, controlling the FP unit available</li>
+<li><tt class="docutils literal"><span class="pre">-mfloat-abi=<fabi></span></tt>, like soft, hard, controlling which registers
+to use for floating-point</li>
+</ul>
+</dd>
+</dl>
+<p>The default is normally the common denominator, so that Clang doesn’t
+generate code that breaks. But that also means you won’t get the best
+code for your specific hardware, which may mean orders of magnitude
+slower than you expect.</p>
+<p>For example, if your target is <tt class="docutils literal"><span class="pre">arm-none-eabi</span></tt>, the default CPU will
+be <tt class="docutils literal"><span class="pre">arm7tdmi</span></tt> using soft float, which is extremely slow on modern cores,
+whereas if your triple is <tt class="docutils literal"><span class="pre">armv7a-none-eabi</span></tt>, it’ll be Cortex-A8 with
+NEON, but still using soft-float, which is much better, but still not
+great.</p>
+</div>
+<div class="section" id="toolchain-options">
+<h3>Toolchain Options<a class="headerlink" href="#toolchain-options" title="Permalink to this headline">¶</a></h3>
+<p>There are three main options to control access to your cross-compiler:
+<tt class="docutils literal"><span class="pre">--sysroot</span></tt>, <tt class="docutils literal"><span class="pre">-I</span></tt>, and <tt class="docutils literal"><span class="pre">-L</span></tt>. The two last ones are well known,
+but they’re particularly important for additional libraries
+and headers that are specific to your target.</p>
+<p>There are two main ways to have a cross-compiler:</p>
+<ol class="arabic">
+<li><p class="first">When you have extracted your cross-compiler from a zip file into
+a directory, you have to use <tt class="docutils literal"><span class="pre">--sysroot=<path></span></tt>. The path is the
+root directory where you have unpacked your file, and Clang will
+look for the directories <tt class="docutils literal"><span class="pre">bin</span></tt>, <tt class="docutils literal"><span class="pre">lib</span></tt>, <tt class="docutils literal"><span class="pre">include</span></tt> in there.</p>
+<p>In this case, your setup should be pretty much done (if no
+additional headers or libraries are needed), as Clang will find
+all binaries it needs (assembler, linker, etc) in there.</p>
+</li>
+<li><p class="first">When you have installed via a package manager (modern Linux
+distributions have cross-compiler packages available), make
+sure the target triple you set is <em>also</em> the prefix of your
+cross-compiler toolchain.</p>
+<p>In this case, Clang will find the other binaries (assembler,
+linker), but not always where the target headers and libraries
+are. People add system-specific clues to Clang often, but as
+things change, it’s more likely that it won’t find than the
+other way around.</p>
+<p>So, here, you’ll be a lot safer if you specify the include/library
+directories manually (via <tt class="docutils literal"><span class="pre">-I</span></tt> and <tt class="docutils literal"><span class="pre">-L</span></tt>).</p>
+</li>
+</ol>
+</div>
+</div>
+<div class="section" id="target-specific-libraries">
+<h2>Target-Specific Libraries<a class="headerlink" href="#target-specific-libraries" title="Permalink to this headline">¶</a></h2>
+<p>All libraries that you compile as part of your build will be
+cross-compiled to your target, and your build system will probably
+find them in the right place. But all dependencies that are
+normally checked against (like <tt class="docutils literal"><span class="pre">libxml</span></tt> or <tt class="docutils literal"><span class="pre">libz</span></tt> etc) will match
+against the host platform, not the target.</p>
+<p>So, if the build system is not aware that you want to cross-compile
+your code, it will get every dependency wrong, and your compilation
+will fail during build time, not configure time.</p>
+<p>Also, finding the libraries for your target are not as easy
+as for your host machine. There aren’t many cross-libraries available
+as packages to most OS’s, so you’ll have to either cross-compile them
+from source, or download the package for your target platform,
+extract the libraries and headers, put them in specific directories
+and add <tt class="docutils literal"><span class="pre">-I</span></tt> and <tt class="docutils literal"><span class="pre">-L</span></tt> pointing to them.</p>
+<p>Also, some libraries have different dependencies on different targets,
+so configuration tools to find dependencies in the host can get the
+list wrong for the target platform. This means that the configuration
+of your build can get things wrong when setting their own library
+paths, and you’ll have to augment it via additional flags (configure,
+Make, CMake, etc).</p>
+<div class="section" id="multilibs">
+<h3>Multilibs<a class="headerlink" href="#multilibs" title="Permalink to this headline">¶</a></h3>
+<p>When you want to cross-compile to more than one configuration, for
+example hard-float-ARM and soft-float-ARM, you’ll have to have multiple
+copies of your libraries and (possibly) headers.</p>
+<p>Some Linux distributions have support for Multilib, which handle that
+for you in an easier way, but if you’re not careful and, for instance,
+forget to specify <tt class="docutils literal"><span class="pre">-ccc-gcc-name</span> <span class="pre">armv7l-linux-gnueabihf-gcc</span></tt> (which
+uses hard-float), Clang will pick the <tt class="docutils literal"><span class="pre">armv7l-linux-gnueabi-ld</span></tt>
+(which uses soft-float) and linker errors will happen.</p>
+<p>The same is true if you’re compiling for different ABIs, like <tt class="docutils literal"><span class="pre">gnueabi</span></tt>
+and <tt class="docutils literal"><span class="pre">androideabi</span></tt>, and might even link and run, but produce run-time
+errors, which are much harder to track down and fix.</p>
+</div>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="AttributeReference.html">Attributes in Clang</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ThreadSafetyAnalysis.html">Thread Safety Analysis</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/DataFlowSanitizer.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/DataFlowSanitizer.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/DataFlowSanitizer.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/DataFlowSanitizer.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,221 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>DataFlowSanitizer — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="DataFlowSanitizer Design Document" href="DataFlowSanitizerDesign.html" />
+ <link rel="prev" title="UndefinedBehaviorSanitizer" href="UndefinedBehaviorSanitizer.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>DataFlowSanitizer</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="UndefinedBehaviorSanitizer.html">UndefinedBehaviorSanitizer</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="DataFlowSanitizerDesign.html">DataFlowSanitizer Design Document</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="dataflowsanitizer">
+<h1>DataFlowSanitizer<a class="headerlink" href="#dataflowsanitizer" title="Permalink to this headline">¶</a></h1>
+<div class="toctree-wrapper compound">
+</div>
+<div class="contents local topic" id="contents">
+<ul class="simple">
+<li><a class="reference internal" href="#introduction" id="id1">Introduction</a></li>
+<li><a class="reference internal" href="#usage" id="id2">Usage</a><ul>
+<li><a class="reference internal" href="#abi-list" id="id3">ABI List</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#example" id="id4">Example</a></li>
+<li><a class="reference internal" href="#current-status" id="id5">Current status</a></li>
+<li><a class="reference internal" href="#design" id="id6">Design</a></li>
+</ul>
+</div>
+<div class="section" id="introduction">
+<h2><a class="toc-backref" href="#id1">Introduction</a><a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>DataFlowSanitizer is a generalised dynamic data flow analysis.</p>
+<p>Unlike other Sanitizer tools, this tool is not designed to detect a
+specific class of bugs on its own. Instead, it provides a generic
+dynamic data flow analysis framework to be used by clients to help
+detect application-specific issues within their own code.</p>
+</div>
+<div class="section" id="usage">
+<h2><a class="toc-backref" href="#id2">Usage</a><a class="headerlink" href="#usage" title="Permalink to this headline">¶</a></h2>
+<p>With no program changes, applying DataFlowSanitizer to a program
+will not alter its behavior. To use DataFlowSanitizer, the program
+uses API functions to apply tags to data to cause it to be tracked, and to
+check the tag of a specific data item. DataFlowSanitizer manages
+the propagation of tags through the program according to its data flow.</p>
+<p>The APIs are defined in the header file <tt class="docutils literal"><span class="pre">sanitizer/dfsan_interface.h</span></tt>.
+For further information about each function, please refer to the header
+file.</p>
+<div class="section" id="abi-list">
+<h3><a class="toc-backref" href="#id3">ABI List</a><a class="headerlink" href="#abi-list" title="Permalink to this headline">¶</a></h3>
+<p>DataFlowSanitizer uses a list of functions known as an ABI list to decide
+whether a call to a specific function should use the operating system’s native
+ABI or whether it should use a variant of this ABI that also propagates labels
+through function parameters and return values. The ABI list file also controls
+how labels are propagated in the former case. DataFlowSanitizer comes with a
+default ABI list which is intended to eventually cover the glibc library on
+Linux but it may become necessary for users to extend the ABI list in cases
+where a particular library or function cannot be instrumented (e.g. because
+it is implemented in assembly or another language which DataFlowSanitizer does
+not support) or a function is called from a library or function which cannot
+be instrumented.</p>
+<p>DataFlowSanitizer’s ABI list file is a <a class="reference internal" href="SanitizerSpecialCaseList.html"><em>Sanitizer special case list</em></a>.
+The pass treats every function in the <tt class="docutils literal"><span class="pre">uninstrumented</span></tt> category in the
+ABI list file as conforming to the native ABI. Unless the ABI list contains
+additional categories for those functions, a call to one of those functions
+will produce a warning message, as the labelling behavior of the function
+is unknown. The other supported categories are <tt class="docutils literal"><span class="pre">discard</span></tt>, <tt class="docutils literal"><span class="pre">functional</span></tt>
+and <tt class="docutils literal"><span class="pre">custom</span></tt>.</p>
+<ul class="simple">
+<li><tt class="docutils literal"><span class="pre">discard</span></tt> – To the extent that this function writes to (user-accessible)
+memory, it also updates labels in shadow memory (this condition is trivially
+satisfied for functions which do not write to user-accessible memory). Its
+return value is unlabelled.</li>
+<li><tt class="docutils literal"><span class="pre">functional</span></tt> – Like <tt class="docutils literal"><span class="pre">discard</span></tt>, except that the label of its return value
+is the union of the label of its arguments.</li>
+<li><tt class="docutils literal"><span class="pre">custom</span></tt> – Instead of calling the function, a custom wrapper <tt class="docutils literal"><span class="pre">__dfsw_F</span></tt>
+is called, where <tt class="docutils literal"><span class="pre">F</span></tt> is the name of the function. This function may wrap
+the original function or provide its own implementation. This category is
+generally used for uninstrumentable functions which write to user-accessible
+memory or which have more complex label propagation behavior. The signature
+of <tt class="docutils literal"><span class="pre">__dfsw_F</span></tt> is based on that of <tt class="docutils literal"><span class="pre">F</span></tt> with each argument having a
+label of type <tt class="docutils literal"><span class="pre">dfsan_label</span></tt> appended to the argument list. If <tt class="docutils literal"><span class="pre">F</span></tt>
+is of non-void return type a final argument of type <tt class="docutils literal"><span class="pre">dfsan_label</span> <span class="pre">*</span></tt>
+is appended to which the custom function can store the label for the
+return value. For example:</li>
+</ul>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">void</span> <span class="nf">f</span><span class="p">(</span><span class="kt">int</span> <span class="n">x</span><span class="p">);</span>
+<span class="kt">void</span> <span class="nf">__dfsw_f</span><span class="p">(</span><span class="kt">int</span> <span class="n">x</span><span class="p">,</span> <span class="n">dfsan_label</span> <span class="n">x_label</span><span class="p">);</span>
+
+<span class="kt">void</span> <span class="o">*</span><span class="nf">memcpy</span><span class="p">(</span><span class="kt">void</span> <span class="o">*</span><span class="n">dest</span><span class="p">,</span> <span class="k">const</span> <span class="kt">void</span> <span class="o">*</span><span class="n">src</span><span class="p">,</span> <span class="kt">size_t</span> <span class="n">n</span><span class="p">);</span>
+<span class="kt">void</span> <span class="o">*</span><span class="nf">__dfsw_memcpy</span><span class="p">(</span><span class="kt">void</span> <span class="o">*</span><span class="n">dest</span><span class="p">,</span> <span class="k">const</span> <span class="kt">void</span> <span class="o">*</span><span class="n">src</span><span class="p">,</span> <span class="kt">size_t</span> <span class="n">n</span><span class="p">,</span>
+ <span class="n">dfsan_label</span> <span class="n">dest_label</span><span class="p">,</span> <span class="n">dfsan_label</span> <span class="n">src_label</span><span class="p">,</span>
+ <span class="n">dfsan_label</span> <span class="n">n_label</span><span class="p">,</span> <span class="n">dfsan_label</span> <span class="o">*</span><span class="n">ret_label</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>If a function defined in the translation unit being compiled belongs to the
+<tt class="docutils literal"><span class="pre">uninstrumented</span></tt> category, it will be compiled so as to conform to the
+native ABI. Its arguments will be assumed to be unlabelled, but it will
+propagate labels in shadow memory.</p>
+<p>For example:</p>
+<div class="highlight-none"><div class="highlight"><pre># main is called by the C runtime using the native ABI.
+fun:main=uninstrumented
+fun:main=discard
+
+# malloc only writes to its internal data structures, not user-accessible memory.
+fun:malloc=uninstrumented
+fun:malloc=discard
+
+# tolower is a pure function.
+fun:tolower=uninstrumented
+fun:tolower=functional
+
+# memcpy needs to copy the shadow from the source to the destination region.
+# This is done in a custom function.
+fun:memcpy=uninstrumented
+fun:memcpy=custom
+</pre></div>
+</div>
+</div>
+</div>
+<div class="section" id="example">
+<h2><a class="toc-backref" href="#id4">Example</a><a class="headerlink" href="#example" title="Permalink to this headline">¶</a></h2>
+<p>The following program demonstrates label propagation by checking that
+the correct labels are propagated.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#include <sanitizer/dfsan_interface.h></span>
+<span class="cp">#include <assert.h></span>
+
+<span class="kt">int</span> <span class="nf">main</span><span class="p">(</span><span class="kt">void</span><span class="p">)</span> <span class="p">{</span>
+ <span class="kt">int</span> <span class="n">i</span> <span class="o">=</span> <span class="mi">1</span><span class="p">;</span>
+ <span class="n">dfsan_label</span> <span class="n">i_label</span> <span class="o">=</span> <span class="n">dfsan_create_label</span><span class="p">(</span><span class="s">"i"</span><span class="p">,</span> <span class="mi">0</span><span class="p">);</span>
+ <span class="n">dfsan_set_label</span><span class="p">(</span><span class="n">i_label</span><span class="p">,</span> <span class="o">&</span><span class="n">i</span><span class="p">,</span> <span class="k">sizeof</span><span class="p">(</span><span class="n">i</span><span class="p">));</span>
+
+ <span class="kt">int</span> <span class="n">j</span> <span class="o">=</span> <span class="mi">2</span><span class="p">;</span>
+ <span class="n">dfsan_label</span> <span class="n">j_label</span> <span class="o">=</span> <span class="n">dfsan_create_label</span><span class="p">(</span><span class="s">"j"</span><span class="p">,</span> <span class="mi">0</span><span class="p">);</span>
+ <span class="n">dfsan_set_label</span><span class="p">(</span><span class="n">j_label</span><span class="p">,</span> <span class="o">&</span><span class="n">j</span><span class="p">,</span> <span class="k">sizeof</span><span class="p">(</span><span class="n">j</span><span class="p">));</span>
+
+ <span class="kt">int</span> <span class="n">k</span> <span class="o">=</span> <span class="mi">3</span><span class="p">;</span>
+ <span class="n">dfsan_label</span> <span class="n">k_label</span> <span class="o">=</span> <span class="n">dfsan_create_label</span><span class="p">(</span><span class="s">"k"</span><span class="p">,</span> <span class="mi">0</span><span class="p">);</span>
+ <span class="n">dfsan_set_label</span><span class="p">(</span><span class="n">k_label</span><span class="p">,</span> <span class="o">&</span><span class="n">k</span><span class="p">,</span> <span class="k">sizeof</span><span class="p">(</span><span class="n">k</span><span class="p">));</span>
+
+ <span class="n">dfsan_label</span> <span class="n">ij_label</span> <span class="o">=</span> <span class="n">dfsan_get_label</span><span class="p">(</span><span class="n">i</span> <span class="o">+</span> <span class="n">j</span><span class="p">);</span>
+ <span class="n">assert</span><span class="p">(</span><span class="n">dfsan_has_label</span><span class="p">(</span><span class="n">ij_label</span><span class="p">,</span> <span class="n">i_label</span><span class="p">));</span>
+ <span class="n">assert</span><span class="p">(</span><span class="n">dfsan_has_label</span><span class="p">(</span><span class="n">ij_label</span><span class="p">,</span> <span class="n">j_label</span><span class="p">));</span>
+ <span class="n">assert</span><span class="p">(</span><span class="o">!</span><span class="n">dfsan_has_label</span><span class="p">(</span><span class="n">ij_label</span><span class="p">,</span> <span class="n">k_label</span><span class="p">));</span>
+
+ <span class="n">dfsan_label</span> <span class="n">ijk_label</span> <span class="o">=</span> <span class="n">dfsan_get_label</span><span class="p">(</span><span class="n">i</span> <span class="o">+</span> <span class="n">j</span> <span class="o">+</span> <span class="n">k</span><span class="p">);</span>
+ <span class="n">assert</span><span class="p">(</span><span class="n">dfsan_has_label</span><span class="p">(</span><span class="n">ijk_label</span><span class="p">,</span> <span class="n">i_label</span><span class="p">));</span>
+ <span class="n">assert</span><span class="p">(</span><span class="n">dfsan_has_label</span><span class="p">(</span><span class="n">ijk_label</span><span class="p">,</span> <span class="n">j_label</span><span class="p">));</span>
+ <span class="n">assert</span><span class="p">(</span><span class="n">dfsan_has_label</span><span class="p">(</span><span class="n">ijk_label</span><span class="p">,</span> <span class="n">k_label</span><span class="p">));</span>
+
+ <span class="k">return</span> <span class="mi">0</span><span class="p">;</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="current-status">
+<h2><a class="toc-backref" href="#id5">Current status</a><a class="headerlink" href="#current-status" title="Permalink to this headline">¶</a></h2>
+<p>DataFlowSanitizer is a work in progress, currently under development for
+x86_64 Linux.</p>
+</div>
+<div class="section" id="design">
+<h2><a class="toc-backref" href="#id6">Design</a><a class="headerlink" href="#design" title="Permalink to this headline">¶</a></h2>
+<p>Please refer to the <a class="reference internal" href="DataFlowSanitizerDesign.html"><em>design document</em></a>.</p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="UndefinedBehaviorSanitizer.html">UndefinedBehaviorSanitizer</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="DataFlowSanitizerDesign.html">DataFlowSanitizer Design Document</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/DataFlowSanitizerDesign.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/DataFlowSanitizerDesign.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/DataFlowSanitizerDesign.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/DataFlowSanitizerDesign.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,291 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>DataFlowSanitizer Design Document — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="up" title="DataFlowSanitizer" href="DataFlowSanitizer.html" />
+ <link rel="next" title="LeakSanitizer" href="LeakSanitizer.html" />
+ <link rel="prev" title="DataFlowSanitizer" href="DataFlowSanitizer.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>DataFlowSanitizer Design Document</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="DataFlowSanitizer.html">DataFlowSanitizer</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="LeakSanitizer.html">LeakSanitizer</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="dataflowsanitizer-design-document">
+<h1>DataFlowSanitizer Design Document<a class="headerlink" href="#dataflowsanitizer-design-document" title="Permalink to this headline">¶</a></h1>
+<p>This document sets out the design for DataFlowSanitizer, a general
+dynamic data flow analysis. Unlike other Sanitizer tools, this tool is
+not designed to detect a specific class of bugs on its own. Instead,
+it provides a generic dynamic data flow analysis framework to be used
+by clients to help detect application-specific issues within their
+own code.</p>
+<p>DataFlowSanitizer is a program instrumentation which can associate
+a number of taint labels with any data stored in any memory region
+accessible by the program. The analysis is dynamic, which means that
+it operates on a running program, and tracks how the labels propagate
+through that program. The tool shall support a large (>100) number
+of labels, such that programs which operate on large numbers of data
+items may be analysed with each data item being tracked separately.</p>
+<div class="section" id="use-cases">
+<h2>Use Cases<a class="headerlink" href="#use-cases" title="Permalink to this headline">¶</a></h2>
+<p>This instrumentation can be used as a tool to help monitor how data
+flows from a program’s inputs (sources) to its outputs (sinks).
+This has applications from a privacy/security perspective in that
+one can audit how a sensitive data item is used within a program and
+ensure it isn’t exiting the program anywhere it shouldn’t be.</p>
+</div>
+<div class="section" id="interface">
+<h2>Interface<a class="headerlink" href="#interface" title="Permalink to this headline">¶</a></h2>
+<p>A number of functions are provided which will create taint labels,
+attach labels to memory regions and extract the set of labels
+associated with a specific memory region. These functions are declared
+in the header file <tt class="docutils literal"><span class="pre">sanitizer/dfsan_interface.h</span></tt>.</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="c1">/// Creates and returns a base label with the given description and user data.</span>
+<span class="n">dfsan_label</span> <span class="nf">dfsan_create_label</span><span class="p">(</span><span class="k">const</span> <span class="kt">char</span> <span class="o">*</span><span class="n">desc</span><span class="p">,</span> <span class="kt">void</span> <span class="o">*</span><span class="n">userdata</span><span class="p">);</span>
+
+<span class="c1">/// Sets the label for each address in [addr,addr+size) to \c label.</span>
+<span class="kt">void</span> <span class="nf">dfsan_set_label</span><span class="p">(</span><span class="n">dfsan_label</span> <span class="n">label</span><span class="p">,</span> <span class="kt">void</span> <span class="o">*</span><span class="n">addr</span><span class="p">,</span> <span class="kt">size_t</span> <span class="n">size</span><span class="p">);</span>
+
+<span class="c1">/// Sets the label for each address in [addr,addr+size) to the union of the</span>
+<span class="c1">/// current label for that address and \c label.</span>
+<span class="kt">void</span> <span class="nf">dfsan_add_label</span><span class="p">(</span><span class="n">dfsan_label</span> <span class="n">label</span><span class="p">,</span> <span class="kt">void</span> <span class="o">*</span><span class="n">addr</span><span class="p">,</span> <span class="kt">size_t</span> <span class="n">size</span><span class="p">);</span>
+
+<span class="c1">/// Retrieves the label associated with the given data.</span>
+<span class="c1">///</span>
+<span class="c1">/// The type of 'data' is arbitrary. The function accepts a value of any type,</span>
+<span class="c1">/// which can be truncated or extended (implicitly or explicitly) as necessary.</span>
+<span class="c1">/// The truncation/extension operations will preserve the label of the original</span>
+<span class="c1">/// value.</span>
+<span class="n">dfsan_label</span> <span class="nf">dfsan_get_label</span><span class="p">(</span><span class="kt">long</span> <span class="n">data</span><span class="p">);</span>
+
+<span class="c1">/// Retrieves a pointer to the dfsan_label_info struct for the given label.</span>
+<span class="k">const</span> <span class="k">struct</span> <span class="n">dfsan_label_info</span> <span class="o">*</span><span class="nf">dfsan_get_label_info</span><span class="p">(</span><span class="n">dfsan_label</span> <span class="n">label</span><span class="p">);</span>
+
+<span class="c1">/// Returns whether the given label label contains the label elem.</span>
+<span class="kt">int</span> <span class="nf">dfsan_has_label</span><span class="p">(</span><span class="n">dfsan_label</span> <span class="n">label</span><span class="p">,</span> <span class="n">dfsan_label</span> <span class="n">elem</span><span class="p">);</span>
+
+<span class="c1">/// If the given label label contains a label with the description desc, returns</span>
+<span class="c1">/// that label, else returns 0.</span>
+<span class="n">dfsan_label</span> <span class="nf">dfsan_has_label_with_desc</span><span class="p">(</span><span class="n">dfsan_label</span> <span class="n">label</span><span class="p">,</span> <span class="k">const</span> <span class="kt">char</span> <span class="o">*</span><span class="n">desc</span><span class="p">);</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="taint-label-representation">
+<h2>Taint label representation<a class="headerlink" href="#taint-label-representation" title="Permalink to this headline">¶</a></h2>
+<p>As stated above, the tool must track a large number of taint
+labels. This poses an implementation challenge, as most multiple-label
+tainting systems assign one label per bit to shadow storage, and
+union taint labels using a bitwise or operation. This will not scale
+to clients which use hundreds or thousands of taint labels, as the
+label union operation becomes O(n) in the number of supported labels,
+and data associated with it will quickly dominate the live variable
+set, causing register spills and hampering performance.</p>
+<p>Instead, a low overhead approach is proposed which is best-case O(log<sub>2</sub> n) during execution. The underlying assumption is that
+the required space of label unions is sparse, which is a reasonable
+assumption to make given that we are optimizing for the case where
+applications mostly copy data from one place to another, without often
+invoking the need for an actual union operation. The representation
+of a taint label is a 16-bit integer, and new labels are allocated
+sequentially from a pool. The label identifier 0 is special, and means
+that the data item is unlabelled.</p>
+<p>When a label union operation is requested at a join point (any
+arithmetic or logical operation with two or more operands, such as
+addition), the code checks whether a union is required, whether the
+same union has been requested before, and whether one union label
+subsumes the other. If so, it returns the previously allocated union
+label. If not, it allocates a new union label from the same pool used
+for new labels.</p>
+<p>Specifically, the instrumentation pass will insert code like this
+to decide the union label <tt class="docutils literal"><span class="pre">lu</span></tt> for a pair of labels <tt class="docutils literal"><span class="pre">l1</span></tt>
+and <tt class="docutils literal"><span class="pre">l2</span></tt>:</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="k">if</span> <span class="p">(</span><span class="n">l1</span> <span class="o">==</span> <span class="n">l2</span><span class="p">)</span>
+ <span class="n">lu</span> <span class="o">=</span> <span class="n">l1</span><span class="p">;</span>
+<span class="k">else</span>
+ <span class="n">lu</span> <span class="o">=</span> <span class="n">__dfsan_union</span><span class="p">(</span><span class="n">l1</span><span class="p">,</span> <span class="n">l2</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>The equality comparison is outlined, to provide an early exit in
+the common cases where the program is processing unlabelled data, or
+where the two data items have the same label. <tt class="docutils literal"><span class="pre">__dfsan_union</span></tt> is
+a runtime library function which performs all other union computation.</p>
+<p>Further optimizations are possible, for example if <tt class="docutils literal"><span class="pre">l1</span></tt> is known
+at compile time to be zero (e.g. it is derived from a constant),
+<tt class="docutils literal"><span class="pre">l2</span></tt> can be used for <tt class="docutils literal"><span class="pre">lu</span></tt>, and vice versa.</p>
+</div>
+<div class="section" id="memory-layout-and-label-management">
+<h2>Memory layout and label management<a class="headerlink" href="#memory-layout-and-label-management" title="Permalink to this headline">¶</a></h2>
+<p>The following is the current memory layout for Linux/x86_64:</p>
+<table border="1" class="docutils">
+<colgroup>
+<col width="30%" />
+<col width="30%" />
+<col width="40%" />
+</colgroup>
+<thead valign="bottom">
+<tr class="row-odd"><th class="head">Start</th>
+<th class="head">End</th>
+<th class="head">Use</th>
+</tr>
+</thead>
+<tbody valign="top">
+<tr class="row-even"><td>0x700000008000</td>
+<td>0x800000000000</td>
+<td>application memory</td>
+</tr>
+<tr class="row-odd"><td>0x200200000000</td>
+<td>0x700000008000</td>
+<td>unused</td>
+</tr>
+<tr class="row-even"><td>0x200000000000</td>
+<td>0x200200000000</td>
+<td>union table</td>
+</tr>
+<tr class="row-odd"><td>0x000000010000</td>
+<td>0x200000000000</td>
+<td>shadow memory</td>
+</tr>
+<tr class="row-even"><td>0x000000000000</td>
+<td>0x000000010000</td>
+<td>reserved by kernel</td>
+</tr>
+</tbody>
+</table>
+<p>Each byte of application memory corresponds to two bytes of shadow
+memory, which are used to store its taint label. As for LLVM SSA
+registers, we have not found it necessary to associate a label with
+each byte or bit of data, as some other tools do. Instead, labels are
+associated directly with registers. Loads will result in a union of
+all shadow labels corresponding to bytes loaded (which most of the
+time will be short circuited by the initial comparison) and stores will
+result in a copy of the label to the shadow of all bytes stored to.</p>
+</div>
+<div class="section" id="propagating-labels-through-arguments">
+<h2>Propagating labels through arguments<a class="headerlink" href="#propagating-labels-through-arguments" title="Permalink to this headline">¶</a></h2>
+<p>In order to propagate labels through function arguments and return values,
+DataFlowSanitizer changes the ABI of each function in the translation unit.
+There are currently two supported ABIs:</p>
+<ul class="simple">
+<li>Args – Argument and return value labels are passed through additional
+arguments and by modifying the return type.</li>
+<li>TLS – Argument and return value labels are passed through TLS variables
+<tt class="docutils literal"><span class="pre">__dfsan_arg_tls</span></tt> and <tt class="docutils literal"><span class="pre">__dfsan_retval_tls</span></tt>.</li>
+</ul>
+<p>The main advantage of the TLS ABI is that it is more tolerant of ABI mismatches
+(TLS storage is not shared with any other form of storage, whereas extra
+arguments may be stored in registers which under the native ABI are not used
+for parameter passing and thus could contain arbitrary values). On the other
+hand the args ABI is more efficient and allows ABI mismatches to be more easily
+identified by checking for nonzero labels in nominally unlabelled programs.</p>
+</div>
+<div class="section" id="implementing-the-abi-list">
+<h2>Implementing the ABI list<a class="headerlink" href="#implementing-the-abi-list" title="Permalink to this headline">¶</a></h2>
+<p>The <a class="reference external" href="DataFlowSanitizer.html#abi-list">ABI list</a> provides a list of functions
+which conform to the native ABI, each of which is callable from an instrumented
+program. This is implemented by replacing each reference to a native ABI
+function with a reference to a function which uses the instrumented ABI.
+Such functions are automatically-generated wrappers for the native functions.
+For example, given the ABI list example provided in the user manual, the
+following wrappers will be generated under the args ABI:</p>
+<div class="highlight-llvm"><div class="highlight"><pre><span class="k">define</span> <span class="k">linkonce_odr</span> <span class="p">{</span> <span class="k">i8</span><span class="p">*,</span> <span class="k">i16</span> <span class="p">}</span> <span class="vg">@"dfsw$malloc"</span><span class="p">(</span><span class="k">i64</span> <span class="nv-Anonymous">%0</span><span class="p">,</span> <span class="k">i16</span> <span class="nv-Anonymous">%1</span><span class="p">)</span> <span class="p">{</span>
+<span class="nl">entry:</span>
+ <span class="nv-Anonymous">%2</span> <span class="p">=</span> <span class="k">call</span> <span class="k">i8</span><span class="p">*</span> <span class="vg">@malloc</span><span class="p">(</span><span class="k">i64</span> <span class="nv-Anonymous">%0</span><span class="p">)</span>
+ <span class="nv-Anonymous">%3</span> <span class="p">=</span> <span class="k">insertvalue</span> <span class="p">{</span> <span class="k">i8</span><span class="p">*,</span> <span class="k">i16</span> <span class="p">}</span> <span class="k">undef</span><span class="p">,</span> <span class="k">i8</span><span class="p">*</span> <span class="nv-Anonymous">%2</span><span class="p">,</span> <span class="m">0</span>
+ <span class="nv-Anonymous">%4</span> <span class="p">=</span> <span class="k">insertvalue</span> <span class="p">{</span> <span class="k">i8</span><span class="p">*,</span> <span class="k">i16</span> <span class="p">}</span> <span class="nv-Anonymous">%3</span><span class="p">,</span> <span class="k">i16</span> <span class="m">0</span><span class="p">,</span> <span class="m">1</span>
+ <span class="k">ret</span> <span class="p">{</span> <span class="k">i8</span><span class="p">*,</span> <span class="k">i16</span> <span class="p">}</span> <span class="nv-Anonymous">%4</span>
+<span class="p">}</span>
+
+<span class="k">define</span> <span class="k">linkonce_odr</span> <span class="p">{</span> <span class="k">i32</span><span class="p">,</span> <span class="k">i16</span> <span class="p">}</span> <span class="vg">@"dfsw$tolower"</span><span class="p">(</span><span class="k">i32</span> <span class="nv-Anonymous">%0</span><span class="p">,</span> <span class="k">i16</span> <span class="nv-Anonymous">%1</span><span class="p">)</span> <span class="p">{</span>
+<span class="nl">entry:</span>
+ <span class="nv-Anonymous">%2</span> <span class="p">=</span> <span class="k">call</span> <span class="k">i32</span> <span class="vg">@tolower</span><span class="p">(</span><span class="k">i32</span> <span class="nv-Anonymous">%0</span><span class="p">)</span>
+ <span class="nv-Anonymous">%3</span> <span class="p">=</span> <span class="k">insertvalue</span> <span class="p">{</span> <span class="k">i32</span><span class="p">,</span> <span class="k">i16</span> <span class="p">}</span> <span class="k">undef</span><span class="p">,</span> <span class="k">i32</span> <span class="nv-Anonymous">%2</span><span class="p">,</span> <span class="m">0</span>
+ <span class="nv-Anonymous">%4</span> <span class="p">=</span> <span class="k">insertvalue</span> <span class="p">{</span> <span class="k">i32</span><span class="p">,</span> <span class="k">i16</span> <span class="p">}</span> <span class="nv-Anonymous">%3</span><span class="p">,</span> <span class="k">i16</span> <span class="nv-Anonymous">%1</span><span class="p">,</span> <span class="m">1</span>
+ <span class="k">ret</span> <span class="p">{</span> <span class="k">i32</span><span class="p">,</span> <span class="k">i16</span> <span class="p">}</span> <span class="nv-Anonymous">%4</span>
+<span class="p">}</span>
+
+<span class="k">define</span> <span class="k">linkonce_odr</span> <span class="p">{</span> <span class="k">i8</span><span class="p">*,</span> <span class="k">i16</span> <span class="p">}</span> <span class="vg">@"dfsw$memcpy"</span><span class="p">(</span><span class="k">i8</span><span class="p">*</span> <span class="nv-Anonymous">%0</span><span class="p">,</span> <span class="k">i8</span><span class="p">*</span> <span class="nv-Anonymous">%1</span><span class="p">,</span> <span class="k">i64</span> <span class="nv-Anonymous">%2</span><span class="p">,</span> <span class="k">i16</span> <span class="nv-Anonymous">%3</span><span class="p">,</span> <span class="k">i16</span> <span class="nv-Anonymous">%4</span><span class="p">,</span> <span class="k">i16</span> <span class="nv-Anonymous">%5</span><span class="p">)</span> <span class="p">{</span>
+<span class="nl">entry:</span>
+ <span class="nv">%labelreturn</span> <span class="p">=</span> <span class="k">alloca</span> <span class="k">i16</span>
+ <span class="nv-Anonymous">%6</span> <span class="p">=</span> <span class="k">call</span> <span class="k">i8</span><span class="p">*</span> <span class="vg">@__dfsw_memcpy</span><span class="p">(</span><span class="k">i8</span><span class="p">*</span> <span class="nv-Anonymous">%0</span><span class="p">,</span> <span class="k">i8</span><span class="p">*</span> <span class="nv-Anonymous">%1</span><span class="p">,</span> <span class="k">i64</span> <span class="nv-Anonymous">%2</span><span class="p">,</span> <span class="k">i16</span> <span class="nv-Anonymous">%3</span><span class="p">,</span> <span class="k">i16</span> <span class="nv-Anonymous">%4</span><span class="p">,</span> <span class="k">i16</span> <span class="nv-Anonymous">%5</span><span class="p">,</span> <span class="k">i16</span><span class="p">*</span> <span class="nv">%labelreturn</span><span class="p">)</span>
+ <span class="nv-Anonymous">%7</span> <span class="p">=</span> <span class="k">load</span> <span class="k">i16</span><span class="p">*</span> <span class="nv">%labelreturn</span>
+ <span class="nv-Anonymous">%8</span> <span class="p">=</span> <span class="k">insertvalue</span> <span class="p">{</span> <span class="k">i8</span><span class="p">*,</span> <span class="k">i16</span> <span class="p">}</span> <span class="k">undef</span><span class="p">,</span> <span class="k">i8</span><span class="p">*</span> <span class="nv-Anonymous">%6</span><span class="p">,</span> <span class="m">0</span>
+ <span class="nv-Anonymous">%9</span> <span class="p">=</span> <span class="k">insertvalue</span> <span class="p">{</span> <span class="k">i8</span><span class="p">*,</span> <span class="k">i16</span> <span class="p">}</span> <span class="nv-Anonymous">%8</span><span class="p">,</span> <span class="k">i16</span> <span class="nv-Anonymous">%7</span><span class="p">,</span> <span class="m">1</span>
+ <span class="k">ret</span> <span class="p">{</span> <span class="k">i8</span><span class="p">*,</span> <span class="k">i16</span> <span class="p">}</span> <span class="nv-Anonymous">%9</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>As an optimization, direct calls to native ABI functions will call the
+native ABI function directly and the pass will compute the appropriate label
+internally. This has the advantage of reducing the number of union operations
+required when the return value label is known to be zero (i.e. <tt class="docutils literal"><span class="pre">discard</span></tt>
+functions, or <tt class="docutils literal"><span class="pre">functional</span></tt> functions with known unlabelled arguments).</p>
+</div>
+<div class="section" id="checking-abi-consistency">
+<h2>Checking ABI Consistency<a class="headerlink" href="#checking-abi-consistency" title="Permalink to this headline">¶</a></h2>
+<p>DFSan changes the ABI of each function in the module. This makes it possible
+for a function with the native ABI to be called with the instrumented ABI,
+or vice versa, thus possibly invoking undefined behavior. A simple way
+of statically detecting instances of this problem is to prepend the prefix
+“dfs$” to the name of each instrumented-ABI function.</p>
+<p>This will not catch every such problem; in particular function pointers passed
+across the instrumented-native barrier cannot be used on the other side.
+These problems could potentially be caught dynamically.</p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="DataFlowSanitizer.html">DataFlowSanitizer</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="LeakSanitizer.html">LeakSanitizer</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/DriverInternals.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/DriverInternals.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/DriverInternals.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/DriverInternals.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,452 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Driver Design & Internals — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Pretokenized Headers (PTH)" href="PTHInternals.html" />
+ <link rel="prev" title="âClangâ CFE Internals Manual" href="InternalsManual.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Driver Design & Internals</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="InternalsManual.html">“Clang” CFE Internals Manual</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="PTHInternals.html">Pretokenized Headers (PTH)</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="driver-design-internals">
+<h1>Driver Design & Internals<a class="headerlink" href="#driver-design-internals" title="Permalink to this headline">¶</a></h1>
+<div class="contents local topic" id="contents">
+<ul class="simple">
+<li><a class="reference internal" href="#introduction" id="id2">Introduction</a></li>
+<li><a class="reference internal" href="#features-and-goals" id="id3">Features and Goals</a><ul>
+<li><a class="reference internal" href="#gcc-compatibility" id="id4">GCC Compatibility</a></li>
+<li><a class="reference internal" href="#flexible" id="id5">Flexible</a></li>
+<li><a class="reference internal" href="#low-overhead" id="id6">Low Overhead</a></li>
+<li><a class="reference internal" href="#simple" id="id7">Simple</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#internal-design-and-implementation" id="id8">Internal Design and Implementation</a><ul>
+<li><a class="reference internal" href="#internals-introduction" id="id9">Internals Introduction</a></li>
+<li><a class="reference internal" href="#design-overview" id="id10">Design Overview</a></li>
+<li><a class="reference internal" href="#driver-stages" id="id11">Driver Stages</a></li>
+<li><a class="reference internal" href="#additional-notes" id="id12">Additional Notes</a><ul>
+<li><a class="reference internal" href="#the-compilation-object" id="id13">The Compilation Object</a></li>
+<li><a class="reference internal" href="#unified-parsing-pipelining" id="id14">Unified Parsing & Pipelining</a></li>
+<li><a class="reference internal" href="#toolchain-argument-translation" id="id15">ToolChain Argument Translation</a></li>
+<li><a class="reference internal" href="#unused-argument-warnings" id="id16">Unused Argument Warnings</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#relation-to-gcc-driver-concepts" id="id17">Relation to GCC Driver Concepts</a></li>
+</ul>
+</li>
+</ul>
+</div>
+<div class="section" id="introduction">
+<h2><a class="toc-backref" href="#id2">Introduction</a><a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>This document describes the Clang driver. The purpose of this document
+is to describe both the motivation and design goals for the driver, as
+well as details of the internal implementation.</p>
+</div>
+<div class="section" id="features-and-goals">
+<h2><a class="toc-backref" href="#id3">Features and Goals</a><a class="headerlink" href="#features-and-goals" title="Permalink to this headline">¶</a></h2>
+<p>The Clang driver is intended to be a production quality compiler driver
+providing access to the Clang compiler and tools, with a command line
+interface which is compatible with the gcc driver.</p>
+<p>Although the driver is part of and driven by the Clang project, it is
+logically a separate tool which shares many of the same goals as Clang:</p>
+<div class="contents local topic" id="features">
+<p class="topic-title first">Features</p>
+<ul class="simple">
+<li><a class="reference internal" href="#gcc-compatibility" id="id18">GCC Compatibility</a></li>
+<li><a class="reference internal" href="#flexible" id="id19">Flexible</a></li>
+<li><a class="reference internal" href="#low-overhead" id="id20">Low Overhead</a></li>
+<li><a class="reference internal" href="#simple" id="id21">Simple</a></li>
+</ul>
+</div>
+<div class="section" id="gcc-compatibility">
+<h3><a class="toc-backref" href="#id18">GCC Compatibility</a><a class="headerlink" href="#gcc-compatibility" title="Permalink to this headline">¶</a></h3>
+<p>The number one goal of the driver is to ease the adoption of Clang by
+allowing users to drop Clang into a build system which was designed to
+call GCC. Although this makes the driver much more complicated than
+might otherwise be necessary, we decided that being very compatible with
+the gcc command line interface was worth it in order to allow users to
+quickly test clang on their projects.</p>
+</div>
+<div class="section" id="flexible">
+<h3><a class="toc-backref" href="#id19">Flexible</a><a class="headerlink" href="#flexible" title="Permalink to this headline">¶</a></h3>
+<p>The driver was designed to be flexible and easily accommodate new uses
+as we grow the clang and LLVM infrastructure. As one example, the driver
+can easily support the introduction of tools which have an integrated
+assembler; something we hope to add to LLVM in the future.</p>
+<p>Similarly, most of the driver functionality is kept in a library which
+can be used to build other tools which want to implement or accept a gcc
+like interface.</p>
+</div>
+<div class="section" id="low-overhead">
+<h3><a class="toc-backref" href="#id20">Low Overhead</a><a class="headerlink" href="#low-overhead" title="Permalink to this headline">¶</a></h3>
+<p>The driver should have as little overhead as possible. In practice, we
+found that the gcc driver by itself incurred a small but meaningful
+overhead when compiling many small files. The driver doesn’t do much
+work compared to a compilation, but we have tried to keep it as
+efficient as possible by following a few simple principles:</p>
+<ul class="simple">
+<li>Avoid memory allocation and string copying when possible.</li>
+<li>Don’t parse arguments more than once.</li>
+<li>Provide a few simple interfaces for efficiently searching arguments.</li>
+</ul>
+</div>
+<div class="section" id="simple">
+<h3><a class="toc-backref" href="#id21">Simple</a><a class="headerlink" href="#simple" title="Permalink to this headline">¶</a></h3>
+<p>Finally, the driver was designed to be “as simple as possible”, given
+the other goals. Notably, trying to be completely compatible with the
+gcc driver adds a significant amount of complexity. However, the design
+of the driver attempts to mitigate this complexity by dividing the
+process into a number of independent stages instead of a single
+monolithic task.</p>
+</div>
+</div>
+<div class="section" id="internal-design-and-implementation">
+<h2><a class="toc-backref" href="#id8">Internal Design and Implementation</a><a class="headerlink" href="#internal-design-and-implementation" title="Permalink to this headline">¶</a></h2>
+<div class="contents local topic" id="id1">
+<ul class="simple">
+<li><a class="reference internal" href="#internals-introduction" id="id22">Internals Introduction</a></li>
+<li><a class="reference internal" href="#design-overview" id="id23">Design Overview</a></li>
+<li><a class="reference internal" href="#driver-stages" id="id24">Driver Stages</a></li>
+<li><a class="reference internal" href="#additional-notes" id="id25">Additional Notes</a></li>
+<li><a class="reference internal" href="#relation-to-gcc-driver-concepts" id="id26">Relation to GCC Driver Concepts</a></li>
+</ul>
+</div>
+<div class="section" id="internals-introduction">
+<h3><a class="toc-backref" href="#id22">Internals Introduction</a><a class="headerlink" href="#internals-introduction" title="Permalink to this headline">¶</a></h3>
+<p>In order to satisfy the stated goals, the driver was designed to
+completely subsume the functionality of the gcc executable; that is, the
+driver should not need to delegate to gcc to perform subtasks. On
+Darwin, this implies that the Clang driver also subsumes the gcc
+driver-driver, which is used to implement support for building universal
+images (binaries and object files). This also implies that the driver
+should be able to call the language specific compilers (e.g. cc1)
+directly, which means that it must have enough information to forward
+command line arguments to child processes correctly.</p>
+</div>
+<div class="section" id="design-overview">
+<h3><a class="toc-backref" href="#id23">Design Overview</a><a class="headerlink" href="#design-overview" title="Permalink to this headline">¶</a></h3>
+<p>The diagram below shows the significant components of the driver
+architecture and how they relate to one another. The orange components
+represent concrete data structures built by the driver, the green
+components indicate conceptually distinct stages which manipulate these
+data structures, and the blue components are important helper classes.</p>
+<img alt="Driver Architecture Diagram" class="align-center" src="_images/DriverArchitecture.png" />
+</div>
+<div class="section" id="driver-stages">
+<h3><a class="toc-backref" href="#id24">Driver Stages</a><a class="headerlink" href="#driver-stages" title="Permalink to this headline">¶</a></h3>
+<p>The driver functionality is conceptually divided into five stages:</p>
+<ol class="arabic">
+<li><p class="first"><strong>Parse: Option Parsing</strong></p>
+<p>The command line argument strings are decomposed into arguments
+(<tt class="docutils literal"><span class="pre">Arg</span></tt> instances). The driver expects to understand all available
+options, although there is some facility for just passing certain
+classes of options through (like <tt class="docutils literal"><span class="pre">-Wl,</span></tt>).</p>
+<p>Each argument corresponds to exactly one abstract <tt class="docutils literal"><span class="pre">Option</span></tt>
+definition, which describes how the option is parsed along with some
+additional metadata. The Arg instances themselves are lightweight and
+merely contain enough information for clients to determine which
+option they correspond to and their values (if they have additional
+parameters).</p>
+<p>For example, a command line like “-Ifoo -I foo” would parse to two
+Arg instances (a JoinedArg and a SeparateArg instance), but each
+would refer to the same Option.</p>
+<p>Options are lazily created in order to avoid populating all Option
+classes when the driver is loaded. Most of the driver code only needs
+to deal with options by their unique ID (e.g., <tt class="docutils literal"><span class="pre">options::OPT_I</span></tt>),</p>
+<p>Arg instances themselves do not generally store the values of
+parameters. In many cases, this would simply result in creating
+unnecessary string copies. Instead, Arg instances are always embedded
+inside an ArgList structure, which contains the original vector of
+argument strings. Each Arg itself only needs to contain an index into
+this vector instead of storing its values directly.</p>
+<p>The clang driver can dump the results of this stage using the
+<tt class="docutils literal"><span class="pre">-###</span></tt> flag (which must precede any actual command
+line arguments). For example:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> clang -### -Xarch_i386 -fomit-frame-pointer -Wa,-fast -Ifoo -I foo t.c
+<span class="go">Option 0 - Name: "-Xarch_", Values: {"i386", "-fomit-frame-pointer"}</span>
+<span class="go">Option 1 - Name: "-Wa,", Values: {"-fast"}</span>
+<span class="go">Option 2 - Name: "-I", Values: {"foo"}</span>
+<span class="go">Option 3 - Name: "-I", Values: {"foo"}</span>
+<span class="go">Option 4 - Name: "<input>", Values: {"t.c"}</span>
+</pre></div>
+</div>
+<p>After this stage is complete the command line should be broken down
+into well defined option objects with their appropriate parameters.
+Subsequent stages should rarely, if ever, need to do any string
+processing.</p>
+</li>
+<li><p class="first"><strong>Pipeline: Compilation Action Construction</strong></p>
+<p>Once the arguments are parsed, the tree of subprocess jobs needed for
+the desired compilation sequence are constructed. This involves
+determining the input files and their types, what work is to be done
+on them (preprocess, compile, assemble, link, etc.), and constructing
+a list of Action instances for each task. The result is a list of one
+or more top-level actions, each of which generally corresponds to a
+single output (for example, an object or linked executable).</p>
+<p>The majority of Actions correspond to actual tasks, however there are
+two special Actions. The first is InputAction, which simply serves to
+adapt an input argument for use as an input to other Actions. The
+second is BindArchAction, which conceptually alters the architecture
+to be used for all of its input Actions.</p>
+<p>The clang driver can dump the results of this stage using the
+<tt class="docutils literal"><span class="pre">-ccc-print-phases</span></tt> flag. For example:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> clang -ccc-print-phases -x c t.c -x assembler t.s
+<span class="go">0: input, "t.c", c</span>
+<span class="go">1: preprocessor, {0}, cpp-output</span>
+<span class="go">2: compiler, {1}, assembler</span>
+<span class="go">3: assembler, {2}, object</span>
+<span class="go">4: input, "t.s", assembler</span>
+<span class="go">5: assembler, {4}, object</span>
+<span class="go">6: linker, {3, 5}, image</span>
+</pre></div>
+</div>
+<p>Here the driver is constructing seven distinct actions, four to
+compile the “t.c” input into an object file, two to assemble the
+“t.s” input, and one to link them together.</p>
+<p>A rather different compilation pipeline is shown here; in this
+example there are two top level actions to compile the input files
+into two separate object files, where each object file is built using
+<tt class="docutils literal"><span class="pre">lipo</span></tt> to merge results built for two separate architectures.</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> clang -ccc-print-phases -c -arch i386 -arch x86_64 t0.c t1.c
+<span class="go">0: input, "t0.c", c</span>
+<span class="go">1: preprocessor, {0}, cpp-output</span>
+<span class="go">2: compiler, {1}, assembler</span>
+<span class="go">3: assembler, {2}, object</span>
+<span class="go">4: bind-arch, "i386", {3}, object</span>
+<span class="go">5: bind-arch, "x86_64", {3}, object</span>
+<span class="go">6: lipo, {4, 5}, object</span>
+<span class="go">7: input, "t1.c", c</span>
+<span class="go">8: preprocessor, {7}, cpp-output</span>
+<span class="go">9: compiler, {8}, assembler</span>
+<span class="go">10: assembler, {9}, object</span>
+<span class="go">11: bind-arch, "i386", {10}, object</span>
+<span class="go">12: bind-arch, "x86_64", {10}, object</span>
+<span class="go">13: lipo, {11, 12}, object</span>
+</pre></div>
+</div>
+<p>After this stage is complete the compilation process is divided into
+a simple set of actions which need to be performed to produce
+intermediate or final outputs (in some cases, like <tt class="docutils literal"><span class="pre">-fsyntax-only</span></tt>,
+there is no “real” final output). Phases are well known compilation
+steps, such as “preprocess”, “compile”, “assemble”, “link”, etc.</p>
+</li>
+<li><p class="first"><strong>Bind: Tool & Filename Selection</strong></p>
+<p>This stage (in conjunction with the Translate stage) turns the tree
+of Actions into a list of actual subprocess to run. Conceptually, the
+driver performs a top down matching to assign Action(s) to Tools. The
+ToolChain is responsible for selecting the tool to perform a
+particular action; once selected the driver interacts with the tool
+to see if it can match additional actions (for example, by having an
+integrated preprocessor).</p>
+<p>Once Tools have been selected for all actions, the driver determines
+how the tools should be connected (for example, using an inprocess
+module, pipes, temporary files, or user provided filenames). If an
+output file is required, the driver also computes the appropriate
+file name (the suffix and file location depend on the input types and
+options such as <tt class="docutils literal"><span class="pre">-save-temps</span></tt>).</p>
+<p>The driver interacts with a ToolChain to perform the Tool bindings.
+Each ToolChain contains information about all the tools needed for
+compilation for a particular architecture, platform, and operating
+system. A single driver invocation may query multiple ToolChains
+during one compilation in order to interact with tools for separate
+architectures.</p>
+<p>The results of this stage are not computed directly, but the driver
+can print the results via the <tt class="docutils literal"><span class="pre">-ccc-print-bindings</span></tt> option. For
+example:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> clang -ccc-print-bindings -arch i386 -arch ppc t0.c
+<span class="gp">#</span> <span class="s2">"i386-apple-darwin9"</span> - <span class="s2">"clang"</span>, inputs: <span class="o">[</span><span class="s2">"t0.c"</span><span class="o">]</span>, output: <span class="s2">"/tmp/cc-Sn4RKF.s"</span>
+<span class="gp">#</span> <span class="s2">"i386-apple-darwin9"</span> - <span class="s2">"darwin::Assemble"</span>, inputs: <span class="o">[</span><span class="s2">"/tmp/cc-Sn4RKF.s"</span><span class="o">]</span>, output: <span class="s2">"/tmp/cc-gvSnbS.o"</span>
+<span class="gp">#</span> <span class="s2">"i386-apple-darwin9"</span> - <span class="s2">"darwin::Link"</span>, inputs: <span class="o">[</span><span class="s2">"/tmp/cc-gvSnbS.o"</span><span class="o">]</span>, output: <span class="s2">"/tmp/cc-jgHQxi.out"</span>
+<span class="gp">#</span> <span class="s2">"ppc-apple-darwin9"</span> - <span class="s2">"gcc::Compile"</span>, inputs: <span class="o">[</span><span class="s2">"t0.c"</span><span class="o">]</span>, output: <span class="s2">"/tmp/cc-Q0bTox.s"</span>
+<span class="gp">#</span> <span class="s2">"ppc-apple-darwin9"</span> - <span class="s2">"gcc::Assemble"</span>, inputs: <span class="o">[</span><span class="s2">"/tmp/cc-Q0bTox.s"</span><span class="o">]</span>, output: <span class="s2">"/tmp/cc-WCdicw.o"</span>
+<span class="gp">#</span> <span class="s2">"ppc-apple-darwin9"</span> - <span class="s2">"gcc::Link"</span>, inputs: <span class="o">[</span><span class="s2">"/tmp/cc-WCdicw.o"</span><span class="o">]</span>, output: <span class="s2">"/tmp/cc-HHBEBh.out"</span>
+<span class="gp">#</span> <span class="s2">"i386-apple-darwin9"</span> - <span class="s2">"darwin::Lipo"</span>, inputs: <span class="o">[</span><span class="s2">"/tmp/cc-jgHQxi.out"</span>, <span class="s2">"/tmp/cc-HHBEBh.out"</span><span class="o">]</span>, output: <span class="s2">"a.out"</span>
+</pre></div>
+</div>
+<p>This shows the tool chain, tool, inputs and outputs which have been
+bound for this compilation sequence. Here clang is being used to
+compile t0.c on the i386 architecture and darwin specific versions of
+the tools are being used to assemble and link the result, but generic
+gcc versions of the tools are being used on PowerPC.</p>
+</li>
+<li><p class="first"><strong>Translate: Tool Specific Argument Translation</strong></p>
+<p>Once a Tool has been selected to perform a particular Action, the
+Tool must construct concrete Commands which will be executed during
+compilation. The main work is in translating from the gcc style
+command line options to whatever options the subprocess expects.</p>
+<p>Some tools, such as the assembler, only interact with a handful of
+arguments and just determine the path of the executable to call and
+pass on their input and output arguments. Others, like the compiler
+or the linker, may translate a large number of arguments in addition.</p>
+<p>The ArgList class provides a number of simple helper methods to
+assist with translating arguments; for example, to pass on only the
+last of arguments corresponding to some option, or all arguments for
+an option.</p>
+<p>The result of this stage is a list of Commands (executable paths and
+argument strings) to execute.</p>
+</li>
+<li><p class="first"><strong>Execute</strong></p>
+<p>Finally, the compilation pipeline is executed. This is mostly
+straightforward, although there is some interaction with options like
+<tt class="docutils literal"><span class="pre">-pipe</span></tt>, <tt class="docutils literal"><span class="pre">-pass-exit-codes</span></tt> and <tt class="docutils literal"><span class="pre">-time</span></tt>.</p>
+</li>
+</ol>
+</div>
+<div class="section" id="additional-notes">
+<h3><a class="toc-backref" href="#id25">Additional Notes</a><a class="headerlink" href="#additional-notes" title="Permalink to this headline">¶</a></h3>
+<div class="section" id="the-compilation-object">
+<h4><a class="toc-backref" href="#id13">The Compilation Object</a><a class="headerlink" href="#the-compilation-object" title="Permalink to this headline">¶</a></h4>
+<p>The driver constructs a Compilation object for each set of command line
+arguments. The Driver itself is intended to be invariant during
+construction of a Compilation; an IDE should be able to construct a
+single long lived driver instance to use for an entire build, for
+example.</p>
+<p>The Compilation object holds information that is particular to each
+compilation sequence. For example, the list of used temporary files
+(which must be removed once compilation is finished) and result files
+(which should be removed if compilation fails).</p>
+</div>
+<div class="section" id="unified-parsing-pipelining">
+<h4><a class="toc-backref" href="#id14">Unified Parsing & Pipelining</a><a class="headerlink" href="#unified-parsing-pipelining" title="Permalink to this headline">¶</a></h4>
+<p>Parsing and pipelining both occur without reference to a Compilation
+instance. This is by design; the driver expects that both of these
+phases are platform neutral, with a few very well defined exceptions
+such as whether the platform uses a driver driver.</p>
+</div>
+<div class="section" id="toolchain-argument-translation">
+<h4><a class="toc-backref" href="#id15">ToolChain Argument Translation</a><a class="headerlink" href="#toolchain-argument-translation" title="Permalink to this headline">¶</a></h4>
+<p>In order to match gcc very closely, the clang driver currently allows
+tool chains to perform their own translation of the argument list (into
+a new ArgList data structure). Although this allows the clang driver to
+match gcc easily, it also makes the driver operation much harder to
+understand (since the Tools stop seeing some arguments the user
+provided, and see new ones instead).</p>
+<p>For example, on Darwin <tt class="docutils literal"><span class="pre">-gfull</span></tt> gets translated into two separate
+arguments, <tt class="docutils literal"><span class="pre">-g</span></tt> and <tt class="docutils literal"><span class="pre">-fno-eliminate-unused-debug-symbols</span></tt>. Trying to
+write Tool logic to do something with <tt class="docutils literal"><span class="pre">-gfull</span></tt> will not work, because
+Tool argument translation is done after the arguments have been
+translated.</p>
+<p>A long term goal is to remove this tool chain specific translation, and
+instead force each tool to change its own logic to do the right thing on
+the untranslated original arguments.</p>
+</div>
+<div class="section" id="unused-argument-warnings">
+<h4><a class="toc-backref" href="#id16">Unused Argument Warnings</a><a class="headerlink" href="#unused-argument-warnings" title="Permalink to this headline">¶</a></h4>
+<p>The driver operates by parsing all arguments but giving Tools the
+opportunity to choose which arguments to pass on. One downside of this
+infrastructure is that if the user misspells some option, or is confused
+about which options to use, some command line arguments the user really
+cared about may go unused. This problem is particularly important when
+using clang as a compiler, since the clang compiler does not support
+anywhere near all the options that gcc does, and we want to make sure
+users know which ones are being used.</p>
+<p>To support this, the driver maintains a bit associated with each
+argument of whether it has been used (at all) during the compilation.
+This bit usually doesn’t need to be set by hand, as the key ArgList
+accessors will set it automatically.</p>
+<p>When a compilation is successful (there are no errors), the driver
+checks the bit and emits an “unused argument” warning for any arguments
+which were never accessed. This is conservative (the argument may not
+have been used to do what the user wanted) but still catches the most
+obvious cases.</p>
+</div>
+</div>
+<div class="section" id="relation-to-gcc-driver-concepts">
+<h3><a class="toc-backref" href="#id26">Relation to GCC Driver Concepts</a><a class="headerlink" href="#relation-to-gcc-driver-concepts" title="Permalink to this headline">¶</a></h3>
+<p>For those familiar with the gcc driver, this section provides a brief
+overview of how things from the gcc driver map to the clang driver.</p>
+<ul>
+<li><p class="first"><strong>Driver Driver</strong></p>
+<p>The driver driver is fully integrated into the clang driver. The
+driver simply constructs additional Actions to bind the architecture
+during the <em>Pipeline</em> phase. The tool chain specific argument
+translation is responsible for handling <tt class="docutils literal"><span class="pre">-Xarch_</span></tt>.</p>
+<p>The one caveat is that this approach requires <tt class="docutils literal"><span class="pre">-Xarch_</span></tt> not be used
+to alter the compilation itself (for example, one cannot provide
+<tt class="docutils literal"><span class="pre">-S</span></tt> as an <tt class="docutils literal"><span class="pre">-Xarch_</span></tt> argument). The driver attempts to reject
+such invocations, and overall there isn’t a good reason to abuse
+<tt class="docutils literal"><span class="pre">-Xarch_</span></tt> to that end in practice.</p>
+<p>The upside is that the clang driver is more efficient and does little
+extra work to support universal builds. It also provides better error
+reporting and UI consistency.</p>
+</li>
+<li><p class="first"><strong>Specs</strong></p>
+<p>The clang driver has no direct correspondent for “specs”. The
+majority of the functionality that is embedded in specs is in the
+Tool specific argument translation routines. The parts of specs which
+control the compilation pipeline are generally part of the <em>Pipeline</em>
+stage.</p>
+</li>
+<li><p class="first"><strong>Toolchains</strong></p>
+<p>The gcc driver has no direct understanding of tool chains. Each gcc
+binary roughly corresponds to the information which is embedded
+inside a single ToolChain.</p>
+<p>The clang driver is intended to be portable and support complex
+compilation environments. All platform and tool chain specific code
+should be protected behind either abstract or well defined interfaces
+(such as whether the platform supports use as a driver driver).</p>
+</li>
+</ul>
+</div>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="InternalsManual.html">“Clang” CFE Internals Manual</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="PTHInternals.html">Pretokenized Headers (PTH)</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/ExternalClangExamples.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/ExternalClangExamples.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/ExternalClangExamples.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/ExternalClangExamples.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,141 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>External Clang Examples — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Introduction to the Clang AST" href="IntroductionToTheClangAST.html" />
+ <link rel="prev" title="Choosing the Right Interface for Your Application" href="Tooling.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>External Clang Examples</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="Tooling.html">Choosing the Right Interface for Your Application</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="IntroductionToTheClangAST.html">Introduction to the Clang AST</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="external-clang-examples">
+<h1>External Clang Examples<a class="headerlink" href="#external-clang-examples" title="Permalink to this headline">¶</a></h1>
+<div class="section" id="introduction">
+<h2>Introduction<a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>This page provides some examples of the kinds of things that people have
+done with Clang that might serve as useful guides (or starting points) from
+which to develop your own tools. They may be helpful even for something as
+banal (but necessary) as how to set up your build to integrate Clang.</p>
+<p>Clang’s library-based design is deliberately aimed at facilitating use by
+external projects, and we are always interested in improving Clang to
+better serve our external users. Some typical categories of applications
+where Clang is used are:</p>
+<ul class="simple">
+<li>Static analysis.</li>
+<li>Documentation/cross-reference generation.</li>
+</ul>
+<p>If you know of (or wrote!) a tool or project using Clang, please send an
+email to Clang’s <a class="reference external" href="http://lists.llvm.org/mailman/listinfo/cfe-dev">development discussion mailing list</a> to have it added.
+(or if you are already a Clang contributor, feel free to directly commit
+additions). Since the primary purpose of this page is to provide examples
+that can help developers, generally they must have code available.</p>
+</div>
+<div class="section" id="list-of-projects-and-tools">
+<h2>List of projects and tools<a class="headerlink" href="#list-of-projects-and-tools" title="Permalink to this headline">¶</a></h2>
+<dl class="docutils">
+<dt><a class="reference external" href="https://github.com/Andersbakken/rtags/">https://github.com/Andersbakken/rtags/</a></dt>
+<dd>“RTags is a client/server application that indexes c/c++ code and keeps
+a persistent in-memory database of references, symbolnames, completions
+etc.”</dd>
+<dt><a class="reference external" href="http://rprichard.github.com/sourceweb/">http://rprichard.github.com/sourceweb/</a></dt>
+<dd>“A C/C++ source code indexer and navigator”</dd>
+<dt><a class="reference external" href="https://github.com/etaoins/qconnectlint">https://github.com/etaoins/qconnectlint</a></dt>
+<dd>“qconnectlint is a Clang tool for statically verifying the consistency
+of signal and slot connections made with Qt’s <tt class="docutils literal"><span class="pre">QObject::connect</span></tt>.”</dd>
+<dt><a class="reference external" href="https://github.com/woboq/woboq_codebrowser">https://github.com/woboq/woboq_codebrowser</a></dt>
+<dd>“The Woboq Code Browser is a web-based code browser for C/C++ projects.
+Check out <a class="reference external" href="http://code.woboq.org/">http://code.woboq.org/</a> for an example!”</dd>
+<dt><a class="reference external" href="https://github.com/mozilla/dxr">https://github.com/mozilla/dxr</a></dt>
+<dd>“DXR is a source code cross-reference tool that uses static analysis
+data collected by instrumented compilers.”</dd>
+<dt><a class="reference external" href="https://github.com/eschulte/clang-mutate">https://github.com/eschulte/clang-mutate</a></dt>
+<dd>“This tool performs a number of operations on C-language source files.”</dd>
+<dt><a class="reference external" href="https://github.com/gmarpons/Crisp">https://github.com/gmarpons/Crisp</a></dt>
+<dd>“A coding rule validation add-on for LLVM/clang. Crisp rules are written
+in Prolog. A high-level declarative DSL to easily write new rules is under
+development. It will be called CRISP, an acronym for <em>Coding Rules in
+Sugared Prolog</em>.”</dd>
+<dt><a class="reference external" href="https://github.com/drothlis/clang-ctags">https://github.com/drothlis/clang-ctags</a></dt>
+<dd>“Generate tag file for C++ source code.”</dd>
+<dt><a class="reference external" href="https://github.com/exclipy/clang_indexer">https://github.com/exclipy/clang_indexer</a></dt>
+<dd>“This is an indexer for C and C++ based on the libclang library.”</dd>
+<dt><a class="reference external" href="https://github.com/holtgrewe/linty">https://github.com/holtgrewe/linty</a></dt>
+<dd>“Linty - C/C++ Style Checking with Python & libclang.”</dd>
+<dt><a class="reference external" href="https://github.com/axw/cmonster">https://github.com/axw/cmonster</a></dt>
+<dd>“cmonster is a Python wrapper for the Clang C++ parser.”</dd>
+<dt><a class="reference external" href="https://github.com/rizsotto/Constantine">https://github.com/rizsotto/Constantine</a></dt>
+<dd>“Constantine is a toy project to learn how to write clang plugin.
+Implements pseudo const analysis. Generates warnings about variables,
+which were declared without const qualifier.”</dd>
+<dt><a class="reference external" href="https://github.com/jessevdk/cldoc">https://github.com/jessevdk/cldoc</a></dt>
+<dd>“cldoc is a Clang based documentation generator for C and C++.
+cldoc tries to solve the issue of writing C/C++ software documentation
+with a modern, non-intrusive and robust approach.”</dd>
+<dt><a class="reference external" href="https://github.com/AlexDenisov/ToyClangPlugin">https://github.com/AlexDenisov/ToyClangPlugin</a></dt>
+<dd>“The simplest Clang plugin implementing a semantic check for Objective-C.
+This example shows how to use the <tt class="docutils literal"><span class="pre">DiagnosticsEngine</span></tt> (emit warnings,
+errors, fixit hints). See also <a class="reference external" href="http://l.rw.rw/clang_plugin">http://l.rw.rw/clang_plugin</a> for
+step-by-step instructions.”</dd>
+</dl>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="Tooling.html">Choosing the Right Interface for Your Application</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="IntroductionToTheClangAST.html">Introduction to the Clang AST</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/FAQ.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/FAQ.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/FAQ.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/FAQ.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,127 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Frequently Asked Questions (FAQ) — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Choosing the Right Interface for Your Application" href="Tooling.html" />
+ <link rel="prev" title="clang - the Clang C, C++, and Objective-C compiler" href="CommandGuide/clang.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Frequently Asked Questions (FAQ)</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="CommandGuide/clang.html">clang - the Clang C, C++, and Objective-C compiler</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="Tooling.html">Choosing the Right Interface for Your Application</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="frequently-asked-questions-faq">
+<h1>Frequently Asked Questions (FAQ)<a class="headerlink" href="#frequently-asked-questions-faq" title="Permalink to this headline">¶</a></h1>
+<div class="contents local topic" id="contents">
+<ul class="simple">
+<li><a class="reference internal" href="#driver" id="id1">Driver</a><ul>
+<li><a class="reference internal" href="#i-run-clang-cc1-and-get-weird-errors-about-missing-headers" id="id2">I run <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span> <span class="pre">...</span></tt> and get weird errors about missing headers</a></li>
+<li><a class="reference internal" href="#i-get-errors-about-some-headers-being-missing-stddef-h-stdarg-h" id="id3">I get errors about some headers being missing (<tt class="docutils literal"><span class="pre">stddef.h</span></tt>, <tt class="docutils literal"><span class="pre">stdarg.h</span></tt>)</a></li>
+</ul>
+</li>
+</ul>
+</div>
+<div class="section" id="driver">
+<h2><a class="toc-backref" href="#id1">Driver</a><a class="headerlink" href="#driver" title="Permalink to this headline">¶</a></h2>
+<div class="section" id="i-run-clang-cc1-and-get-weird-errors-about-missing-headers">
+<h3><a class="toc-backref" href="#id2">I run <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span> <span class="pre">...</span></tt> and get weird errors about missing headers</a><a class="headerlink" href="#i-run-clang-cc1-and-get-weird-errors-about-missing-headers" title="Permalink to this headline">¶</a></h3>
+<p>Given this source file:</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="cp">#include <stdio.h></span>
+
+<span class="kt">int</span> <span class="nf">main</span><span class="p">()</span> <span class="p">{</span>
+ <span class="n">printf</span><span class="p">(</span><span class="s">"Hello world</span><span class="se">\n</span><span class="s">"</span><span class="p">);</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>If you run:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> clang -cc1 hello.c
+<span class="go">hello.c:1:10: fatal error: 'stdio.h' file not found</span>
+<span class="gp">#</span>include <stdio.h>
+<span class="go"> ^</span>
+<span class="go">1 error generated.</span>
+</pre></div>
+</div>
+<p><tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span></tt> is the frontend, <tt class="docutils literal"><span class="pre">clang</span></tt> is the <a class="reference internal" href="DriverInternals.html"><em>driver</em></a>. The driver invokes the frontend with options appropriate
+for your system. To see these options, run:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> clang -### -c hello.c
+</pre></div>
+</div>
+<p>Some clang command line options are driver-only options, some are frontend-only
+options. Frontend-only options are intended to be used only by clang developers.
+Users should not run <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span></tt> directly, because <tt class="docutils literal"><span class="pre">-cc1</span></tt> options are not
+guaranteed to be stable.</p>
+<p>If you want to use a frontend-only option (“a <tt class="docutils literal"><span class="pre">-cc1</span></tt> option”), for example
+<tt class="docutils literal"><span class="pre">-ast-dump</span></tt>, then you need to take the <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span></tt> line generated by the
+driver and add the option you need. Alternatively, you can run
+<tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-Xclang</span> <span class="pre"><option></span> <span class="pre">...</span></tt> to force the driver pass <tt class="docutils literal"><span class="pre"><option></span></tt> to
+<tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span></tt>.</p>
+</div>
+<div class="section" id="i-get-errors-about-some-headers-being-missing-stddef-h-stdarg-h">
+<h3><a class="toc-backref" href="#id3">I get errors about some headers being missing (<tt class="docutils literal"><span class="pre">stddef.h</span></tt>, <tt class="docutils literal"><span class="pre">stdarg.h</span></tt>)</a><a class="headerlink" href="#i-get-errors-about-some-headers-being-missing-stddef-h-stdarg-h" title="Permalink to this headline">¶</a></h3>
+<p>Some header files (<tt class="docutils literal"><span class="pre">stddef.h</span></tt>, <tt class="docutils literal"><span class="pre">stdarg.h</span></tt>, and others) are shipped with
+Clang — these are called builtin includes. Clang searches for them in a
+directory relative to the location of the <tt class="docutils literal"><span class="pre">clang</span></tt> binary. If you moved the
+<tt class="docutils literal"><span class="pre">clang</span></tt> binary, you need to move the builtin headers, too.</p>
+<p>More information can be found in the <a class="reference internal" href="LibTooling.html#libtooling-builtin-includes"><em>Builtin includes</em></a>
+section.</p>
+</div>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="CommandGuide/clang.html">clang - the Clang C, C++, and Objective-C compiler</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="Tooling.html">Choosing the Right Interface for Your Application</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/HowToSetupToolingForLLVM.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/HowToSetupToolingForLLVM.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/HowToSetupToolingForLLVM.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/HowToSetupToolingForLLVM.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,237 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>How To Setup Clang Tooling For LLVM — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="JSON Compilation Database Format Specification" href="JSONCompilationDatabase.html" />
+ <link rel="prev" title="Matching the Clang AST" href="LibASTMatchers.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>How To Setup Clang Tooling For LLVM</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="LibASTMatchers.html">Matching the Clang AST</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="JSONCompilationDatabase.html">JSON Compilation Database Format Specification</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="how-to-setup-clang-tooling-for-llvm">
+<h1>How To Setup Clang Tooling For LLVM<a class="headerlink" href="#how-to-setup-clang-tooling-for-llvm" title="Permalink to this headline">¶</a></h1>
+<p>Clang Tooling provides infrastructure to write tools that need syntactic
+and semantic information about a program. This term also relates to a set
+of specific tools using this infrastructure (e.g. <tt class="docutils literal"><span class="pre">clang-check</span></tt>). This
+document provides information on how to set up and use Clang Tooling for
+the LLVM source code.</p>
+<div class="section" id="introduction">
+<h2>Introduction<a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>Clang Tooling needs a compilation database to figure out specific build
+options for each file. Currently it can create a compilation database
+from the <tt class="docutils literal"><span class="pre">compilation_commands.json</span></tt> file, generated by CMake. When
+invoking clang tools, you can either specify a path to a build directory
+using a command line parameter <tt class="docutils literal"><span class="pre">-p</span></tt> or let Clang Tooling find this
+file in your source tree. In either case you need to configure your
+build using CMake to use clang tools.</p>
+</div>
+<div class="section" id="setup-clang-tooling-using-cmake-and-make">
+<h2>Setup Clang Tooling Using CMake and Make<a class="headerlink" href="#setup-clang-tooling-using-cmake-and-make" title="Permalink to this headline">¶</a></h2>
+<p>If you intend to use make to build LLVM, you should have CMake 2.8.6 or
+later installed (can be found <a class="reference external" href="http://cmake.org">here</a>).</p>
+<p>First, you need to generate Makefiles for LLVM with CMake. You need to
+make a build directory and run CMake from it:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> mkdir your/build/directory
+<span class="gp">$</span> <span class="nb">cd </span>your/build/directory
+<span class="gp">$</span> cmake -DCMAKE_EXPORT_COMPILE_COMMANDS<span class="o">=</span>ON path/to/llvm/sources
+</pre></div>
+</div>
+<p>If you want to use clang instead of GCC, you can add
+<tt class="docutils literal"><span class="pre">-DCMAKE_C_COMPILER=/path/to/clang</span> <span class="pre">-DCMAKE_CXX_COMPILER=/path/to/clang++</span></tt>.
+You can also use <tt class="docutils literal"><span class="pre">ccmake</span></tt>, which provides a curses interface to configure
+CMake variables for lazy people.</p>
+<p>As a result, the new <tt class="docutils literal"><span class="pre">compile_commands.json</span></tt> file should appear in the
+current directory. You should link it to the LLVM source tree so that
+Clang Tooling is able to use it:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> ln -s <span class="nv">$PWD</span>/compile_commands.json path/to/llvm/source/
+</pre></div>
+</div>
+<p>Now you are ready to build and test LLVM using make:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> make check-all
+</pre></div>
+</div>
+</div>
+<div class="section" id="using-clang-tools">
+<h2>Using Clang Tools<a class="headerlink" href="#using-clang-tools" title="Permalink to this headline">¶</a></h2>
+<p>After you completed the previous steps, you are ready to run clang tools. If
+you have a recent clang installed, you should have <tt class="docutils literal"><span class="pre">clang-check</span></tt> in
+<tt class="docutils literal"><span class="pre">$PATH</span></tt>. Try to run it on any <tt class="docutils literal"><span class="pre">.cpp</span></tt> file inside the LLVM source tree:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> clang-check tools/clang/lib/Tooling/CompilationDatabase.cpp
+</pre></div>
+</div>
+<p>If you’re using vim, it’s convenient to have clang-check integrated. Put
+this into your <tt class="docutils literal"><span class="pre">.vimrc</span></tt>:</p>
+<div class="highlight-python"><div class="highlight"><pre>function! ClangCheckImpl(cmd)
+ if &autowrite | wall | endif
+ echo "Running " . a:cmd . " ..."
+ let l:output = system(a:cmd)
+ cexpr l:output
+ cwindow
+ let w:quickfix_title = a:cmd
+ if v:shell_error != 0
+ cc
+ endif
+ let g:clang_check_last_cmd = a:cmd
+endfunction
+
+function! ClangCheck()
+ let l:filename = expand('%')
+ if l:filename =~ '\.\(cpp\|cxx\|cc\|c\)$'
+ call ClangCheckImpl("clang-check " . l:filename)
+ elseif exists("g:clang_check_last_cmd")
+ call ClangCheckImpl(g:clang_check_last_cmd)
+ else
+ echo "Can't detect file's compilation arguments and no previous clang-check invocation!"
+ endif
+endfunction
+
+nmap <silent> <F5> :call ClangCheck()<CR><CR>
+</pre></div>
+</div>
+<p>When editing a .cpp/.cxx/.cc/.c file, hit F5 to reparse the file. In
+case the current file has a different extension (for example, .h), F5
+will re-run the last clang-check invocation made from this vim instance
+(if any). The output will go into the error window, which is opened
+automatically when clang-check finds errors, and can be re-opened with
+<tt class="docutils literal"><span class="pre">:cope</span></tt>.</p>
+<p>Other <tt class="docutils literal"><span class="pre">clang-check</span></tt> options that can be useful when working with clang
+AST:</p>
+<ul class="simple">
+<li><tt class="docutils literal"><span class="pre">-ast-print</span></tt> — Build ASTs and then pretty-print them.</li>
+<li><tt class="docutils literal"><span class="pre">-ast-dump</span></tt> — Build ASTs and then debug dump them.</li>
+<li><tt class="docutils literal"><span class="pre">-ast-dump-filter=<string></span></tt> — Use with <tt class="docutils literal"><span class="pre">-ast-dump</span></tt> or <tt class="docutils literal"><span class="pre">-ast-print</span></tt> to
+dump/print only AST declaration nodes having a certain substring in a
+qualified name. Use <tt class="docutils literal"><span class="pre">-ast-list</span></tt> to list all filterable declaration node
+names.</li>
+<li><tt class="docutils literal"><span class="pre">-ast-list</span></tt> — Build ASTs and print the list of declaration node qualified
+names.</li>
+</ul>
+<p>Examples:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> clang-check tools/clang/tools/clang-check/ClangCheck.cpp -ast-dump -ast-dump-filter ActionFactory::newASTConsumer
+<span class="go">Processing: tools/clang/tools/clang-check/ClangCheck.cpp.</span>
+<span class="go">Dumping ::ActionFactory::newASTConsumer:</span>
+<span class="go">clang::ASTConsumer *newASTConsumer() (CompoundStmt 0x44da290 </home/alexfh/local/llvm/tools/clang/tools/clang-check/ClangCheck.cpp:64:40, line:72:3></span>
+<span class="go"> (IfStmt 0x44d97c8 <line:65:5, line:66:45></span>
+<span class="go"> <<<NULL>>></span>
+<span class="go"> (ImplicitCastExpr 0x44d96d0 <line:65:9> '_Bool':'_Bool' <UserDefinedConversion></span>
+<span class="go">...</span>
+<span class="gp">$</span> clang-check tools/clang/tools/clang-check/ClangCheck.cpp -ast-print -ast-dump-filter ActionFactory::newASTConsumer
+<span class="go">Processing: tools/clang/tools/clang-check/ClangCheck.cpp.</span>
+<span class="go">Printing <anonymous namespace>::ActionFactory::newASTConsumer:</span>
+<span class="go">clang::ASTConsumer *newASTConsumer() {</span>
+<span class="go"> if (this->ASTList.operator _Bool())</span>
+<span class="go"> return clang::CreateASTDeclNodeLister();</span>
+<span class="go"> if (this->ASTDump.operator _Bool())</span>
+<span class="go"> return clang::CreateASTDumper(this->ASTDumpFilter);</span>
+<span class="go"> if (this->ASTPrint.operator _Bool())</span>
+<span class="go"> return clang::CreateASTPrinter(&llvm::outs(), this->ASTDumpFilter);</span>
+<span class="go"> return new clang::ASTConsumer();</span>
+<span class="go">}</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="experimental-using-ninja-build-system">
+<h2>(Experimental) Using Ninja Build System<a class="headerlink" href="#experimental-using-ninja-build-system" title="Permalink to this headline">¶</a></h2>
+<p>Optionally you can use the <a class="reference external" href="https://github.com/martine/ninja">Ninja</a>
+build system instead of make. It is aimed at making your builds faster.
+Currently this step will require building Ninja from sources.</p>
+<p>To take advantage of using Clang Tools along with Ninja build you need
+at least CMake 2.8.9.</p>
+<p>Clone the Ninja git repository and build Ninja from sources:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> git clone git://github.com/martine/ninja.git
+<span class="gp">$</span> <span class="nb">cd </span>ninja/
+<span class="gp">$</span> ./bootstrap.py
+</pre></div>
+</div>
+<p>This will result in a single binary <tt class="docutils literal"><span class="pre">ninja</span></tt> in the current directory.
+It doesn’t require installation and can just be copied to any location
+inside <tt class="docutils literal"><span class="pre">$PATH</span></tt>, say <tt class="docutils literal"><span class="pre">/usr/local/bin/</span></tt>:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> sudo cp ninja /usr/local/bin/
+<span class="gp">$</span> sudo chmod a+rx /usr/local/bin/ninja
+</pre></div>
+</div>
+<p>After doing all of this, you’ll need to generate Ninja build files for
+LLVM with CMake. You need to make a build directory and run CMake from
+it:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> mkdir your/build/directory
+<span class="gp">$</span> <span class="nb">cd </span>your/build/directory
+<span class="gp">$</span> cmake -G Ninja -DCMAKE_EXPORT_COMPILE_COMMANDS<span class="o">=</span>ON path/to/llvm/sources
+</pre></div>
+</div>
+<p>If you want to use clang instead of GCC, you can add
+<tt class="docutils literal"><span class="pre">-DCMAKE_C_COMPILER=/path/to/clang</span> <span class="pre">-DCMAKE_CXX_COMPILER=/path/to/clang++</span></tt>.
+You can also use <tt class="docutils literal"><span class="pre">ccmake</span></tt>, which provides a curses interface to configure
+CMake variables in an interactive manner.</p>
+<p>As a result, the new <tt class="docutils literal"><span class="pre">compile_commands.json</span></tt> file should appear in the
+current directory. You should link it to the LLVM source tree so that
+Clang Tooling is able to use it:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> ln -s <span class="nv">$PWD</span>/compile_commands.json path/to/llvm/source/
+</pre></div>
+</div>
+<p>Now you are ready to build and test LLVM using Ninja:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> ninja check-all
+</pre></div>
+</div>
+<p>Other target names can be used in the same way as with make.</p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="LibASTMatchers.html">Matching the Clang AST</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="JSONCompilationDatabase.html">JSON Compilation Database Format Specification</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/InternalsManual.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/InternalsManual.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/InternalsManual.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/InternalsManual.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,2018 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>âClangâ CFE Internals Manual — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Driver Design & Internals" href="DriverInternals.html" />
+ <link rel="prev" title="Clang-Format Style Options" href="ClangFormatStyleOptions.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>âClangâ CFE Internals Manual</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="ClangFormatStyleOptions.html">Clang-Format Style Options</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="DriverInternals.html">Driver Design & Internals</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="clang-cfe-internals-manual">
+<h1>“Clang” CFE Internals Manual<a class="headerlink" href="#clang-cfe-internals-manual" title="Permalink to this headline">¶</a></h1>
+<div class="contents local topic" id="contents">
+<ul class="simple">
+<li><a class="reference internal" href="#introduction" id="id3">Introduction</a></li>
+<li><a class="reference internal" href="#llvm-support-library" id="id4">LLVM Support Library</a></li>
+<li><a class="reference internal" href="#the-clang-basic-library" id="id5">The Clang “Basic” Library</a><ul>
+<li><a class="reference internal" href="#the-diagnostics-subsystem" id="id6">The Diagnostics Subsystem</a><ul>
+<li><a class="reference internal" href="#the-diagnostic-kinds-td-files" id="id7">The <tt class="docutils literal"><span class="pre">Diagnostic*Kinds.td</span></tt> files</a></li>
+<li><a class="reference internal" href="#the-format-string" id="id8">The Format String</a></li>
+<li><a class="reference internal" href="#formatting-a-diagnostic-argument" id="id9">Formatting a Diagnostic Argument</a></li>
+<li><a class="reference internal" href="#producing-the-diagnostic" id="id10">Producing the Diagnostic</a></li>
+<li><a class="reference internal" href="#fix-it-hints" id="id11">Fix-It Hints</a></li>
+<li><a class="reference internal" href="#the-diagnosticclient-interface" id="id12">The <tt class="docutils literal"><span class="pre">DiagnosticClient</span></tt> Interface</a></li>
+<li><a class="reference internal" href="#adding-translations-to-clang" id="id13">Adding Translations to Clang</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#the-sourcelocation-and-sourcemanager-classes" id="id14">The <tt class="docutils literal"><span class="pre">SourceLocation</span></tt> and <tt class="docutils literal"><span class="pre">SourceManager</span></tt> classes</a></li>
+<li><a class="reference internal" href="#sourcerange-and-charsourcerange" id="id15"><tt class="docutils literal"><span class="pre">SourceRange</span></tt> and <tt class="docutils literal"><span class="pre">CharSourceRange</span></tt></a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#the-driver-library" id="id16">The Driver Library</a></li>
+<li><a class="reference internal" href="#precompiled-headers" id="id17">Precompiled Headers</a></li>
+<li><a class="reference internal" href="#the-frontend-library" id="id18">The Frontend Library</a></li>
+<li><a class="reference internal" href="#the-lexer-and-preprocessor-library" id="id19">The Lexer and Preprocessor Library</a><ul>
+<li><a class="reference internal" href="#the-token-class" id="id20">The Token class</a></li>
+<li><a class="reference internal" href="#annotation-tokens" id="id21">Annotation Tokens</a></li>
+<li><a class="reference internal" href="#the-lexer-class" id="id22">The <tt class="docutils literal"><span class="pre">Lexer</span></tt> class</a></li>
+<li><a class="reference internal" href="#the-tokenlexer-class" id="id23">The <tt class="docutils literal"><span class="pre">TokenLexer</span></tt> class</a></li>
+<li><a class="reference internal" href="#the-multipleincludeopt-class" id="id24">The <tt class="docutils literal"><span class="pre">MultipleIncludeOpt</span></tt> class</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#the-parser-library" id="id25">The Parser Library</a></li>
+<li><a class="reference internal" href="#the-ast-library" id="id26">The AST Library</a><ul>
+<li><a class="reference internal" href="#the-type-class-and-its-subclasses" id="id27">The <tt class="docutils literal"><span class="pre">Type</span></tt> class and its subclasses</a><ul>
+<li><a class="reference internal" href="#canonical-types" id="id28">Canonical Types</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#the-qualtype-class" id="id29">The <tt class="docutils literal"><span class="pre">QualType</span></tt> class</a></li>
+<li><a class="reference internal" href="#declaration-names" id="id30">Declaration names</a></li>
+<li><a class="reference internal" href="#declaration-contexts" id="id31">Declaration contexts</a><ul>
+<li><a class="reference internal" href="#redeclarations-and-overloads" id="id32">Redeclarations and Overloads</a></li>
+<li><a class="reference internal" href="#lexical-and-semantic-contexts" id="id33">Lexical and Semantic Contexts</a></li>
+<li><a class="reference internal" href="#transparent-declaration-contexts" id="id34">Transparent Declaration Contexts</a></li>
+<li><a class="reference internal" href="#multiply-defined-declaration-contexts" id="id35">Multiply-Defined Declaration Contexts</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#the-cfg-class" id="id36">The <tt class="docutils literal"><span class="pre">CFG</span></tt> class</a><ul>
+<li><a class="reference internal" href="#basic-blocks" id="id37">Basic Blocks</a></li>
+<li><a class="reference internal" href="#entry-and-exit-blocks" id="id38">Entry and Exit Blocks</a></li>
+<li><a class="reference internal" href="#conditional-control-flow" id="id39">Conditional Control-Flow</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#constant-folding-in-the-clang-ast" id="id40">Constant Folding in the Clang AST</a><ul>
+<li><a class="reference internal" href="#implementation-approach" id="id41">Implementation Approach</a></li>
+<li><a class="reference internal" href="#extensions" id="id42">Extensions</a></li>
+</ul>
+</li>
+</ul>
+</li>
+<li><a class="reference internal" href="#the-sema-library" id="id43">The Sema Library</a></li>
+<li><a class="reference internal" href="#the-codegen-library" id="id44">The CodeGen Library</a></li>
+<li><a class="reference internal" href="#how-to-change-clang" id="id45">How to change Clang</a><ul>
+<li><a class="reference internal" href="#how-to-add-an-attribute" id="id46">How to add an attribute</a><ul>
+<li><a class="reference internal" href="#attribute-basics" id="id47">Attribute Basics</a></li>
+<li><a class="reference internal" href="#include-clang-basic-attr-td" id="id48"><tt class="docutils literal"><span class="pre">include/clang/Basic/Attr.td</span></tt></a><ul>
+<li><a class="reference internal" href="#spellings" id="id49">Spellings</a></li>
+<li><a class="reference internal" href="#subjects" id="id50">Subjects</a></li>
+<li><a class="reference internal" href="#documentation" id="id51">Documentation</a></li>
+<li><a class="reference internal" href="#arguments" id="id52">Arguments</a></li>
+<li><a class="reference internal" href="#other-properties" id="id53">Other Properties</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#boilerplate" id="id54">Boilerplate</a></li>
+<li><a class="reference internal" href="#semantic-handling" id="id55">Semantic handling</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#how-to-add-an-expression-or-statement" id="id56">How to add an expression or statement</a></li>
+</ul>
+</li>
+</ul>
+</div>
+<div class="section" id="introduction">
+<h2><a class="toc-backref" href="#id3">Introduction</a><a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>This document describes some of the more important APIs and internal design
+decisions made in the Clang C front-end. The purpose of this document is to
+both capture some of this high level information and also describe some of the
+design decisions behind it. This is meant for people interested in hacking on
+Clang, not for end-users. The description below is categorized by libraries,
+and does not describe any of the clients of the libraries.</p>
+</div>
+<div class="section" id="llvm-support-library">
+<h2><a class="toc-backref" href="#id4">LLVM Support Library</a><a class="headerlink" href="#llvm-support-library" title="Permalink to this headline">¶</a></h2>
+<p>The LLVM <tt class="docutils literal"><span class="pre">libSupport</span></tt> library provides many underlying libraries and
+<a class="reference external" href="http://llvm.org/docs/ProgrammersManual.html">data-structures</a>, including
+command line option processing, various containers and a system abstraction
+layer, which is used for file system access.</p>
+</div>
+<div class="section" id="the-clang-basic-library">
+<h2><a class="toc-backref" href="#id5">The Clang “Basic” Library</a><a class="headerlink" href="#the-clang-basic-library" title="Permalink to this headline">¶</a></h2>
+<p>This library certainly needs a better name. The “basic” library contains a
+number of low-level utilities for tracking and manipulating source buffers,
+locations within the source buffers, diagnostics, tokens, target abstraction,
+and information about the subset of the language being compiled for.</p>
+<p>Part of this infrastructure is specific to C (such as the <tt class="docutils literal"><span class="pre">TargetInfo</span></tt>
+class), other parts could be reused for other non-C-based languages
+(<tt class="docutils literal"><span class="pre">SourceLocation</span></tt>, <tt class="docutils literal"><span class="pre">SourceManager</span></tt>, <tt class="docutils literal"><span class="pre">Diagnostics</span></tt>, <tt class="docutils literal"><span class="pre">FileManager</span></tt>).
+When and if there is future demand we can figure out if it makes sense to
+introduce a new library, move the general classes somewhere else, or introduce
+some other solution.</p>
+<p>We describe the roles of these classes in order of their dependencies.</p>
+<div class="section" id="the-diagnostics-subsystem">
+<h3><a class="toc-backref" href="#id6">The Diagnostics Subsystem</a><a class="headerlink" href="#the-diagnostics-subsystem" title="Permalink to this headline">¶</a></h3>
+<p>The Clang Diagnostics subsystem is an important part of how the compiler
+communicates with the human. Diagnostics are the warnings and errors produced
+when the code is incorrect or dubious. In Clang, each diagnostic produced has
+(at the minimum) a unique ID, an English translation associated with it, a
+<a class="reference internal" href="#sourcelocation"><em>SourceLocation</em></a> to “put the caret”, and a severity
+(e.g., <tt class="docutils literal"><span class="pre">WARNING</span></tt> or <tt class="docutils literal"><span class="pre">ERROR</span></tt>). They can also optionally include a number of
+arguments to the dianostic (which fill in “%0“‘s in the string) as well as a
+number of source ranges that related to the diagnostic.</p>
+<p>In this section, we’ll be giving examples produced by the Clang command line
+driver, but diagnostics can be <a class="reference internal" href="#diagnosticclient"><em>rendered in many different ways</em></a> depending on how the <tt class="docutils literal"><span class="pre">DiagnosticClient</span></tt> interface is
+implemented. A representative example of a diagnostic is:</p>
+<div class="highlight-c++"><div class="highlight"><pre>t.c:38:15: error: invalid operands to binary expression ('int *' and '_Complex float')
+P = (P-42) + Gamma*4;
+ ~~~~~~ ^ ~~~~~~~
+</pre></div>
+</div>
+<p>In this example, you can see the English translation, the severity (error), you
+can see the source location (the caret (“<tt class="docutils literal"><span class="pre">^</span></tt>”) and file/line/column info),
+the source ranges “<tt class="docutils literal"><span class="pre">~~~~</span></tt>”, arguments to the diagnostic (“<tt class="docutils literal"><span class="pre">int*</span></tt>” and
+“<tt class="docutils literal"><span class="pre">_Complex</span> <span class="pre">float</span></tt>”). You’ll have to believe me that there is a unique ID
+backing the diagnostic :).</p>
+<p>Getting all of this to happen has several steps and involves many moving
+pieces, this section describes them and talks about best practices when adding
+a new diagnostic.</p>
+<div class="section" id="the-diagnostic-kinds-td-files">
+<h4><a class="toc-backref" href="#id7">The <tt class="docutils literal"><span class="pre">Diagnostic*Kinds.td</span></tt> files</a><a class="headerlink" href="#the-diagnostic-kinds-td-files" title="Permalink to this headline">¶</a></h4>
+<p>Diagnostics are created by adding an entry to one of the
+<tt class="docutils literal"><span class="pre">clang/Basic/Diagnostic*Kinds.td</span></tt> files, depending on what library will be
+using it. From this file, <strong class="program">tblgen</strong> generates the unique ID of the
+diagnostic, the severity of the diagnostic and the English translation + format
+string.</p>
+<p>There is little sanity with the naming of the unique ID’s right now. Some
+start with <tt class="docutils literal"><span class="pre">err_</span></tt>, <tt class="docutils literal"><span class="pre">warn_</span></tt>, <tt class="docutils literal"><span class="pre">ext_</span></tt> to encode the severity into the name.
+Since the enum is referenced in the C++ code that produces the diagnostic, it
+is somewhat useful for it to be reasonably short.</p>
+<p>The severity of the diagnostic comes from the set {<tt class="docutils literal"><span class="pre">NOTE</span></tt>, <tt class="docutils literal"><span class="pre">REMARK</span></tt>,
+<tt class="docutils literal"><span class="pre">WARNING</span></tt>,
+<tt class="docutils literal"><span class="pre">EXTENSION</span></tt>, <tt class="docutils literal"><span class="pre">EXTWARN</span></tt>, <tt class="docutils literal"><span class="pre">ERROR</span></tt>}. The <tt class="docutils literal"><span class="pre">ERROR</span></tt> severity is used for
+diagnostics indicating the program is never acceptable under any circumstances.
+When an error is emitted, the AST for the input code may not be fully built.
+The <tt class="docutils literal"><span class="pre">EXTENSION</span></tt> and <tt class="docutils literal"><span class="pre">EXTWARN</span></tt> severities are used for extensions to the
+language that Clang accepts. This means that Clang fully understands and can
+represent them in the AST, but we produce diagnostics to tell the user their
+code is non-portable. The difference is that the former are ignored by
+default, and the later warn by default. The <tt class="docutils literal"><span class="pre">WARNING</span></tt> severity is used for
+constructs that are valid in the currently selected source language but that
+are dubious in some way. The <tt class="docutils literal"><span class="pre">REMARK</span></tt> severity provides generic information
+about the compilation that is not necessarily related to any dubious code. The
+<tt class="docutils literal"><span class="pre">NOTE</span></tt> level is used to staple more information onto previous diagnostics.</p>
+<p>These <em>severities</em> are mapped into a smaller set (the <tt class="docutils literal"><span class="pre">Diagnostic::Level</span></tt>
+enum, {<tt class="docutils literal"><span class="pre">Ignored</span></tt>, <tt class="docutils literal"><span class="pre">Note</span></tt>, <tt class="docutils literal"><span class="pre">Remark</span></tt>, <tt class="docutils literal"><span class="pre">Warning</span></tt>, <tt class="docutils literal"><span class="pre">Error</span></tt>, <tt class="docutils literal"><span class="pre">Fatal</span></tt>}) of
+output
+<em>levels</em> by the diagnostics subsystem based on various configuration options.
+Clang internally supports a fully fine grained mapping mechanism that allows
+you to map almost any diagnostic to the output level that you want. The only
+diagnostics that cannot be mapped are <tt class="docutils literal"><span class="pre">NOTE</span></tt>s, which always follow the
+severity of the previously emitted diagnostic and <tt class="docutils literal"><span class="pre">ERROR</span></tt>s, which can only
+be mapped to <tt class="docutils literal"><span class="pre">Fatal</span></tt> (it is not possible to turn an error into a warning, for
+example).</p>
+<p>Diagnostic mappings are used in many ways. For example, if the user specifies
+<tt class="docutils literal"><span class="pre">-pedantic</span></tt>, <tt class="docutils literal"><span class="pre">EXTENSION</span></tt> maps to <tt class="docutils literal"><span class="pre">Warning</span></tt>, if they specify
+<tt class="docutils literal"><span class="pre">-pedantic-errors</span></tt>, it turns into <tt class="docutils literal"><span class="pre">Error</span></tt>. This is used to implement
+options like <tt class="docutils literal"><span class="pre">-Wunused_macros</span></tt>, <tt class="docutils literal"><span class="pre">-Wundef</span></tt> etc.</p>
+<p>Mapping to <tt class="docutils literal"><span class="pre">Fatal</span></tt> should only be used for diagnostics that are considered so
+severe that error recovery won’t be able to recover sensibly from them (thus
+spewing a ton of bogus errors). One example of this class of error are failure
+to <tt class="docutils literal"><span class="pre">#include</span></tt> a file.</p>
+</div>
+<div class="section" id="the-format-string">
+<h4><a class="toc-backref" href="#id8">The Format String</a><a class="headerlink" href="#the-format-string" title="Permalink to this headline">¶</a></h4>
+<p>The format string for the diagnostic is very simple, but it has some power. It
+takes the form of a string in English with markers that indicate where and how
+arguments to the diagnostic are inserted and formatted. For example, here are
+some simple format strings:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="s">"binary integer literals are an extension"</span>
+<span class="s">"format string contains '</span><span class="se">\\</span><span class="s">0' within the string body"</span>
+<span class="s">"more '%%' conversions than data arguments"</span>
+<span class="s">"invalid operands to binary expression (%0 and %1)"</span>
+<span class="s">"overloaded '%0' must be a %select{unary|binary|unary or binary}2 operator"</span>
+ <span class="s">" (has %1 parameter%s1)"</span>
+</pre></div>
+</div>
+<p>These examples show some important points of format strings. You can use any
+plain ASCII character in the diagnostic string except “<tt class="docutils literal"><span class="pre">%</span></tt>” without a
+problem, but these are C strings, so you have to use and be aware of all the C
+escape sequences (as in the second example). If you want to produce a “<tt class="docutils literal"><span class="pre">%</span></tt>”
+in the output, use the “<tt class="docutils literal"><span class="pre">%%</span></tt>” escape sequence, like the third diagnostic.
+Finally, Clang uses the “<tt class="docutils literal"><span class="pre">%...[digit]</span></tt>” sequences to specify where and how
+arguments to the diagnostic are formatted.</p>
+<p>Arguments to the diagnostic are numbered according to how they are specified by
+the C++ code that <a class="reference internal" href="#internals-producing-diag"><em>produces them</em></a>, and are
+referenced by <tt class="docutils literal"><span class="pre">%0</span></tt> .. <tt class="docutils literal"><span class="pre">%9</span></tt>. If you have more than 10 arguments to your
+diagnostic, you are doing something wrong :). Unlike <tt class="docutils literal"><span class="pre">printf</span></tt>, there is no
+requirement that arguments to the diagnostic end up in the output in the same
+order as they are specified, you could have a format string with “<tt class="docutils literal"><span class="pre">%1</span> <span class="pre">%0</span></tt>”
+that swaps them, for example. The text in between the percent and digit are
+formatting instructions. If there are no instructions, the argument is just
+turned into a string and substituted in.</p>
+<p>Here are some “best practices” for writing the English format string:</p>
+<ul class="simple">
+<li>Keep the string short. It should ideally fit in the 80 column limit of the
+<tt class="docutils literal"><span class="pre">DiagnosticKinds.td</span></tt> file. This avoids the diagnostic wrapping when
+printed, and forces you to think about the important point you are conveying
+with the diagnostic.</li>
+<li>Take advantage of location information. The user will be able to see the
+line and location of the caret, so you don’t need to tell them that the
+problem is with the 4th argument to the function: just point to it.</li>
+<li>Do not capitalize the diagnostic string, and do not end it with a period.</li>
+<li>If you need to quote something in the diagnostic string, use single quotes.</li>
+</ul>
+<p>Diagnostics should never take random English strings as arguments: you
+shouldn’t use “<tt class="docutils literal"><span class="pre">you</span> <span class="pre">have</span> <span class="pre">a</span> <span class="pre">problem</span> <span class="pre">with</span> <span class="pre">%0</span></tt>” and pass in things like “<tt class="docutils literal"><span class="pre">your</span>
+<span class="pre">argument</span></tt>” or “<tt class="docutils literal"><span class="pre">your</span> <span class="pre">return</span> <span class="pre">value</span></tt>” as arguments. Doing this prevents
+<a class="reference internal" href="#internals-diag-translation"><em>translating</em></a> the Clang diagnostics to other
+languages (because they’ll get random English words in their otherwise
+localized diagnostic). The exceptions to this are C/C++ language keywords
+(e.g., <tt class="docutils literal"><span class="pre">auto</span></tt>, <tt class="docutils literal"><span class="pre">const</span></tt>, <tt class="docutils literal"><span class="pre">mutable</span></tt>, etc) and C/C++ operators (<tt class="docutils literal"><span class="pre">/=</span></tt>).
+Note that things like “pointer” and “reference” are not keywords. On the other
+hand, you <em>can</em> include anything that comes from the user’s source code,
+including variable names, types, labels, etc. The “<tt class="docutils literal"><span class="pre">select</span></tt>” format can be
+used to achieve this sort of thing in a localizable way, see below.</p>
+</div>
+<div class="section" id="formatting-a-diagnostic-argument">
+<h4><a class="toc-backref" href="#id9">Formatting a Diagnostic Argument</a><a class="headerlink" href="#formatting-a-diagnostic-argument" title="Permalink to this headline">¶</a></h4>
+<p>Arguments to diagnostics are fully typed internally, and come from a couple
+different classes: integers, types, names, and random strings. Depending on
+the class of the argument, it can be optionally formatted in different ways.
+This gives the <tt class="docutils literal"><span class="pre">DiagnosticClient</span></tt> information about what the argument means
+without requiring it to use a specific presentation (consider this MVC for
+Clang :).</p>
+<p>Here are the different diagnostic argument formats currently supported by
+Clang:</p>
+<p><strong>“s” format</strong></p>
+<dl class="docutils">
+<dt>Example:</dt>
+<dd><tt class="docutils literal"><span class="pre">"requires</span> <span class="pre">%1</span> <span class="pre">parameter%s1"</span></tt></dd>
+<dt>Class:</dt>
+<dd>Integers</dd>
+<dt>Description:</dt>
+<dd>This is a simple formatter for integers that is useful when producing English
+diagnostics. When the integer is 1, it prints as nothing. When the integer
+is not 1, it prints as “<tt class="docutils literal"><span class="pre">s</span></tt>”. This allows some simple grammatical forms to
+be to be handled correctly, and eliminates the need to use gross things like
+<tt class="docutils literal"><span class="pre">"requires</span> <span class="pre">%1</span> <span class="pre">parameter(s)"</span></tt>.</dd>
+</dl>
+<p><strong>“select” format</strong></p>
+<dl class="docutils">
+<dt>Example:</dt>
+<dd><tt class="docutils literal"><span class="pre">"must</span> <span class="pre">be</span> <span class="pre">a</span> <span class="pre">%select{unary|binary|unary</span> <span class="pre">or</span> <span class="pre">binary}2</span> <span class="pre">operator"</span></tt></dd>
+<dt>Class:</dt>
+<dd>Integers</dd>
+<dt>Description:</dt>
+<dd>This format specifier is used to merge multiple related diagnostics together
+into one common one, without requiring the difference to be specified as an
+English string argument. Instead of specifying the string, the diagnostic
+gets an integer argument and the format string selects the numbered option.
+In this case, the “<tt class="docutils literal"><span class="pre">%2</span></tt>” value must be an integer in the range [0..2]. If
+it is 0, it prints “unary”, if it is 1 it prints “binary” if it is 2, it
+prints “unary or binary”. This allows other language translations to
+substitute reasonable words (or entire phrases) based on the semantics of the
+diagnostic instead of having to do things textually. The selected string
+does undergo formatting.</dd>
+</dl>
+<p><strong>“plural” format</strong></p>
+<dl class="docutils">
+<dt>Example:</dt>
+<dd><tt class="docutils literal"><span class="pre">"you</span> <span class="pre">have</span> <span class="pre">%1</span> <span class="pre">%plural{1:mouse|:mice}1</span> <span class="pre">connected</span> <span class="pre">to</span> <span class="pre">your</span> <span class="pre">computer"</span></tt></dd>
+<dt>Class:</dt>
+<dd>Integers</dd>
+<dt>Description:</dt>
+<dd><p class="first">This is a formatter for complex plural forms. It is designed to handle even
+the requirements of languages with very complex plural forms, as many Baltic
+languages have. The argument consists of a series of expression/form pairs,
+separated by ”:”, where the first form whose expression evaluates to true is
+the result of the modifier.</p>
+<p>An expression can be empty, in which case it is always true. See the example
+at the top. Otherwise, it is a series of one or more numeric conditions,
+separated by ”,”. If any condition matches, the expression matches. Each
+numeric condition can take one of three forms.</p>
+<ul class="simple">
+<li>number: A simple decimal number matches if the argument is the same as the
+number. Example: <tt class="docutils literal"><span class="pre">"%plural{1:mouse|:mice}4"</span></tt></li>
+<li>range: A range in square brackets matches if the argument is within the
+range. Then range is inclusive on both ends. Example:
+<tt class="docutils literal"><span class="pre">"%plural{0:none|1:one|[2,5]:some|:many}2"</span></tt></li>
+<li>modulo: A modulo operator is followed by a number, and equals sign and
+either a number or a range. The tests are the same as for plain numbers
+and ranges, but the argument is taken modulo the number first. Example:
+<tt class="docutils literal"><span class="pre">"%plural{%100=0:even</span> <span class="pre">hundred|%100=[1,50]:lower</span> <span class="pre">half|:everything</span> <span class="pre">else}1"</span></tt></li>
+</ul>
+<p class="last">The parser is very unforgiving. A syntax error, even whitespace, will abort,
+as will a failure to match the argument against any expression.</p>
+</dd>
+</dl>
+<p><strong>“ordinal” format</strong></p>
+<dl class="docutils">
+<dt>Example:</dt>
+<dd><tt class="docutils literal"><span class="pre">"ambiguity</span> <span class="pre">in</span> <span class="pre">%ordinal0</span> <span class="pre">argument"</span></tt></dd>
+<dt>Class:</dt>
+<dd>Integers</dd>
+<dt>Description:</dt>
+<dd>This is a formatter which represents the argument number as an ordinal: the
+value <tt class="docutils literal"><span class="pre">1</span></tt> becomes <tt class="docutils literal"><span class="pre">1st</span></tt>, <tt class="docutils literal"><span class="pre">3</span></tt> becomes <tt class="docutils literal"><span class="pre">3rd</span></tt>, and so on. Values less
+than <tt class="docutils literal"><span class="pre">1</span></tt> are not supported. This formatter is currently hard-coded to use
+English ordinals.</dd>
+</dl>
+<p><strong>“objcclass” format</strong></p>
+<dl class="docutils">
+<dt>Example:</dt>
+<dd><tt class="docutils literal"><span class="pre">"method</span> <span class="pre">%objcclass0</span> <span class="pre">not</span> <span class="pre">found"</span></tt></dd>
+<dt>Class:</dt>
+<dd><tt class="docutils literal"><span class="pre">DeclarationName</span></tt></dd>
+<dt>Description:</dt>
+<dd>This is a simple formatter that indicates the <tt class="docutils literal"><span class="pre">DeclarationName</span></tt> corresponds
+to an Objective-C class method selector. As such, it prints the selector
+with a leading “<tt class="docutils literal"><span class="pre">+</span></tt>”.</dd>
+</dl>
+<p><strong>“objcinstance” format</strong></p>
+<dl class="docutils">
+<dt>Example:</dt>
+<dd><tt class="docutils literal"><span class="pre">"method</span> <span class="pre">%objcinstance0</span> <span class="pre">not</span> <span class="pre">found"</span></tt></dd>
+<dt>Class:</dt>
+<dd><tt class="docutils literal"><span class="pre">DeclarationName</span></tt></dd>
+<dt>Description:</dt>
+<dd>This is a simple formatter that indicates the <tt class="docutils literal"><span class="pre">DeclarationName</span></tt> corresponds
+to an Objective-C instance method selector. As such, it prints the selector
+with a leading “<tt class="docutils literal"><span class="pre">-</span></tt>”.</dd>
+</dl>
+<p><strong>“q” format</strong></p>
+<dl class="docutils">
+<dt>Example:</dt>
+<dd><tt class="docutils literal"><span class="pre">"candidate</span> <span class="pre">found</span> <span class="pre">by</span> <span class="pre">name</span> <span class="pre">lookup</span> <span class="pre">is</span> <span class="pre">%q0"</span></tt></dd>
+<dt>Class:</dt>
+<dd><tt class="docutils literal"><span class="pre">NamedDecl</span> <span class="pre">*</span></tt></dd>
+<dt>Description:</dt>
+<dd>This formatter indicates that the fully-qualified name of the declaration
+should be printed, e.g., “<tt class="docutils literal"><span class="pre">std::vector</span></tt>” rather than “<tt class="docutils literal"><span class="pre">vector</span></tt>”.</dd>
+</dl>
+<p><strong>“diff” format</strong></p>
+<dl class="docutils">
+<dt>Example:</dt>
+<dd><tt class="docutils literal"><span class="pre">"no</span> <span class="pre">known</span> <span class="pre">conversion</span> <span class="pre">%diff{from</span> <span class="pre">$</span> <span class="pre">to</span> <span class="pre">$|from</span> <span class="pre">argument</span> <span class="pre">type</span> <span class="pre">to</span> <span class="pre">parameter</span> <span class="pre">type}1,2"</span></tt></dd>
+<dt>Class:</dt>
+<dd><tt class="docutils literal"><span class="pre">QualType</span></tt></dd>
+<dt>Description:</dt>
+<dd>This formatter takes two <tt class="docutils literal"><span class="pre">QualType</span></tt>s and attempts to print a template
+difference between the two. If tree printing is off, the text inside the
+braces before the pipe is printed, with the formatted text replacing the $.
+If tree printing is on, the text after the pipe is printed and a type tree is
+printed after the diagnostic message.</dd>
+</dl>
+<p>It is really easy to add format specifiers to the Clang diagnostics system, but
+they should be discussed before they are added. If you are creating a lot of
+repetitive diagnostics and/or have an idea for a useful formatter, please bring
+it up on the cfe-dev mailing list.</p>
+</div>
+<div class="section" id="producing-the-diagnostic">
+<span id="internals-producing-diag"></span><h4><a class="toc-backref" href="#id10">Producing the Diagnostic</a><a class="headerlink" href="#producing-the-diagnostic" title="Permalink to this headline">¶</a></h4>
+<p>Now that you’ve created the diagnostic in the <tt class="docutils literal"><span class="pre">Diagnostic*Kinds.td</span></tt> file, you
+need to write the code that detects the condition in question and emits the new
+diagnostic. Various components of Clang (e.g., the preprocessor, <tt class="docutils literal"><span class="pre">Sema</span></tt>,
+etc.) provide a helper function named “<tt class="docutils literal"><span class="pre">Diag</span></tt>”. It creates a diagnostic and
+accepts the arguments, ranges, and other information that goes along with it.</p>
+<p>For example, the binary expression error comes from code like this:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">if</span> <span class="p">(</span><span class="n">various</span> <span class="n">things</span> <span class="n">that</span> <span class="n">are</span> <span class="n">bad</span><span class="p">)</span>
+ <span class="n">Diag</span><span class="p">(</span><span class="n">Loc</span><span class="p">,</span> <span class="n">diag</span><span class="o">::</span><span class="n">err_typecheck_invalid_operands</span><span class="p">)</span>
+ <span class="o"><<</span> <span class="n">lex</span><span class="o">-></span><span class="n">getType</span><span class="p">()</span> <span class="o"><<</span> <span class="n">rex</span><span class="o">-></span><span class="n">getType</span><span class="p">()</span>
+ <span class="o"><<</span> <span class="n">lex</span><span class="o">-></span><span class="n">getSourceRange</span><span class="p">()</span> <span class="o"><<</span> <span class="n">rex</span><span class="o">-></span><span class="n">getSourceRange</span><span class="p">();</span>
+</pre></div>
+</div>
+<p>This shows that use of the <tt class="docutils literal"><span class="pre">Diag</span></tt> method: it takes a location (a
+<a class="reference internal" href="#sourcelocation"><em>SourceLocation</em></a> object) and a diagnostic enum value
+(which matches the name from <tt class="docutils literal"><span class="pre">Diagnostic*Kinds.td</span></tt>). If the diagnostic takes
+arguments, they are specified with the <tt class="docutils literal"><span class="pre"><<</span></tt> operator: the first argument
+becomes <tt class="docutils literal"><span class="pre">%0</span></tt>, the second becomes <tt class="docutils literal"><span class="pre">%1</span></tt>, etc. The diagnostic interface
+allows you to specify arguments of many different types, including <tt class="docutils literal"><span class="pre">int</span></tt> and
+<tt class="docutils literal"><span class="pre">unsigned</span></tt> for integer arguments, <tt class="docutils literal"><span class="pre">const</span> <span class="pre">char*</span></tt> and <tt class="docutils literal"><span class="pre">std::string</span></tt> for
+string arguments, <tt class="docutils literal"><span class="pre">DeclarationName</span></tt> and <tt class="docutils literal"><span class="pre">const</span> <span class="pre">IdentifierInfo</span> <span class="pre">*</span></tt> for names,
+<tt class="docutils literal"><span class="pre">QualType</span></tt> for types, etc. <tt class="docutils literal"><span class="pre">SourceRange</span></tt>s are also specified with the
+<tt class="docutils literal"><span class="pre"><<</span></tt> operator, but do not have a specific ordering requirement.</p>
+<p>As you can see, adding and producing a diagnostic is pretty straightforward.
+The hard part is deciding exactly what you need to say to help the user,
+picking a suitable wording, and providing the information needed to format it
+correctly. The good news is that the call site that issues a diagnostic should
+be completely independent of how the diagnostic is formatted and in what
+language it is rendered.</p>
+</div>
+<div class="section" id="fix-it-hints">
+<h4><a class="toc-backref" href="#id11">Fix-It Hints</a><a class="headerlink" href="#fix-it-hints" title="Permalink to this headline">¶</a></h4>
+<p>In some cases, the front end emits diagnostics when it is clear that some small
+change to the source code would fix the problem. For example, a missing
+semicolon at the end of a statement or a use of deprecated syntax that is
+easily rewritten into a more modern form. Clang tries very hard to emit the
+diagnostic and recover gracefully in these and other cases.</p>
+<p>However, for these cases where the fix is obvious, the diagnostic can be
+annotated with a hint (referred to as a “fix-it hint”) that describes how to
+change the code referenced by the diagnostic to fix the problem. For example,
+it might add the missing semicolon at the end of the statement or rewrite the
+use of a deprecated construct into something more palatable. Here is one such
+example from the C++ front end, where we warn about the right-shift operator
+changing meaning from C++98 to C++11:</p>
+<div class="highlight-c++"><div class="highlight"><pre>test.cpp:3:7: warning: use of right-shift operator ('>>') in template argument
+ will require parentheses in C++11
+A<100 >> 2> *a;
+ ^
+ ( )
+</pre></div>
+</div>
+<p>Here, the fix-it hint is suggesting that parentheses be added, and showing
+exactly where those parentheses would be inserted into the source code. The
+fix-it hints themselves describe what changes to make to the source code in an
+abstract manner, which the text diagnostic printer renders as a line of
+“insertions” below the caret line. <a class="reference internal" href="#diagnosticclient"><em>Other diagnostic clients</em></a> might choose to render the code differently (e.g., as
+markup inline) or even give the user the ability to automatically fix the
+problem.</p>
+<p>Fix-it hints on errors and warnings need to obey these rules:</p>
+<ul class="simple">
+<li>Since they are automatically applied if <tt class="docutils literal"><span class="pre">-Xclang</span> <span class="pre">-fixit</span></tt> is passed to the
+driver, they should only be used when it’s very likely they match the user’s
+intent.</li>
+<li>Clang must recover from errors as if the fix-it had been applied.</li>
+</ul>
+<p>If a fix-it can’t obey these rules, put the fix-it on a note. Fix-its on notes
+are not applied automatically.</p>
+<p>All fix-it hints are described by the <tt class="docutils literal"><span class="pre">FixItHint</span></tt> class, instances of which
+should be attached to the diagnostic using the <tt class="docutils literal"><span class="pre"><<</span></tt> operator in the same way
+that highlighted source ranges and arguments are passed to the diagnostic.
+Fix-it hints can be created with one of three constructors:</p>
+<ul>
+<li><p class="first"><tt class="docutils literal"><span class="pre">FixItHint::CreateInsertion(Loc,</span> <span class="pre">Code)</span></tt></p>
+<blockquote>
+<div><p>Specifies that the given <tt class="docutils literal"><span class="pre">Code</span></tt> (a string) should be inserted before the
+source location <tt class="docutils literal"><span class="pre">Loc</span></tt>.</p>
+</div></blockquote>
+</li>
+<li><p class="first"><tt class="docutils literal"><span class="pre">FixItHint::CreateRemoval(Range)</span></tt></p>
+<blockquote>
+<div><p>Specifies that the code in the given source <tt class="docutils literal"><span class="pre">Range</span></tt> should be removed.</p>
+</div></blockquote>
+</li>
+<li><p class="first"><tt class="docutils literal"><span class="pre">FixItHint::CreateReplacement(Range,</span> <span class="pre">Code)</span></tt></p>
+<blockquote>
+<div><p>Specifies that the code in the given source <tt class="docutils literal"><span class="pre">Range</span></tt> should be removed,
+and replaced with the given <tt class="docutils literal"><span class="pre">Code</span></tt> string.</p>
+</div></blockquote>
+</li>
+</ul>
+</div>
+<div class="section" id="the-diagnosticclient-interface">
+<span id="diagnosticclient"></span><h4><a class="toc-backref" href="#id12">The <tt class="docutils literal"><span class="pre">DiagnosticClient</span></tt> Interface</a><a class="headerlink" href="#the-diagnosticclient-interface" title="Permalink to this headline">¶</a></h4>
+<p>Once code generates a diagnostic with all of the arguments and the rest of the
+relevant information, Clang needs to know what to do with it. As previously
+mentioned, the diagnostic machinery goes through some filtering to map a
+severity onto a diagnostic level, then (assuming the diagnostic is not mapped
+to “<tt class="docutils literal"><span class="pre">Ignore</span></tt>”) it invokes an object that implements the <tt class="docutils literal"><span class="pre">DiagnosticClient</span></tt>
+interface with the information.</p>
+<p>It is possible to implement this interface in many different ways. For
+example, the normal Clang <tt class="docutils literal"><span class="pre">DiagnosticClient</span></tt> (named
+<tt class="docutils literal"><span class="pre">TextDiagnosticPrinter</span></tt>) turns the arguments into strings (according to the
+various formatting rules), prints out the file/line/column information and the
+string, then prints out the line of code, the source ranges, and the caret.
+However, this behavior isn’t required.</p>
+<p>Another implementation of the <tt class="docutils literal"><span class="pre">DiagnosticClient</span></tt> interface is the
+<tt class="docutils literal"><span class="pre">TextDiagnosticBuffer</span></tt> class, which is used when Clang is in <tt class="docutils literal"><span class="pre">-verify</span></tt>
+mode. Instead of formatting and printing out the diagnostics, this
+implementation just captures and remembers the diagnostics as they fly by.
+Then <tt class="docutils literal"><span class="pre">-verify</span></tt> compares the list of produced diagnostics to the list of
+expected ones. If they disagree, it prints out its own output. Full
+documentation for the <tt class="docutils literal"><span class="pre">-verify</span></tt> mode can be found in the Clang API
+documentation for <a class="reference external" href="/doxygen/classclang_1_1VerifyDiagnosticConsumer.html#details">VerifyDiagnosticConsumer</a>.</p>
+<p>There are many other possible implementations of this interface, and this is
+why we prefer diagnostics to pass down rich structured information in
+arguments. For example, an HTML output might want declaration names be
+linkified to where they come from in the source. Another example is that a GUI
+might let you click on typedefs to expand them. This application would want to
+pass significantly more information about types through to the GUI than a
+simple flat string. The interface allows this to happen.</p>
+</div>
+<div class="section" id="adding-translations-to-clang">
+<span id="internals-diag-translation"></span><h4><a class="toc-backref" href="#id13">Adding Translations to Clang</a><a class="headerlink" href="#adding-translations-to-clang" title="Permalink to this headline">¶</a></h4>
+<p>Not possible yet! Diagnostic strings should be written in UTF-8, the client can
+translate to the relevant code page if needed. Each translation completely
+replaces the format string for the diagnostic.</p>
+</div>
+</div>
+<div class="section" id="the-sourcelocation-and-sourcemanager-classes">
+<span id="sourcemanager"></span><span id="sourcelocation"></span><h3><a class="toc-backref" href="#id14">The <tt class="docutils literal"><span class="pre">SourceLocation</span></tt> and <tt class="docutils literal"><span class="pre">SourceManager</span></tt> classes</a><a class="headerlink" href="#the-sourcelocation-and-sourcemanager-classes" title="Permalink to this headline">¶</a></h3>
+<p>Strangely enough, the <tt class="docutils literal"><span class="pre">SourceLocation</span></tt> class represents a location within the
+source code of the program. Important design points include:</p>
+<ol class="arabic simple">
+<li><tt class="docutils literal"><span class="pre">sizeof(SourceLocation)</span></tt> must be extremely small, as these are embedded
+into many AST nodes and are passed around often. Currently it is 32 bits.</li>
+<li><tt class="docutils literal"><span class="pre">SourceLocation</span></tt> must be a simple value object that can be efficiently
+copied.</li>
+<li>We should be able to represent a source location for any byte of any input
+file. This includes in the middle of tokens, in whitespace, in trigraphs,
+etc.</li>
+<li>A <tt class="docutils literal"><span class="pre">SourceLocation</span></tt> must encode the current <tt class="docutils literal"><span class="pre">#include</span></tt> stack that was
+active when the location was processed. For example, if the location
+corresponds to a token, it should contain the set of <tt class="docutils literal"><span class="pre">#include</span></tt>s active
+when the token was lexed. This allows us to print the <tt class="docutils literal"><span class="pre">#include</span></tt> stack
+for a diagnostic.</li>
+<li><tt class="docutils literal"><span class="pre">SourceLocation</span></tt> must be able to describe macro expansions, capturing both
+the ultimate instantiation point and the source of the original character
+data.</li>
+</ol>
+<p>In practice, the <tt class="docutils literal"><span class="pre">SourceLocation</span></tt> works together with the <tt class="docutils literal"><span class="pre">SourceManager</span></tt>
+class to encode two pieces of information about a location: its spelling
+location and its instantiation location. For most tokens, these will be the
+same. However, for a macro expansion (or tokens that came from a <tt class="docutils literal"><span class="pre">_Pragma</span></tt>
+directive) these will describe the location of the characters corresponding to
+the token and the location where the token was used (i.e., the macro
+instantiation point or the location of the <tt class="docutils literal"><span class="pre">_Pragma</span></tt> itself).</p>
+<p>The Clang front-end inherently depends on the location of a token being tracked
+correctly. If it is ever incorrect, the front-end may get confused and die.
+The reason for this is that the notion of the “spelling” of a <tt class="docutils literal"><span class="pre">Token</span></tt> in
+Clang depends on being able to find the original input characters for the
+token. This concept maps directly to the “spelling location” for the token.</p>
+</div>
+<div class="section" id="sourcerange-and-charsourcerange">
+<h3><a class="toc-backref" href="#id15"><tt class="docutils literal"><span class="pre">SourceRange</span></tt> and <tt class="docutils literal"><span class="pre">CharSourceRange</span></tt></a><a class="headerlink" href="#sourcerange-and-charsourcerange" title="Permalink to this headline">¶</a></h3>
+<p>Clang represents most source ranges by [first, last], where “first” and “last”
+each point to the beginning of their respective tokens. For example consider
+the <tt class="docutils literal"><span class="pre">SourceRange</span></tt> of the following statement:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">x</span> <span class="o">=</span> <span class="n">foo</span> <span class="o">+</span> <span class="n">bar</span><span class="p">;</span>
+<span class="o">^</span><span class="n">first</span> <span class="o">^</span><span class="n">last</span>
+</pre></div>
+</div>
+<p>To map from this representation to a character-based representation, the “last”
+location needs to be adjusted to point to (or past) the end of that token with
+either <tt class="docutils literal"><span class="pre">Lexer::MeasureTokenLength()</span></tt> or <tt class="docutils literal"><span class="pre">Lexer::getLocForEndOfToken()</span></tt>. For
+the rare cases where character-level source ranges information is needed we use
+the <tt class="docutils literal"><span class="pre">CharSourceRange</span></tt> class.</p>
+</div>
+</div>
+<div class="section" id="the-driver-library">
+<h2><a class="toc-backref" href="#id16">The Driver Library</a><a class="headerlink" href="#the-driver-library" title="Permalink to this headline">¶</a></h2>
+<p>The clang Driver and library are documented <a class="reference internal" href="DriverInternals.html"><em>here</em></a>.</p>
+</div>
+<div class="section" id="precompiled-headers">
+<h2><a class="toc-backref" href="#id17">Precompiled Headers</a><a class="headerlink" href="#precompiled-headers" title="Permalink to this headline">¶</a></h2>
+<p>Clang supports two implementations of precompiled headers. The default
+implementation, precompiled headers (<a class="reference internal" href="PCHInternals.html"><em>PCH</em></a>) uses a
+serialized representation of Clang’s internal data structures, encoded with the
+<a class="reference external" href="http://llvm.org/docs/BitCodeFormat.html">LLVM bitstream format</a>.
+Pretokenized headers (<a class="reference internal" href="PTHInternals.html"><em>PTH</em></a>), on the other hand, contain a
+serialized representation of the tokens encountered when preprocessing a header
+(and anything that header includes).</p>
+</div>
+<div class="section" id="the-frontend-library">
+<h2><a class="toc-backref" href="#id18">The Frontend Library</a><a class="headerlink" href="#the-frontend-library" title="Permalink to this headline">¶</a></h2>
+<p>The Frontend library contains functionality useful for building tools on top of
+the Clang libraries, for example several methods for outputting diagnostics.</p>
+</div>
+<div class="section" id="the-lexer-and-preprocessor-library">
+<h2><a class="toc-backref" href="#id19">The Lexer and Preprocessor Library</a><a class="headerlink" href="#the-lexer-and-preprocessor-library" title="Permalink to this headline">¶</a></h2>
+<p>The Lexer library contains several tightly-connected classes that are involved
+with the nasty process of lexing and preprocessing C source code. The main
+interface to this library for outside clients is the large <tt class="docutils literal"><span class="pre">Preprocessor</span></tt>
+class. It contains the various pieces of state that are required to coherently
+read tokens out of a translation unit.</p>
+<p>The core interface to the <tt class="docutils literal"><span class="pre">Preprocessor</span></tt> object (once it is set up) is the
+<tt class="docutils literal"><span class="pre">Preprocessor::Lex</span></tt> method, which returns the next <a class="reference internal" href="#token"><em>Token</em></a> from
+the preprocessor stream. There are two types of token providers that the
+preprocessor is capable of reading from: a buffer lexer (provided by the
+<a class="reference internal" href="#lexer"><em>Lexer</em></a> class) and a buffered token stream (provided by the
+<a class="reference internal" href="#tokenlexer"><em>TokenLexer</em></a> class).</p>
+<div class="section" id="the-token-class">
+<span id="token"></span><h3><a class="toc-backref" href="#id20">The Token class</a><a class="headerlink" href="#the-token-class" title="Permalink to this headline">¶</a></h3>
+<p>The <tt class="docutils literal"><span class="pre">Token</span></tt> class is used to represent a single lexed token. Tokens are
+intended to be used by the lexer/preprocess and parser libraries, but are not
+intended to live beyond them (for example, they should not live in the ASTs).</p>
+<p>Tokens most often live on the stack (or some other location that is efficient
+to access) as the parser is running, but occasionally do get buffered up. For
+example, macro definitions are stored as a series of tokens, and the C++
+front-end periodically needs to buffer tokens up for tentative parsing and
+various pieces of look-ahead. As such, the size of a <tt class="docutils literal"><span class="pre">Token</span></tt> matters. On a
+32-bit system, <tt class="docutils literal"><span class="pre">sizeof(Token)</span></tt> is currently 16 bytes.</p>
+<p>Tokens occur in two forms: <a class="reference internal" href="#annotationtoken"><em>annotation tokens</em></a> and
+normal tokens. Normal tokens are those returned by the lexer, annotation
+tokens represent semantic information and are produced by the parser, replacing
+normal tokens in the token stream. Normal tokens contain the following
+information:</p>
+<ul class="simple">
+<li><strong>A SourceLocation</strong> — This indicates the location of the start of the
+token.</li>
+<li><strong>A length</strong> — This stores the length of the token as stored in the
+<tt class="docutils literal"><span class="pre">SourceBuffer</span></tt>. For tokens that include them, this length includes
+trigraphs and escaped newlines which are ignored by later phases of the
+compiler. By pointing into the original source buffer, it is always possible
+to get the original spelling of a token completely accurately.</li>
+<li><strong>IdentifierInfo</strong> — If a token takes the form of an identifier, and if
+identifier lookup was enabled when the token was lexed (e.g., the lexer was
+not reading in “raw” mode) this contains a pointer to the unique hash value
+for the identifier. Because the lookup happens before keyword
+identification, this field is set even for language keywords like “<tt class="docutils literal"><span class="pre">for</span></tt>”.</li>
+<li><strong>TokenKind</strong> — This indicates the kind of token as classified by the
+lexer. This includes things like <tt class="docutils literal"><span class="pre">tok::starequal</span></tt> (for the “<tt class="docutils literal"><span class="pre">*=</span></tt>”
+operator), <tt class="docutils literal"><span class="pre">tok::ampamp</span></tt> for the “<tt class="docutils literal"><span class="pre">&&</span></tt>” token, and keyword values (e.g.,
+<tt class="docutils literal"><span class="pre">tok::kw_for</span></tt>) for identifiers that correspond to keywords. Note that
+some tokens can be spelled multiple ways. For example, C++ supports
+“operator keywords”, where things like “<tt class="docutils literal"><span class="pre">and</span></tt>” are treated exactly like the
+“<tt class="docutils literal"><span class="pre">&&</span></tt>” operator. In these cases, the kind value is set to <tt class="docutils literal"><span class="pre">tok::ampamp</span></tt>,
+which is good for the parser, which doesn’t have to consider both forms. For
+something that cares about which form is used (e.g., the preprocessor
+“stringize” operator) the spelling indicates the original form.</li>
+<li><strong>Flags</strong> — There are currently four flags tracked by the
+lexer/preprocessor system on a per-token basis:<ol class="arabic">
+<li><strong>StartOfLine</strong> — This was the first token that occurred on its input
+source line.</li>
+<li><strong>LeadingSpace</strong> — There was a space character either immediately before
+the token or transitively before the token as it was expanded through a
+macro. The definition of this flag is very closely defined by the
+stringizing requirements of the preprocessor.</li>
+<li><strong>DisableExpand</strong> — This flag is used internally to the preprocessor to
+represent identifier tokens which have macro expansion disabled. This
+prevents them from being considered as candidates for macro expansion ever
+in the future.</li>
+<li><strong>NeedsCleaning</strong> — This flag is set if the original spelling for the
+token includes a trigraph or escaped newline. Since this is uncommon,
+many pieces of code can fast-path on tokens that did not need cleaning.</li>
+</ol>
+</li>
+</ul>
+<p>One interesting (and somewhat unusual) aspect of normal tokens is that they
+don’t contain any semantic information about the lexed value. For example, if
+the token was a pp-number token, we do not represent the value of the number
+that was lexed (this is left for later pieces of code to decide).
+Additionally, the lexer library has no notion of typedef names vs variable
+names: both are returned as identifiers, and the parser is left to decide
+whether a specific identifier is a typedef or a variable (tracking this
+requires scope information among other things). The parser can do this
+translation by replacing tokens returned by the preprocessor with “Annotation
+Tokens”.</p>
+</div>
+<div class="section" id="annotation-tokens">
+<span id="annotationtoken"></span><h3><a class="toc-backref" href="#id21">Annotation Tokens</a><a class="headerlink" href="#annotation-tokens" title="Permalink to this headline">¶</a></h3>
+<p>Annotation tokens are tokens that are synthesized by the parser and injected
+into the preprocessor’s token stream (replacing existing tokens) to record
+semantic information found by the parser. For example, if “<tt class="docutils literal"><span class="pre">foo</span></tt>” is found
+to be a typedef, the “<tt class="docutils literal"><span class="pre">foo</span></tt>” <tt class="docutils literal"><span class="pre">tok::identifier</span></tt> token is replaced with an
+<tt class="docutils literal"><span class="pre">tok::annot_typename</span></tt>. This is useful for a couple of reasons: 1) this makes
+it easy to handle qualified type names (e.g., “<tt class="docutils literal"><span class="pre">foo::bar::baz<42>::t</span></tt>”) in
+C++ as a single “token” in the parser. 2) if the parser backtracks, the
+reparse does not need to redo semantic analysis to determine whether a token
+sequence is a variable, type, template, etc.</p>
+<p>Annotation tokens are created by the parser and reinjected into the parser’s
+token stream (when backtracking is enabled). Because they can only exist in
+tokens that the preprocessor-proper is done with, it doesn’t need to keep
+around flags like “start of line” that the preprocessor uses to do its job.
+Additionally, an annotation token may “cover” a sequence of preprocessor tokens
+(e.g., “<tt class="docutils literal"><span class="pre">a::b::c</span></tt>” is five preprocessor tokens). As such, the valid fields
+of an annotation token are different than the fields for a normal token (but
+they are multiplexed into the normal <tt class="docutils literal"><span class="pre">Token</span></tt> fields):</p>
+<ul class="simple">
+<li><strong>SourceLocation “Location”</strong> — The <tt class="docutils literal"><span class="pre">SourceLocation</span></tt> for the annotation
+token indicates the first token replaced by the annotation token. In the
+example above, it would be the location of the “<tt class="docutils literal"><span class="pre">a</span></tt>” identifier.</li>
+<li><strong>SourceLocation “AnnotationEndLoc”</strong> — This holds the location of the last
+token replaced with the annotation token. In the example above, it would be
+the location of the “<tt class="docutils literal"><span class="pre">c</span></tt>” identifier.</li>
+<li><strong>void* “AnnotationValue”</strong> — This contains an opaque object that the
+parser gets from <tt class="docutils literal"><span class="pre">Sema</span></tt>. The parser merely preserves the information for
+<tt class="docutils literal"><span class="pre">Sema</span></tt> to later interpret based on the annotation token kind.</li>
+<li><strong>TokenKind “Kind”</strong> — This indicates the kind of Annotation token this is.
+See below for the different valid kinds.</li>
+</ul>
+<p>Annotation tokens currently come in three kinds:</p>
+<ol class="arabic simple">
+<li><strong>tok::annot_typename</strong>: This annotation token represents a resolved
+typename token that is potentially qualified. The <tt class="docutils literal"><span class="pre">AnnotationValue</span></tt> field
+contains the <tt class="docutils literal"><span class="pre">QualType</span></tt> returned by <tt class="docutils literal"><span class="pre">Sema::getTypeName()</span></tt>, possibly with
+source location information attached.</li>
+<li><strong>tok::annot_cxxscope</strong>: This annotation token represents a C++ scope
+specifier, such as “<tt class="docutils literal"><span class="pre">A::B::</span></tt>”. This corresponds to the grammar
+productions “<em>::</em>” and “<em>:: [opt] nested-name-specifier</em>”. The
+<tt class="docutils literal"><span class="pre">AnnotationValue</span></tt> pointer is a <tt class="docutils literal"><span class="pre">NestedNameSpecifier</span> <span class="pre">*</span></tt> returned by the
+<tt class="docutils literal"><span class="pre">Sema::ActOnCXXGlobalScopeSpecifier</span></tt> and
+<tt class="docutils literal"><span class="pre">Sema::ActOnCXXNestedNameSpecifier</span></tt> callbacks.</li>
+<li><strong>tok::annot_template_id</strong>: This annotation token represents a C++
+template-id such as “<tt class="docutils literal"><span class="pre">foo<int,</span> <span class="pre">4></span></tt>”, where “<tt class="docutils literal"><span class="pre">foo</span></tt>” is the name of a
+template. The <tt class="docutils literal"><span class="pre">AnnotationValue</span></tt> pointer is a pointer to a <tt class="docutils literal"><span class="pre">malloc</span></tt>‘d
+<tt class="docutils literal"><span class="pre">TemplateIdAnnotation</span></tt> object. Depending on the context, a parsed
+template-id that names a type might become a typename annotation token (if
+all we care about is the named type, e.g., because it occurs in a type
+specifier) or might remain a template-id token (if we want to retain more
+source location information or produce a new type, e.g., in a declaration of
+a class template specialization). template-id annotation tokens that refer
+to a type can be “upgraded” to typename annotation tokens by the parser.</li>
+</ol>
+<p>As mentioned above, annotation tokens are not returned by the preprocessor,
+they are formed on demand by the parser. This means that the parser has to be
+aware of cases where an annotation could occur and form it where appropriate.
+This is somewhat similar to how the parser handles Translation Phase 6 of C99:
+String Concatenation (see C99 5.1.1.2). In the case of string concatenation,
+the preprocessor just returns distinct <tt class="docutils literal"><span class="pre">tok::string_literal</span></tt> and
+<tt class="docutils literal"><span class="pre">tok::wide_string_literal</span></tt> tokens and the parser eats a sequence of them
+wherever the grammar indicates that a string literal can occur.</p>
+<p>In order to do this, whenever the parser expects a <tt class="docutils literal"><span class="pre">tok::identifier</span></tt> or
+<tt class="docutils literal"><span class="pre">tok::coloncolon</span></tt>, it should call the <tt class="docutils literal"><span class="pre">TryAnnotateTypeOrScopeToken</span></tt> or
+<tt class="docutils literal"><span class="pre">TryAnnotateCXXScopeToken</span></tt> methods to form the annotation token. These
+methods will maximally form the specified annotation tokens and replace the
+current token with them, if applicable. If the current tokens is not valid for
+an annotation token, it will remain an identifier or “<tt class="docutils literal"><span class="pre">::</span></tt>” token.</p>
+</div>
+<div class="section" id="the-lexer-class">
+<span id="lexer"></span><h3><a class="toc-backref" href="#id22">The <tt class="docutils literal"><span class="pre">Lexer</span></tt> class</a><a class="headerlink" href="#the-lexer-class" title="Permalink to this headline">¶</a></h3>
+<p>The <tt class="docutils literal"><span class="pre">Lexer</span></tt> class provides the mechanics of lexing tokens out of a source
+buffer and deciding what they mean. The <tt class="docutils literal"><span class="pre">Lexer</span></tt> is complicated by the fact
+that it operates on raw buffers that have not had spelling eliminated (this is
+a necessity to get decent performance), but this is countered with careful
+coding as well as standard performance techniques (for example, the comment
+handling code is vectorized on X86 and PowerPC hosts).</p>
+<p>The lexer has a couple of interesting modal features:</p>
+<ul class="simple">
+<li>The lexer can operate in “raw” mode. This mode has several features that
+make it possible to quickly lex the file (e.g., it stops identifier lookup,
+doesn’t specially handle preprocessor tokens, handles EOF differently, etc).
+This mode is used for lexing within an “<tt class="docutils literal"><span class="pre">#if</span> <span class="pre">0</span></tt>” block, for example.</li>
+<li>The lexer can capture and return comments as tokens. This is required to
+support the <tt class="docutils literal"><span class="pre">-C</span></tt> preprocessor mode, which passes comments through, and is
+used by the diagnostic checker to identifier expect-error annotations.</li>
+<li>The lexer can be in <tt class="docutils literal"><span class="pre">ParsingFilename</span></tt> mode, which happens when
+preprocessing after reading a <tt class="docutils literal"><span class="pre">#include</span></tt> directive. This mode changes the
+parsing of “<tt class="docutils literal"><span class="pre"><</span></tt>” to return an “angled string” instead of a bunch of tokens
+for each thing within the filename.</li>
+<li>When parsing a preprocessor directive (after “<tt class="docutils literal"><span class="pre">#</span></tt>”) the
+<tt class="docutils literal"><span class="pre">ParsingPreprocessorDirective</span></tt> mode is entered. This changes the parser to
+return EOD at a newline.</li>
+<li>The <tt class="docutils literal"><span class="pre">Lexer</span></tt> uses a <tt class="docutils literal"><span class="pre">LangOptions</span></tt> object to know whether trigraphs are
+enabled, whether C++ or ObjC keywords are recognized, etc.</li>
+</ul>
+<p>In addition to these modes, the lexer keeps track of a couple of other features
+that are local to a lexed buffer, which change as the buffer is lexed:</p>
+<ul class="simple">
+<li>The <tt class="docutils literal"><span class="pre">Lexer</span></tt> uses <tt class="docutils literal"><span class="pre">BufferPtr</span></tt> to keep track of the current character being
+lexed.</li>
+<li>The <tt class="docutils literal"><span class="pre">Lexer</span></tt> uses <tt class="docutils literal"><span class="pre">IsAtStartOfLine</span></tt> to keep track of whether the next
+lexed token will start with its “start of line” bit set.</li>
+<li>The <tt class="docutils literal"><span class="pre">Lexer</span></tt> keeps track of the current “<tt class="docutils literal"><span class="pre">#if</span></tt>” directives that are active
+(which can be nested).</li>
+<li>The <tt class="docutils literal"><span class="pre">Lexer</span></tt> keeps track of an <a class="reference internal" href="#multipleincludeopt"><em>MultipleIncludeOpt</em></a> object, which is used to detect whether the buffer uses
+the standard “<tt class="docutils literal"><span class="pre">#ifndef</span> <span class="pre">XX</span></tt> / <tt class="docutils literal"><span class="pre">#define</span> <span class="pre">XX</span></tt>” idiom to prevent multiple
+inclusion. If a buffer does, subsequent includes can be ignored if the
+“<tt class="docutils literal"><span class="pre">XX</span></tt>” macro is defined.</li>
+</ul>
+</div>
+<div class="section" id="the-tokenlexer-class">
+<span id="tokenlexer"></span><h3><a class="toc-backref" href="#id23">The <tt class="docutils literal"><span class="pre">TokenLexer</span></tt> class</a><a class="headerlink" href="#the-tokenlexer-class" title="Permalink to this headline">¶</a></h3>
+<p>The <tt class="docutils literal"><span class="pre">TokenLexer</span></tt> class is a token provider that returns tokens from a list of
+tokens that came from somewhere else. It typically used for two things: 1)
+returning tokens from a macro definition as it is being expanded 2) returning
+tokens from an arbitrary buffer of tokens. The later use is used by
+<tt class="docutils literal"><span class="pre">_Pragma</span></tt> and will most likely be used to handle unbounded look-ahead for the
+C++ parser.</p>
+</div>
+<div class="section" id="the-multipleincludeopt-class">
+<span id="multipleincludeopt"></span><h3><a class="toc-backref" href="#id24">The <tt class="docutils literal"><span class="pre">MultipleIncludeOpt</span></tt> class</a><a class="headerlink" href="#the-multipleincludeopt-class" title="Permalink to this headline">¶</a></h3>
+<p>The <tt class="docutils literal"><span class="pre">MultipleIncludeOpt</span></tt> class implements a really simple little state
+machine that is used to detect the standard “<tt class="docutils literal"><span class="pre">#ifndef</span> <span class="pre">XX</span></tt> / <tt class="docutils literal"><span class="pre">#define</span> <span class="pre">XX</span></tt>”
+idiom that people typically use to prevent multiple inclusion of headers. If a
+buffer uses this idiom and is subsequently <tt class="docutils literal"><span class="pre">#include</span></tt>‘d, the preprocessor can
+simply check to see whether the guarding condition is defined or not. If so,
+the preprocessor can completely ignore the include of the header.</p>
+</div>
+</div>
+<div class="section" id="the-parser-library">
+<span id="parser"></span><h2><a class="toc-backref" href="#id25">The Parser Library</a><a class="headerlink" href="#the-parser-library" title="Permalink to this headline">¶</a></h2>
+<p>This library contains a recursive-descent parser that polls tokens from the
+preprocessor and notifies a client of the parsing progress.</p>
+<p>Historically, the parser used to talk to an abstract <tt class="docutils literal"><span class="pre">Action</span></tt> interface that
+had virtual methods for parse events, for example <tt class="docutils literal"><span class="pre">ActOnBinOp()</span></tt>. When Clang
+grew C++ support, the parser stopped supporting general <tt class="docutils literal"><span class="pre">Action</span></tt> clients –
+it now always talks to the <a class="reference internal" href="#sema"><em>Sema libray</em></a>. However, the Parser
+still accesses AST objects only through opaque types like <tt class="docutils literal"><span class="pre">ExprResult</span></tt> and
+<tt class="docutils literal"><span class="pre">StmtResult</span></tt>. Only <a class="reference internal" href="#sema"><em>Sema</em></a> looks at the AST node contents of these
+wrappers.</p>
+</div>
+<div class="section" id="the-ast-library">
+<span id="ast"></span><h2><a class="toc-backref" href="#id26">The AST Library</a><a class="headerlink" href="#the-ast-library" title="Permalink to this headline">¶</a></h2>
+<div class="section" id="the-type-class-and-its-subclasses">
+<span id="type"></span><h3><a class="toc-backref" href="#id27">The <tt class="docutils literal"><span class="pre">Type</span></tt> class and its subclasses</a><a class="headerlink" href="#the-type-class-and-its-subclasses" title="Permalink to this headline">¶</a></h3>
+<p>The <tt class="docutils literal"><span class="pre">Type</span></tt> class (and its subclasses) are an important part of the AST.
+Types are accessed through the <tt class="docutils literal"><span class="pre">ASTContext</span></tt> class, which implicitly creates
+and uniques them as they are needed. Types have a couple of non-obvious
+features: 1) they do not capture type qualifiers like <tt class="docutils literal"><span class="pre">const</span></tt> or <tt class="docutils literal"><span class="pre">volatile</span></tt>
+(see <a class="reference internal" href="#qualtype"><em>QualType</em></a>), and 2) they implicitly capture typedef
+information. Once created, types are immutable (unlike decls).</p>
+<p>Typedefs in C make semantic analysis a bit more complex than it would be without
+them. The issue is that we want to capture typedef information and represent it
+in the AST perfectly, but the semantics of operations need to “see through”
+typedefs. For example, consider this code:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">void</span> <span class="nf">func</span><span class="p">()</span> <span class="p">{</span>
+ <span class="k">typedef</span> <span class="kt">int</span> <span class="n">foo</span><span class="p">;</span>
+ <span class="n">foo</span> <span class="n">X</span><span class="p">,</span> <span class="o">*</span><span class="n">Y</span><span class="p">;</span>
+ <span class="k">typedef</span> <span class="n">foo</span> <span class="o">*</span><span class="n">bar</span><span class="p">;</span>
+ <span class="n">bar</span> <span class="n">Z</span><span class="p">;</span>
+ <span class="o">*</span><span class="n">X</span><span class="p">;</span> <span class="c1">// error</span>
+ <span class="o">**</span><span class="n">Y</span><span class="p">;</span> <span class="c1">// error</span>
+ <span class="o">**</span><span class="n">Z</span><span class="p">;</span> <span class="c1">// error</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>The code above is illegal, and thus we expect there to be diagnostics emitted
+on the annotated lines. In this example, we expect to get:</p>
+<div class="highlight-c++"><div class="highlight"><pre>test.c:6:1: error: indirection requires pointer operand ('foo' invalid)
+ *X; // error
+ ^~
+test.c:7:1: error: indirection requires pointer operand ('foo' invalid)
+ **Y; // error
+ ^~~
+test.c:8:1: error: indirection requires pointer operand ('foo' invalid)
+ **Z; // error
+ ^~~
+</pre></div>
+</div>
+<p>While this example is somewhat silly, it illustrates the point: we want to
+retain typedef information where possible, so that we can emit errors about
+“<tt class="docutils literal"><span class="pre">std::string</span></tt>” instead of “<tt class="docutils literal"><span class="pre">std::basic_string<char,</span> <span class="pre">std:...</span></tt>”. Doing this
+requires properly keeping typedef information (for example, the type of <tt class="docutils literal"><span class="pre">X</span></tt>
+is “<tt class="docutils literal"><span class="pre">foo</span></tt>”, not “<tt class="docutils literal"><span class="pre">int</span></tt>”), and requires properly propagating it through the
+various operators (for example, the type of <tt class="docutils literal"><span class="pre">*Y</span></tt> is “<tt class="docutils literal"><span class="pre">foo</span></tt>”, not
+“<tt class="docutils literal"><span class="pre">int</span></tt>”). In order to retain this information, the type of these expressions
+is an instance of the <tt class="docutils literal"><span class="pre">TypedefType</span></tt> class, which indicates that the type of
+these expressions is a typedef for “<tt class="docutils literal"><span class="pre">foo</span></tt>”.</p>
+<p>Representing types like this is great for diagnostics, because the
+user-specified type is always immediately available. There are two problems
+with this: first, various semantic checks need to make judgements about the
+<em>actual structure</em> of a type, ignoring typedefs. Second, we need an efficient
+way to query whether two types are structurally identical to each other,
+ignoring typedefs. The solution to both of these problems is the idea of
+canonical types.</p>
+<div class="section" id="canonical-types">
+<h4><a class="toc-backref" href="#id28">Canonical Types</a><a class="headerlink" href="#canonical-types" title="Permalink to this headline">¶</a></h4>
+<p>Every instance of the <tt class="docutils literal"><span class="pre">Type</span></tt> class contains a canonical type pointer. For
+simple types with no typedefs involved (e.g., “<tt class="docutils literal"><span class="pre">int</span></tt>”, “<tt class="docutils literal"><span class="pre">int*</span></tt>”,
+“<tt class="docutils literal"><span class="pre">int**</span></tt>”), the type just points to itself. For types that have a typedef
+somewhere in their structure (e.g., “<tt class="docutils literal"><span class="pre">foo</span></tt>”, “<tt class="docutils literal"><span class="pre">foo*</span></tt>”, “<tt class="docutils literal"><span class="pre">foo**</span></tt>”,
+“<tt class="docutils literal"><span class="pre">bar</span></tt>”), the canonical type pointer points to their structurally equivalent
+type without any typedefs (e.g., “<tt class="docutils literal"><span class="pre">int</span></tt>”, “<tt class="docutils literal"><span class="pre">int*</span></tt>”, “<tt class="docutils literal"><span class="pre">int**</span></tt>”, and
+“<tt class="docutils literal"><span class="pre">int*</span></tt>” respectively).</p>
+<p>This design provides a constant time operation (dereferencing the canonical type
+pointer) that gives us access to the structure of types. For example, we can
+trivially tell that “<tt class="docutils literal"><span class="pre">bar</span></tt>” and “<tt class="docutils literal"><span class="pre">foo*</span></tt>” are the same type by dereferencing
+their canonical type pointers and doing a pointer comparison (they both point
+to the single “<tt class="docutils literal"><span class="pre">int*</span></tt>” type).</p>
+<p>Canonical types and typedef types bring up some complexities that must be
+carefully managed. Specifically, the <tt class="docutils literal"><span class="pre">isa</span></tt>/<tt class="docutils literal"><span class="pre">cast</span></tt>/<tt class="docutils literal"><span class="pre">dyn_cast</span></tt> operators
+generally shouldn’t be used in code that is inspecting the AST. For example,
+when type checking the indirection operator (unary “<tt class="docutils literal"><span class="pre">*</span></tt>” on a pointer), the
+type checker must verify that the operand has a pointer type. It would not be
+correct to check that with “<tt class="docutils literal"><span class="pre">isa<PointerType>(SubExpr->getType())</span></tt>”, because
+this predicate would fail if the subexpression had a typedef type.</p>
+<p>The solution to this problem are a set of helper methods on <tt class="docutils literal"><span class="pre">Type</span></tt>, used to
+check their properties. In this case, it would be correct to use
+“<tt class="docutils literal"><span class="pre">SubExpr->getType()->isPointerType()</span></tt>” to do the check. This predicate will
+return true if the <em>canonical type is a pointer</em>, which is true any time the
+type is structurally a pointer type. The only hard part here is remembering
+not to use the <tt class="docutils literal"><span class="pre">isa</span></tt>/<tt class="docutils literal"><span class="pre">cast</span></tt>/<tt class="docutils literal"><span class="pre">dyn_cast</span></tt> operations.</p>
+<p>The second problem we face is how to get access to the pointer type once we
+know it exists. To continue the example, the result type of the indirection
+operator is the pointee type of the subexpression. In order to determine the
+type, we need to get the instance of <tt class="docutils literal"><span class="pre">PointerType</span></tt> that best captures the
+typedef information in the program. If the type of the expression is literally
+a <tt class="docutils literal"><span class="pre">PointerType</span></tt>, we can return that, otherwise we have to dig through the
+typedefs to find the pointer type. For example, if the subexpression had type
+“<tt class="docutils literal"><span class="pre">foo*</span></tt>”, we could return that type as the result. If the subexpression had
+type “<tt class="docutils literal"><span class="pre">bar</span></tt>”, we want to return “<tt class="docutils literal"><span class="pre">foo*</span></tt>” (note that we do <em>not</em> want
+“<tt class="docutils literal"><span class="pre">int*</span></tt>”). In order to provide all of this, <tt class="docutils literal"><span class="pre">Type</span></tt> has a
+<tt class="docutils literal"><span class="pre">getAsPointerType()</span></tt> method that checks whether the type is structurally a
+<tt class="docutils literal"><span class="pre">PointerType</span></tt> and, if so, returns the best one. If not, it returns a null
+pointer.</p>
+<p>This structure is somewhat mystical, but after meditating on it, it will make
+sense to you :).</p>
+</div>
+</div>
+<div class="section" id="the-qualtype-class">
+<span id="qualtype"></span><h3><a class="toc-backref" href="#id29">The <tt class="docutils literal"><span class="pre">QualType</span></tt> class</a><a class="headerlink" href="#the-qualtype-class" title="Permalink to this headline">¶</a></h3>
+<p>The <tt class="docutils literal"><span class="pre">QualType</span></tt> class is designed as a trivial value class that is small,
+passed by-value and is efficient to query. The idea of <tt class="docutils literal"><span class="pre">QualType</span></tt> is that it
+stores the type qualifiers (<tt class="docutils literal"><span class="pre">const</span></tt>, <tt class="docutils literal"><span class="pre">volatile</span></tt>, <tt class="docutils literal"><span class="pre">restrict</span></tt>, plus some
+extended qualifiers required by language extensions) separately from the types
+themselves. <tt class="docutils literal"><span class="pre">QualType</span></tt> is conceptually a pair of “<tt class="docutils literal"><span class="pre">Type*</span></tt>” and the bits
+for these type qualifiers.</p>
+<p>By storing the type qualifiers as bits in the conceptual pair, it is extremely
+efficient to get the set of qualifiers on a <tt class="docutils literal"><span class="pre">QualType</span></tt> (just return the field
+of the pair), add a type qualifier (which is a trivial constant-time operation
+that sets a bit), and remove one or more type qualifiers (just return a
+<tt class="docutils literal"><span class="pre">QualType</span></tt> with the bitfield set to empty).</p>
+<p>Further, because the bits are stored outside of the type itself, we do not need
+to create duplicates of types with different sets of qualifiers (i.e. there is
+only a single heap allocated “<tt class="docutils literal"><span class="pre">int</span></tt>” type: “<tt class="docutils literal"><span class="pre">const</span> <span class="pre">int</span></tt>” and “<tt class="docutils literal"><span class="pre">volatile</span>
+<span class="pre">const</span> <span class="pre">int</span></tt>” both point to the same heap allocated “<tt class="docutils literal"><span class="pre">int</span></tt>” type). This
+reduces the heap size used to represent bits and also means we do not have to
+consider qualifiers when uniquing types (<a class="reference internal" href="#type"><em>Type</em></a> does not even
+contain qualifiers).</p>
+<p>In practice, the two most common type qualifiers (<tt class="docutils literal"><span class="pre">const</span></tt> and <tt class="docutils literal"><span class="pre">restrict</span></tt>)
+are stored in the low bits of the pointer to the <tt class="docutils literal"><span class="pre">Type</span></tt> object, together with
+a flag indicating whether extended qualifiers are present (which must be
+heap-allocated). This means that <tt class="docutils literal"><span class="pre">QualType</span></tt> is exactly the same size as a
+pointer.</p>
+</div>
+<div class="section" id="declaration-names">
+<span id="declarationname"></span><h3><a class="toc-backref" href="#id30">Declaration names</a><a class="headerlink" href="#declaration-names" title="Permalink to this headline">¶</a></h3>
+<p>The <tt class="docutils literal"><span class="pre">DeclarationName</span></tt> class represents the name of a declaration in Clang.
+Declarations in the C family of languages can take several different forms.
+Most declarations are named by simple identifiers, e.g., “<tt class="docutils literal"><span class="pre">f</span></tt>” and “<tt class="docutils literal"><span class="pre">x</span></tt>” in
+the function declaration <tt class="docutils literal"><span class="pre">f(int</span> <span class="pre">x)</span></tt>. In C++, declaration names can also name
+class constructors (“<tt class="docutils literal"><span class="pre">Class</span></tt>” in <tt class="docutils literal"><span class="pre">struct</span> <span class="pre">Class</span> <span class="pre">{</span> <span class="pre">Class();</span> <span class="pre">}</span></tt>), class
+destructors (“<tt class="docutils literal"><span class="pre">~Class</span></tt>”), overloaded operator names (“<tt class="docutils literal"><span class="pre">operator+</span></tt>”), and
+conversion functions (“<tt class="docutils literal"><span class="pre">operator</span> <span class="pre">void</span> <span class="pre">const</span> <span class="pre">*</span></tt>”). In Objective-C,
+declaration names can refer to the names of Objective-C methods, which involve
+the method name and the parameters, collectively called a <em>selector</em>, e.g.,
+“<tt class="docutils literal"><span class="pre">setWidth:height:</span></tt>”. Since all of these kinds of entities — variables,
+functions, Objective-C methods, C++ constructors, destructors, and operators
+— are represented as subclasses of Clang’s common <tt class="docutils literal"><span class="pre">NamedDecl</span></tt> class,
+<tt class="docutils literal"><span class="pre">DeclarationName</span></tt> is designed to efficiently represent any kind of name.</p>
+<p>Given a <tt class="docutils literal"><span class="pre">DeclarationName</span></tt> <tt class="docutils literal"><span class="pre">N</span></tt>, <tt class="docutils literal"><span class="pre">N.getNameKind()</span></tt> will produce a value
+that describes what kind of name <tt class="docutils literal"><span class="pre">N</span></tt> stores. There are 10 options (all of
+the names are inside the <tt class="docutils literal"><span class="pre">DeclarationName</span></tt> class).</p>
+<p><tt class="docutils literal"><span class="pre">Identifier</span></tt></p>
+<blockquote>
+<div>The name is a simple identifier. Use <tt class="docutils literal"><span class="pre">N.getAsIdentifierInfo()</span></tt> to retrieve
+the corresponding <tt class="docutils literal"><span class="pre">IdentifierInfo*</span></tt> pointing to the actual identifier.</div></blockquote>
+<p><tt class="docutils literal"><span class="pre">ObjCZeroArgSelector</span></tt>, <tt class="docutils literal"><span class="pre">ObjCOneArgSelector</span></tt>, <tt class="docutils literal"><span class="pre">ObjCMultiArgSelector</span></tt></p>
+<blockquote>
+<div>The name is an Objective-C selector, which can be retrieved as a <tt class="docutils literal"><span class="pre">Selector</span></tt>
+instance via <tt class="docutils literal"><span class="pre">N.getObjCSelector()</span></tt>. The three possible name kinds for
+Objective-C reflect an optimization within the <tt class="docutils literal"><span class="pre">DeclarationName</span></tt> class:
+both zero- and one-argument selectors are stored as a masked
+<tt class="docutils literal"><span class="pre">IdentifierInfo</span></tt> pointer, and therefore require very little space, since
+zero- and one-argument selectors are far more common than multi-argument
+selectors (which use a different structure).</div></blockquote>
+<p><tt class="docutils literal"><span class="pre">CXXConstructorName</span></tt></p>
+<blockquote>
+<div>The name is a C++ constructor name. Use <tt class="docutils literal"><span class="pre">N.getCXXNameType()</span></tt> to retrieve
+the <a class="reference internal" href="#qualtype"><em>type</em></a> that this constructor is meant to construct. The
+type is always the canonical type, since all constructors for a given type
+have the same name.</div></blockquote>
+<p><tt class="docutils literal"><span class="pre">CXXDestructorName</span></tt></p>
+<blockquote>
+<div>The name is a C++ destructor name. Use <tt class="docutils literal"><span class="pre">N.getCXXNameType()</span></tt> to retrieve
+the <a class="reference internal" href="#qualtype"><em>type</em></a> whose destructor is being named. This type is
+always a canonical type.</div></blockquote>
+<p><tt class="docutils literal"><span class="pre">CXXConversionFunctionName</span></tt></p>
+<blockquote>
+<div>The name is a C++ conversion function. Conversion functions are named
+according to the type they convert to, e.g., “<tt class="docutils literal"><span class="pre">operator</span> <span class="pre">void</span> <span class="pre">const</span> <span class="pre">*</span></tt>”.
+Use <tt class="docutils literal"><span class="pre">N.getCXXNameType()</span></tt> to retrieve the type that this conversion function
+converts to. This type is always a canonical type.</div></blockquote>
+<p><tt class="docutils literal"><span class="pre">CXXOperatorName</span></tt></p>
+<blockquote>
+<div>The name is a C++ overloaded operator name. Overloaded operators are named
+according to their spelling, e.g., “<tt class="docutils literal"><span class="pre">operator+</span></tt>” or “<tt class="docutils literal"><span class="pre">operator</span> <span class="pre">new</span> <span class="pre">[]</span></tt>”.
+Use <tt class="docutils literal"><span class="pre">N.getCXXOverloadedOperator()</span></tt> to retrieve the overloaded operator (a
+value of type <tt class="docutils literal"><span class="pre">OverloadedOperatorKind</span></tt>).</div></blockquote>
+<p><tt class="docutils literal"><span class="pre">CXXLiteralOperatorName</span></tt></p>
+<blockquote>
+<div>The name is a C++11 user defined literal operator. User defined
+Literal operators are named according to the suffix they define,
+e.g., “<tt class="docutils literal"><span class="pre">_foo</span></tt>” for “<tt class="docutils literal"><span class="pre">operator</span> <span class="pre">""</span> <span class="pre">_foo</span></tt>”. Use
+<tt class="docutils literal"><span class="pre">N.getCXXLiteralIdentifier()</span></tt> to retrieve the corresponding
+<tt class="docutils literal"><span class="pre">IdentifierInfo*</span></tt> pointing to the identifier.</div></blockquote>
+<p><tt class="docutils literal"><span class="pre">CXXUsingDirective</span></tt></p>
+<blockquote>
+<div>The name is a C++ using directive. Using directives are not really
+NamedDecls, in that they all have the same name, but they are
+implemented as such in order to store them in DeclContext
+effectively.</div></blockquote>
+<p><tt class="docutils literal"><span class="pre">DeclarationName</span></tt>s are cheap to create, copy, and compare. They require
+only a single pointer’s worth of storage in the common cases (identifiers,
+zero- and one-argument Objective-C selectors) and use dense, uniqued storage
+for the other kinds of names. Two <tt class="docutils literal"><span class="pre">DeclarationName</span></tt>s can be compared for
+equality (<tt class="docutils literal"><span class="pre">==</span></tt>, <tt class="docutils literal"><span class="pre">!=</span></tt>) using a simple bitwise comparison, can be ordered
+with <tt class="docutils literal"><span class="pre"><</span></tt>, <tt class="docutils literal"><span class="pre">></span></tt>, <tt class="docutils literal"><span class="pre"><=</span></tt>, and <tt class="docutils literal"><span class="pre">>=</span></tt> (which provide a lexicographical ordering
+for normal identifiers but an unspecified ordering for other kinds of names),
+and can be placed into LLVM <tt class="docutils literal"><span class="pre">DenseMap</span></tt>s and <tt class="docutils literal"><span class="pre">DenseSet</span></tt>s.</p>
+<p><tt class="docutils literal"><span class="pre">DeclarationName</span></tt> instances can be created in different ways depending on
+what kind of name the instance will store. Normal identifiers
+(<tt class="docutils literal"><span class="pre">IdentifierInfo</span></tt> pointers) and Objective-C selectors (<tt class="docutils literal"><span class="pre">Selector</span></tt>) can be
+implicitly converted to <tt class="docutils literal"><span class="pre">DeclarationNames</span></tt>. Names for C++ constructors,
+destructors, conversion functions, and overloaded operators can be retrieved
+from the <tt class="docutils literal"><span class="pre">DeclarationNameTable</span></tt>, an instance of which is available as
+<tt class="docutils literal"><span class="pre">ASTContext::DeclarationNames</span></tt>. The member functions
+<tt class="docutils literal"><span class="pre">getCXXConstructorName</span></tt>, <tt class="docutils literal"><span class="pre">getCXXDestructorName</span></tt>,
+<tt class="docutils literal"><span class="pre">getCXXConversionFunctionName</span></tt>, and <tt class="docutils literal"><span class="pre">getCXXOperatorName</span></tt>, respectively,
+return <tt class="docutils literal"><span class="pre">DeclarationName</span></tt> instances for the four kinds of C++ special function
+names.</p>
+</div>
+<div class="section" id="declaration-contexts">
+<span id="declcontext"></span><h3><a class="toc-backref" href="#id31">Declaration contexts</a><a class="headerlink" href="#declaration-contexts" title="Permalink to this headline">¶</a></h3>
+<p>Every declaration in a program exists within some <em>declaration context</em>, such
+as a translation unit, namespace, class, or function. Declaration contexts in
+Clang are represented by the <tt class="docutils literal"><span class="pre">DeclContext</span></tt> class, from which the various
+declaration-context AST nodes (<tt class="docutils literal"><span class="pre">TranslationUnitDecl</span></tt>, <tt class="docutils literal"><span class="pre">NamespaceDecl</span></tt>,
+<tt class="docutils literal"><span class="pre">RecordDecl</span></tt>, <tt class="docutils literal"><span class="pre">FunctionDecl</span></tt>, etc.) will derive. The <tt class="docutils literal"><span class="pre">DeclContext</span></tt> class
+provides several facilities common to each declaration context:</p>
+<p>Source-centric vs. Semantics-centric View of Declarations</p>
+<blockquote>
+<div><tt class="docutils literal"><span class="pre">DeclContext</span></tt> provides two views of the declarations stored within a
+declaration context. The source-centric view accurately represents the
+program source code as written, including multiple declarations of entities
+where present (see the section <a class="reference internal" href="#redeclarations"><em>Redeclarations and Overloads</em></a>), while the semantics-centric view represents the program
+semantics. The two views are kept synchronized by semantic analysis while
+the ASTs are being constructed.</div></blockquote>
+<p>Storage of declarations within that context</p>
+<blockquote>
+<div>Every declaration context can contain some number of declarations. For
+example, a C++ class (represented by <tt class="docutils literal"><span class="pre">RecordDecl</span></tt>) contains various member
+functions, fields, nested types, and so on. All of these declarations will
+be stored within the <tt class="docutils literal"><span class="pre">DeclContext</span></tt>, and one can iterate over the
+declarations via [<tt class="docutils literal"><span class="pre">DeclContext::decls_begin()</span></tt>,
+<tt class="docutils literal"><span class="pre">DeclContext::decls_end()</span></tt>). This mechanism provides the source-centric
+view of declarations in the context.</div></blockquote>
+<p>Lookup of declarations within that context</p>
+<blockquote>
+<div>The <tt class="docutils literal"><span class="pre">DeclContext</span></tt> structure provides efficient name lookup for names within
+that declaration context. For example, if <tt class="docutils literal"><span class="pre">N</span></tt> is a namespace we can look
+for the name <tt class="docutils literal"><span class="pre">N::f</span></tt> using <tt class="docutils literal"><span class="pre">DeclContext::lookup</span></tt>. The lookup itself is
+based on a lazily-constructed array (for declaration contexts with a small
+number of declarations) or hash table (for declaration contexts with more
+declarations). The lookup operation provides the semantics-centric view of
+the declarations in the context.</div></blockquote>
+<p>Ownership of declarations</p>
+<blockquote>
+<div>The <tt class="docutils literal"><span class="pre">DeclContext</span></tt> owns all of the declarations that were declared within
+its declaration context, and is responsible for the management of their
+memory as well as their (de-)serialization.</div></blockquote>
+<p>All declarations are stored within a declaration context, and one can query
+information about the context in which each declaration lives. One can
+retrieve the <tt class="docutils literal"><span class="pre">DeclContext</span></tt> that contains a particular <tt class="docutils literal"><span class="pre">Decl</span></tt> using
+<tt class="docutils literal"><span class="pre">Decl::getDeclContext</span></tt>. However, see the section
+<a class="reference internal" href="#lexicalandsemanticcontexts"><em>Lexical and Semantic Contexts</em></a> for more information about how to interpret
+this context information.</p>
+<div class="section" id="redeclarations-and-overloads">
+<span id="redeclarations"></span><h4><a class="toc-backref" href="#id32">Redeclarations and Overloads</a><a class="headerlink" href="#redeclarations-and-overloads" title="Permalink to this headline">¶</a></h4>
+<p>Within a translation unit, it is common for an entity to be declared several
+times. For example, we might declare a function “<tt class="docutils literal"><span class="pre">f</span></tt>” and then later
+re-declare it as part of an inlined definition:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">void</span> <span class="nf">f</span><span class="p">(</span><span class="kt">int</span> <span class="n">x</span><span class="p">,</span> <span class="kt">int</span> <span class="n">y</span><span class="p">,</span> <span class="kt">int</span> <span class="n">z</span> <span class="o">=</span> <span class="mi">1</span><span class="p">);</span>
+
+<span class="kr">inline</span> <span class="kt">void</span> <span class="nf">f</span><span class="p">(</span><span class="kt">int</span> <span class="n">x</span><span class="p">,</span> <span class="kt">int</span> <span class="n">y</span><span class="p">,</span> <span class="kt">int</span> <span class="n">z</span><span class="p">)</span> <span class="p">{</span> <span class="cm">/* ... */</span> <span class="p">}</span>
+</pre></div>
+</div>
+<p>The representation of “<tt class="docutils literal"><span class="pre">f</span></tt>” differs in the source-centric and
+semantics-centric views of a declaration context. In the source-centric view,
+all redeclarations will be present, in the order they occurred in the source
+code, making this view suitable for clients that wish to see the structure of
+the source code. In the semantics-centric view, only the most recent “<tt class="docutils literal"><span class="pre">f</span></tt>”
+will be found by the lookup, since it effectively replaces the first
+declaration of “<tt class="docutils literal"><span class="pre">f</span></tt>”.</p>
+<p>In the semantics-centric view, overloading of functions is represented
+explicitly. For example, given two declarations of a function “<tt class="docutils literal"><span class="pre">g</span></tt>” that are
+overloaded, e.g.,</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">void</span> <span class="nf">g</span><span class="p">();</span>
+<span class="kt">void</span> <span class="nf">g</span><span class="p">(</span><span class="kt">int</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>the <tt class="docutils literal"><span class="pre">DeclContext::lookup</span></tt> operation will return a
+<tt class="docutils literal"><span class="pre">DeclContext::lookup_result</span></tt> that contains a range of iterators over
+declarations of “<tt class="docutils literal"><span class="pre">g</span></tt>”. Clients that perform semantic analysis on a program
+that is not concerned with the actual source code will primarily use this
+semantics-centric view.</p>
+</div>
+<div class="section" id="lexical-and-semantic-contexts">
+<span id="lexicalandsemanticcontexts"></span><h4><a class="toc-backref" href="#id33">Lexical and Semantic Contexts</a><a class="headerlink" href="#lexical-and-semantic-contexts" title="Permalink to this headline">¶</a></h4>
+<p>Each declaration has two potentially different declaration contexts: a
+<em>lexical</em> context, which corresponds to the source-centric view of the
+declaration context, and a <em>semantic</em> context, which corresponds to the
+semantics-centric view. The lexical context is accessible via
+<tt class="docutils literal"><span class="pre">Decl::getLexicalDeclContext</span></tt> while the semantic context is accessible via
+<tt class="docutils literal"><span class="pre">Decl::getDeclContext</span></tt>, both of which return <tt class="docutils literal"><span class="pre">DeclContext</span></tt> pointers. For
+most declarations, the two contexts are identical. For example:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">class</span> <span class="nc">X</span> <span class="p">{</span>
+<span class="nl">public:</span>
+ <span class="kt">void</span> <span class="n">f</span><span class="p">(</span><span class="kt">int</span> <span class="n">x</span><span class="p">);</span>
+<span class="p">};</span>
+</pre></div>
+</div>
+<p>Here, the semantic and lexical contexts of <tt class="docutils literal"><span class="pre">X::f</span></tt> are the <tt class="docutils literal"><span class="pre">DeclContext</span></tt>
+associated with the class <tt class="docutils literal"><span class="pre">X</span></tt> (itself stored as a <tt class="docutils literal"><span class="pre">RecordDecl</span></tt> AST node).
+However, we can now define <tt class="docutils literal"><span class="pre">X::f</span></tt> out-of-line:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">void</span> <span class="n">X</span><span class="o">::</span><span class="n">f</span><span class="p">(</span><span class="kt">int</span> <span class="n">x</span> <span class="o">=</span> <span class="mi">17</span><span class="p">)</span> <span class="p">{</span> <span class="cm">/* ... */</span> <span class="p">}</span>
+</pre></div>
+</div>
+<p>This definition of “<tt class="docutils literal"><span class="pre">f</span></tt>” has different lexical and semantic contexts. The
+lexical context corresponds to the declaration context in which the actual
+declaration occurred in the source code, e.g., the translation unit containing
+<tt class="docutils literal"><span class="pre">X</span></tt>. Thus, this declaration of <tt class="docutils literal"><span class="pre">X::f</span></tt> can be found by traversing the
+declarations provided by [<tt class="docutils literal"><span class="pre">decls_begin()</span></tt>, <tt class="docutils literal"><span class="pre">decls_end()</span></tt>) in the
+translation unit.</p>
+<p>The semantic context of <tt class="docutils literal"><span class="pre">X::f</span></tt> corresponds to the class <tt class="docutils literal"><span class="pre">X</span></tt>, since this
+member function is (semantically) a member of <tt class="docutils literal"><span class="pre">X</span></tt>. Lookup of the name <tt class="docutils literal"><span class="pre">f</span></tt>
+into the <tt class="docutils literal"><span class="pre">DeclContext</span></tt> associated with <tt class="docutils literal"><span class="pre">X</span></tt> will then return the definition
+of <tt class="docutils literal"><span class="pre">X::f</span></tt> (including information about the default argument).</p>
+</div>
+<div class="section" id="transparent-declaration-contexts">
+<h4><a class="toc-backref" href="#id34">Transparent Declaration Contexts</a><a class="headerlink" href="#transparent-declaration-contexts" title="Permalink to this headline">¶</a></h4>
+<p>In C and C++, there are several contexts in which names that are logically
+declared inside another declaration will actually “leak” out into the enclosing
+scope from the perspective of name lookup. The most obvious instance of this
+behavior is in enumeration types, e.g.,</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">enum</span> <span class="n">Color</span> <span class="p">{</span>
+ <span class="n">Red</span><span class="p">,</span>
+ <span class="n">Green</span><span class="p">,</span>
+ <span class="n">Blue</span>
+<span class="p">};</span>
+</pre></div>
+</div>
+<p>Here, <tt class="docutils literal"><span class="pre">Color</span></tt> is an enumeration, which is a declaration context that contains
+the enumerators <tt class="docutils literal"><span class="pre">Red</span></tt>, <tt class="docutils literal"><span class="pre">Green</span></tt>, and <tt class="docutils literal"><span class="pre">Blue</span></tt>. Thus, traversing the list of
+declarations contained in the enumeration <tt class="docutils literal"><span class="pre">Color</span></tt> will yield <tt class="docutils literal"><span class="pre">Red</span></tt>,
+<tt class="docutils literal"><span class="pre">Green</span></tt>, and <tt class="docutils literal"><span class="pre">Blue</span></tt>. However, outside of the scope of <tt class="docutils literal"><span class="pre">Color</span></tt> one can
+name the enumerator <tt class="docutils literal"><span class="pre">Red</span></tt> without qualifying the name, e.g.,</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">Color</span> <span class="n">c</span> <span class="o">=</span> <span class="n">Red</span><span class="p">;</span>
+</pre></div>
+</div>
+<p>There are other entities in C++ that provide similar behavior. For example,
+linkage specifications that use curly braces:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">extern</span> <span class="s">"C"</span> <span class="p">{</span>
+ <span class="kt">void</span> <span class="n">f</span><span class="p">(</span><span class="kt">int</span><span class="p">);</span>
+ <span class="kt">void</span> <span class="nf">g</span><span class="p">(</span><span class="kt">int</span><span class="p">);</span>
+<span class="p">}</span>
+<span class="c1">// f and g are visible here</span>
+</pre></div>
+</div>
+<p>For source-level accuracy, we treat the linkage specification and enumeration
+type as a declaration context in which its enclosed declarations (“<tt class="docutils literal"><span class="pre">Red</span></tt>”,
+“<tt class="docutils literal"><span class="pre">Green</span></tt>”, and “<tt class="docutils literal"><span class="pre">Blue</span></tt>”; “<tt class="docutils literal"><span class="pre">f</span></tt>” and “<tt class="docutils literal"><span class="pre">g</span></tt>”) are declared. However, these
+declarations are visible outside of the scope of the declaration context.</p>
+<p>These language features (and several others, described below) have roughly the
+same set of requirements: declarations are declared within a particular lexical
+context, but the declarations are also found via name lookup in scopes
+enclosing the declaration itself. This feature is implemented via
+<em>transparent</em> declaration contexts (see
+<tt class="docutils literal"><span class="pre">DeclContext::isTransparentContext()</span></tt>), whose declarations are visible in the
+nearest enclosing non-transparent declaration context. This means that the
+lexical context of the declaration (e.g., an enumerator) will be the
+transparent <tt class="docutils literal"><span class="pre">DeclContext</span></tt> itself, as will the semantic context, but the
+declaration will be visible in every outer context up to and including the
+first non-transparent declaration context (since transparent declaration
+contexts can be nested).</p>
+<p>The transparent <tt class="docutils literal"><span class="pre">DeclContext</span></tt>s are:</p>
+<ul>
+<li><p class="first">Enumerations (but not C++11 “scoped enumerations”):</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">enum</span> <span class="n">Color</span> <span class="p">{</span>
+ <span class="n">Red</span><span class="p">,</span>
+ <span class="n">Green</span><span class="p">,</span>
+ <span class="n">Blue</span>
+<span class="p">};</span>
+<span class="c1">// Red, Green, and Blue are in scope</span>
+</pre></div>
+</div>
+</li>
+<li><p class="first">C++ linkage specifications:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">extern</span> <span class="s">"C"</span> <span class="p">{</span>
+ <span class="kt">void</span> <span class="n">f</span><span class="p">(</span><span class="kt">int</span><span class="p">);</span>
+ <span class="kt">void</span> <span class="nf">g</span><span class="p">(</span><span class="kt">int</span><span class="p">);</span>
+<span class="p">}</span>
+<span class="c1">// f and g are in scope</span>
+</pre></div>
+</div>
+</li>
+<li><p class="first">Anonymous unions and structs:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">struct</span> <span class="n">LookupTable</span> <span class="p">{</span>
+ <span class="kt">bool</span> <span class="n">IsVector</span><span class="p">;</span>
+ <span class="k">union</span> <span class="p">{</span>
+ <span class="n">std</span><span class="o">::</span><span class="n">vector</span><span class="o"><</span><span class="n">Item</span><span class="o">></span> <span class="o">*</span><span class="n">Vector</span><span class="p">;</span>
+ <span class="n">std</span><span class="o">::</span><span class="n">set</span><span class="o"><</span><span class="n">Item</span><span class="o">></span> <span class="o">*</span><span class="n">Set</span><span class="p">;</span>
+ <span class="p">};</span>
+<span class="p">};</span>
+
+<span class="n">LookupTable</span> <span class="n">LT</span><span class="p">;</span>
+<span class="n">LT</span><span class="p">.</span><span class="n">Vector</span> <span class="o">=</span> <span class="mi">0</span><span class="p">;</span> <span class="c1">// Okay: finds Vector inside the unnamed union</span>
+</pre></div>
+</div>
+</li>
+<li><p class="first">C++11 inline namespaces:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">namespace</span> <span class="n">mylib</span> <span class="p">{</span>
+ <span class="kr">inline</span> <span class="k">namespace</span> <span class="n">debug</span> <span class="p">{</span>
+ <span class="k">class</span> <span class="nc">X</span><span class="p">;</span>
+ <span class="p">}</span>
+<span class="p">}</span>
+<span class="n">mylib</span><span class="o">::</span><span class="n">X</span> <span class="o">*</span><span class="n">xp</span><span class="p">;</span> <span class="c1">// okay: mylib::X refers to mylib::debug::X</span>
+</pre></div>
+</div>
+</li>
+</ul>
+</div>
+<div class="section" id="multiply-defined-declaration-contexts">
+<span id="multideclcontext"></span><h4><a class="toc-backref" href="#id35">Multiply-Defined Declaration Contexts</a><a class="headerlink" href="#multiply-defined-declaration-contexts" title="Permalink to this headline">¶</a></h4>
+<p>C++ namespaces have the interesting — and, so far, unique — property that
+the namespace can be defined multiple times, and the declarations provided by
+each namespace definition are effectively merged (from the semantic point of
+view). For example, the following two code snippets are semantically
+indistinguishable:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="c1">// Snippet #1:</span>
+<span class="k">namespace</span> <span class="n">N</span> <span class="p">{</span>
+ <span class="kt">void</span> <span class="n">f</span><span class="p">();</span>
+<span class="p">}</span>
+<span class="k">namespace</span> <span class="n">N</span> <span class="p">{</span>
+ <span class="kt">void</span> <span class="n">f</span><span class="p">(</span><span class="kt">int</span><span class="p">);</span>
+<span class="p">}</span>
+
+<span class="c1">// Snippet #2:</span>
+<span class="k">namespace</span> <span class="n">N</span> <span class="p">{</span>
+ <span class="kt">void</span> <span class="n">f</span><span class="p">();</span>
+ <span class="kt">void</span> <span class="nf">f</span><span class="p">(</span><span class="kt">int</span><span class="p">);</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>In Clang’s representation, the source-centric view of declaration contexts will
+actually have two separate <tt class="docutils literal"><span class="pre">NamespaceDecl</span></tt> nodes in Snippet #1, each of which
+is a declaration context that contains a single declaration of “<tt class="docutils literal"><span class="pre">f</span></tt>”.
+However, the semantics-centric view provided by name lookup into the namespace
+<tt class="docutils literal"><span class="pre">N</span></tt> for “<tt class="docutils literal"><span class="pre">f</span></tt>” will return a <tt class="docutils literal"><span class="pre">DeclContext::lookup_result</span></tt> that contains a
+range of iterators over declarations of “<tt class="docutils literal"><span class="pre">f</span></tt>”.</p>
+<p><tt class="docutils literal"><span class="pre">DeclContext</span></tt> manages multiply-defined declaration contexts internally. The
+function <tt class="docutils literal"><span class="pre">DeclContext::getPrimaryContext</span></tt> retrieves the “primary” context for
+a given <tt class="docutils literal"><span class="pre">DeclContext</span></tt> instance, which is the <tt class="docutils literal"><span class="pre">DeclContext</span></tt> responsible for
+maintaining the lookup table used for the semantics-centric view. Given a
+DeclContext, one can obtain the set of declaration contexts that are semanticaly
+connected to this declaration context, in source order, including this context
+(which will be the only result, for non-namespace contexts) via
+<tt class="docutils literal"><span class="pre">DeclContext::collectAllContexts</span></tt>. Note that these functions are used
+internally within the lookup and insertion methods of the <tt class="docutils literal"><span class="pre">DeclContext</span></tt>, so
+the vast majority of clients can ignore them.</p>
+</div>
+</div>
+<div class="section" id="the-cfg-class">
+<span id="cfg"></span><h3><a class="toc-backref" href="#id36">The <tt class="docutils literal"><span class="pre">CFG</span></tt> class</a><a class="headerlink" href="#the-cfg-class" title="Permalink to this headline">¶</a></h3>
+<p>The <tt class="docutils literal"><span class="pre">CFG</span></tt> class is designed to represent a source-level control-flow graph
+for a single statement (<tt class="docutils literal"><span class="pre">Stmt*</span></tt>). Typically instances of <tt class="docutils literal"><span class="pre">CFG</span></tt> are
+constructed for function bodies (usually an instance of <tt class="docutils literal"><span class="pre">CompoundStmt</span></tt>), but
+can also be instantiated to represent the control-flow of any class that
+subclasses <tt class="docutils literal"><span class="pre">Stmt</span></tt>, which includes simple expressions. Control-flow graphs
+are especially useful for performing <a class="reference external" href="http://en.wikipedia.org/wiki/Data_flow_analysis#Sensitivities">flow- or path-sensitive</a> program
+analyses on a given function.</p>
+<div class="section" id="basic-blocks">
+<h4><a class="toc-backref" href="#id37">Basic Blocks</a><a class="headerlink" href="#basic-blocks" title="Permalink to this headline">¶</a></h4>
+<p>Concretely, an instance of <tt class="docutils literal"><span class="pre">CFG</span></tt> is a collection of basic blocks. Each basic
+block is an instance of <tt class="docutils literal"><span class="pre">CFGBlock</span></tt>, which simply contains an ordered sequence
+of <tt class="docutils literal"><span class="pre">Stmt*</span></tt> (each referring to statements in the AST). The ordering of
+statements within a block indicates unconditional flow of control from one
+statement to the next. <a class="reference internal" href="#conditionalcontrolflow"><em>Conditional control-flow</em></a> is represented using edges between basic blocks. The
+statements within a given <tt class="docutils literal"><span class="pre">CFGBlock</span></tt> can be traversed using the
+<tt class="docutils literal"><span class="pre">CFGBlock::*iterator</span></tt> interface.</p>
+<p>A <tt class="docutils literal"><span class="pre">CFG</span></tt> object owns the instances of <tt class="docutils literal"><span class="pre">CFGBlock</span></tt> within the control-flow
+graph it represents. Each <tt class="docutils literal"><span class="pre">CFGBlock</span></tt> within a CFG is also uniquely numbered
+(accessible via <tt class="docutils literal"><span class="pre">CFGBlock::getBlockID()</span></tt>). Currently the number is based on
+the ordering the blocks were created, but no assumptions should be made on how
+<tt class="docutils literal"><span class="pre">CFGBlocks</span></tt> are numbered other than their numbers are unique and that they
+are numbered from 0..N-1 (where N is the number of basic blocks in the CFG).</p>
+</div>
+<div class="section" id="entry-and-exit-blocks">
+<h4><a class="toc-backref" href="#id38">Entry and Exit Blocks</a><a class="headerlink" href="#entry-and-exit-blocks" title="Permalink to this headline">¶</a></h4>
+<p>Each instance of <tt class="docutils literal"><span class="pre">CFG</span></tt> contains two special blocks: an <em>entry</em> block
+(accessible via <tt class="docutils literal"><span class="pre">CFG::getEntry()</span></tt>), which has no incoming edges, and an
+<em>exit</em> block (accessible via <tt class="docutils literal"><span class="pre">CFG::getExit()</span></tt>), which has no outgoing edges.
+Neither block contains any statements, and they serve the role of providing a
+clear entrance and exit for a body of code such as a function body. The
+presence of these empty blocks greatly simplifies the implementation of many
+analyses built on top of CFGs.</p>
+</div>
+<div class="section" id="conditional-control-flow">
+<span id="conditionalcontrolflow"></span><h4><a class="toc-backref" href="#id39">Conditional Control-Flow</a><a class="headerlink" href="#conditional-control-flow" title="Permalink to this headline">¶</a></h4>
+<p>Conditional control-flow (such as those induced by if-statements and loops) is
+represented as edges between <tt class="docutils literal"><span class="pre">CFGBlocks</span></tt>. Because different C language
+constructs can induce control-flow, each <tt class="docutils literal"><span class="pre">CFGBlock</span></tt> also records an extra
+<tt class="docutils literal"><span class="pre">Stmt*</span></tt> that represents the <em>terminator</em> of the block. A terminator is
+simply the statement that caused the control-flow, and is used to identify the
+nature of the conditional control-flow between blocks. For example, in the
+case of an if-statement, the terminator refers to the <tt class="docutils literal"><span class="pre">IfStmt</span></tt> object in the
+AST that represented the given branch.</p>
+<p>To illustrate, consider the following code example:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">int</span> <span class="nf">foo</span><span class="p">(</span><span class="kt">int</span> <span class="n">x</span><span class="p">)</span> <span class="p">{</span>
+ <span class="n">x</span> <span class="o">=</span> <span class="n">x</span> <span class="o">+</span> <span class="mi">1</span><span class="p">;</span>
+ <span class="k">if</span> <span class="p">(</span><span class="n">x</span> <span class="o">></span> <span class="mi">2</span><span class="p">)</span>
+ <span class="n">x</span><span class="o">++</span><span class="p">;</span>
+ <span class="k">else</span> <span class="p">{</span>
+ <span class="n">x</span> <span class="o">+=</span> <span class="mi">2</span><span class="p">;</span>
+ <span class="n">x</span> <span class="o">*=</span> <span class="mi">2</span><span class="p">;</span>
+ <span class="p">}</span>
+
+ <span class="k">return</span> <span class="n">x</span><span class="p">;</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>After invoking the parser+semantic analyzer on this code fragment, the AST of
+the body of <tt class="docutils literal"><span class="pre">foo</span></tt> is referenced by a single <tt class="docutils literal"><span class="pre">Stmt*</span></tt>. We can then construct
+an instance of <tt class="docutils literal"><span class="pre">CFG</span></tt> representing the control-flow graph of this function
+body by single call to a static class method:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">Stmt</span> <span class="o">*</span><span class="n">FooBody</span> <span class="o">=</span> <span class="p">...</span>
+<span class="n">std</span><span class="o">::</span><span class="n">unique_ptr</span><span class="o"><</span><span class="n">CFG</span><span class="o">></span> <span class="n">FooCFG</span> <span class="o">=</span> <span class="n">CFG</span><span class="o">::</span><span class="n">buildCFG</span><span class="p">(</span><span class="n">FooBody</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>Along with providing an interface to iterate over its <tt class="docutils literal"><span class="pre">CFGBlocks</span></tt>, the
+<tt class="docutils literal"><span class="pre">CFG</span></tt> class also provides methods that are useful for debugging and
+visualizing CFGs. For example, the method <tt class="docutils literal"><span class="pre">CFG::dump()</span></tt> dumps a
+pretty-printed version of the CFG to standard error. This is especially useful
+when one is using a debugger such as gdb. For example, here is the output of
+<tt class="docutils literal"><span class="pre">FooCFG->dump()</span></tt>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="p">[</span> <span class="n">B5</span> <span class="p">(</span><span class="n">ENTRY</span><span class="p">)</span> <span class="p">]</span>
+ <span class="n">Predecessors</span> <span class="p">(</span><span class="mi">0</span><span class="p">)</span><span class="o">:</span>
+ <span class="n">Successors</span> <span class="p">(</span><span class="mi">1</span><span class="p">)</span><span class="o">:</span> <span class="n">B4</span>
+
+<span class="p">[</span> <span class="n">B4</span> <span class="p">]</span>
+ <span class="mi">1</span><span class="o">:</span> <span class="n">x</span> <span class="o">=</span> <span class="n">x</span> <span class="o">+</span> <span class="mi">1</span>
+ <span class="mi">2</span><span class="o">:</span> <span class="p">(</span><span class="n">x</span> <span class="o">></span> <span class="mi">2</span><span class="p">)</span>
+ <span class="nl">T:</span> <span class="k">if</span> <span class="p">[</span><span class="n">B4</span><span class="mf">.2</span><span class="p">]</span>
+ <span class="n">Predecessors</span> <span class="p">(</span><span class="mi">1</span><span class="p">)</span><span class="o">:</span> <span class="n">B5</span>
+ <span class="n">Successors</span> <span class="p">(</span><span class="mi">2</span><span class="p">)</span><span class="o">:</span> <span class="n">B3</span> <span class="n">B2</span>
+
+<span class="p">[</span> <span class="n">B3</span> <span class="p">]</span>
+ <span class="mi">1</span><span class="o">:</span> <span class="n">x</span><span class="o">++</span>
+ <span class="n">Predecessors</span> <span class="p">(</span><span class="mi">1</span><span class="p">)</span><span class="o">:</span> <span class="n">B4</span>
+ <span class="n">Successors</span> <span class="p">(</span><span class="mi">1</span><span class="p">)</span><span class="o">:</span> <span class="n">B1</span>
+
+<span class="p">[</span> <span class="n">B2</span> <span class="p">]</span>
+ <span class="mi">1</span><span class="o">:</span> <span class="n">x</span> <span class="o">+=</span> <span class="mi">2</span>
+ <span class="mi">2</span><span class="o">:</span> <span class="n">x</span> <span class="o">*=</span> <span class="mi">2</span>
+ <span class="n">Predecessors</span> <span class="p">(</span><span class="mi">1</span><span class="p">)</span><span class="o">:</span> <span class="n">B4</span>
+ <span class="n">Successors</span> <span class="p">(</span><span class="mi">1</span><span class="p">)</span><span class="o">:</span> <span class="n">B1</span>
+
+<span class="p">[</span> <span class="n">B1</span> <span class="p">]</span>
+ <span class="mi">1</span><span class="o">:</span> <span class="k">return</span> <span class="n">x</span><span class="p">;</span>
+ <span class="n">Predecessors</span> <span class="p">(</span><span class="mi">2</span><span class="p">)</span><span class="o">:</span> <span class="n">B2</span> <span class="n">B3</span>
+ <span class="n">Successors</span> <span class="p">(</span><span class="mi">1</span><span class="p">)</span><span class="o">:</span> <span class="n">B0</span>
+
+<span class="p">[</span> <span class="n">B0</span> <span class="p">(</span><span class="n">EXIT</span><span class="p">)</span> <span class="p">]</span>
+ <span class="n">Predecessors</span> <span class="p">(</span><span class="mi">1</span><span class="p">)</span><span class="o">:</span> <span class="n">B1</span>
+ <span class="n">Successors</span> <span class="p">(</span><span class="mi">0</span><span class="p">)</span><span class="o">:</span>
+</pre></div>
+</div>
+<p>For each block, the pretty-printed output displays for each block the number of
+<em>predecessor</em> blocks (blocks that have outgoing control-flow to the given
+block) and <em>successor</em> blocks (blocks that have control-flow that have incoming
+control-flow from the given block). We can also clearly see the special entry
+and exit blocks at the beginning and end of the pretty-printed output. For the
+entry block (block B5), the number of predecessor blocks is 0, while for the
+exit block (block B0) the number of successor blocks is 0.</p>
+<p>The most interesting block here is B4, whose outgoing control-flow represents
+the branching caused by the sole if-statement in <tt class="docutils literal"><span class="pre">foo</span></tt>. Of particular
+interest is the second statement in the block, <tt class="docutils literal"><span class="pre">(x</span> <span class="pre">></span> <span class="pre">2)</span></tt>, and the terminator,
+printed as <tt class="docutils literal"><span class="pre">if</span> <span class="pre">[B4.2]</span></tt>. The second statement represents the evaluation of
+the condition of the if-statement, which occurs before the actual branching of
+control-flow. Within the <tt class="docutils literal"><span class="pre">CFGBlock</span></tt> for B4, the <tt class="docutils literal"><span class="pre">Stmt*</span></tt> for the second
+statement refers to the actual expression in the AST for <tt class="docutils literal"><span class="pre">(x</span> <span class="pre">></span> <span class="pre">2)</span></tt>. Thus
+pointers to subclasses of <tt class="docutils literal"><span class="pre">Expr</span></tt> can appear in the list of statements in a
+block, and not just subclasses of <tt class="docutils literal"><span class="pre">Stmt</span></tt> that refer to proper C statements.</p>
+<p>The terminator of block B4 is a pointer to the <tt class="docutils literal"><span class="pre">IfStmt</span></tt> object in the AST.
+The pretty-printer outputs <tt class="docutils literal"><span class="pre">if</span> <span class="pre">[B4.2]</span></tt> because the condition expression of
+the if-statement has an actual place in the basic block, and thus the
+terminator is essentially <em>referring</em> to the expression that is the second
+statement of block B4 (i.e., B4.2). In this manner, conditions for
+control-flow (which also includes conditions for loops and switch statements)
+are hoisted into the actual basic block.</p>
+</div>
+</div>
+<div class="section" id="constant-folding-in-the-clang-ast">
+<h3><a class="toc-backref" href="#id40">Constant Folding in the Clang AST</a><a class="headerlink" href="#constant-folding-in-the-clang-ast" title="Permalink to this headline">¶</a></h3>
+<p>There are several places where constants and constant folding matter a lot to
+the Clang front-end. First, in general, we prefer the AST to retain the source
+code as close to how the user wrote it as possible. This means that if they
+wrote “<tt class="docutils literal"><span class="pre">5+4</span></tt>”, we want to keep the addition and two constants in the AST, we
+don’t want to fold to “<tt class="docutils literal"><span class="pre">9</span></tt>”. This means that constant folding in various
+ways turns into a tree walk that needs to handle the various cases.</p>
+<p>However, there are places in both C and C++ that require constants to be
+folded. For example, the C standard defines what an “integer constant
+expression” (i-c-e) is with very precise and specific requirements. The
+language then requires i-c-e’s in a lot of places (for example, the size of a
+bitfield, the value for a case statement, etc). For these, we have to be able
+to constant fold the constants, to do semantic checks (e.g., verify bitfield
+size is non-negative and that case statements aren’t duplicated). We aim for
+Clang to be very pedantic about this, diagnosing cases when the code does not
+use an i-c-e where one is required, but accepting the code unless running with
+<tt class="docutils literal"><span class="pre">-pedantic-errors</span></tt>.</p>
+<p>Things get a little bit more tricky when it comes to compatibility with
+real-world source code. Specifically, GCC has historically accepted a huge
+superset of expressions as i-c-e’s, and a lot of real world code depends on
+this unfortuate accident of history (including, e.g., the glibc system
+headers). GCC accepts anything its “fold” optimizer is capable of reducing to
+an integer constant, which means that the definition of what it accepts changes
+as its optimizer does. One example is that GCC accepts things like “<tt class="docutils literal"><span class="pre">case</span>
+<span class="pre">X-X:</span></tt>” even when <tt class="docutils literal"><span class="pre">X</span></tt> is a variable, because it can fold this to 0.</p>
+<p>Another issue are how constants interact with the extensions we support, such
+as <tt class="docutils literal"><span class="pre">__builtin_constant_p</span></tt>, <tt class="docutils literal"><span class="pre">__builtin_inf</span></tt>, <tt class="docutils literal"><span class="pre">__extension__</span></tt> and many
+others. C99 obviously does not specify the semantics of any of these
+extensions, and the definition of i-c-e does not include them. However, these
+extensions are often used in real code, and we have to have a way to reason
+about them.</p>
+<p>Finally, this is not just a problem for semantic analysis. The code generator
+and other clients have to be able to fold constants (e.g., to initialize global
+variables) and has to handle a superset of what C99 allows. Further, these
+clients can benefit from extended information. For example, we know that
+“<tt class="docutils literal"><span class="pre">foo()</span> <span class="pre">||</span> <span class="pre">1</span></tt>” always evaluates to <tt class="docutils literal"><span class="pre">true</span></tt>, but we can’t replace the
+expression with <tt class="docutils literal"><span class="pre">true</span></tt> because it has side effects.</p>
+<div class="section" id="implementation-approach">
+<h4><a class="toc-backref" href="#id41">Implementation Approach</a><a class="headerlink" href="#implementation-approach" title="Permalink to this headline">¶</a></h4>
+<p>After trying several different approaches, we’ve finally converged on a design
+(Note, at the time of this writing, not all of this has been implemented,
+consider this a design goal!). Our basic approach is to define a single
+recursive method evaluation method (<tt class="docutils literal"><span class="pre">Expr::Evaluate</span></tt>), which is implemented
+in <tt class="docutils literal"><span class="pre">AST/ExprConstant.cpp</span></tt>. Given an expression with “scalar” type (integer,
+fp, complex, or pointer) this method returns the following information:</p>
+<ul class="simple">
+<li>Whether the expression is an integer constant expression, a general constant
+that was folded but has no side effects, a general constant that was folded
+but that does have side effects, or an uncomputable/unfoldable value.</li>
+<li>If the expression was computable in any way, this method returns the
+<tt class="docutils literal"><span class="pre">APValue</span></tt> for the result of the expression.</li>
+<li>If the expression is not evaluatable at all, this method returns information
+on one of the problems with the expression. This includes a
+<tt class="docutils literal"><span class="pre">SourceLocation</span></tt> for where the problem is, and a diagnostic ID that explains
+the problem. The diagnostic should have <tt class="docutils literal"><span class="pre">ERROR</span></tt> type.</li>
+<li>If the expression is not an integer constant expression, this method returns
+information on one of the problems with the expression. This includes a
+<tt class="docutils literal"><span class="pre">SourceLocation</span></tt> for where the problem is, and a diagnostic ID that
+explains the problem. The diagnostic should have <tt class="docutils literal"><span class="pre">EXTENSION</span></tt> type.</li>
+</ul>
+<p>This information gives various clients the flexibility that they want, and we
+will eventually have some helper methods for various extensions. For example,
+<tt class="docutils literal"><span class="pre">Sema</span></tt> should have a <tt class="docutils literal"><span class="pre">Sema::VerifyIntegerConstantExpression</span></tt> method, which
+calls <tt class="docutils literal"><span class="pre">Evaluate</span></tt> on the expression. If the expression is not foldable, the
+error is emitted, and it would return <tt class="docutils literal"><span class="pre">true</span></tt>. If the expression is not an
+i-c-e, the <tt class="docutils literal"><span class="pre">EXTENSION</span></tt> diagnostic is emitted. Finally it would return
+<tt class="docutils literal"><span class="pre">false</span></tt> to indicate that the AST is OK.</p>
+<p>Other clients can use the information in other ways, for example, codegen can
+just use expressions that are foldable in any way.</p>
+</div>
+<div class="section" id="extensions">
+<h4><a class="toc-backref" href="#id42">Extensions</a><a class="headerlink" href="#extensions" title="Permalink to this headline">¶</a></h4>
+<p>This section describes how some of the various extensions Clang supports
+interacts with constant evaluation:</p>
+<ul class="simple">
+<li><tt class="docutils literal"><span class="pre">__extension__</span></tt>: The expression form of this extension causes any
+evaluatable subexpression to be accepted as an integer constant expression.</li>
+<li><tt class="docutils literal"><span class="pre">__builtin_constant_p</span></tt>: This returns true (as an integer constant
+expression) if the operand evaluates to either a numeric value (that is, not
+a pointer cast to integral type) of integral, enumeration, floating or
+complex type, or if it evaluates to the address of the first character of a
+string literal (possibly cast to some other type). As a special case, if
+<tt class="docutils literal"><span class="pre">__builtin_constant_p</span></tt> is the (potentially parenthesized) condition of a
+conditional operator expression (“<tt class="docutils literal"><span class="pre">?:</span></tt>”), only the true side of the
+conditional operator is considered, and it is evaluated with full constant
+folding.</li>
+<li><tt class="docutils literal"><span class="pre">__builtin_choose_expr</span></tt>: The condition is required to be an integer
+constant expression, but we accept any constant as an “extension of an
+extension”. This only evaluates one operand depending on which way the
+condition evaluates.</li>
+<li><tt class="docutils literal"><span class="pre">__builtin_classify_type</span></tt>: This always returns an integer constant
+expression.</li>
+<li><tt class="docutils literal"><span class="pre">__builtin_inf,</span> <span class="pre">nan,</span> <span class="pre">...</span></tt>: These are treated just like a floating-point
+literal.</li>
+<li><tt class="docutils literal"><span class="pre">__builtin_abs,</span> <span class="pre">copysign,</span> <span class="pre">...</span></tt>: These are constant folded as general
+constant expressions.</li>
+<li><tt class="docutils literal"><span class="pre">__builtin_strlen</span></tt> and <tt class="docutils literal"><span class="pre">strlen</span></tt>: These are constant folded as integer
+constant expressions if the argument is a string literal.</li>
+</ul>
+</div>
+</div>
+</div>
+<div class="section" id="the-sema-library">
+<span id="sema"></span><h2><a class="toc-backref" href="#id43">The Sema Library</a><a class="headerlink" href="#the-sema-library" title="Permalink to this headline">¶</a></h2>
+<p>This library is called by the <a class="reference internal" href="#parser"><em>Parser library</em></a> during parsing to
+do semantic analysis of the input. For valid programs, Sema builds an AST for
+parsed constructs.</p>
+</div>
+<div class="section" id="the-codegen-library">
+<span id="codegen"></span><h2><a class="toc-backref" href="#id44">The CodeGen Library</a><a class="headerlink" href="#the-codegen-library" title="Permalink to this headline">¶</a></h2>
+<p>CodeGen takes an <a class="reference internal" href="#ast"><em>AST</em></a> as input and produces <a class="reference external" href="//llvm.org/docs/LangRef.html">LLVM IR code</a> from it.</p>
+</div>
+<div class="section" id="how-to-change-clang">
+<h2><a class="toc-backref" href="#id45">How to change Clang</a><a class="headerlink" href="#how-to-change-clang" title="Permalink to this headline">¶</a></h2>
+<div class="section" id="how-to-add-an-attribute">
+<h3><a class="toc-backref" href="#id46">How to add an attribute</a><a class="headerlink" href="#how-to-add-an-attribute" title="Permalink to this headline">¶</a></h3>
+<p>Attributes are a form of metadata that can be attached to a program construct,
+allowing the programmer to pass semantic information along to the compiler for
+various uses. For example, attributes may be used to alter the code generation
+for a program construct, or to provide extra semantic information for static
+analysis. This document explains how to add a custom attribute to Clang.
+Documentation on existing attributes can be found <a class="reference external" href="//clang.llvm.org/docs/AttributeReference.html">here</a>.</p>
+<div class="section" id="attribute-basics">
+<h4><a class="toc-backref" href="#id47">Attribute Basics</a><a class="headerlink" href="#attribute-basics" title="Permalink to this headline">¶</a></h4>
+<p>Attributes in Clang are handled in three stages: parsing into a parsed attribute
+representation, conversion from a parsed attribute into a semantic attribute,
+and then the semantic handling of the attribute.</p>
+<p>Parsing of the attribute is determined by the various syntactic forms attributes
+can take, such as GNU, C++11, and Microsoft style attributes, as well as other
+information provided by the table definition of the attribute. Ultimately, the
+parsed representation of an attribute object is an <tt class="docutils literal"><span class="pre">AttributeList</span></tt> object.
+These parsed attributes chain together as a list of parsed attributes attached
+to a declarator or declaration specifier. The parsing of attributes is handled
+automatically by Clang, except for attributes spelled as keywords. When
+implementing a keyword attribute, the parsing of the keyword and creation of the
+<tt class="docutils literal"><span class="pre">AttributeList</span></tt> object must be done manually.</p>
+<p>Eventually, <tt class="docutils literal"><span class="pre">Sema::ProcessDeclAttributeList()</span></tt> is called with a <tt class="docutils literal"><span class="pre">Decl</span></tt> and
+an <tt class="docutils literal"><span class="pre">AttributeList</span></tt>, at which point the parsed attribute can be transformed
+into a semantic attribute. The process by which a parsed attribute is converted
+into a semantic attribute depends on the attribute definition and semantic
+requirements of the attribute. The end result, however, is that the semantic
+attribute object is attached to the <tt class="docutils literal"><span class="pre">Decl</span></tt> object, and can be obtained by a
+call to <tt class="docutils literal"><span class="pre">Decl::getAttr<T>()</span></tt>.</p>
+<p>The structure of the semantic attribute is also governed by the attribute
+definition given in Attr.td. This definition is used to automatically generate
+functionality used for the implementation of the attribute, such as a class
+derived from <tt class="docutils literal"><span class="pre">clang::Attr</span></tt>, information for the parser to use, automated
+semantic checking for some attributes, etc.</p>
+</div>
+<div class="section" id="include-clang-basic-attr-td">
+<h4><a class="toc-backref" href="#id48"><tt class="docutils literal"><span class="pre">include/clang/Basic/Attr.td</span></tt></a><a class="headerlink" href="#include-clang-basic-attr-td" title="Permalink to this headline">¶</a></h4>
+<p>The first step to adding a new attribute to Clang is to add its definition to
+<a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/Attr.td?view=markup">include/clang/Basic/Attr.td</a>.
+This tablegen definition must derive from the <tt class="docutils literal"><span class="pre">Attr</span></tt> (tablegen, not
+semantic) type, or one of its derivatives. Most attributes will derive from the
+<tt class="docutils literal"><span class="pre">InheritableAttr</span></tt> type, which specifies that the attribute can be inherited by
+later redeclarations of the <tt class="docutils literal"><span class="pre">Decl</span></tt> it is associated with.
+<tt class="docutils literal"><span class="pre">InheritableParamAttr</span></tt> is similar to <tt class="docutils literal"><span class="pre">InheritableAttr</span></tt>, except that the
+attribute is written on a parameter instead of a declaration. If the attribute
+is intended to apply to a type instead of a declaration, such an attribute
+should derive from <tt class="docutils literal"><span class="pre">TypeAttr</span></tt>, and will generally not be given an AST
+representation. (Note that this document does not cover the creation of type
+attributes.) An attribute that inherits from <tt class="docutils literal"><span class="pre">IgnoredAttr</span></tt> is parsed, but will
+generate an ignored attribute diagnostic when used, which may be useful when an
+attribute is supported by another vendor but not supported by clang.</p>
+<p>The definition will specify several key pieces of information, such as the
+semantic name of the attribute, the spellings the attribute supports, the
+arguments the attribute expects, and more. Most members of the <tt class="docutils literal"><span class="pre">Attr</span></tt> tablegen
+type do not require definitions in the derived definition as the default
+suffice. However, every attribute must specify at least a spelling list, a
+subject list, and a documentation list.</p>
+<div class="section" id="spellings">
+<h5><a class="toc-backref" href="#id49">Spellings</a><a class="headerlink" href="#spellings" title="Permalink to this headline">¶</a></h5>
+<p>All attributes are required to specify a spelling list that denotes the ways in
+which the attribute can be spelled. For instance, a single semantic attribute
+may have a keyword spelling, as well as a C++11 spelling and a GNU spelling. An
+empty spelling list is also permissible and may be useful for attributes which
+are created implicitly. The following spellings are accepted:</p>
+<blockquote>
+<div><table border="1" class="docutils">
+<colgroup>
+<col width="16%" />
+<col width="84%" />
+</colgroup>
+<thead valign="bottom">
+<tr class="row-odd"><th class="head">Spelling</th>
+<th class="head">Description</th>
+</tr>
+</thead>
+<tbody valign="top">
+<tr class="row-even"><td><tt class="docutils literal"><span class="pre">GNU</span></tt></td>
+<td>Spelled with a GNU-style <tt class="docutils literal"><span class="pre">__attribute__((attr))</span></tt> syntax and
+placement.</td>
+</tr>
+<tr class="row-odd"><td><tt class="docutils literal"><span class="pre">CXX11</span></tt></td>
+<td>Spelled with a C++-style <tt class="docutils literal"><span class="pre">[[attr]]</span></tt> syntax. If the attribute
+is meant to be used by Clang, it should set the namespace to
+<tt class="docutils literal"><span class="pre">"clang"</span></tt>.</td>
+</tr>
+<tr class="row-even"><td><tt class="docutils literal"><span class="pre">Declspec</span></tt></td>
+<td>Spelled with a Microsoft-style <tt class="docutils literal"><span class="pre">__declspec(attr)</span></tt> syntax.</td>
+</tr>
+<tr class="row-odd"><td><tt class="docutils literal"><span class="pre">Keyword</span></tt></td>
+<td>The attribute is spelled as a keyword, and required custom
+parsing.</td>
+</tr>
+<tr class="row-even"><td><tt class="docutils literal"><span class="pre">GCC</span></tt></td>
+<td>Specifies two spellings: the first is a GNU-style spelling, and
+the second is a C++-style spelling with the <tt class="docutils literal"><span class="pre">gnu</span></tt> namespace.
+Attributes should only specify this spelling for attributes
+supported by GCC.</td>
+</tr>
+<tr class="row-odd"><td><tt class="docutils literal"><span class="pre">Pragma</span></tt></td>
+<td>The attribute is spelled as a <tt class="docutils literal"><span class="pre">#pragma</span></tt>, and requires custom
+processing within the preprocessor. If the attribute is meant to
+be used by Clang, it should set the namespace to <tt class="docutils literal"><span class="pre">"clang"</span></tt>.
+Note that this spelling is not used for declaration attributes.</td>
+</tr>
+</tbody>
+</table>
+</div></blockquote>
+</div>
+<div class="section" id="subjects">
+<h5><a class="toc-backref" href="#id50">Subjects</a><a class="headerlink" href="#subjects" title="Permalink to this headline">¶</a></h5>
+<p>Attributes appertain to one or more <tt class="docutils literal"><span class="pre">Decl</span></tt> subjects. If the attribute attempts
+to attach to a subject that is not in the subject list, a diagnostic is issued
+automatically. Whether the diagnostic is a warning or an error depends on how
+the attribute’s <tt class="docutils literal"><span class="pre">SubjectList</span></tt> is defined, but the default behavior is to warn.
+The diagnostics displayed to the user are automatically determined based on the
+subjects in the list, but a custom diagnostic parameter can also be specified in
+the <tt class="docutils literal"><span class="pre">SubjectList</span></tt>. The diagnostics generated for subject list violations are
+either <tt class="docutils literal"><span class="pre">diag::warn_attribute_wrong_decl_type</span></tt> or
+<tt class="docutils literal"><span class="pre">diag::err_attribute_wrong_decl_type</span></tt>, and the parameter enumeration is found
+in <a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Sema/AttributeList.h?view=markup">include/clang/Sema/AttributeList.h</a>
+If a previously unused Decl node is added to the <tt class="docutils literal"><span class="pre">SubjectList</span></tt>, the logic used
+to automatically determine the diagnostic parameter in <a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/utils/TableGen/ClangAttrEmitter.cpp?view=markup">utils/TableGen/ClangAttrEmitter.cpp</a>
+may need to be updated.</p>
+<p>By default, all subjects in the SubjectList must either be a Decl node defined
+in <tt class="docutils literal"><span class="pre">DeclNodes.td</span></tt>, or a statement node defined in <tt class="docutils literal"><span class="pre">StmtNodes.td</span></tt>. However,
+more complex subjects can be created by creating a <tt class="docutils literal"><span class="pre">SubsetSubject</span></tt> object.
+Each such object has a base subject which it appertains to (which must be a
+Decl or Stmt node, and not a SubsetSubject node), and some custom code which is
+called when determining whether an attribute appertains to the subject. For
+instance, a <tt class="docutils literal"><span class="pre">NonBitField</span></tt> SubsetSubject appertains to a <tt class="docutils literal"><span class="pre">FieldDecl</span></tt>, and
+tests whether the given FieldDecl is a bit field. When a SubsetSubject is
+specified in a SubjectList, a custom diagnostic parameter must also be provided.</p>
+<p>Diagnostic checking for attribute subject lists is automated except when
+<tt class="docutils literal"><span class="pre">HasCustomParsing</span></tt> is set to <tt class="docutils literal"><span class="pre">1</span></tt>.</p>
+</div>
+<div class="section" id="documentation">
+<h5><a class="toc-backref" href="#id51">Documentation</a><a class="headerlink" href="#documentation" title="Permalink to this headline">¶</a></h5>
+<p>All attributes must have some form of documentation associated with them.
+Documentation is table generated on the public web server by a server-side
+process that runs daily. Generally, the documentation for an attribute is a
+stand-alone definition in <a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/AttdDocs.td?view=markup">include/clang/Basic/AttrDocs.td</a>
+that is named after the attribute being documented.</p>
+<p>If the attribute is not for public consumption, or is an implicitly-created
+attribute that has no visible spelling, the documentation list can specify the
+<tt class="docutils literal"><span class="pre">Undocumented</span></tt> object. Otherwise, the attribute should have its documentation
+added to AttrDocs.td.</p>
+<p>Documentation derives from the <tt class="docutils literal"><span class="pre">Documentation</span></tt> tablegen type. All derived
+types must specify a documentation category and the actual documentation itself.
+Additionally, it can specify a custom heading for the attribute, though a
+default heading will be chosen when possible.</p>
+<p>There are four predefined documentation categories: <tt class="docutils literal"><span class="pre">DocCatFunction</span></tt> for
+attributes that appertain to function-like subjects, <tt class="docutils literal"><span class="pre">DocCatVariable</span></tt> for
+attributes that appertain to variable-like subjects, <tt class="docutils literal"><span class="pre">DocCatType</span></tt> for type
+attributes, and <tt class="docutils literal"><span class="pre">DocCatStmt</span></tt> for statement attributes. A custom documentation
+category should be used for groups of attributes with similar functionality.
+Custom categories are good for providing overview information for the attributes
+grouped under it. For instance, the consumed annotation attributes define a
+custom category, <tt class="docutils literal"><span class="pre">DocCatConsumed</span></tt>, that explains what consumed annotations are
+at a high level.</p>
+<p>Documentation content (whether it is for an attribute or a category) is written
+using reStructuredText (RST) syntax.</p>
+<p>After writing the documentation for the attribute, it should be locally tested
+to ensure that there are no issues generating the documentation on the server.
+Local testing requires a fresh build of clang-tblgen. To generate the attribute
+documentation, execute the following command:</p>
+<div class="highlight-python"><div class="highlight"><pre><span class="n">clang</span><span class="o">-</span><span class="n">tblgen</span> <span class="o">-</span><span class="n">gen</span><span class="o">-</span><span class="n">attr</span><span class="o">-</span><span class="n">docs</span> <span class="o">-</span><span class="n">I</span> <span class="o">/</span><span class="n">path</span><span class="o">/</span><span class="n">to</span><span class="o">/</span><span class="n">clang</span><span class="o">/</span><span class="n">include</span> <span class="o">/</span><span class="n">path</span><span class="o">/</span><span class="n">to</span><span class="o">/</span><span class="n">clang</span><span class="o">/</span><span class="n">include</span><span class="o">/</span><span class="n">clang</span><span class="o">/</span><span class="n">Basic</span><span class="o">/</span><span class="n">Attr</span><span class="o">.</span><span class="n">td</span> <span class="o">-</span><span class="
n">o</span> <span class="o">/</span><span class="n">path</span><span class="o">/</span><span class="n">to</span><span class="o">/</span><span class="n">clang</span><span class="o">/</span><span class="n">docs</span><span class="o">/</span><span class="n">AttributeReference</span><span class="o">.</span><span class="n">rst</span>
+</pre></div>
+</div>
+<p>When testing locally, <em>do not</em> commit changes to <tt class="docutils literal"><span class="pre">AttributeReference.rst</span></tt>.
+This file is generated by the server automatically, and any changes made to this
+file will be overwritten.</p>
+</div>
+<div class="section" id="arguments">
+<h5><a class="toc-backref" href="#id52">Arguments</a><a class="headerlink" href="#arguments" title="Permalink to this headline">¶</a></h5>
+<p>Attributes may optionally specify a list of arguments that can be passed to the
+attribute. Attribute arguments specify both the parsed form and the semantic
+form of the attribute. For example, if <tt class="docutils literal"><span class="pre">Args</span></tt> is
+<tt class="docutils literal"><span class="pre">[StringArgument<"Arg1">,</span> <span class="pre">IntArgument<"Arg2">]</span></tt> then
+<tt class="docutils literal"><span class="pre">__attribute__((myattribute("Hello",</span> <span class="pre">3)))</span></tt> will be a valid use; it requires
+two arguments while parsing, and the Attr subclass’ constructor for the
+semantic attribute will require a string and integer argument.</p>
+<p>All arguments have a name and a flag that specifies whether the argument is
+optional. The associated C++ type of the argument is determined by the argument
+definition type. If the existing argument types are insufficient, new types can
+be created, but it requires modifying <a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/utils/TableGen/ClangAttrEmitter.cpp?view=markup">utils/TableGen/ClangAttrEmitter.cpp</a>
+to properly support the type.</p>
+</div>
+<div class="section" id="other-properties">
+<h5><a class="toc-backref" href="#id53">Other Properties</a><a class="headerlink" href="#other-properties" title="Permalink to this headline">¶</a></h5>
+<p>The <tt class="docutils literal"><span class="pre">Attr</span></tt> definition has other members which control the behavior of the
+attribute. Many of them are special-purpose and beyond the scope of this
+document, however a few deserve mention.</p>
+<p>If the parsed form of the attribute is more complex, or differs from the
+semantic form, the <tt class="docutils literal"><span class="pre">HasCustomParsing</span></tt> bit can be set to <tt class="docutils literal"><span class="pre">1</span></tt> for the class,
+and the parsing code in <a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Parse/ParseDecl.cpp?view=markup">Parser::ParseGNUAttributeArgs()</a>
+can be updated for the special case. Note that this only applies to arguments
+with a GNU spelling – attributes with a __declspec spelling currently ignore
+this flag and are handled by <tt class="docutils literal"><span class="pre">Parser::ParseMicrosoftDeclSpec</span></tt>.</p>
+<p>Note that setting this member to 1 will opt out of common attribute semantic
+handling, requiring extra implementation efforts to ensure the attribute
+appertains to the appropriate subject, etc.</p>
+<p>If the attribute should not be propagated from from a template declaration to an
+instantiation of the template, set the <tt class="docutils literal"><span class="pre">Clone</span></tt> member to 0. By default, all
+attributes will be cloned to template instantiations.</p>
+<p>Attributes that do not require an AST node should set the <tt class="docutils literal"><span class="pre">ASTNode</span></tt> field to
+<tt class="docutils literal"><span class="pre">0</span></tt> to avoid polluting the AST. Note that anything inheriting from
+<tt class="docutils literal"><span class="pre">TypeAttr</span></tt> or <tt class="docutils literal"><span class="pre">IgnoredAttr</span></tt> automatically do not generate an AST node. All
+other attributes generate an AST node by default. The AST node is the semantic
+representation of the attribute.</p>
+<p>The <tt class="docutils literal"><span class="pre">LangOpts</span></tt> field specifies a list of language options required by the
+attribute. For instance, all of the CUDA-specific attributes specify <tt class="docutils literal"><span class="pre">[CUDA]</span></tt>
+for the <tt class="docutils literal"><span class="pre">LangOpts</span></tt> field, and when the CUDA language option is not enabled, an
+“attribute ignored” warning diagnostic is emitted. Since language options are
+not table generated nodes, new language options must be created manually and
+should specify the spelling used by <tt class="docutils literal"><span class="pre">LangOptions</span></tt> class.</p>
+<p>Custom accessors can be generated for an attribute based on the spelling list
+for that attribute. For instance, if an attribute has two different spellings:
+‘Foo’ and ‘Bar’, accessors can be created:
+<tt class="docutils literal"><span class="pre">[Accessor<"isFoo",</span> <span class="pre">[GNU<"Foo">]>,</span> <span class="pre">Accessor<"isBar",</span> <span class="pre">[GNU<"Bar">]>]</span></tt>
+These accessors will be generated on the semantic form of the attribute,
+accepting no arguments and returning a <tt class="docutils literal"><span class="pre">bool</span></tt>.</p>
+<p>Attributes that do not require custom semantic handling should set the
+<tt class="docutils literal"><span class="pre">SemaHandler</span></tt> field to <tt class="docutils literal"><span class="pre">0</span></tt>. Note that anything inheriting from
+<tt class="docutils literal"><span class="pre">IgnoredAttr</span></tt> automatically do not get a semantic handler. All other
+attributes are assumed to use a semantic handler by default. Attributes
+without a semantic handler are not given a parsed attribute <tt class="docutils literal"><span class="pre">Kind</span></tt> enumerator.</p>
+<p>Target-specific attributes may share a spelling with other attributes in
+different targets. For instance, the ARM and MSP430 targets both have an
+attribute spelled <tt class="docutils literal"><span class="pre">GNU<"interrupt"></span></tt>, but with different parsing and semantic
+requirements. To support this feature, an attribute inheriting from
+<tt class="docutils literal"><span class="pre">TargetSpecificAttribute</span></tt> may specify a <tt class="docutils literal"><span class="pre">ParseKind</span></tt> field. This field
+should be the same value between all arguments sharing a spelling, and
+corresponds to the parsed attribute’s <tt class="docutils literal"><span class="pre">Kind</span></tt> enumerator. This allows
+attributes to share a parsed attribute kind, but have distinct semantic
+attribute classes. For instance, <tt class="docutils literal"><span class="pre">AttributeList::AT_Interrupt</span></tt> is the shared
+parsed attribute kind, but ARMInterruptAttr and MSP430InterruptAttr are the
+semantic attributes generated.</p>
+<p>By default, when declarations are merging attributes, an attribute will not be
+duplicated. However, if an attribute can be duplicated during this merging
+stage, set <tt class="docutils literal"><span class="pre">DuplicatesAllowedWhileMerging</span></tt> to <tt class="docutils literal"><span class="pre">1</span></tt>, and the attribute will
+be merged.</p>
+<p>By default, attribute arguments are parsed in an evaluated context. If the
+arguments for an attribute should be parsed in an unevaluated context (akin to
+the way the argument to a <tt class="docutils literal"><span class="pre">sizeof</span></tt> expression is parsed), set
+<tt class="docutils literal"><span class="pre">ParseArgumentsAsUnevaluated</span></tt> to <tt class="docutils literal"><span class="pre">1</span></tt>.</p>
+<p>If additional functionality is desired for the semantic form of the attribute,
+the <tt class="docutils literal"><span class="pre">AdditionalMembers</span></tt> field specifies code to be copied verbatim into the
+semantic attribute class object, with <tt class="docutils literal"><span class="pre">public</span></tt> access.</p>
+</div>
+</div>
+<div class="section" id="boilerplate">
+<h4><a class="toc-backref" href="#id54">Boilerplate</a><a class="headerlink" href="#boilerplate" title="Permalink to this headline">¶</a></h4>
+<p>All semantic processing of declaration attributes happens in <a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Sema/SemaDeclAttr.cpp?view=markup">lib/Sema/SemaDeclAttr.cpp</a>,
+and generally starts in the <tt class="docutils literal"><span class="pre">ProcessDeclAttribute()</span></tt> function. If the
+attribute is a “simple” attribute – meaning that it requires no custom semantic
+processing aside from what is automatically provided, add a call to
+<tt class="docutils literal"><span class="pre">handleSimpleAttribute<YourAttr>(S,</span> <span class="pre">D,</span> <span class="pre">Attr);</span></tt> to the switch statement.
+Otherwise, write a new <tt class="docutils literal"><span class="pre">handleYourAttr()</span></tt> function, and add that to the switch
+statement. Please do not implement handling logic directly in the <tt class="docutils literal"><span class="pre">case</span></tt> for
+the attribute.</p>
+<p>Unless otherwise specified by the attribute definition, common semantic checking
+of the parsed attribute is handled automatically. This includes diagnosing
+parsed attributes that do not appertain to the given <tt class="docutils literal"><span class="pre">Decl</span></tt>, ensuring the
+correct minimum number of arguments are passed, etc.</p>
+<p>If the attribute adds additional warnings, define a <tt class="docutils literal"><span class="pre">DiagGroup</span></tt> in
+<a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/DiagnosticGroups.td?view=markup">include/clang/Basic/DiagnosticGroups.td</a>
+named after the attribute’s <tt class="docutils literal"><span class="pre">Spelling</span></tt> with “_”s replaced by “-“s. If there
+is only a single diagnostic, it is permissible to use <tt class="docutils literal"><span class="pre">InGroup<DiagGroup<"your-attribute">></span></tt>
+directly in <a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td?view=markup">DiagnosticSemaKinds.td</a></p>
+<p>All semantic diagnostics generated for your attribute, including automatically-
+generated ones (such as subjects and argument counts), should have a
+corresponding test case.</p>
+</div>
+<div class="section" id="semantic-handling">
+<h4><a class="toc-backref" href="#id55">Semantic handling</a><a class="headerlink" href="#semantic-handling" title="Permalink to this headline">¶</a></h4>
+<p>Most attributes are implemented to have some effect on the compiler. For
+instance, to modify the way code is generated, or to add extra semantic checks
+for an analysis pass, etc. Having added the attribute definition and conversion
+to the semantic representation for the attribute, what remains is to implement
+the custom logic requiring use of the attribute.</p>
+<p>The <tt class="docutils literal"><span class="pre">clang::Decl</span></tt> object can be queried for the presence or absence of an
+attribute using <tt class="docutils literal"><span class="pre">hasAttr<T>()</span></tt>. To obtain a pointer to the semantic
+representation of the attribute, <tt class="docutils literal"><span class="pre">getAttr<T></span></tt> may be used.</p>
+</div>
+</div>
+<div class="section" id="how-to-add-an-expression-or-statement">
+<h3><a class="toc-backref" href="#id56">How to add an expression or statement</a><a class="headerlink" href="#how-to-add-an-expression-or-statement" title="Permalink to this headline">¶</a></h3>
+<p>Expressions and statements are one of the most fundamental constructs within a
+compiler, because they interact with many different parts of the AST, semantic
+analysis, and IR generation. Therefore, adding a new expression or statement
+kind into Clang requires some care. The following list details the various
+places in Clang where an expression or statement needs to be introduced, along
+with patterns to follow to ensure that the new expression or statement works
+well across all of the C languages. We focus on expressions, but statements
+are similar.</p>
+<ol class="arabic simple">
+<li>Introduce parsing actions into the parser. Recursive-descent parsing is
+mostly self-explanatory, but there are a few things that are worth keeping
+in mind:<ul>
+<li>Keep as much source location information as possible! You’ll want it later
+to produce great diagnostics and support Clang’s various features that map
+between source code and the AST.</li>
+<li>Write tests for all of the “bad” parsing cases, to make sure your recovery
+is good. If you have matched delimiters (e.g., parentheses, square
+brackets, etc.), use <tt class="docutils literal"><span class="pre">Parser::BalancedDelimiterTracker</span></tt> to give nice
+diagnostics when things go wrong.</li>
+</ul>
+</li>
+<li>Introduce semantic analysis actions into <tt class="docutils literal"><span class="pre">Sema</span></tt>. Semantic analysis should
+always involve two functions: an <tt class="docutils literal"><span class="pre">ActOnXXX</span></tt> function that will be called
+directly from the parser, and a <tt class="docutils literal"><span class="pre">BuildXXX</span></tt> function that performs the
+actual semantic analysis and will (eventually!) build the AST node. It’s
+fairly common for the <tt class="docutils literal"><span class="pre">ActOnCXX</span></tt> function to do very little (often just
+some minor translation from the parser’s representation to <tt class="docutils literal"><span class="pre">Sema</span></tt>‘s
+representation of the same thing), but the separation is still important:
+C++ template instantiation, for example, should always call the <tt class="docutils literal"><span class="pre">BuildXXX</span></tt>
+variant. Several notes on semantic analysis before we get into construction
+of the AST:<ul>
+<li>Your expression probably involves some types and some subexpressions.
+Make sure to fully check that those types, and the types of those
+subexpressions, meet your expectations. Add implicit conversions where
+necessary to make sure that all of the types line up exactly the way you
+want them. Write extensive tests to check that you’re getting good
+diagnostics for mistakes and that you can use various forms of
+subexpressions with your expression.</li>
+<li>When type-checking a type or subexpression, make sure to first check
+whether the type is “dependent” (<tt class="docutils literal"><span class="pre">Type::isDependentType()</span></tt>) or whether a
+subexpression is type-dependent (<tt class="docutils literal"><span class="pre">Expr::isTypeDependent()</span></tt>). If any of
+these return <tt class="docutils literal"><span class="pre">true</span></tt>, then you’re inside a template and you can’t do much
+type-checking now. That’s normal, and your AST node (when you get there)
+will have to deal with this case. At this point, you can write tests that
+use your expression within templates, but don’t try to instantiate the
+templates.</li>
+<li>For each subexpression, be sure to call <tt class="docutils literal"><span class="pre">Sema::CheckPlaceholderExpr()</span></tt>
+to deal with “weird” expressions that don’t behave well as subexpressions.
+Then, determine whether you need to perform lvalue-to-rvalue conversions
+(<tt class="docutils literal"><span class="pre">Sema::DefaultLvalueConversions</span></tt>) or the usual unary conversions
+(<tt class="docutils literal"><span class="pre">Sema::UsualUnaryConversions</span></tt>), for places where the subexpression is
+producing a value you intend to use.</li>
+<li>Your <tt class="docutils literal"><span class="pre">BuildXXX</span></tt> function will probably just return <tt class="docutils literal"><span class="pre">ExprError()</span></tt> at
+this point, since you don’t have an AST. That’s perfectly fine, and
+shouldn’t impact your testing.</li>
+</ul>
+</li>
+<li>Introduce an AST node for your new expression. This starts with declaring
+the node in <tt class="docutils literal"><span class="pre">include/Basic/StmtNodes.td</span></tt> and creating a new class for your
+expression in the appropriate <tt class="docutils literal"><span class="pre">include/AST/Expr*.h</span></tt> header. It’s best to
+look at the class for a similar expression to get ideas, and there are some
+specific things to watch for:<ul>
+<li>If you need to allocate memory, use the <tt class="docutils literal"><span class="pre">ASTContext</span></tt> allocator to
+allocate memory. Never use raw <tt class="docutils literal"><span class="pre">malloc</span></tt> or <tt class="docutils literal"><span class="pre">new</span></tt>, and never hold any
+resources in an AST node, because the destructor of an AST node is never
+called.</li>
+<li>Make sure that <tt class="docutils literal"><span class="pre">getSourceRange()</span></tt> covers the exact source range of your
+expression. This is needed for diagnostics and for IDE support.</li>
+<li>Make sure that <tt class="docutils literal"><span class="pre">children()</span></tt> visits all of the subexpressions. This is
+important for a number of features (e.g., IDE support, C++ variadic
+templates). If you have sub-types, you’ll also need to visit those
+sub-types in <tt class="docutils literal"><span class="pre">RecursiveASTVisitor</span></tt>.</li>
+<li>Add printing support (<tt class="docutils literal"><span class="pre">StmtPrinter.cpp</span></tt>) for your expression.</li>
+<li>Add profiling support (<tt class="docutils literal"><span class="pre">StmtProfile.cpp</span></tt>) for your AST node, noting the
+distinguishing (non-source location) characteristics of an instance of
+your expression. Omitting this step will lead to hard-to-diagnose
+failures regarding matching of template declarations.</li>
+<li>Add serialization support (<tt class="docutils literal"><span class="pre">ASTReaderStmt.cpp</span></tt>, <tt class="docutils literal"><span class="pre">ASTWriterStmt.cpp</span></tt>)
+for your AST node.</li>
+</ul>
+</li>
+<li>Teach semantic analysis to build your AST node. At this point, you can wire
+up your <tt class="docutils literal"><span class="pre">Sema::BuildXXX</span></tt> function to actually create your AST. A few
+things to check at this point:<ul>
+<li>If your expression can construct a new C++ class or return a new
+Objective-C object, be sure to update and then call
+<tt class="docutils literal"><span class="pre">Sema::MaybeBindToTemporary</span></tt> for your just-created AST node to be sure
+that the object gets properly destructed. An easy way to test this is to
+return a C++ class with a private destructor: semantic analysis should
+flag an error here with the attempt to call the destructor.</li>
+<li>Inspect the generated AST by printing it using <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span> <span class="pre">-ast-print</span></tt>,
+to make sure you’re capturing all of the important information about how
+the AST was written.</li>
+<li>Inspect the generated AST under <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span> <span class="pre">-ast-dump</span></tt> to verify that
+all of the types in the generated AST line up the way you want them.
+Remember that clients of the AST should never have to “think” to
+understand what’s going on. For example, all implicit conversions should
+show up explicitly in the AST.</li>
+<li>Write tests that use your expression as a subexpression of other,
+well-known expressions. Can you call a function using your expression as
+an argument? Can you use the ternary operator?</li>
+</ul>
+</li>
+<li>Teach code generation to create IR to your AST node. This step is the first
+(and only) that requires knowledge of LLVM IR. There are several things to
+keep in mind:<ul>
+<li>Code generation is separated into scalar/aggregate/complex and
+lvalue/rvalue paths, depending on what kind of result your expression
+produces. On occasion, this requires some careful factoring of code to
+avoid duplication.</li>
+<li><tt class="docutils literal"><span class="pre">CodeGenFunction</span></tt> contains functions <tt class="docutils literal"><span class="pre">ConvertType</span></tt> and
+<tt class="docutils literal"><span class="pre">ConvertTypeForMem</span></tt> that convert Clang’s types (<tt class="docutils literal"><span class="pre">clang::Type*</span></tt> or
+<tt class="docutils literal"><span class="pre">clang::QualType</span></tt>) to LLVM types. Use the former for values, and the
+later for memory locations: test with the C++ “<tt class="docutils literal"><span class="pre">bool</span></tt>” type to check
+this. If you find that you are having to use LLVM bitcasts to make the
+subexpressions of your expression have the type that your expression
+expects, STOP! Go fix semantic analysis and the AST so that you don’t
+need these bitcasts.</li>
+<li>The <tt class="docutils literal"><span class="pre">CodeGenFunction</span></tt> class has a number of helper functions to make
+certain operations easy, such as generating code to produce an lvalue or
+an rvalue, or to initialize a memory location with a given value. Prefer
+to use these functions rather than directly writing loads and stores,
+because these functions take care of some of the tricky details for you
+(e.g., for exceptions).</li>
+<li>If your expression requires some special behavior in the event of an
+exception, look at the <tt class="docutils literal"><span class="pre">push*Cleanup</span></tt> functions in <tt class="docutils literal"><span class="pre">CodeGenFunction</span></tt>
+to introduce a cleanup. You shouldn’t have to deal with
+exception-handling directly.</li>
+<li>Testing is extremely important in IR generation. Use <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span>
+<span class="pre">-emit-llvm</span></tt> and <a class="reference external" href="http://llvm.org/docs/CommandGuide/FileCheck.html">FileCheck</a> to verify that you’re
+generating the right IR.</li>
+</ul>
+</li>
+<li>Teach template instantiation how to cope with your AST node, which requires
+some fairly simple code:<ul>
+<li>Make sure that your expression’s constructor properly computes the flags
+for type dependence (i.e., the type your expression produces can change
+from one instantiation to the next), value dependence (i.e., the constant
+value your expression produces can change from one instantiation to the
+next), instantiation dependence (i.e., a template parameter occurs
+anywhere in your expression), and whether your expression contains a
+parameter pack (for variadic templates). Often, computing these flags
+just means combining the results from the various types and
+subexpressions.</li>
+<li>Add <tt class="docutils literal"><span class="pre">TransformXXX</span></tt> and <tt class="docutils literal"><span class="pre">RebuildXXX</span></tt> functions to the <tt class="docutils literal"><span class="pre">TreeTransform</span></tt>
+class template in <tt class="docutils literal"><span class="pre">Sema</span></tt>. <tt class="docutils literal"><span class="pre">TransformXXX</span></tt> should (recursively)
+transform all of the subexpressions and types within your expression,
+using <tt class="docutils literal"><span class="pre">getDerived().TransformYYY</span></tt>. If all of the subexpressions and
+types transform without error, it will then call the <tt class="docutils literal"><span class="pre">RebuildXXX</span></tt>
+function, which will in turn call <tt class="docutils literal"><span class="pre">getSema().BuildXXX</span></tt> to perform
+semantic analysis and build your expression.</li>
+<li>To test template instantiation, take those tests you wrote to make sure
+that you were type checking with type-dependent expressions and dependent
+types (from step #2) and instantiate those templates with various types,
+some of which type-check and some that don’t, and test the error messages
+in each case.</li>
+</ul>
+</li>
+<li>There are some “extras” that make other features work better. It’s worth
+handling these extras to give your expression complete integration into
+Clang:<ul>
+<li>Add code completion support for your expression in
+<tt class="docutils literal"><span class="pre">SemaCodeComplete.cpp</span></tt>.</li>
+<li>If your expression has types in it, or has any “interesting” features
+other than subexpressions, extend libclang’s <tt class="docutils literal"><span class="pre">CursorVisitor</span></tt> to provide
+proper visitation for your expression, enabling various IDE features such
+as syntax highlighting, cross-referencing, and so on. The
+<tt class="docutils literal"><span class="pre">c-index-test</span></tt> helper program can be used to test these features.</li>
+</ul>
+</li>
+</ol>
+</div>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="ClangFormatStyleOptions.html">Clang-Format Style Options</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="DriverInternals.html">Driver Design & Internals</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/IntroductionToTheClangAST.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/IntroductionToTheClangAST.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/IntroductionToTheClangAST.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/IntroductionToTheClangAST.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,179 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Introduction to the Clang AST — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="LibTooling" href="LibTooling.html" />
+ <link rel="prev" title="External Clang Examples" href="ExternalClangExamples.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Introduction to the Clang AST</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="ExternalClangExamples.html">External Clang Examples</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="LibTooling.html">LibTooling</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="introduction-to-the-clang-ast">
+<h1>Introduction to the Clang AST<a class="headerlink" href="#introduction-to-the-clang-ast" title="Permalink to this headline">¶</a></h1>
+<p>This document gives a gentle introduction to the mysteries of the Clang
+AST. It is targeted at developers who either want to contribute to
+Clang, or use tools that work based on Clang’s AST, like the AST
+matchers.</p>
+<center><iframe width="560" height="315" src="http://www.youtube.com/embed/VqCkCDFLSsc?vq=hd720" frameborder="0" allowfullscreen></iframe></center><p><a class="reference external" href="http://llvm.org/devmtg/2013-04/klimek-slides.pdf">Slides</a></p>
+<div class="section" id="introduction">
+<h2>Introduction<a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>Clang’s AST is different from ASTs produced by some other compilers in
+that it closely resembles both the written C++ code and the C++
+standard. For example, parenthesis expressions and compile time
+constants are available in an unreduced form in the AST. This makes
+Clang’s AST a good fit for refactoring tools.</p>
+<p>Documentation for all Clang AST nodes is available via the generated
+<a class="reference external" href="http://clang.llvm.org/doxygen">Doxygen</a>. The doxygen online
+documentation is also indexed by your favorite search engine, which will
+make a search for clang and the AST node’s class name usually turn up
+the doxygen of the class you’re looking for (for example, search for:
+clang ParenExpr).</p>
+</div>
+<div class="section" id="examining-the-ast">
+<h2>Examining the AST<a class="headerlink" href="#examining-the-ast" title="Permalink to this headline">¶</a></h2>
+<p>A good way to familarize yourself with the Clang AST is to actually look
+at it on some simple example code. Clang has a builtin AST-dump mode,
+which can be enabled with the flag <tt class="docutils literal"><span class="pre">-ast-dump</span></tt>.</p>
+<p>Let’s look at a simple example AST:</p>
+<div class="highlight-python"><div class="highlight"><pre>$ cat test.cc
+int f(int x) {
+ int result = (x / 42);
+ return result;
+}
+
+# Clang by default is a frontend for many tools; -Xclang is used to pass
+# options directly to the C++ frontend.
+$ clang -Xclang -ast-dump -fsyntax-only test.cc
+TranslationUnitDecl 0x5aea0d0 <<invalid sloc>>
+... cutting out internal declarations of clang ...
+`-FunctionDecl 0x5aeab50 <test.cc:1:1, line:4:1> f 'int (int)'
+ |-ParmVarDecl 0x5aeaa90 <line:1:7, col:11> x 'int'
+ `-CompoundStmt 0x5aead88 <col:14, line:4:1>
+ |-DeclStmt 0x5aead10 <line:2:3, col:24>
+ | `-VarDecl 0x5aeac10 <col:3, col:23> result 'int'
+ | `-ParenExpr 0x5aeacf0 <col:16, col:23> 'int'
+ | `-BinaryOperator 0x5aeacc8 <col:17, col:21> 'int' '/'
+ | |-ImplicitCastExpr 0x5aeacb0 <col:17> 'int' <LValueToRValue>
+ | | `-DeclRefExpr 0x5aeac68 <col:17> 'int' lvalue ParmVar 0x5aeaa90 'x' 'int'
+ | `-IntegerLiteral 0x5aeac90 <col:21> 'int' 42
+ `-ReturnStmt 0x5aead68 <line:3:3, col:10>
+ `-ImplicitCastExpr 0x5aead50 <col:10> 'int' <LValueToRValue>
+ `-DeclRefExpr 0x5aead28 <col:10> 'int' lvalue Var 0x5aeac10 'result' 'int'
+</pre></div>
+</div>
+<p>The toplevel declaration in
+a translation unit is always the <a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1TranslationUnitDecl.html">translation unit
+declaration</a>.
+In this example, our first user written declaration is the <a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1FunctionDecl.html">function
+declaration</a>
+of “<tt class="docutils literal"><span class="pre">f</span></tt>”. The body of “<tt class="docutils literal"><span class="pre">f</span></tt>” is a <a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1CompoundStmt.html">compound
+statement</a>,
+whose child nodes are a <a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1DeclStmt.html">declaration
+statement</a>
+that declares our result variable, and the <a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1ReturnStmt.html">return
+statement</a>.</p>
+</div>
+<div class="section" id="ast-context">
+<h2>AST Context<a class="headerlink" href="#ast-context" title="Permalink to this headline">¶</a></h2>
+<p>All information about the AST for a translation unit is bundled up in
+the class
+<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1ASTContext.html">ASTContext</a>.
+It allows traversal of the whole translation unit starting from
+<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1ASTContext.html#abd909fb01ef10cfd0244832a67b1dd64">getTranslationUnitDecl</a>,
+or to access Clang’s <a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1ASTContext.html#a4f95adb9958e22fbe55212ae6482feb4">table of
+identifiers</a>
+for the parsed translation unit.</p>
+</div>
+<div class="section" id="ast-nodes">
+<h2>AST Nodes<a class="headerlink" href="#ast-nodes" title="Permalink to this headline">¶</a></h2>
+<p>Clang’s AST nodes are modeled on a class hierarchy that does not have a
+common ancestor. Instead, there are multiple larger hierarchies for
+basic node types like
+<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1Decl.html">Decl</a> and
+<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1Stmt.html">Stmt</a>. Many
+important AST nodes derive from
+<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1Type.html">Type</a>,
+<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1Decl.html">Decl</a>,
+<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1DeclContext.html">DeclContext</a>
+or <a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1Stmt.html">Stmt</a>, with
+some classes deriving from both Decl and DeclContext.</p>
+<p>There are also a multitude of nodes in the AST that are not part of a
+larger hierarchy, and are only reachable from specific other nodes, like
+<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1CXXBaseSpecifier.html">CXXBaseSpecifier</a>.</p>
+<p>Thus, to traverse the full AST, one starts from the
+<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1TranslationUnitDecl.html">TranslationUnitDecl</a>
+and then recursively traverses everything that can be reached from that
+node - this information has to be encoded for each specific node type.
+This algorithm is encoded in the
+<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1RecursiveASTVisitor.html">RecursiveASTVisitor</a>.
+See the <a class="reference external" href="http://clang.llvm.org/docs/RAVFrontendAction.html">RecursiveASTVisitor
+tutorial</a>.</p>
+<p>The two most basic nodes in the Clang AST are statements
+(<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1Stmt.html">Stmt</a>) and
+declarations
+(<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1Decl.html">Decl</a>). Note
+that expressions
+(<a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>) are
+also statements in Clang’s AST.</p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="ExternalClangExamples.html">External Clang Examples</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="LibTooling.html">LibTooling</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/JSONCompilationDatabase.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/JSONCompilationDatabase.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/JSONCompilationDatabase.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/JSONCompilationDatabase.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,149 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>JSON Compilation Database Format Specification — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Overview" href="ClangTools.html" />
+ <link rel="prev" title="How To Setup Clang Tooling For LLVM" href="HowToSetupToolingForLLVM.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>JSON Compilation Database Format Specification</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="HowToSetupToolingForLLVM.html">How To Setup Clang Tooling For LLVM</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ClangTools.html">Overview</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="json-compilation-database-format-specification">
+<h1>JSON Compilation Database Format Specification<a class="headerlink" href="#json-compilation-database-format-specification" title="Permalink to this headline">¶</a></h1>
+<p>This document describes a format for specifying how to replay single
+compilations independently of the build system.</p>
+<div class="section" id="background">
+<h2>Background<a class="headerlink" href="#background" title="Permalink to this headline">¶</a></h2>
+<p>Tools based on the C++ Abstract Syntax Tree need full information how to
+parse a translation unit. Usually this information is implicitly
+available in the build system, but running tools as part of the build
+system is not necessarily the best solution:</p>
+<ul class="simple">
+<li>Build systems are inherently change driven, so running multiple tools
+over the same code base without changing the code does not fit into
+the architecture of many build systems.</li>
+<li>Figuring out whether things have changed is often an IO bound
+process; this makes it hard to build low latency end user tools based
+on the build system.</li>
+<li>Build systems are inherently sequential in the build graph, for
+example due to generated source code. While tools that run
+independently of the build still need the generated source code to
+exist, running tools multiple times over unchanging source does not
+require serialization of the runs according to the build dependency
+graph.</li>
+</ul>
+</div>
+<div class="section" id="supported-systems">
+<h2>Supported Systems<a class="headerlink" href="#supported-systems" title="Permalink to this headline">¶</a></h2>
+<p>Currently <a class="reference external" href="http://cmake.org">CMake</a> (since 2.8.5) supports generation
+of compilation databases for Unix Makefile builds (Ninja builds in the
+works) with the option <tt class="docutils literal"><span class="pre">CMAKE_EXPORT_COMPILE_COMMANDS</span></tt>.</p>
+<p>For projects on Linux, there is an alternative to intercept compiler
+calls with a tool called <a class="reference external" href="https://github.com/rizsotto/Bear">Bear</a>.</p>
+<p>Clang’s tooling interface supports reading compilation databases; see
+the <a class="reference internal" href="LibTooling.html"><em>LibTooling documentation</em></a>. libclang and its
+python bindings also support this (since clang 3.2); see
+<a class="reference external" href="/doxygen/group__COMPILATIONDB.html">CXCompilationDatabase.h</a>.</p>
+</div>
+<div class="section" id="format">
+<h2>Format<a class="headerlink" href="#format" title="Permalink to this headline">¶</a></h2>
+<p>A compilation database is a JSON file, which consist of an array of
+“command objects”, where each command object specifies one way a
+translation unit is compiled in the project.</p>
+<p>Each command object contains the translation unit’s main file, the
+working directory of the compile run and the actual compile command.</p>
+<p>Example:</p>
+<div class="highlight-python"><div class="highlight"><pre><span class="p">[</span>
+ <span class="p">{</span> <span class="s">"directory"</span><span class="p">:</span> <span class="s">"/home/user/llvm/build"</span><span class="p">,</span>
+ <span class="s">"command"</span><span class="p">:</span> <span class="s">"/usr/bin/clang++ -Irelative -DSOMEDEF=</span><span class="se">\"</span><span class="s">With spaces, quotes and </span><span class="se">\\</span><span class="s">-es.</span><span class="se">\"</span><span class="s"> -c -o file.o file.cc"</span><span class="p">,</span>
+ <span class="s">"file"</span><span class="p">:</span> <span class="s">"file.cc"</span> <span class="p">},</span>
+ <span class="o">...</span>
+<span class="p">]</span>
+</pre></div>
+</div>
+<p>The contracts for each field in the command object are:</p>
+<ul class="simple">
+<li><strong>directory:</strong> The working directory of the compilation. All paths
+specified in the <strong>command</strong> or <strong>file</strong> fields must be either
+absolute or relative to this directory.</li>
+<li><strong>file:</strong> The main translation unit source processed by this
+compilation step. This is used by tools as the key into the
+compilation database. There can be multiple command objects for the
+same file, for example if the same source file is compiled with
+different configurations.</li>
+<li><strong>command:</strong> The compile command executed. After JSON unescaping,
+this must be a valid command to rerun the exact compilation step for
+the translation unit in the environment the build system uses.
+Parameters use shell quoting and shell escaping of quotes, with ‘<tt class="docutils literal"><span class="pre">"</span></tt>‘
+and ‘<tt class="docutils literal"><span class="pre">\</span></tt>‘ being the only special characters. Shell expansion is not
+supported.</li>
+</ul>
+</div>
+<div class="section" id="build-system-integration">
+<h2>Build System Integration<a class="headerlink" href="#build-system-integration" title="Permalink to this headline">¶</a></h2>
+<p>The convention is to name the file compile_commands.json and put it at
+the top of the build directory. Clang tools are pointed to the top of
+the build directory to detect the file and use the compilation database
+to parse C++ code in the source tree.</p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="HowToSetupToolingForLLVM.html">How To Setup Clang Tooling For LLVM</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ClangTools.html">Overview</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/LanguageExtensions.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/LanguageExtensions.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/LanguageExtensions.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/LanguageExtensions.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,1934 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Clang Language Extensions — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Objective-C Literals" href="ObjectiveCLiterals.html" />
+ <link rel="prev" title="Clang Compiler Userâs Manual" href="UsersManual.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Clang Language Extensions</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="UsersManual.html">Clang Compiler User’s Manual</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ObjectiveCLiterals.html">Objective-C Literals</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="clang-language-extensions">
+<h1>Clang Language Extensions<a class="headerlink" href="#clang-language-extensions" title="Permalink to this headline">¶</a></h1>
+<div class="contents local topic" id="contents">
+<ul class="simple">
+<li><a class="reference internal" href="#introduction" id="id1">Introduction</a></li>
+<li><a class="reference internal" href="#feature-checking-macros" id="id2">Feature Checking Macros</a></li>
+<li><a class="reference internal" href="#include-file-checking-macros" id="id3">Include File Checking Macros</a></li>
+<li><a class="reference internal" href="#builtin-macros" id="id4">Builtin Macros</a></li>
+<li><a class="reference internal" href="#vectors-and-extended-vectors" id="id5">Vectors and Extended Vectors</a></li>
+<li><a class="reference internal" href="#messages-on-deprecated-and-unavailable-attributes" id="id6">Messages on <tt class="docutils literal"><span class="pre">deprecated</span></tt> and <tt class="docutils literal"><span class="pre">unavailable</span></tt> Attributes</a></li>
+<li><a class="reference internal" href="#attributes-on-enumerators" id="id7">Attributes on Enumerators</a></li>
+<li><a class="reference internal" href="#user-specified-system-frameworks" id="id8">‘User-Specified’ System Frameworks</a></li>
+<li><a class="reference internal" href="#checks-for-standard-language-features" id="id9">Checks for Standard Language Features</a></li>
+<li><a class="reference internal" href="#checks-for-type-trait-primitives" id="id10">Checks for Type Trait Primitives</a></li>
+<li><a class="reference internal" href="#blocks" id="id11">Blocks</a></li>
+<li><a class="reference internal" href="#objective-c-features" id="id12">Objective-C Features</a></li>
+<li><a class="reference internal" href="#initializer-lists-for-complex-numbers-in-c" id="id13">Initializer lists for complex numbers in C</a></li>
+<li><a class="reference internal" href="#builtin-functions" id="id14">Builtin Functions</a></li>
+<li><a class="reference internal" href="#non-standard-c-11-attributes" id="id15">Non-standard C++11 Attributes</a></li>
+<li><a class="reference internal" href="#target-specific-extensions" id="id16">Target-Specific Extensions</a></li>
+<li><a class="reference internal" href="#extensions-for-static-analysis" id="id17">Extensions for Static Analysis</a></li>
+<li><a class="reference internal" href="#extensions-for-dynamic-analysis" id="id18">Extensions for Dynamic Analysis</a></li>
+<li><a class="reference internal" href="#extensions-for-selectively-disabling-optimization" id="id19">Extensions for selectively disabling optimization</a></li>
+<li><a class="reference internal" href="#extensions-for-loop-hint-optimizations" id="id20">Extensions for loop hint optimizations</a></li>
+</ul>
+</div>
+<div class="toctree-wrapper compound">
+</div>
+<div class="section" id="introduction">
+<h2><a class="toc-backref" href="#id1">Introduction</a><a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>This document describes the language extensions provided by Clang. In addition
+to the language extensions listed here, Clang aims to support a broad range of
+GCC extensions. Please see the <a class="reference external" href="http://gcc.gnu.org/onlinedocs/gcc/C-Extensions.html">GCC manual</a> for more information on
+these extensions.</p>
+</div>
+<div class="section" id="feature-checking-macros">
+<span id="langext-feature-check"></span><h2><a class="toc-backref" href="#id2">Feature Checking Macros</a><a class="headerlink" href="#feature-checking-macros" title="Permalink to this headline">¶</a></h2>
+<p>Language extensions can be very useful, but only if you know you can depend on
+them. In order to allow fine-grain features checks, we support three builtin
+function-like macros. This allows you to directly test for a feature in your
+code without having to resort to something like autoconf or fragile “compiler
+version checks”.</p>
+<div class="section" id="has-builtin">
+<h3><tt class="docutils literal"><span class="pre">__has_builtin</span></tt><a class="headerlink" href="#has-builtin" title="Permalink to this headline">¶</a></h3>
+<p>This function-like macro takes a single identifier argument that is the name of
+a builtin function. It evaluates to 1 if the builtin is supported or 0 if not.
+It can be used like this:</p>
+<div class="highlight-c++"><div class="highlight"><pre>#ifndef __has_builtin // Optional of course.
+ #define __has_builtin(x) 0 // Compatibility with non-clang compilers.
+#endif
+
+...
+#if __has_builtin(__builtin_trap)
+ __builtin_trap();
+#else
+ abort();
+#endif
+...
+</pre></div>
+</div>
+</div>
+<div class="section" id="has-feature-and-has-extension">
+<span id="langext-has-feature-has-extension"></span><h3><tt class="docutils literal"><span class="pre">__has_feature</span></tt> and <tt class="docutils literal"><span class="pre">__has_extension</span></tt><a class="headerlink" href="#has-feature-and-has-extension" title="Permalink to this headline">¶</a></h3>
+<p>These function-like macros take a single identifier argument that is the name
+of a feature. <tt class="docutils literal"><span class="pre">__has_feature</span></tt> evaluates to 1 if the feature is both
+supported by Clang and standardized in the current language standard or 0 if
+not (but see <a class="reference internal" href="#langext-has-feature-back-compat"><em>below</em></a>), while
+<tt class="docutils literal"><span class="pre">__has_extension</span></tt> evaluates to 1 if the feature is supported by Clang in the
+current language (either as a language extension or a standard language
+feature) or 0 if not. They can be used like this:</p>
+<div class="highlight-c++"><div class="highlight"><pre>#ifndef __has_feature // Optional of course.
+ #define __has_feature(x) 0 // Compatibility with non-clang compilers.
+#endif
+#ifndef __has_extension
+ #define __has_extension __has_feature // Compatibility with pre-3.0 compilers.
+#endif
+
+...
+#if __has_feature(cxx_rvalue_references)
+// This code will only be compiled with the -std=c++11 and -std=gnu++11
+// options, because rvalue references are only standardized in C++11.
+#endif
+
+#if __has_extension(cxx_rvalue_references)
+// This code will be compiled with the -std=c++11, -std=gnu++11, -std=c++98
+// and -std=gnu++98 options, because rvalue references are supported as a
+// language extension in C++98.
+#endif
+</pre></div>
+</div>
+<p id="langext-has-feature-back-compat">For backward compatibility, <tt class="docutils literal"><span class="pre">__has_feature</span></tt> can also be used to test
+for support for non-standardized features, i.e. features not prefixed <tt class="docutils literal"><span class="pre">c_</span></tt>,
+<tt class="docutils literal"><span class="pre">cxx_</span></tt> or <tt class="docutils literal"><span class="pre">objc_</span></tt>.</p>
+<p>Another use of <tt class="docutils literal"><span class="pre">__has_feature</span></tt> is to check for compiler features not related
+to the language standard, such as e.g. <a class="reference internal" href="AddressSanitizer.html"><em>AddressSanitizer</em></a>.</p>
+<p>If the <tt class="docutils literal"><span class="pre">-pedantic-errors</span></tt> option is given, <tt class="docutils literal"><span class="pre">__has_extension</span></tt> is equivalent
+to <tt class="docutils literal"><span class="pre">__has_feature</span></tt>.</p>
+<p>The feature tag is described along with the language feature below.</p>
+<p>The feature name or extension name can also be specified with a preceding and
+following <tt class="docutils literal"><span class="pre">__</span></tt> (double underscore) to avoid interference from a macro with
+the same name. For instance, <tt class="docutils literal"><span class="pre">__cxx_rvalue_references__</span></tt> can be used instead
+of <tt class="docutils literal"><span class="pre">cxx_rvalue_references</span></tt>.</p>
+</div>
+<div class="section" id="has-cpp-attribute">
+<h3><tt class="docutils literal"><span class="pre">__has_cpp_attribute</span></tt><a class="headerlink" href="#has-cpp-attribute" title="Permalink to this headline">¶</a></h3>
+<p>This function-like macro takes a single argument that is the name of a
+C++11-style attribute. The argument can either be a single identifier, or a
+scoped identifier. If the attribute is supported, a nonzero value is returned.
+If the attribute is a standards-based attribute, this macro returns a nonzero
+value based on the year and month in which the attribute was voted into the
+working draft. If the attribute is not supported by the current compliation
+target, this macro evaluates to 0. It can be used like this:</p>
+<div class="highlight-c++"><div class="highlight"><pre>#ifndef __has_cpp_attribute // Optional of course.
+ #define __has_cpp_attribute(x) 0 // Compatibility with non-clang compilers.
+#endif
+
+...
+#if __has_cpp_attribute(clang::fallthrough)
+#define FALLTHROUGH [[clang::fallthrough]]
+#else
+#define FALLTHROUGH
+#endif
+...
+</pre></div>
+</div>
+<p>The attribute identifier (but not scope) can also be specified with a preceding
+and following <tt class="docutils literal"><span class="pre">__</span></tt> (double underscore) to avoid interference from a macro with
+the same name. For instance, <tt class="docutils literal"><span class="pre">gnu::__const__</span></tt> can be used instead of
+<tt class="docutils literal"><span class="pre">gnu::const</span></tt>.</p>
+</div>
+<div class="section" id="has-attribute">
+<h3><tt class="docutils literal"><span class="pre">__has_attribute</span></tt><a class="headerlink" href="#has-attribute" title="Permalink to this headline">¶</a></h3>
+<p>This function-like macro takes a single identifier argument that is the name of
+a GNU-style attribute. It evaluates to 1 if the attribute is supported by the
+current compilation target, or 0 if not. It can be used like this:</p>
+<div class="highlight-c++"><div class="highlight"><pre>#ifndef __has_attribute // Optional of course.
+ #define __has_attribute(x) 0 // Compatibility with non-clang compilers.
+#endif
+
+...
+#if __has_attribute(always_inline)
+#define ALWAYS_INLINE __attribute__((always_inline))
+#else
+#define ALWAYS_INLINE
+#endif
+...
+</pre></div>
+</div>
+<p>The attribute name can also be specified with a preceding and following <tt class="docutils literal"><span class="pre">__</span></tt>
+(double underscore) to avoid interference from a macro with the same name. For
+instance, <tt class="docutils literal"><span class="pre">__always_inline__</span></tt> can be used instead of <tt class="docutils literal"><span class="pre">always_inline</span></tt>.</p>
+</div>
+<div class="section" id="has-declspec-attribute">
+<h3><tt class="docutils literal"><span class="pre">__has_declspec_attribute</span></tt><a class="headerlink" href="#has-declspec-attribute" title="Permalink to this headline">¶</a></h3>
+<p>This function-like macro takes a single identifier argument that is the name of
+an attribute implemented as a Microsoft-style <tt class="docutils literal"><span class="pre">__declspec</span></tt> attribute. It
+evaluates to 1 if the attribute is supported by the current compilation target,
+or 0 if not. It can be used like this:</p>
+<div class="highlight-c++"><div class="highlight"><pre>#ifndef __has_declspec_attribute // Optional of course.
+ #define __has_declspec_attribute(x) 0 // Compatibility with non-clang compilers.
+#endif
+
+...
+#if __has_declspec_attribute(dllexport)
+#define DLLEXPORT __declspec(dllexport)
+#else
+#define DLLEXPORT
+#endif
+...
+</pre></div>
+</div>
+<p>The attribute name can also be specified with a preceding and following <tt class="docutils literal"><span class="pre">__</span></tt>
+(double underscore) to avoid interference from a macro with the same name. For
+instance, <tt class="docutils literal"><span class="pre">__dllexport__</span></tt> can be used instead of <tt class="docutils literal"><span class="pre">dllexport</span></tt>.</p>
+</div>
+<div class="section" id="is-identifier">
+<h3><tt class="docutils literal"><span class="pre">__is_identifier</span></tt><a class="headerlink" href="#is-identifier" title="Permalink to this headline">¶</a></h3>
+<p>This function-like macro takes a single identifier argument that might be either
+a reserved word or a regular identifier. It evaluates to 1 if the argument is just
+a regular identifier and not a reserved word, in the sense that it can then be
+used as the name of a user-defined function or variable. Otherwise it evaluates
+to 0. It can be used like this:</p>
+<div class="highlight-c++"><div class="highlight"><pre>...
+#ifdef __is_identifier // Compatibility with non-clang compilers.
+ #if __is_identifier(__wchar_t)
+ typedef wchar_t __wchar_t;
+ #endif
+#endif
+
+__wchar_t WideCharacter;
+...
+</pre></div>
+</div>
+</div>
+</div>
+<div class="section" id="include-file-checking-macros">
+<h2><a class="toc-backref" href="#id3">Include File Checking Macros</a><a class="headerlink" href="#include-file-checking-macros" title="Permalink to this headline">¶</a></h2>
+<p>Not all developments systems have the same include files. The
+<a class="reference internal" href="#langext-has-include"><em>__has_include</em></a> and <a class="reference internal" href="#langext-has-include-next"><em>__has_include_next</em></a> macros allow
+you to check for the existence of an include file before doing a possibly
+failing <tt class="docutils literal"><span class="pre">#include</span></tt> directive. Include file checking macros must be used
+as expressions in <tt class="docutils literal"><span class="pre">#if</span></tt> or <tt class="docutils literal"><span class="pre">#elif</span></tt> preprocessing directives.</p>
+<div class="section" id="has-include">
+<span id="langext-has-include"></span><h3><tt class="docutils literal"><span class="pre">__has_include</span></tt><a class="headerlink" href="#has-include" title="Permalink to this headline">¶</a></h3>
+<p>This function-like macro takes a single file name string argument that is the
+name of an include file. It evaluates to 1 if the file can be found using the
+include paths, or 0 otherwise:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="c1">// Note the two possible file name string formats.</span>
+<span class="cp">#if __has_include("myinclude.h") && __has_include(<stdint.h>)</span>
+<span class="cp"># include "myinclude.h"</span>
+<span class="cp">#endif</span>
+</pre></div>
+</div>
+<p>To test for this feature, use <tt class="docutils literal"><span class="pre">#if</span> <span class="pre">defined(__has_include)</span></tt>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="c1">// To avoid problem with non-clang compilers not having this macro.</span>
+<span class="cp">#if defined(__has_include)</span>
+<span class="cp">#if __has_include("myinclude.h")</span>
+<span class="cp"># include "myinclude.h"</span>
+<span class="cp">#endif</span>
+<span class="cp">#endif</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="has-include-next">
+<span id="langext-has-include-next"></span><h3><tt class="docutils literal"><span class="pre">__has_include_next</span></tt><a class="headerlink" href="#has-include-next" title="Permalink to this headline">¶</a></h3>
+<p>This function-like macro takes a single file name string argument that is the
+name of an include file. It is like <tt class="docutils literal"><span class="pre">__has_include</span></tt> except that it looks for
+the second instance of the given file found in the include paths. It evaluates
+to 1 if the second instance of the file can be found using the include paths,
+or 0 otherwise:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="c1">// Note the two possible file name string formats.</span>
+<span class="cp">#if __has_include_next("myinclude.h") && __has_include_next(<stdint.h>)</span>
+<span class="cp"># include_next "myinclude.h"</span>
+<span class="cp">#endif</span>
+
+<span class="c1">// To avoid problem with non-clang compilers not having this macro.</span>
+<span class="cp">#if defined(__has_include_next)</span>
+<span class="cp">#if __has_include_next("myinclude.h")</span>
+<span class="cp"># include_next "myinclude.h"</span>
+<span class="cp">#endif</span>
+<span class="cp">#endif</span>
+</pre></div>
+</div>
+<p>Note that <tt class="docutils literal"><span class="pre">__has_include_next</span></tt>, like the GNU extension <tt class="docutils literal"><span class="pre">#include_next</span></tt>
+directive, is intended for use in headers only, and will issue a warning if
+used in the top-level compilation file. A warning will also be issued if an
+absolute path is used in the file argument.</p>
+</div>
+<div class="section" id="has-warning">
+<h3><tt class="docutils literal"><span class="pre">__has_warning</span></tt><a class="headerlink" href="#has-warning" title="Permalink to this headline">¶</a></h3>
+<p>This function-like macro takes a string literal that represents a command line
+option for a warning and returns true if that is a valid warning option.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#if __has_warning("-Wformat")</span>
+<span class="p">...</span>
+<span class="cp">#endif</span>
+</pre></div>
+</div>
+</div>
+</div>
+<div class="section" id="builtin-macros">
+<h2><a class="toc-backref" href="#id4">Builtin Macros</a><a class="headerlink" href="#builtin-macros" title="Permalink to this headline">¶</a></h2>
+<dl class="docutils">
+<dt><tt class="docutils literal"><span class="pre">__BASE_FILE__</span></tt></dt>
+<dd>Defined to a string that contains the name of the main input file passed to
+Clang.</dd>
+<dt><tt class="docutils literal"><span class="pre">__COUNTER__</span></tt></dt>
+<dd>Defined to an integer value that starts at zero and is incremented each time
+the <tt class="docutils literal"><span class="pre">__COUNTER__</span></tt> macro is expanded.</dd>
+<dt><tt class="docutils literal"><span class="pre">__INCLUDE_LEVEL__</span></tt></dt>
+<dd>Defined to an integral value that is the include depth of the file currently
+being translated. For the main file, this value is zero.</dd>
+<dt><tt class="docutils literal"><span class="pre">__TIMESTAMP__</span></tt></dt>
+<dd>Defined to the date and time of the last modification of the current source
+file.</dd>
+<dt><tt class="docutils literal"><span class="pre">__clang__</span></tt></dt>
+<dd>Defined when compiling with Clang</dd>
+<dt><tt class="docutils literal"><span class="pre">__clang_major__</span></tt></dt>
+<dd>Defined to the major marketing version number of Clang (e.g., the 2 in
+2.0.1). Note that marketing version numbers should not be used to check for
+language features, as different vendors use different numbering schemes.
+Instead, use the <a class="reference internal" href="#langext-feature-check"><em>Feature Checking Macros</em></a>.</dd>
+<dt><tt class="docutils literal"><span class="pre">__clang_minor__</span></tt></dt>
+<dd>Defined to the minor version number of Clang (e.g., the 0 in 2.0.1). Note
+that marketing version numbers should not be used to check for language
+features, as different vendors use different numbering schemes. Instead, use
+the <a class="reference internal" href="#langext-feature-check"><em>Feature Checking Macros</em></a>.</dd>
+<dt><tt class="docutils literal"><span class="pre">__clang_patchlevel__</span></tt></dt>
+<dd>Defined to the marketing patch level of Clang (e.g., the 1 in 2.0.1).</dd>
+<dt><tt class="docutils literal"><span class="pre">__clang_version__</span></tt></dt>
+<dd>Defined to a string that captures the Clang marketing version, including the
+Subversion tag or revision number, e.g., “<tt class="docutils literal"><span class="pre">1.5</span> <span class="pre">(trunk</span> <span class="pre">102332)</span></tt>”.</dd>
+</dl>
+</div>
+<div class="section" id="vectors-and-extended-vectors">
+<span id="langext-vectors"></span><h2><a class="toc-backref" href="#id5">Vectors and Extended Vectors</a><a class="headerlink" href="#vectors-and-extended-vectors" title="Permalink to this headline">¶</a></h2>
+<p>Supports the GCC, OpenCL, AltiVec and NEON vector extensions.</p>
+<p>OpenCL vector types are created using <tt class="docutils literal"><span class="pre">ext_vector_type</span></tt> attribute. It
+support for <tt class="docutils literal"><span class="pre">V.xyzw</span></tt> syntax and other tidbits as seen in OpenCL. An example
+is:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">typedef</span> <span class="kt">float</span> <span class="n">float4</span> <span class="nf">__attribute__</span><span class="p">((</span><span class="n">ext_vector_type</span><span class="p">(</span><span class="mi">4</span><span class="p">)));</span>
+<span class="k">typedef</span> <span class="kt">float</span> <span class="n">float2</span> <span class="nf">__attribute__</span><span class="p">((</span><span class="n">ext_vector_type</span><span class="p">(</span><span class="mi">2</span><span class="p">)));</span>
+
+<span class="n">float4</span> <span class="nf">foo</span><span class="p">(</span><span class="n">float2</span> <span class="n">a</span><span class="p">,</span> <span class="n">float2</span> <span class="n">b</span><span class="p">)</span> <span class="p">{</span>
+ <span class="n">float4</span> <span class="n">c</span><span class="p">;</span>
+ <span class="n">c</span><span class="p">.</span><span class="n">xz</span> <span class="o">=</span> <span class="n">a</span><span class="p">;</span>
+ <span class="n">c</span><span class="p">.</span><span class="n">yw</span> <span class="o">=</span> <span class="n">b</span><span class="p">;</span>
+ <span class="k">return</span> <span class="n">c</span><span class="p">;</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>Query for this feature with <tt class="docutils literal"><span class="pre">__has_extension(attribute_ext_vector_type)</span></tt>.</p>
+<p>Giving <tt class="docutils literal"><span class="pre">-faltivec</span></tt> option to clang enables support for AltiVec vector syntax
+and functions. For example:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">vector</span> <span class="kt">float</span> <span class="nf">foo</span><span class="p">(</span><span class="n">vector</span> <span class="kt">int</span> <span class="n">a</span><span class="p">)</span> <span class="p">{</span>
+ <span class="n">vector</span> <span class="kt">int</span> <span class="n">b</span><span class="p">;</span>
+ <span class="n">b</span> <span class="o">=</span> <span class="n">vec_add</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">a</span><span class="p">)</span> <span class="o">+</span> <span class="n">a</span><span class="p">;</span>
+ <span class="k">return</span> <span class="p">(</span><span class="n">vector</span> <span class="kt">float</span><span class="p">)</span><span class="n">b</span><span class="p">;</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>NEON vector types are created using <tt class="docutils literal"><span class="pre">neon_vector_type</span></tt> and
+<tt class="docutils literal"><span class="pre">neon_polyvector_type</span></tt> attributes. For example:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">typedef</span> <span class="n">__attribute__</span><span class="p">((</span><span class="n">neon_vector_type</span><span class="p">(</span><span class="mi">8</span><span class="p">)))</span> <span class="kt">int8_t</span> <span class="kt">int8x8_t</span><span class="p">;</span>
+<span class="k">typedef</span> <span class="n">__attribute__</span><span class="p">((</span><span class="n">neon_polyvector_type</span><span class="p">(</span><span class="mi">16</span><span class="p">)))</span> <span class="kt">poly8_t</span> <span class="kt">poly8x16_t</span><span class="p">;</span>
+
+<span class="kt">int8x8_t</span> <span class="nf">foo</span><span class="p">(</span><span class="kt">int8x8_t</span> <span class="n">a</span><span class="p">)</span> <span class="p">{</span>
+ <span class="kt">int8x8_t</span> <span class="n">v</span><span class="p">;</span>
+ <span class="n">v</span> <span class="o">=</span> <span class="n">a</span><span class="p">;</span>
+ <span class="k">return</span> <span class="n">v</span><span class="p">;</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<div class="section" id="vector-literals">
+<h3>Vector Literals<a class="headerlink" href="#vector-literals" title="Permalink to this headline">¶</a></h3>
+<p>Vector literals can be used to create vectors from a set of scalars, or
+vectors. Either parentheses or braces form can be used. In the parentheses
+form the number of literal values specified must be one, i.e. referring to a
+scalar value, or must match the size of the vector type being created. If a
+single scalar literal value is specified, the scalar literal value will be
+replicated to all the components of the vector type. In the brackets form any
+number of literals can be specified. For example:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">typedef</span> <span class="kt">int</span> <span class="n">v4si</span> <span class="nf">__attribute__</span><span class="p">((</span><span class="n">__vector_size__</span><span class="p">(</span><span class="mi">16</span><span class="p">)));</span>
+<span class="k">typedef</span> <span class="kt">float</span> <span class="n">float4</span> <span class="nf">__attribute__</span><span class="p">((</span><span class="n">ext_vector_type</span><span class="p">(</span><span class="mi">4</span><span class="p">)));</span>
+<span class="k">typedef</span> <span class="kt">float</span> <span class="n">float2</span> <span class="nf">__attribute__</span><span class="p">((</span><span class="n">ext_vector_type</span><span class="p">(</span><span class="mi">2</span><span class="p">)));</span>
+
+<span class="n">v4si</span> <span class="n">vsi</span> <span class="o">=</span> <span class="p">(</span><span class="n">v4si</span><span class="p">){</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">4</span><span class="p">};</span>
+<span class="n">float4</span> <span class="n">vf</span> <span class="o">=</span> <span class="p">(</span><span class="n">float4</span><span class="p">)(</span><span class="mf">1.0f</span><span class="p">,</span> <span class="mf">2.0f</span><span class="p">,</span> <span class="mf">3.0f</span><span class="p">,</span> <span class="mf">4.0f</span><span class="p">);</span>
+<span class="n">vector</span> <span class="kt">int</span> <span class="n">vi1</span> <span class="o">=</span> <span class="p">(</span><span class="n">vector</span> <span class="kt">int</span><span class="p">)(</span><span class="mi">1</span><span class="p">);</span> <span class="c1">// vi1 will be (1, 1, 1, 1).</span>
+<span class="n">vector</span> <span class="kt">int</span> <span class="n">vi2</span> <span class="o">=</span> <span class="p">(</span><span class="n">vector</span> <span class="kt">int</span><span class="p">){</span><span class="mi">1</span><span class="p">};</span> <span class="c1">// vi2 will be (1, 0, 0, 0).</span>
+<span class="n">vector</span> <span class="kt">int</span> <span class="n">vi3</span> <span class="o">=</span> <span class="p">(</span><span class="n">vector</span> <span class="kt">int</span><span class="p">)(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">);</span> <span class="c1">// error</span>
+<span class="n">vector</span> <span class="kt">int</span> <span class="n">vi4</span> <span class="o">=</span> <span class="p">(</span><span class="n">vector</span> <span class="kt">int</span><span class="p">){</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">};</span> <span class="c1">// vi4 will be (1, 2, 0, 0).</span>
+<span class="n">vector</span> <span class="kt">int</span> <span class="n">vi5</span> <span class="o">=</span> <span class="p">(</span><span class="n">vector</span> <span class="kt">int</span><span class="p">)(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">4</span><span class="p">);</span>
+<span class="n">float4</span> <span class="n">vf</span> <span class="o">=</span> <span class="p">(</span><span class="n">float4</span><span class="p">)((</span><span class="n">float2</span><span class="p">)(</span><span class="mf">1.0f</span><span class="p">,</span> <span class="mf">2.0f</span><span class="p">),</span> <span class="p">(</span><span class="n">float2</span><span class="p">)(</span><span class="mf">3.0f</span><span class="p">,</span> <span class="mf">4.0f</span><span class="p">));</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="vector-operations">
+<h3>Vector Operations<a class="headerlink" href="#vector-operations" title="Permalink to this headline">¶</a></h3>
+<p>The table below shows the support for each operation by vector extension. A
+dash indicates that an operation is not accepted according to a corresponding
+specification.</p>
+<table border="1" class="docutils">
+<colgroup>
+<col width="52%" />
+<col width="12%" />
+<col width="12%" />
+<col width="12%" />
+<col width="12%" />
+</colgroup>
+<thead valign="bottom">
+<tr class="row-odd"><th class="head">Operator</th>
+<th class="head">OpenCL</th>
+<th class="head">AltiVec</th>
+<th class="head">GCC</th>
+<th class="head">NEON</th>
+</tr>
+</thead>
+<tbody valign="top">
+<tr class="row-even"><td>[]</td>
+<td>yes</td>
+<td>yes</td>
+<td>yes</td>
+<td>–</td>
+</tr>
+<tr class="row-odd"><td>unary operators +, –</td>
+<td>yes</td>
+<td>yes</td>
+<td>yes</td>
+<td>–</td>
+</tr>
+<tr class="row-even"><td>++, – –</td>
+<td>yes</td>
+<td>yes</td>
+<td>yes</td>
+<td>–</td>
+</tr>
+<tr class="row-odd"><td>+,–,*,/,%</td>
+<td>yes</td>
+<td>yes</td>
+<td>yes</td>
+<td>–</td>
+</tr>
+<tr class="row-even"><td>bitwise operators &,|,^,~</td>
+<td>yes</td>
+<td>yes</td>
+<td>yes</td>
+<td>–</td>
+</tr>
+<tr class="row-odd"><td>>>,<<</td>
+<td>yes</td>
+<td>yes</td>
+<td>yes</td>
+<td>–</td>
+</tr>
+<tr class="row-even"><td>!, &&, ||</td>
+<td>yes</td>
+<td>–</td>
+<td>–</td>
+<td>–</td>
+</tr>
+<tr class="row-odd"><td>==, !=, >, <, >=, <=</td>
+<td>yes</td>
+<td>yes</td>
+<td>–</td>
+<td>–</td>
+</tr>
+<tr class="row-even"><td>=</td>
+<td>yes</td>
+<td>yes</td>
+<td>yes</td>
+<td>yes</td>
+</tr>
+<tr class="row-odd"><td>:?</td>
+<td>yes</td>
+<td>–</td>
+<td>–</td>
+<td>–</td>
+</tr>
+<tr class="row-even"><td>sizeof</td>
+<td>yes</td>
+<td>yes</td>
+<td>yes</td>
+<td>yes</td>
+</tr>
+<tr class="row-odd"><td>C-style cast</td>
+<td>yes</td>
+<td>yes</td>
+<td>yes</td>
+<td>no</td>
+</tr>
+<tr class="row-even"><td>reinterpret_cast</td>
+<td>yes</td>
+<td>no</td>
+<td>yes</td>
+<td>no</td>
+</tr>
+<tr class="row-odd"><td>static_cast</td>
+<td>yes</td>
+<td>no</td>
+<td>yes</td>
+<td>no</td>
+</tr>
+<tr class="row-even"><td>const_cast</td>
+<td>no</td>
+<td>no</td>
+<td>no</td>
+<td>no</td>
+</tr>
+</tbody>
+</table>
+<p>See also <a class="reference internal" href="#langext-builtin-shufflevector"><em>__builtin_shufflevector</em></a>, <a class="reference internal" href="#langext-builtin-convertvector"><em>__builtin_convertvector</em></a>.</p>
+</div>
+</div>
+<div class="section" id="messages-on-deprecated-and-unavailable-attributes">
+<h2><a class="toc-backref" href="#id6">Messages on <tt class="docutils literal"><span class="pre">deprecated</span></tt> and <tt class="docutils literal"><span class="pre">unavailable</span></tt> Attributes</a><a class="headerlink" href="#messages-on-deprecated-and-unavailable-attributes" title="Permalink to this headline">¶</a></h2>
+<p>An optional string message can be added to the <tt class="docutils literal"><span class="pre">deprecated</span></tt> and
+<tt class="docutils literal"><span class="pre">unavailable</span></tt> attributes. For example:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">void</span> <span class="nf">explode</span><span class="p">(</span><span class="kt">void</span><span class="p">)</span> <span class="n">__attribute__</span><span class="p">((</span><span class="n">deprecated</span><span class="p">(</span><span class="s">"extremely unsafe, use 'combust' instead!!!"</span><span class="p">)));</span>
+</pre></div>
+</div>
+<p>If the deprecated or unavailable declaration is used, the message will be
+incorporated into the appropriate diagnostic:</p>
+<div class="highlight-c++"><div class="highlight"><pre>harmless.c:4:3: warning: 'explode' is deprecated: extremely unsafe, use 'combust' instead!!!
+ [-Wdeprecated-declarations]
+ explode();
+ ^
+</pre></div>
+</div>
+<p>Query for this feature with
+<tt class="docutils literal"><span class="pre">__has_extension(attribute_deprecated_with_message)</span></tt> and
+<tt class="docutils literal"><span class="pre">__has_extension(attribute_unavailable_with_message)</span></tt>.</p>
+</div>
+<div class="section" id="attributes-on-enumerators">
+<h2><a class="toc-backref" href="#id7">Attributes on Enumerators</a><a class="headerlink" href="#attributes-on-enumerators" title="Permalink to this headline">¶</a></h2>
+<p>Clang allows attributes to be written on individual enumerators. This allows
+enumerators to be deprecated, made unavailable, etc. The attribute must appear
+after the enumerator name and before any initializer, like so:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">enum</span> <span class="n">OperationMode</span> <span class="p">{</span>
+ <span class="n">OM_Invalid</span><span class="p">,</span>
+ <span class="n">OM_Normal</span><span class="p">,</span>
+ <span class="n">OM_Terrified</span> <span class="n">__attribute__</span><span class="p">((</span><span class="n">deprecated</span><span class="p">)),</span>
+ <span class="n">OM_AbortOnError</span> <span class="n">__attribute__</span><span class="p">((</span><span class="n">deprecated</span><span class="p">))</span> <span class="o">=</span> <span class="mi">4</span>
+<span class="p">};</span>
+</pre></div>
+</div>
+<p>Attributes on the <tt class="docutils literal"><span class="pre">enum</span></tt> declaration do not apply to individual enumerators.</p>
+<p>Query for this feature with <tt class="docutils literal"><span class="pre">__has_extension(enumerator_attributes)</span></tt>.</p>
+</div>
+<div class="section" id="user-specified-system-frameworks">
+<h2><a class="toc-backref" href="#id8">‘User-Specified’ System Frameworks</a><a class="headerlink" href="#user-specified-system-frameworks" title="Permalink to this headline">¶</a></h2>
+<p>Clang provides a mechanism by which frameworks can be built in such a way that
+they will always be treated as being “system frameworks”, even if they are not
+present in a system framework directory. This can be useful to system
+framework developers who want to be able to test building other applications
+with development builds of their framework, including the manner in which the
+compiler changes warning behavior for system headers.</p>
+<p>Framework developers can opt-in to this mechanism by creating a
+“<tt class="docutils literal"><span class="pre">.system_framework</span></tt>” file at the top-level of their framework. That is, the
+framework should have contents like:</p>
+<div class="highlight-none"><div class="highlight"><pre>.../TestFramework.framework
+.../TestFramework.framework/.system_framework
+.../TestFramework.framework/Headers
+.../TestFramework.framework/Headers/TestFramework.h
+...
+</pre></div>
+</div>
+<p>Clang will treat the presence of this file as an indicator that the framework
+should be treated as a system framework, regardless of how it was found in the
+framework search path. For consistency, we recommend that such files never be
+included in installed versions of the framework.</p>
+</div>
+<div class="section" id="checks-for-standard-language-features">
+<h2><a class="toc-backref" href="#id9">Checks for Standard Language Features</a><a class="headerlink" href="#checks-for-standard-language-features" title="Permalink to this headline">¶</a></h2>
+<p>The <tt class="docutils literal"><span class="pre">__has_feature</span></tt> macro can be used to query if certain standard language
+features are enabled. The <tt class="docutils literal"><span class="pre">__has_extension</span></tt> macro can be used to query if
+language features are available as an extension when compiling for a standard
+which does not provide them. The features which can be tested are listed here.</p>
+<p>Since Clang 3.4, the C++ SD-6 feature test macros are also supported.
+These are macros with names of the form <tt class="docutils literal"><span class="pre">__cpp_<feature_name></span></tt>, and are
+intended to be a portable way to query the supported features of the compiler.
+See <a class="reference external" href="http://clang.llvm.org/cxx_status.html#ts">the C++ status page</a> for
+information on the version of SD-6 supported by each Clang release, and the
+macros provided by that revision of the recommendations.</p>
+<div class="section" id="c-98">
+<h3>C++98<a class="headerlink" href="#c-98" title="Permalink to this headline">¶</a></h3>
+<p>The features listed below are part of the C++98 standard. These features are
+enabled by default when compiling C++ code.</p>
+<div class="section" id="c-exceptions">
+<h4>C++ exceptions<a class="headerlink" href="#c-exceptions" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_exceptions)</span></tt> to determine if C++ exceptions have been
+enabled. For example, compiling code with <tt class="docutils literal"><span class="pre">-fno-exceptions</span></tt> disables C++
+exceptions.</p>
+</div>
+<div class="section" id="c-rtti">
+<h4>C++ RTTI<a class="headerlink" href="#c-rtti" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_rtti)</span></tt> to determine if C++ RTTI has been enabled. For
+example, compiling code with <tt class="docutils literal"><span class="pre">-fno-rtti</span></tt> disables the use of RTTI.</p>
+</div>
+</div>
+<div class="section" id="c-11">
+<h3>C++11<a class="headerlink" href="#c-11" title="Permalink to this headline">¶</a></h3>
+<p>The features listed below are part of the C++11 standard. As a result, all
+these features are enabled with the <tt class="docutils literal"><span class="pre">-std=c++11</span></tt> or <tt class="docutils literal"><span class="pre">-std=gnu++11</span></tt> option
+when compiling C++ code.</p>
+<div class="section" id="c-11-sfinae-includes-access-control">
+<h4>C++11 SFINAE includes access control<a class="headerlink" href="#c-11-sfinae-includes-access-control" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_access_control_sfinae)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_access_control_sfinae)</span></tt> to determine whether
+access-control errors (e.g., calling a private constructor) are considered to
+be template argument deduction errors (aka SFINAE errors), per <a class="reference external" href="http://www.open-std.org/jtc1/sc22/wg21/docs/cwg_defects.html#1170">C++ DR1170</a>.</p>
+</div>
+<div class="section" id="c-11-alias-templates">
+<h4>C++11 alias templates<a class="headerlink" href="#c-11-alias-templates" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_alias_templates)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_alias_templates)</span></tt> to determine if support for C++11’s
+alias declarations and alias templates is enabled.</p>
+</div>
+<div class="section" id="c-11-alignment-specifiers">
+<h4>C++11 alignment specifiers<a class="headerlink" href="#c-11-alignment-specifiers" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_alignas)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(cxx_alignas)</span></tt> to
+determine if support for alignment specifiers using <tt class="docutils literal"><span class="pre">alignas</span></tt> is enabled.</p>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_alignof)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(cxx_alignof)</span></tt> to
+determine if support for the <tt class="docutils literal"><span class="pre">alignof</span></tt> keyword is enabled.</p>
+</div>
+<div class="section" id="c-11-attributes">
+<h4>C++11 attributes<a class="headerlink" href="#c-11-attributes" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_attributes)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(cxx_attributes)</span></tt> to
+determine if support for attribute parsing with C++11’s square bracket notation
+is enabled.</p>
+</div>
+<div class="section" id="c-11-generalized-constant-expressions">
+<h4>C++11 generalized constant expressions<a class="headerlink" href="#c-11-generalized-constant-expressions" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_constexpr)</span></tt> to determine if support for generalized
+constant expressions (e.g., <tt class="docutils literal"><span class="pre">constexpr</span></tt>) is enabled.</p>
+</div>
+<div class="section" id="c-11-decltype">
+<h4>C++11 <tt class="docutils literal"><span class="pre">decltype()</span></tt><a class="headerlink" href="#c-11-decltype" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_decltype)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(cxx_decltype)</span></tt> to
+determine if support for the <tt class="docutils literal"><span class="pre">decltype()</span></tt> specifier is enabled. C++11’s
+<tt class="docutils literal"><span class="pre">decltype</span></tt> does not require type-completeness of a function call expression.
+Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_decltype_incomplete_return_types)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_decltype_incomplete_return_types)</span></tt> to determine if
+support for this feature is enabled.</p>
+</div>
+<div class="section" id="c-11-default-template-arguments-in-function-templates">
+<h4>C++11 default template arguments in function templates<a class="headerlink" href="#c-11-default-template-arguments-in-function-templates" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_default_function_template_args)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_default_function_template_args)</span></tt> to determine if support
+for default template arguments in function templates is enabled.</p>
+</div>
+<div class="section" id="c-11-defaulted-functions">
+<h4>C++11 <tt class="docutils literal"><span class="pre">default</span></tt>ed functions<a class="headerlink" href="#c-11-defaulted-functions" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_defaulted_functions)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_defaulted_functions)</span></tt> to determine if support for
+defaulted function definitions (with <tt class="docutils literal"><span class="pre">=</span> <span class="pre">default</span></tt>) is enabled.</p>
+</div>
+<div class="section" id="c-11-delegating-constructors">
+<h4>C++11 delegating constructors<a class="headerlink" href="#c-11-delegating-constructors" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_delegating_constructors)</span></tt> to determine if support for
+delegating constructors is enabled.</p>
+</div>
+<div class="section" id="c-11-deleted-functions">
+<h4>C++11 <tt class="docutils literal"><span class="pre">deleted</span></tt> functions<a class="headerlink" href="#c-11-deleted-functions" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_deleted_functions)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_deleted_functions)</span></tt> to determine if support for deleted
+function definitions (with <tt class="docutils literal"><span class="pre">=</span> <span class="pre">delete</span></tt>) is enabled.</p>
+</div>
+<div class="section" id="c-11-explicit-conversion-functions">
+<h4>C++11 explicit conversion functions<a class="headerlink" href="#c-11-explicit-conversion-functions" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_explicit_conversions)</span></tt> to determine if support for
+<tt class="docutils literal"><span class="pre">explicit</span></tt> conversion functions is enabled.</p>
+</div>
+<div class="section" id="c-11-generalized-initializers">
+<h4>C++11 generalized initializers<a class="headerlink" href="#c-11-generalized-initializers" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_generalized_initializers)</span></tt> to determine if support for
+generalized initializers (using braced lists and <tt class="docutils literal"><span class="pre">std::initializer_list</span></tt>) is
+enabled.</p>
+</div>
+<div class="section" id="c-11-implicit-move-constructors-assignment-operators">
+<h4>C++11 implicit move constructors/assignment operators<a class="headerlink" href="#c-11-implicit-move-constructors-assignment-operators" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_implicit_moves)</span></tt> to determine if Clang will implicitly
+generate move constructors and move assignment operators where needed.</p>
+</div>
+<div class="section" id="c-11-inheriting-constructors">
+<h4>C++11 inheriting constructors<a class="headerlink" href="#c-11-inheriting-constructors" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_inheriting_constructors)</span></tt> to determine if support for
+inheriting constructors is enabled.</p>
+</div>
+<div class="section" id="c-11-inline-namespaces">
+<h4>C++11 inline namespaces<a class="headerlink" href="#c-11-inline-namespaces" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_inline_namespaces)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_inline_namespaces)</span></tt> to determine if support for inline
+namespaces is enabled.</p>
+</div>
+<div class="section" id="c-11-lambdas">
+<h4>C++11 lambdas<a class="headerlink" href="#c-11-lambdas" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_lambdas)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(cxx_lambdas)</span></tt> to
+determine if support for lambdas is enabled.</p>
+</div>
+<div class="section" id="c-11-local-and-unnamed-types-as-template-arguments">
+<h4>C++11 local and unnamed types as template arguments<a class="headerlink" href="#c-11-local-and-unnamed-types-as-template-arguments" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_local_type_template_args)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_local_type_template_args)</span></tt> to determine if support for
+local and unnamed types as template arguments is enabled.</p>
+</div>
+<div class="section" id="c-11-noexcept">
+<h4>C++11 noexcept<a class="headerlink" href="#c-11-noexcept" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_noexcept)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(cxx_noexcept)</span></tt> to
+determine if support for noexcept exception specifications is enabled.</p>
+</div>
+<div class="section" id="c-11-in-class-non-static-data-member-initialization">
+<h4>C++11 in-class non-static data member initialization<a class="headerlink" href="#c-11-in-class-non-static-data-member-initialization" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_nonstatic_member_init)</span></tt> to determine whether in-class
+initialization of non-static data members is enabled.</p>
+</div>
+<div class="section" id="c-11-nullptr">
+<h4>C++11 <tt class="docutils literal"><span class="pre">nullptr</span></tt><a class="headerlink" href="#c-11-nullptr" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_nullptr)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(cxx_nullptr)</span></tt> to
+determine if support for <tt class="docutils literal"><span class="pre">nullptr</span></tt> is enabled.</p>
+</div>
+<div class="section" id="c-11-override-control">
+<h4>C++11 <tt class="docutils literal"><span class="pre">override</span> <span class="pre">control</span></tt><a class="headerlink" href="#c-11-override-control" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_override_control)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_override_control)</span></tt> to determine if support for the
+override control keywords is enabled.</p>
+</div>
+<div class="section" id="c-11-reference-qualified-functions">
+<h4>C++11 reference-qualified functions<a class="headerlink" href="#c-11-reference-qualified-functions" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_reference_qualified_functions)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_reference_qualified_functions)</span></tt> to determine if support
+for reference-qualified functions (e.g., member functions with <tt class="docutils literal"><span class="pre">&</span></tt> or <tt class="docutils literal"><span class="pre">&&</span></tt>
+applied to <tt class="docutils literal"><span class="pre">*this</span></tt>) is enabled.</p>
+</div>
+<div class="section" id="c-11-range-based-for-loop">
+<h4>C++11 range-based <tt class="docutils literal"><span class="pre">for</span></tt> loop<a class="headerlink" href="#c-11-range-based-for-loop" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_range_for)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(cxx_range_for)</span></tt> to
+determine if support for the range-based for loop is enabled.</p>
+</div>
+<div class="section" id="c-11-raw-string-literals">
+<h4>C++11 raw string literals<a class="headerlink" href="#c-11-raw-string-literals" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_raw_string_literals)</span></tt> to determine if support for raw
+string literals (e.g., <tt class="docutils literal"><span class="pre">R"x(foo\bar)x"</span></tt>) is enabled.</p>
+</div>
+<div class="section" id="c-11-rvalue-references">
+<h4>C++11 rvalue references<a class="headerlink" href="#c-11-rvalue-references" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_rvalue_references)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_rvalue_references)</span></tt> to determine if support for rvalue
+references is enabled.</p>
+</div>
+<div class="section" id="c-11-static-assert">
+<h4>C++11 <tt class="docutils literal"><span class="pre">static_assert()</span></tt><a class="headerlink" href="#c-11-static-assert" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_static_assert)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_static_assert)</span></tt> to determine if support for compile-time
+assertions using <tt class="docutils literal"><span class="pre">static_assert</span></tt> is enabled.</p>
+</div>
+<div class="section" id="c-11-thread-local">
+<h4>C++11 <tt class="docutils literal"><span class="pre">thread_local</span></tt><a class="headerlink" href="#c-11-thread-local" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_thread_local)</span></tt> to determine if support for
+<tt class="docutils literal"><span class="pre">thread_local</span></tt> variables is enabled.</p>
+</div>
+<div class="section" id="c-11-type-inference">
+<h4>C++11 type inference<a class="headerlink" href="#c-11-type-inference" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_auto_type)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(cxx_auto_type)</span></tt> to
+determine C++11 type inference is supported using the <tt class="docutils literal"><span class="pre">auto</span></tt> specifier. If
+this is disabled, <tt class="docutils literal"><span class="pre">auto</span></tt> will instead be a storage class specifier, as in C
+or C++98.</p>
+</div>
+<div class="section" id="c-11-strongly-typed-enumerations">
+<h4>C++11 strongly typed enumerations<a class="headerlink" href="#c-11-strongly-typed-enumerations" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_strong_enums)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_strong_enums)</span></tt> to determine if support for strongly
+typed, scoped enumerations is enabled.</p>
+</div>
+<div class="section" id="c-11-trailing-return-type">
+<h4>C++11 trailing return type<a class="headerlink" href="#c-11-trailing-return-type" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_trailing_return)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_trailing_return)</span></tt> to determine if support for the
+alternate function declaration syntax with trailing return type is enabled.</p>
+</div>
+<div class="section" id="c-11-unicode-string-literals">
+<h4>C++11 Unicode string literals<a class="headerlink" href="#c-11-unicode-string-literals" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_unicode_literals)</span></tt> to determine if support for Unicode
+string literals is enabled.</p>
+</div>
+<div class="section" id="c-11-unrestricted-unions">
+<h4>C++11 unrestricted unions<a class="headerlink" href="#c-11-unrestricted-unions" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_unrestricted_unions)</span></tt> to determine if support for
+unrestricted unions is enabled.</p>
+</div>
+<div class="section" id="c-11-user-defined-literals">
+<h4>C++11 user-defined literals<a class="headerlink" href="#c-11-user-defined-literals" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_user_literals)</span></tt> to determine if support for
+user-defined literals is enabled.</p>
+</div>
+<div class="section" id="c-11-variadic-templates">
+<h4>C++11 variadic templates<a class="headerlink" href="#c-11-variadic-templates" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_variadic_templates)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_variadic_templates)</span></tt> to determine if support for
+variadic templates is enabled.</p>
+</div>
+</div>
+<div class="section" id="c-1y">
+<h3>C++1y<a class="headerlink" href="#c-1y" title="Permalink to this headline">¶</a></h3>
+<p>The features listed below are part of the committee draft for the C++1y
+standard. As a result, all these features are enabled with the <tt class="docutils literal"><span class="pre">-std=c++1y</span></tt>
+or <tt class="docutils literal"><span class="pre">-std=gnu++1y</span></tt> option when compiling C++ code.</p>
+<div class="section" id="c-1y-binary-literals">
+<h4>C++1y binary literals<a class="headerlink" href="#c-1y-binary-literals" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_binary_literals)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_binary_literals)</span></tt> to determine whether
+binary literals (for instance, <tt class="docutils literal"><span class="pre">0b10010</span></tt>) are recognized. Clang supports this
+feature as an extension in all language modes.</p>
+</div>
+<div class="section" id="c-1y-contextual-conversions">
+<h4>C++1y contextual conversions<a class="headerlink" href="#c-1y-contextual-conversions" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_contextual_conversions)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_contextual_conversions)</span></tt> to determine if the C++1y rules
+are used when performing an implicit conversion for an array bound in a
+<em>new-expression</em>, the operand of a <em>delete-expression</em>, an integral constant
+expression, or a condition in a <tt class="docutils literal"><span class="pre">switch</span></tt> statement.</p>
+</div>
+<div class="section" id="c-1y-decltype-auto">
+<h4>C++1y decltype(auto)<a class="headerlink" href="#c-1y-decltype-auto" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_decltype_auto)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_decltype_auto)</span></tt> to determine if support
+for the <tt class="docutils literal"><span class="pre">decltype(auto)</span></tt> placeholder type is enabled.</p>
+</div>
+<div class="section" id="c-1y-default-initializers-for-aggregates">
+<h4>C++1y default initializers for aggregates<a class="headerlink" href="#c-1y-default-initializers-for-aggregates" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_aggregate_nsdmi)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_aggregate_nsdmi)</span></tt> to determine if support
+for default initializers in aggregate members is enabled.</p>
+</div>
+<div class="section" id="c-1y-digit-separators">
+<h4>C++1y digit separators<a class="headerlink" href="#c-1y-digit-separators" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__cpp_digit_separators</span></tt> to determine if support for digit separators
+using single quotes (for instance, <tt class="docutils literal"><span class="pre">10'000</span></tt>) is enabled. At this time, there
+is no corresponding <tt class="docutils literal"><span class="pre">__has_feature</span></tt> name</p>
+</div>
+<div class="section" id="c-1y-generalized-lambda-capture">
+<h4>C++1y generalized lambda capture<a class="headerlink" href="#c-1y-generalized-lambda-capture" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_init_captures)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_init_captures)</span></tt> to determine if support for
+lambda captures with explicit initializers is enabled
+(for instance, <tt class="docutils literal"><span class="pre">[n(0)]</span> <span class="pre">{</span> <span class="pre">return</span> <span class="pre">++n;</span> <span class="pre">}</span></tt>).</p>
+</div>
+<div class="section" id="c-1y-generic-lambdas">
+<h4>C++1y generic lambdas<a class="headerlink" href="#c-1y-generic-lambdas" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_generic_lambdas)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_generic_lambdas)</span></tt> to determine if support for generic
+(polymorphic) lambdas is enabled
+(for instance, <tt class="docutils literal"><span class="pre">[]</span> <span class="pre">(auto</span> <span class="pre">x)</span> <span class="pre">{</span> <span class="pre">return</span> <span class="pre">x</span> <span class="pre">+</span> <span class="pre">1;</span> <span class="pre">}</span></tt>).</p>
+</div>
+<div class="section" id="c-1y-relaxed-constexpr">
+<h4>C++1y relaxed constexpr<a class="headerlink" href="#c-1y-relaxed-constexpr" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_relaxed_constexpr)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_relaxed_constexpr)</span></tt> to determine if variable
+declarations, local variable modification, and control flow constructs
+are permitted in <tt class="docutils literal"><span class="pre">constexpr</span></tt> functions.</p>
+</div>
+<div class="section" id="c-1y-return-type-deduction">
+<h4>C++1y return type deduction<a class="headerlink" href="#c-1y-return-type-deduction" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_return_type_deduction)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_return_type_deduction)</span></tt> to determine if support
+for return type deduction for functions (using <tt class="docutils literal"><span class="pre">auto</span></tt> as a return type)
+is enabled.</p>
+</div>
+<div class="section" id="c-1y-runtime-sized-arrays">
+<h4>C++1y runtime-sized arrays<a class="headerlink" href="#c-1y-runtime-sized-arrays" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_runtime_array)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_runtime_array)</span></tt> to determine if support
+for arrays of runtime bound (a restricted form of variable-length arrays)
+is enabled.
+Clang’s implementation of this feature is incomplete.</p>
+</div>
+<div class="section" id="c-1y-variable-templates">
+<h4>C++1y variable templates<a class="headerlink" href="#c-1y-variable-templates" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(cxx_variable_templates)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(cxx_variable_templates)</span></tt> to determine if support for
+templated variable declarations is enabled.</p>
+</div>
+</div>
+<div class="section" id="c11">
+<h3>C11<a class="headerlink" href="#c11" title="Permalink to this headline">¶</a></h3>
+<p>The features listed below are part of the C11 standard. As a result, all these
+features are enabled with the <tt class="docutils literal"><span class="pre">-std=c11</span></tt> or <tt class="docutils literal"><span class="pre">-std=gnu11</span></tt> option when
+compiling C code. Additionally, because these features are all
+backward-compatible, they are available as extensions in all language modes.</p>
+<div class="section" id="c11-alignment-specifiers">
+<h4>C11 alignment specifiers<a class="headerlink" href="#c11-alignment-specifiers" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(c_alignas)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(c_alignas)</span></tt> to determine
+if support for alignment specifiers using <tt class="docutils literal"><span class="pre">_Alignas</span></tt> is enabled.</p>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(c_alignof)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(c_alignof)</span></tt> to determine
+if support for the <tt class="docutils literal"><span class="pre">_Alignof</span></tt> keyword is enabled.</p>
+</div>
+<div class="section" id="c11-atomic-operations">
+<h4>C11 atomic operations<a class="headerlink" href="#c11-atomic-operations" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(c_atomic)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(c_atomic)</span></tt> to determine
+if support for atomic types using <tt class="docutils literal"><span class="pre">_Atomic</span></tt> is enabled. Clang also provides
+<a class="reference internal" href="#langext-c11-atomic"><em>a set of builtins</em></a> which can be used to implement
+the <tt class="docutils literal"><span class="pre"><stdatomic.h></span></tt> operations on <tt class="docutils literal"><span class="pre">_Atomic</span></tt> types. Use
+<tt class="docutils literal"><span class="pre">__has_include(<stdatomic.h>)</span></tt> to determine if C11’s <tt class="docutils literal"><span class="pre"><stdatomic.h></span></tt> header
+is available.</p>
+<p>Clang will use the system’s <tt class="docutils literal"><span class="pre"><stdatomic.h></span></tt> header when one is available, and
+will otherwise use its own. When using its own, implementations of the atomic
+operations are provided as macros. In the cases where C11 also requires a real
+function, this header provides only the declaration of that function (along
+with a shadowing macro implementation), and you must link to a library which
+provides a definition of the function if you use it instead of the macro.</p>
+</div>
+<div class="section" id="c11-generic-selections">
+<h4>C11 generic selections<a class="headerlink" href="#c11-generic-selections" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(c_generic_selections)</span></tt> or
+<tt class="docutils literal"><span class="pre">__has_extension(c_generic_selections)</span></tt> to determine if support for generic
+selections is enabled.</p>
+<p>As an extension, the C11 generic selection expression is available in all
+languages supported by Clang. The syntax is the same as that given in the C11
+standard.</p>
+<p>In C, type compatibility is decided according to the rules given in the
+appropriate standard, but in C++, which lacks the type compatibility rules used
+in C, types are considered compatible only if they are equivalent.</p>
+</div>
+<div class="section" id="c11-static-assert">
+<h4>C11 <tt class="docutils literal"><span class="pre">_Static_assert()</span></tt><a class="headerlink" href="#c11-static-assert" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(c_static_assert)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(c_static_assert)</span></tt>
+to determine if support for compile-time assertions using <tt class="docutils literal"><span class="pre">_Static_assert</span></tt> is
+enabled.</p>
+</div>
+<div class="section" id="c11-thread-local">
+<h4>C11 <tt class="docutils literal"><span class="pre">_Thread_local</span></tt><a class="headerlink" href="#c11-thread-local" title="Permalink to this headline">¶</a></h4>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(c_thread_local)</span></tt> or <tt class="docutils literal"><span class="pre">__has_extension(c_thread_local)</span></tt>
+to determine if support for <tt class="docutils literal"><span class="pre">_Thread_local</span></tt> variables is enabled.</p>
+</div>
+</div>
+<div class="section" id="modules">
+<h3>Modules<a class="headerlink" href="#modules" title="Permalink to this headline">¶</a></h3>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(modules)</span></tt> to determine if Modules have been enabled.
+For example, compiling code with <tt class="docutils literal"><span class="pre">-fmodules</span></tt> enables the use of Modules.</p>
+<p>More information could be found <a class="reference external" href="http://clang.llvm.org/docs/Modules.html">here</a>.</p>
+</div>
+</div>
+<div class="section" id="checks-for-type-trait-primitives">
+<h2><a class="toc-backref" href="#id10">Checks for Type Trait Primitives</a><a class="headerlink" href="#checks-for-type-trait-primitives" title="Permalink to this headline">¶</a></h2>
+<p>Type trait primitives are special builtin constant expressions that can be used
+by the standard C++ library to facilitate or simplify the implementation of
+user-facing type traits in the <type_traits> header.</p>
+<p>They are not intended to be used directly by user code because they are
+implementation-defined and subject to change – as such they’re tied closely to
+the supported set of system headers, currently:</p>
+<ul class="simple">
+<li>LLVM’s own libc++</li>
+<li>GNU libstdc++</li>
+<li>The Microsoft standard C++ library</li>
+</ul>
+<p>Clang supports the <a class="reference external" href="http://gcc.gnu.org/onlinedocs/gcc/Type-Traits.html">GNU C++ type traits</a> and a subset of the
+<a class="reference external" href="http://msdn.microsoft.com/en-us/library/ms177194(v=VS.100).aspx">Microsoft Visual C++ Type traits</a>.</p>
+<p>Feature detection is supported only for some of the primitives at present. User
+code should not use these checks because they bear no direct relation to the
+actual set of type traits supported by the C++ standard library.</p>
+<p>For type trait <tt class="docutils literal"><span class="pre">__X</span></tt>, <tt class="docutils literal"><span class="pre">__has_extension(X)</span></tt> indicates the presence of the
+type trait primitive in the compiler. A simplistic usage example as might be
+seen in standard C++ headers follows:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#if __has_extension(is_convertible_to)</span>
+<span class="k">template</span><span class="o"><</span><span class="k">typename</span> <span class="n">From</span><span class="p">,</span> <span class="k">typename</span> <span class="n">To</span><span class="o">></span>
+<span class="k">struct</span> <span class="n">is_convertible_to</span> <span class="p">{</span>
+ <span class="k">static</span> <span class="k">const</span> <span class="kt">bool</span> <span class="n">value</span> <span class="o">=</span> <span class="n">__is_convertible_to</span><span class="p">(</span><span class="n">From</span><span class="p">,</span> <span class="n">To</span><span class="p">);</span>
+<span class="p">};</span>
+<span class="cp">#else</span>
+<span class="c1">// Emulate type trait for compatibility with other compilers.</span>
+<span class="cp">#endif</span>
+</pre></div>
+</div>
+<p>The following type trait primitives are supported by Clang:</p>
+<ul class="simple">
+<li><tt class="docutils literal"><span class="pre">__has_nothrow_assign</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__has_nothrow_copy</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__has_nothrow_constructor</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__has_trivial_assign</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__has_trivial_copy</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__has_trivial_constructor</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__has_trivial_destructor</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__has_virtual_destructor</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__is_abstract</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__is_base_of</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__is_class</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__is_convertible_to</span></tt> (Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__is_empty</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__is_enum</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__is_interface_class</span></tt> (Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__is_pod</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__is_polymorphic</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__is_union</span></tt> (GNU, Microsoft)</li>
+<li><tt class="docutils literal"><span class="pre">__is_literal(type)</span></tt>: Determines whether the given type is a literal type</li>
+<li><tt class="docutils literal"><span class="pre">__is_final</span></tt>: Determines whether the given type is declared with a
+<tt class="docutils literal"><span class="pre">final</span></tt> class-virt-specifier.</li>
+<li><tt class="docutils literal"><span class="pre">__underlying_type(type)</span></tt>: Retrieves the underlying type for a given
+<tt class="docutils literal"><span class="pre">enum</span></tt> type. This trait is required to implement the C++11 standard
+library.</li>
+<li><tt class="docutils literal"><span class="pre">__is_trivially_assignable(totype,</span> <span class="pre">fromtype)</span></tt>: Determines whether a value
+of type <tt class="docutils literal"><span class="pre">totype</span></tt> can be assigned to from a value of type <tt class="docutils literal"><span class="pre">fromtype</span></tt> such
+that no non-trivial functions are called as part of that assignment. This
+trait is required to implement the C++11 standard library.</li>
+<li><tt class="docutils literal"><span class="pre">__is_trivially_constructible(type,</span> <span class="pre">argtypes...)</span></tt>: Determines whether a
+value of type <tt class="docutils literal"><span class="pre">type</span></tt> can be direct-initialized with arguments of types
+<tt class="docutils literal"><span class="pre">argtypes...</span></tt> such that no non-trivial functions are called as part of
+that initialization. This trait is required to implement the C++11 standard
+library.</li>
+<li><tt class="docutils literal"><span class="pre">__is_destructible</span></tt> (MSVC 2013)</li>
+<li><tt class="docutils literal"><span class="pre">__is_nothrow_destructible</span></tt> (MSVC 2013)</li>
+<li><tt class="docutils literal"><span class="pre">__is_nothrow_assignable</span></tt> (MSVC 2013, clang)</li>
+<li><tt class="docutils literal"><span class="pre">__is_constructible</span></tt> (MSVC 2013, clang)</li>
+<li><tt class="docutils literal"><span class="pre">__is_nothrow_constructible</span></tt> (MSVC 2013, clang)</li>
+</ul>
+</div>
+<div class="section" id="blocks">
+<h2><a class="toc-backref" href="#id11">Blocks</a><a class="headerlink" href="#blocks" title="Permalink to this headline">¶</a></h2>
+<p>The syntax and high level language feature description is in
+<a class="reference internal" href="BlockLanguageSpec.html"><em>BlockLanguageSpec</em></a>. Implementation and ABI details for
+the clang implementation are in <a class="reference internal" href="Block-ABI-Apple.html"><em>Block-ABI-Apple</em></a>.</p>
+<p>Query for this feature with <tt class="docutils literal"><span class="pre">__has_extension(blocks)</span></tt>.</p>
+</div>
+<div class="section" id="objective-c-features">
+<h2><a class="toc-backref" href="#id12">Objective-C Features</a><a class="headerlink" href="#objective-c-features" title="Permalink to this headline">¶</a></h2>
+<div class="section" id="related-result-types">
+<h3>Related result types<a class="headerlink" href="#related-result-types" title="Permalink to this headline">¶</a></h3>
+<p>According to Cocoa conventions, Objective-C methods with certain names
+(“<tt class="docutils literal"><span class="pre">init</span></tt>”, “<tt class="docutils literal"><span class="pre">alloc</span></tt>”, etc.) always return objects that are an instance of
+the receiving class’s type. Such methods are said to have a “related result
+type”, meaning that a message send to one of these methods will have the same
+static type as an instance of the receiver class. For example, given the
+following classes:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="k">@interface</span> <span class="nc">NSObject</span>
+<span class="k">+</span> <span class="p">(</span><span class="kt">id</span><span class="p">)</span><span class="nf">alloc</span><span class="p">;</span>
+<span class="k">-</span> <span class="p">(</span><span class="kt">id</span><span class="p">)</span><span class="nf">init</span><span class="p">;</span>
+<span class="k">@end</span>
+
+<span class="k">@interface</span> <span class="nc">NSArray</span> : <span class="nc">NSObject</span>
+<span class="k">@end</span>
+</pre></div>
+</div>
+<p>and this common initialization pattern</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="n">NSArray</span> <span class="o">*</span><span class="n">array</span> <span class="o">=</span> <span class="p">[[</span><span class="n">NSArray</span> <span class="n">alloc</span><span class="p">]</span> <span class="n">init</span><span class="p">];</span>
+</pre></div>
+</div>
+<p>the type of the expression <tt class="docutils literal"><span class="pre">[NSArray</span> <span class="pre">alloc]</span></tt> is <tt class="docutils literal"><span class="pre">NSArray*</span></tt> because
+<tt class="docutils literal"><span class="pre">alloc</span></tt> implicitly has a related result type. Similarly, the type of the
+expression <tt class="docutils literal"><span class="pre">[[NSArray</span> <span class="pre">alloc]</span> <span class="pre">init]</span></tt> is <tt class="docutils literal"><span class="pre">NSArray*</span></tt>, since <tt class="docutils literal"><span class="pre">init</span></tt> has a
+related result type and its receiver is known to have the type <tt class="docutils literal"><span class="pre">NSArray</span> <span class="pre">*</span></tt>.
+If neither <tt class="docutils literal"><span class="pre">alloc</span></tt> nor <tt class="docutils literal"><span class="pre">init</span></tt> had a related result type, the expressions
+would have had type <tt class="docutils literal"><span class="pre">id</span></tt>, as declared in the method signature.</p>
+<p>A method with a related result type can be declared by using the type
+<tt class="docutils literal"><span class="pre">instancetype</span></tt> as its result type. <tt class="docutils literal"><span class="pre">instancetype</span></tt> is a contextual keyword
+that is only permitted in the result type of an Objective-C method, e.g.</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="k">@interface</span> <span class="nc">A</span>
+<span class="k">+</span> <span class="p">(</span><span class="n">instancetype</span><span class="p">)</span><span class="nf">constructAnA</span><span class="p">;</span>
+<span class="k">@end</span>
+</pre></div>
+</div>
+<p>The related result type can also be inferred for some methods. To determine
+whether a method has an inferred related result type, the first word in the
+camel-case selector (e.g., “<tt class="docutils literal"><span class="pre">init</span></tt>” in “<tt class="docutils literal"><span class="pre">initWithObjects</span></tt>”) is considered,
+and the method will have a related result type if its return type is compatible
+with the type of its class and if:</p>
+<ul class="simple">
+<li>the first word is “<tt class="docutils literal"><span class="pre">alloc</span></tt>” or “<tt class="docutils literal"><span class="pre">new</span></tt>”, and the method is a class method,
+or</li>
+<li>the first word is “<tt class="docutils literal"><span class="pre">autorelease</span></tt>”, “<tt class="docutils literal"><span class="pre">init</span></tt>”, “<tt class="docutils literal"><span class="pre">retain</span></tt>”, or “<tt class="docutils literal"><span class="pre">self</span></tt>”,
+and the method is an instance method.</li>
+</ul>
+<p>If a method with a related result type is overridden by a subclass method, the
+subclass method must also return a type that is compatible with the subclass
+type. For example:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="k">@interface</span> <span class="nc">NSString</span> : <span class="nc">NSObject</span>
+<span class="k">-</span> <span class="p">(</span><span class="n">NSUnrelated</span> <span class="o">*</span><span class="p">)</span><span class="nf">init</span><span class="p">;</span> <span class="c1">// incorrect usage: NSUnrelated is not NSString or a superclass of NSString</span>
+<span class="k">@end</span>
+</pre></div>
+</div>
+<p>Related result types only affect the type of a message send or property access
+via the given method. In all other respects, a method with a related result
+type is treated the same way as method that returns <tt class="docutils literal"><span class="pre">id</span></tt>.</p>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(objc_instancetype)</span></tt> to determine whether the
+<tt class="docutils literal"><span class="pre">instancetype</span></tt> contextual keyword is available.</p>
+</div>
+<div class="section" id="automatic-reference-counting">
+<h3>Automatic reference counting<a class="headerlink" href="#automatic-reference-counting" title="Permalink to this headline">¶</a></h3>
+<p>Clang provides support for <a class="reference internal" href="AutomaticReferenceCounting.html"><em>automated reference counting</em></a> in Objective-C, which eliminates the need
+for manual <tt class="docutils literal"><span class="pre">retain</span></tt>/<tt class="docutils literal"><span class="pre">release</span></tt>/<tt class="docutils literal"><span class="pre">autorelease</span></tt> message sends. There are two
+feature macros associated with automatic reference counting:
+<tt class="docutils literal"><span class="pre">__has_feature(objc_arc)</span></tt> indicates the availability of automated reference
+counting in general, while <tt class="docutils literal"><span class="pre">__has_feature(objc_arc_weak)</span></tt> indicates that
+automated reference counting also includes support for <tt class="docutils literal"><span class="pre">__weak</span></tt> pointers to
+Objective-C objects.</p>
+</div>
+<div class="section" id="enumerations-with-a-fixed-underlying-type">
+<span id="objc-fixed-enum"></span><h3>Enumerations with a fixed underlying type<a class="headerlink" href="#enumerations-with-a-fixed-underlying-type" title="Permalink to this headline">¶</a></h3>
+<p>Clang provides support for C++11 enumerations with a fixed underlying type
+within Objective-C. For example, one can write an enumeration type as:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">typedef</span> <span class="k">enum</span> <span class="o">:</span> <span class="kt">unsigned</span> <span class="kt">char</span> <span class="p">{</span> <span class="n">Red</span><span class="p">,</span> <span class="n">Green</span><span class="p">,</span> <span class="n">Blue</span> <span class="p">}</span> <span class="n">Color</span><span class="p">;</span>
+</pre></div>
+</div>
+<p>This specifies that the underlying type, which is used to store the enumeration
+value, is <tt class="docutils literal"><span class="pre">unsigned</span> <span class="pre">char</span></tt>.</p>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(objc_fixed_enum)</span></tt> to determine whether support for fixed
+underlying types is available in Objective-C.</p>
+</div>
+<div class="section" id="interoperability-with-c-11-lambdas">
+<h3>Interoperability with C++11 lambdas<a class="headerlink" href="#interoperability-with-c-11-lambdas" title="Permalink to this headline">¶</a></h3>
+<p>Clang provides interoperability between C++11 lambdas and blocks-based APIs, by
+permitting a lambda to be implicitly converted to a block pointer with the
+corresponding signature. For example, consider an API such as <tt class="docutils literal"><span class="pre">NSArray</span></tt>‘s
+array-sorting method:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="k">-</span> <span class="p">(</span><span class="n">NSArray</span> <span class="o">*</span><span class="p">)</span><span class="nf">sortedArrayUsingComparator:</span><span class="p">(</span><span class="n">NSComparator</span><span class="p">)</span><span class="nv">cmptr</span><span class="p">;</span>
+</pre></div>
+</div>
+<p><tt class="docutils literal"><span class="pre">NSComparator</span></tt> is simply a typedef for the block pointer <tt class="docutils literal"><span class="pre">NSComparisonResult</span>
+<span class="pre">(^)(id,</span> <span class="pre">id)</span></tt>, and parameters of this type are generally provided with block
+literals as arguments. However, one can also use a C++11 lambda so long as it
+provides the same signature (in this case, accepting two parameters of type
+<tt class="docutils literal"><span class="pre">id</span></tt> and returning an <tt class="docutils literal"><span class="pre">NSComparisonResult</span></tt>):</p>
+<div class="highlight-objc"><div class="highlight"><pre>NSArray *array = @[@"string 1", @"string 21", @"string 12", @"String 11",
+ @"String 02"];
+const NSStringCompareOptions comparisonOptions
+ = NSCaseInsensitiveSearch | NSNumericSearch |
+ NSWidthInsensitiveSearch | NSForcedOrderingSearch;
+NSLocale *currentLocale = [NSLocale currentLocale];
+NSArray *sorted
+ = [array sortedArrayUsingComparator:[=](id s1, id s2) -> NSComparisonResult {
+ NSRange string1Range = NSMakeRange(0, [s1 length]);
+ return [s1 compare:s2 options:comparisonOptions
+ range:string1Range locale:currentLocale];
+ }];
+NSLog(@"sorted: %@", sorted);
+</pre></div>
+</div>
+<p>This code relies on an implicit conversion from the type of the lambda
+expression (an unnamed, local class type called the <em>closure type</em>) to the
+corresponding block pointer type. The conversion itself is expressed by a
+conversion operator in that closure type that produces a block pointer with the
+same signature as the lambda itself, e.g.,</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="n">operator</span> <span class="n">NSComparisonResult</span> <span class="p">(</span><span class="o">^</span><span class="p">)(</span><span class="kt">id</span><span class="p">,</span> <span class="kt">id</span><span class="p">)()</span> <span class="k">const</span><span class="p">;</span>
+</pre></div>
+</div>
+<p>This conversion function returns a new block that simply forwards the two
+parameters to the lambda object (which it captures by copy), then returns the
+result. The returned block is first copied (with <tt class="docutils literal"><span class="pre">Block_copy</span></tt>) and then
+autoreleased. As an optimization, if a lambda expression is immediately
+converted to a block pointer (as in the first example, above), then the block
+is not copied and autoreleased: rather, it is given the same lifetime as a
+block literal written at that point in the program, which avoids the overhead
+of copying a block to the heap in the common case.</p>
+<p>The conversion from a lambda to a block pointer is only available in
+Objective-C++, and not in C++ with blocks, due to its use of Objective-C memory
+management (autorelease).</p>
+</div>
+<div class="section" id="object-literals-and-subscripting">
+<h3>Object Literals and Subscripting<a class="headerlink" href="#object-literals-and-subscripting" title="Permalink to this headline">¶</a></h3>
+<p>Clang provides support for <a class="reference internal" href="ObjectiveCLiterals.html"><em>Object Literals and Subscripting</em></a> in Objective-C, which simplifies common Objective-C
+programming patterns, makes programs more concise, and improves the safety of
+container creation. There are several feature macros associated with object
+literals and subscripting: <tt class="docutils literal"><span class="pre">__has_feature(objc_array_literals)</span></tt> tests the
+availability of array literals; <tt class="docutils literal"><span class="pre">__has_feature(objc_dictionary_literals)</span></tt>
+tests the availability of dictionary literals;
+<tt class="docutils literal"><span class="pre">__has_feature(objc_subscripting)</span></tt> tests the availability of object
+subscripting.</p>
+</div>
+<div class="section" id="objective-c-autosynthesis-of-properties">
+<h3>Objective-C Autosynthesis of Properties<a class="headerlink" href="#objective-c-autosynthesis-of-properties" title="Permalink to this headline">¶</a></h3>
+<p>Clang provides support for autosynthesis of declared properties. Using this
+feature, clang provides default synthesis of those properties not declared
+@dynamic and not having user provided backing getter and setter methods.
+<tt class="docutils literal"><span class="pre">__has_feature(objc_default_synthesize_properties)</span></tt> checks for availability
+of this feature in version of clang being used.</p>
+</div>
+<div class="section" id="objective-c-retaining-behavior-attributes">
+<span id="langext-objc-retain-release"></span><h3>Objective-C retaining behavior attributes<a class="headerlink" href="#objective-c-retaining-behavior-attributes" title="Permalink to this headline">¶</a></h3>
+<p>In Objective-C, functions and methods are generally assumed to follow the
+<a class="reference external" href="http://developer.apple.com/library/mac/#documentation/Cocoa/Conceptual/MemoryMgmt/Articles/mmRules.html">Cocoa Memory Management</a>
+conventions for ownership of object arguments and
+return values. However, there are exceptions, and so Clang provides attributes
+to allow these exceptions to be documented. This are used by ARC and the
+<a class="reference external" href="http://clang-analyzer.llvm.org">static analyzer</a> Some exceptions may be
+better described using the <tt class="docutils literal"><span class="pre">objc_method_family</span></tt> attribute instead.</p>
+<p><strong>Usage</strong>: The <tt class="docutils literal"><span class="pre">ns_returns_retained</span></tt>, <tt class="docutils literal"><span class="pre">ns_returns_not_retained</span></tt>,
+<tt class="docutils literal"><span class="pre">ns_returns_autoreleased</span></tt>, <tt class="docutils literal"><span class="pre">cf_returns_retained</span></tt>, and
+<tt class="docutils literal"><span class="pre">cf_returns_not_retained</span></tt> attributes can be placed on methods and functions
+that return Objective-C or CoreFoundation objects. They are commonly placed at
+the end of a function prototype or method declaration:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="kt">id</span> <span class="nf">foo</span><span class="p">()</span> <span class="n">__attribute__</span><span class="p">((</span><span class="n">ns_returns_retained</span><span class="p">));</span>
+
+<span class="k">-</span> <span class="p">(</span><span class="n">NSString</span> <span class="o">*</span><span class="p">)</span><span class="nf">bar:</span><span class="p">(</span><span class="kt">int</span><span class="p">)</span><span class="nv">x</span> <span class="n">__attribute__</span><span class="p">((</span><span class="n">ns_returns_retained</span><span class="p">));</span>
+</pre></div>
+</div>
+<p>The <tt class="docutils literal"><span class="pre">*_returns_retained</span></tt> attributes specify that the returned object has a +1
+retain count. The <tt class="docutils literal"><span class="pre">*_returns_not_retained</span></tt> attributes specify that the return
+object has a +0 retain count, even if the normal convention for its selector
+would be +1. <tt class="docutils literal"><span class="pre">ns_returns_autoreleased</span></tt> specifies that the returned object is
++0, but is guaranteed to live at least as long as the next flush of an
+autorelease pool.</p>
+<p><strong>Usage</strong>: The <tt class="docutils literal"><span class="pre">ns_consumed</span></tt> and <tt class="docutils literal"><span class="pre">cf_consumed</span></tt> attributes can be placed on
+an parameter declaration; they specify that the argument is expected to have a
++1 retain count, which will be balanced in some way by the function or method.
+The <tt class="docutils literal"><span class="pre">ns_consumes_self</span></tt> attribute can only be placed on an Objective-C
+method; it specifies that the method expects its <tt class="docutils literal"><span class="pre">self</span></tt> parameter to have a
++1 retain count, which it will balance in some way.</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="kt">void</span> <span class="nf">foo</span><span class="p">(</span><span class="n">__attribute__</span><span class="p">((</span><span class="n">ns_consumed</span><span class="p">))</span> <span class="n">NSString</span> <span class="o">*</span><span class="n">string</span><span class="p">);</span>
+
+<span class="k">-</span> <span class="p">(</span><span class="kt">void</span><span class="p">)</span> <span class="nf">bar</span> <span class="n">__attribute__</span><span class="p">((</span><span class="n">ns_consumes_self</span><span class="p">));</span>
+<span class="k">-</span> <span class="p">(</span><span class="kt">void</span><span class="p">)</span> <span class="nf">baz:</span><span class="p">(</span><span class="kt">id</span><span class="p">)</span> <span class="n">__attribute__</span><span class="p">((</span><span class="n">ns_consumed</span><span class="p">))</span> <span class="n">x</span><span class="p">;</span>
+</pre></div>
+</div>
+<p>Further examples of these attributes are available in the static analyzer’s <a class="reference external" href="http://clang-analyzer.llvm.org/annotations.html#cocoa_mem">list of annotations for analysis</a>.</p>
+<p>Query for these features with <tt class="docutils literal"><span class="pre">__has_attribute(ns_consumed)</span></tt>,
+<tt class="docutils literal"><span class="pre">__has_attribute(ns_returns_retained)</span></tt>, etc.</p>
+</div>
+<div class="section" id="objective-c-abi-protocol-qualifier-mangling-of-parameters">
+<h3>Objective-C++ ABI: protocol-qualifier mangling of parameters<a class="headerlink" href="#objective-c-abi-protocol-qualifier-mangling-of-parameters" title="Permalink to this headline">¶</a></h3>
+<p>Starting with LLVM 3.4, Clang produces a new mangling for parameters whose
+type is a qualified-<tt class="docutils literal"><span class="pre">id</span></tt> (e.g., <tt class="docutils literal"><span class="pre">id<Foo></span></tt>). This mangling allows such
+parameters to be differentiated from those with the regular unqualified <tt class="docutils literal"><span class="pre">id</span></tt>
+type.</p>
+<p>This was a non-backward compatible mangling change to the ABI. This change
+allows proper overloading, and also prevents mangling conflicts with template
+parameters of protocol-qualified type.</p>
+<p>Query the presence of this new mangling with
+<tt class="docutils literal"><span class="pre">__has_feature(objc_protocol_qualifier_mangling)</span></tt>.</p>
+</div>
+</div>
+<div class="section" id="initializer-lists-for-complex-numbers-in-c">
+<span id="langext-overloading"></span><h2><a class="toc-backref" href="#id13">Initializer lists for complex numbers in C</a><a class="headerlink" href="#initializer-lists-for-complex-numbers-in-c" title="Permalink to this headline">¶</a></h2>
+<p>clang supports an extension which allows the following in C:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#include <math.h></span>
+<span class="cp">#include <complex.h></span>
+<span class="n">complex</span> <span class="kt">float</span> <span class="n">x</span> <span class="o">=</span> <span class="p">{</span> <span class="mf">1.0f</span><span class="p">,</span> <span class="n">INFINITY</span> <span class="p">};</span> <span class="c1">// Init to (1, Inf)</span>
+</pre></div>
+</div>
+<p>This construct is useful because there is no way to separately initialize the
+real and imaginary parts of a complex variable in standard C, given that clang
+does not support <tt class="docutils literal"><span class="pre">_Imaginary</span></tt>. (Clang also supports the <tt class="docutils literal"><span class="pre">__real__</span></tt> and
+<tt class="docutils literal"><span class="pre">__imag__</span></tt> extensions from gcc, which help in some cases, but are not usable
+in static initializers.)</p>
+<p>Note that this extension does not allow eliding the braces; the meaning of the
+following two lines is different:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">complex</span> <span class="kt">float</span> <span class="n">x</span><span class="p">[]</span> <span class="o">=</span> <span class="p">{</span> <span class="p">{</span> <span class="mf">1.0f</span><span class="p">,</span> <span class="mf">1.0f</span> <span class="p">}</span> <span class="p">};</span> <span class="c1">// [0] = (1, 1)</span>
+<span class="n">complex</span> <span class="kt">float</span> <span class="n">x</span><span class="p">[]</span> <span class="o">=</span> <span class="p">{</span> <span class="mf">1.0f</span><span class="p">,</span> <span class="mf">1.0f</span> <span class="p">};</span> <span class="c1">// [0] = (1, 0), [1] = (1, 0)</span>
+</pre></div>
+</div>
+<p>This extension also works in C++ mode, as far as that goes, but does not apply
+to the C++ <tt class="docutils literal"><span class="pre">std::complex</span></tt>. (In C++11, list initialization allows the same
+syntax to be used with <tt class="docutils literal"><span class="pre">std::complex</span></tt> with the same meaning.)</p>
+</div>
+<div class="section" id="builtin-functions">
+<h2><a class="toc-backref" href="#id14">Builtin Functions</a><a class="headerlink" href="#builtin-functions" title="Permalink to this headline">¶</a></h2>
+<p>Clang supports a number of builtin library functions with the same syntax as
+GCC, including things like <tt class="docutils literal"><span class="pre">__builtin_nan</span></tt>, <tt class="docutils literal"><span class="pre">__builtin_constant_p</span></tt>,
+<tt class="docutils literal"><span class="pre">__builtin_choose_expr</span></tt>, <tt class="docutils literal"><span class="pre">__builtin_types_compatible_p</span></tt>,
+<tt class="docutils literal"><span class="pre">__builtin_assume_aligned</span></tt>, <tt class="docutils literal"><span class="pre">__sync_fetch_and_add</span></tt>, etc. In addition to
+the GCC builtins, Clang supports a number of builtins that GCC does not, which
+are listed here.</p>
+<p>Please note that Clang does not and will not support all of the GCC builtins
+for vector operations. Instead of using builtins, you should use the functions
+defined in target-specific header files like <tt class="docutils literal"><span class="pre"><xmmintrin.h></span></tt>, which define
+portable wrappers for these. Many of the Clang versions of these functions are
+implemented directly in terms of <a class="reference internal" href="#langext-vectors"><em>extended vector support</em></a> instead of builtins, in order to reduce the number of
+builtins that we need to implement.</p>
+<div class="section" id="builtin-assume">
+<h3><tt class="docutils literal"><span class="pre">__builtin_assume</span></tt><a class="headerlink" href="#builtin-assume" title="Permalink to this headline">¶</a></h3>
+<p><tt class="docutils literal"><span class="pre">__builtin_assume</span></tt> is used to provide the optimizer with a boolean
+invariant that is defined to be true.</p>
+<p><strong>Syntax</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">__builtin_assume</span><span class="p">(</span><span class="kt">bool</span><span class="p">)</span>
+</pre></div>
+</div>
+<p><strong>Example of Use</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">int</span> <span class="nf">foo</span><span class="p">(</span><span class="kt">int</span> <span class="n">x</span><span class="p">)</span> <span class="p">{</span>
+ <span class="n">__builtin_assume</span><span class="p">(</span><span class="n">x</span> <span class="o">!=</span> <span class="mi">0</span><span class="p">);</span>
+
+ <span class="c1">// The optimizer may short-circuit this check using the invariant.</span>
+ <span class="k">if</span> <span class="p">(</span><span class="n">x</span> <span class="o">==</span> <span class="mi">0</span><span class="p">)</span>
+ <span class="k">return</span> <span class="n">do_something</span><span class="p">();</span>
+
+ <span class="k">return</span> <span class="n">do_something_else</span><span class="p">();</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p><strong>Description</strong>:</p>
+<p>The boolean argument to this function is defined to be true. The optimizer may
+analyze the form of the expression provided as the argument and deduce from
+that information used to optimize the program. If the condition is violated
+during execution, the behavior is undefined. The argument itself is never
+evaluated, so any side effects of the expression will be discarded.</p>
+<p>Query for this feature with <tt class="docutils literal"><span class="pre">__has_builtin(__builtin_assume)</span></tt>.</p>
+</div>
+<div class="section" id="builtin-readcyclecounter">
+<h3><tt class="docutils literal"><span class="pre">__builtin_readcyclecounter</span></tt><a class="headerlink" href="#builtin-readcyclecounter" title="Permalink to this headline">¶</a></h3>
+<p><tt class="docutils literal"><span class="pre">__builtin_readcyclecounter</span></tt> is used to access the cycle counter register (or
+a similar low-latency, high-accuracy clock) on those targets that support it.</p>
+<p><strong>Syntax</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">__builtin_readcyclecounter</span><span class="p">()</span>
+</pre></div>
+</div>
+<p><strong>Example of Use</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">t0</span> <span class="o">=</span> <span class="n">__builtin_readcyclecounter</span><span class="p">();</span>
+<span class="n">do_something</span><span class="p">();</span>
+<span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">t1</span> <span class="o">=</span> <span class="n">__builtin_readcyclecounter</span><span class="p">();</span>
+<span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">cycles_to_do_something</span> <span class="o">=</span> <span class="n">t1</span> <span class="o">-</span> <span class="n">t0</span><span class="p">;</span> <span class="c1">// assuming no overflow</span>
+</pre></div>
+</div>
+<p><strong>Description</strong>:</p>
+<p>The <tt class="docutils literal"><span class="pre">__builtin_readcyclecounter()</span></tt> builtin returns the cycle counter value,
+which may be either global or process/thread-specific depending on the target.
+As the backing counters often overflow quickly (on the order of seconds) this
+should only be used for timing small intervals. When not supported by the
+target, the return value is always zero. This builtin takes no arguments and
+produces an unsigned long long result.</p>
+<p>Query for this feature with <tt class="docutils literal"><span class="pre">__has_builtin(__builtin_readcyclecounter)</span></tt>. Note
+that even if present, its use may depend on run-time privilege or other OS
+controlled state.</p>
+</div>
+<div class="section" id="builtin-shufflevector">
+<span id="langext-builtin-shufflevector"></span><h3><tt class="docutils literal"><span class="pre">__builtin_shufflevector</span></tt><a class="headerlink" href="#builtin-shufflevector" title="Permalink to this headline">¶</a></h3>
+<p><tt class="docutils literal"><span class="pre">__builtin_shufflevector</span></tt> is used to express generic vector
+permutation/shuffle/swizzle operations. This builtin is also very important
+for the implementation of various target-specific header files like
+<tt class="docutils literal"><span class="pre"><xmmintrin.h></span></tt>.</p>
+<p><strong>Syntax</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">__builtin_shufflevector</span><span class="p">(</span><span class="n">vec1</span><span class="p">,</span> <span class="n">vec2</span><span class="p">,</span> <span class="n">index1</span><span class="p">,</span> <span class="n">index2</span><span class="p">,</span> <span class="p">...)</span>
+</pre></div>
+</div>
+<p><strong>Examples</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="c1">// identity operation - return 4-element vector v1.</span>
+<span class="n">__builtin_shufflevector</span><span class="p">(</span><span class="n">v1</span><span class="p">,</span> <span class="n">v1</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">)</span>
+
+<span class="c1">// "Splat" element 0 of V1 into a 4-element result.</span>
+<span class="n">__builtin_shufflevector</span><span class="p">(</span><span class="n">V1</span><span class="p">,</span> <span class="n">V1</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">)</span>
+
+<span class="c1">// Reverse 4-element vector V1.</span>
+<span class="n">__builtin_shufflevector</span><span class="p">(</span><span class="n">V1</span><span class="p">,</span> <span class="n">V1</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">0</span><span class="p">)</span>
+
+<span class="c1">// Concatenate every other element of 4-element vectors V1 and V2.</span>
+<span class="n">__builtin_shufflevector</span><span class="p">(</span><span class="n">V1</span><span class="p">,</span> <span class="n">V2</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">4</span><span class="p">,</span> <span class="mi">6</span><span class="p">)</span>
+
+<span class="c1">// Concatenate every other element of 8-element vectors V1 and V2.</span>
+<span class="n">__builtin_shufflevector</span><span class="p">(</span><span class="n">V1</span><span class="p">,</span> <span class="n">V2</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">4</span><span class="p">,</span> <span class="mi">6</span><span class="p">,</span> <span class="mi">8</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mi">12</span><span class="p">,</span> <span class="mi">14</span><span class="p">)</span>
+
+<span class="c1">// Shuffle v1 with some elements being undefined</span>
+<span class="n">__builtin_shufflevector</span><span class="p">(</span><span class="n">v1</span><span class="p">,</span> <span class="n">v1</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="o">-</span><span class="mi">1</span><span class="p">)</span>
+</pre></div>
+</div>
+<p><strong>Description</strong>:</p>
+<p>The first two arguments to <tt class="docutils literal"><span class="pre">__builtin_shufflevector</span></tt> are vectors that have
+the same element type. The remaining arguments are a list of integers that
+specify the elements indices of the first two vectors that should be extracted
+and returned in a new vector. These element indices are numbered sequentially
+starting with the first vector, continuing into the second vector. Thus, if
+<tt class="docutils literal"><span class="pre">vec1</span></tt> is a 4-element vector, index 5 would refer to the second element of
+<tt class="docutils literal"><span class="pre">vec2</span></tt>. An index of -1 can be used to indicate that the corresponding element
+in the returned vector is a don’t care and can be optimized by the backend.</p>
+<p>The result of <tt class="docutils literal"><span class="pre">__builtin_shufflevector</span></tt> is a vector with the same element
+type as <tt class="docutils literal"><span class="pre">vec1</span></tt>/<tt class="docutils literal"><span class="pre">vec2</span></tt> but that has an element count equal to the number of
+indices specified.</p>
+<p>Query for this feature with <tt class="docutils literal"><span class="pre">__has_builtin(__builtin_shufflevector)</span></tt>.</p>
+</div>
+<div class="section" id="builtin-convertvector">
+<span id="langext-builtin-convertvector"></span><h3><tt class="docutils literal"><span class="pre">__builtin_convertvector</span></tt><a class="headerlink" href="#builtin-convertvector" title="Permalink to this headline">¶</a></h3>
+<p><tt class="docutils literal"><span class="pre">__builtin_convertvector</span></tt> is used to express generic vector
+type-conversion operations. The input vector and the output vector
+type must have the same number of elements.</p>
+<p><strong>Syntax</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">__builtin_convertvector</span><span class="p">(</span><span class="n">src_vec</span><span class="p">,</span> <span class="n">dst_vec_type</span><span class="p">)</span>
+</pre></div>
+</div>
+<p><strong>Examples</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">typedef</span> <span class="kt">double</span> <span class="n">vector4double</span> <span class="nf">__attribute__</span><span class="p">((</span><span class="n">__vector_size__</span><span class="p">(</span><span class="mi">32</span><span class="p">)));</span>
+<span class="k">typedef</span> <span class="kt">float</span> <span class="n">vector4float</span> <span class="nf">__attribute__</span><span class="p">((</span><span class="n">__vector_size__</span><span class="p">(</span><span class="mi">16</span><span class="p">)));</span>
+<span class="k">typedef</span> <span class="kt">short</span> <span class="n">vector4short</span> <span class="nf">__attribute__</span><span class="p">((</span><span class="n">__vector_size__</span><span class="p">(</span><span class="mi">8</span><span class="p">)));</span>
+<span class="n">vector4float</span> <span class="n">vf</span><span class="p">;</span> <span class="n">vector4short</span> <span class="n">vs</span><span class="p">;</span>
+
+<span class="c1">// convert from a vector of 4 floats to a vector of 4 doubles.</span>
+<span class="n">__builtin_convertvector</span><span class="p">(</span><span class="n">vf</span><span class="p">,</span> <span class="n">vector4double</span><span class="p">)</span>
+<span class="c1">// equivalent to:</span>
+<span class="p">(</span><span class="n">vector4double</span><span class="p">)</span> <span class="p">{</span> <span class="p">(</span><span class="kt">double</span><span class="p">)</span> <span class="n">vf</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="p">(</span><span class="kt">double</span><span class="p">)</span> <span class="n">vf</span><span class="p">[</span><span class="mi">1</span><span class="p">],</span> <span class="p">(</span><span class="kt">double</span><span class="p">)</span> <span class="n">vf</span><span class="p">[</span><span class="mi">2</span><span class="p">],</span> <span class="p">(</span><span class="kt">double</span><span class="p">)</span> <span class="n">vf</span><span class="p">[</span><span class="mi">3</span><span class="p">]</span> <span class="p">}</span>
+
+<span class="c1">// convert from a vector of 4 shorts to a vector of 4 floats.</span>
+<span class="n">__builtin_convertvector</span><span class="p">(</span><span class="n">vs</span><span class="p">,</span> <span class="n">vector4float</span><span class="p">)</span>
+<span class="c1">// equivalent to:</span>
+<span class="p">(</span><span class="n">vector4float</span><span class="p">)</span> <span class="p">{</span> <span class="p">(</span><span class="kt">float</span><span class="p">)</span> <span class="n">vs</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="p">(</span><span class="kt">float</span><span class="p">)</span> <span class="n">vs</span><span class="p">[</span><span class="mi">1</span><span class="p">],</span> <span class="p">(</span><span class="kt">float</span><span class="p">)</span> <span class="n">vs</span><span class="p">[</span><span class="mi">2</span><span class="p">],</span> <span class="p">(</span><span class="kt">float</span><span class="p">)</span> <span class="n">vs</span><span class="p">[</span><span class="mi">3</span><span class="p">]</span> <span class="p">}</span>
+</pre></div>
+</div>
+<p><strong>Description</strong>:</p>
+<p>The first argument to <tt class="docutils literal"><span class="pre">__builtin_convertvector</span></tt> is a vector, and the second
+argument is a vector type with the same number of elements as the first
+argument.</p>
+<p>The result of <tt class="docutils literal"><span class="pre">__builtin_convertvector</span></tt> is a vector with the same element
+type as the second argument, with a value defined in terms of the action of a
+C-style cast applied to each element of the first argument.</p>
+<p>Query for this feature with <tt class="docutils literal"><span class="pre">__has_builtin(__builtin_convertvector)</span></tt>.</p>
+</div>
+<div class="section" id="builtin-unreachable">
+<h3><tt class="docutils literal"><span class="pre">__builtin_unreachable</span></tt><a class="headerlink" href="#builtin-unreachable" title="Permalink to this headline">¶</a></h3>
+<p><tt class="docutils literal"><span class="pre">__builtin_unreachable</span></tt> is used to indicate that a specific point in the
+program cannot be reached, even if the compiler might otherwise think it can.
+This is useful to improve optimization and eliminates certain warnings. For
+example, without the <tt class="docutils literal"><span class="pre">__builtin_unreachable</span></tt> in the example below, the
+compiler assumes that the inline asm can fall through and prints a “function
+declared ‘<tt class="docutils literal"><span class="pre">noreturn</span></tt>‘ should not return” warning.</p>
+<p><strong>Syntax</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">__builtin_unreachable</span><span class="p">()</span>
+</pre></div>
+</div>
+<p><strong>Example of use</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">void</span> <span class="nf">myabort</span><span class="p">(</span><span class="kt">void</span><span class="p">)</span> <span class="n">__attribute__</span><span class="p">((</span><span class="n">noreturn</span><span class="p">));</span>
+<span class="kt">void</span> <span class="nf">myabort</span><span class="p">(</span><span class="kt">void</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">asm</span><span class="p">(</span><span class="s">"int3"</span><span class="p">);</span>
+ <span class="n">__builtin_unreachable</span><span class="p">();</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p><strong>Description</strong>:</p>
+<p>The <tt class="docutils literal"><span class="pre">__builtin_unreachable()</span></tt> builtin has completely undefined behavior.
+Since it has undefined behavior, it is a statement that it is never reached and
+the optimizer can take advantage of this to produce better code. This builtin
+takes no arguments and produces a void result.</p>
+<p>Query for this feature with <tt class="docutils literal"><span class="pre">__has_builtin(__builtin_unreachable)</span></tt>.</p>
+</div>
+<div class="section" id="builtin-unpredictable">
+<h3><tt class="docutils literal"><span class="pre">__builtin_unpredictable</span></tt><a class="headerlink" href="#builtin-unpredictable" title="Permalink to this headline">¶</a></h3>
+<p><tt class="docutils literal"><span class="pre">__builtin_unpredictable</span></tt> is used to indicate that a branch condition is
+unpredictable by hardware mechanisms such as branch prediction logic.</p>
+<p><strong>Syntax</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">__builtin_unpredictable</span><span class="p">(</span><span class="kt">long</span> <span class="kt">long</span><span class="p">)</span>
+</pre></div>
+</div>
+<p><strong>Example of use</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">if</span> <span class="p">(</span><span class="n">__builtin_unpredictable</span><span class="p">(</span><span class="n">x</span> <span class="o">></span> <span class="mi">0</span><span class="p">))</span> <span class="p">{</span>
+ <span class="n">foo</span><span class="p">();</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p><strong>Description</strong>:</p>
+<p>The <tt class="docutils literal"><span class="pre">__builtin_unpredictable()</span></tt> builtin is expected to be used with control
+flow conditions such as in <tt class="docutils literal"><span class="pre">if</span></tt> and <tt class="docutils literal"><span class="pre">switch</span></tt> statements.</p>
+<p>Query for this feature with <tt class="docutils literal"><span class="pre">__has_builtin(__builtin_unpredictable)</span></tt>.</p>
+</div>
+<div class="section" id="sync-swap">
+<h3><tt class="docutils literal"><span class="pre">__sync_swap</span></tt><a class="headerlink" href="#sync-swap" title="Permalink to this headline">¶</a></h3>
+<p><tt class="docutils literal"><span class="pre">__sync_swap</span></tt> is used to atomically swap integers or pointers in memory.</p>
+<p><strong>Syntax</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">type</span> <span class="n">__sync_swap</span><span class="p">(</span><span class="n">type</span> <span class="o">*</span><span class="n">ptr</span><span class="p">,</span> <span class="n">type</span> <span class="n">value</span><span class="p">,</span> <span class="p">...)</span>
+</pre></div>
+</div>
+<p><strong>Example of Use</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">int</span> <span class="n">old_value</span> <span class="o">=</span> <span class="n">__sync_swap</span><span class="p">(</span><span class="o">&</span><span class="n">value</span><span class="p">,</span> <span class="n">new_value</span><span class="p">);</span>
+</pre></div>
+</div>
+<p><strong>Description</strong>:</p>
+<p>The <tt class="docutils literal"><span class="pre">__sync_swap()</span></tt> builtin extends the existing <tt class="docutils literal"><span class="pre">__sync_*()</span></tt> family of
+atomic intrinsics to allow code to atomically swap the current value with the
+new value. More importantly, it helps developers write more efficient and
+correct code by avoiding expensive loops around
+<tt class="docutils literal"><span class="pre">__sync_bool_compare_and_swap()</span></tt> or relying on the platform specific
+implementation details of <tt class="docutils literal"><span class="pre">__sync_lock_test_and_set()</span></tt>. The
+<tt class="docutils literal"><span class="pre">__sync_swap()</span></tt> builtin is a full barrier.</p>
+</div>
+<div class="section" id="builtin-addressof">
+<h3><tt class="docutils literal"><span class="pre">__builtin_addressof</span></tt><a class="headerlink" href="#builtin-addressof" title="Permalink to this headline">¶</a></h3>
+<p><tt class="docutils literal"><span class="pre">__builtin_addressof</span></tt> performs the functionality of the built-in <tt class="docutils literal"><span class="pre">&</span></tt>
+operator, ignoring any <tt class="docutils literal"><span class="pre">operator&</span></tt> overload. This is useful in constant
+expressions in C++11, where there is no other way to take the address of an
+object that overloads <tt class="docutils literal"><span class="pre">operator&</span></tt>.</p>
+<p><strong>Example of use</strong>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">template</span><span class="o"><</span><span class="k">typename</span> <span class="n">T</span><span class="o">></span> <span class="n">constexpr</span> <span class="n">T</span> <span class="o">*</span><span class="n">addressof</span><span class="p">(</span><span class="n">T</span> <span class="o">&</span><span class="n">value</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">return</span> <span class="n">__builtin_addressof</span><span class="p">(</span><span class="n">value</span><span class="p">);</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="builtin-operator-new-and-builtin-operator-delete">
+<h3><tt class="docutils literal"><span class="pre">__builtin_operator_new</span></tt> and <tt class="docutils literal"><span class="pre">__builtin_operator_delete</span></tt><a class="headerlink" href="#builtin-operator-new-and-builtin-operator-delete" title="Permalink to this headline">¶</a></h3>
+<p><tt class="docutils literal"><span class="pre">__builtin_operator_new</span></tt> allocates memory just like a non-placement non-class
+<em>new-expression</em>. This is exactly like directly calling the normal
+non-placement <tt class="docutils literal"><span class="pre">::operator</span> <span class="pre">new</span></tt>, except that it allows certain optimizations
+that the C++ standard does not permit for a direct function call to
+<tt class="docutils literal"><span class="pre">::operator</span> <span class="pre">new</span></tt> (in particular, removing <tt class="docutils literal"><span class="pre">new</span></tt> / <tt class="docutils literal"><span class="pre">delete</span></tt> pairs and
+merging allocations).</p>
+<p>Likewise, <tt class="docutils literal"><span class="pre">__builtin_operator_delete</span></tt> deallocates memory just like a
+non-class <em>delete-expression</em>, and is exactly like directly calling the normal
+<tt class="docutils literal"><span class="pre">::operator</span> <span class="pre">delete</span></tt>, except that it permits optimizations. Only the unsized
+form of <tt class="docutils literal"><span class="pre">__builtin_operator_delete</span></tt> is currently available.</p>
+<p>These builtins are intended for use in the implementation of <tt class="docutils literal"><span class="pre">std::allocator</span></tt>
+and other similar allocation libraries, and are only available in C++.</p>
+</div>
+<div class="section" id="multiprecision-arithmetic-builtins">
+<h3>Multiprecision Arithmetic Builtins<a class="headerlink" href="#multiprecision-arithmetic-builtins" title="Permalink to this headline">¶</a></h3>
+<p>Clang provides a set of builtins which expose multiprecision arithmetic in a
+manner amenable to C. They all have the following form:</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="kt">unsigned</span> <span class="n">x</span> <span class="o">=</span> <span class="p">...,</span> <span class="n">y</span> <span class="o">=</span> <span class="p">...,</span> <span class="n">carryin</span> <span class="o">=</span> <span class="p">...,</span> <span class="n">carryout</span><span class="p">;</span>
+<span class="kt">unsigned</span> <span class="n">sum</span> <span class="o">=</span> <span class="n">__builtin_addc</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">,</span> <span class="n">carryin</span><span class="p">,</span> <span class="o">&</span><span class="n">carryout</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>Thus one can form a multiprecision addition chain in the following manner:</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="kt">unsigned</span> <span class="o">*</span><span class="n">x</span><span class="p">,</span> <span class="o">*</span><span class="n">y</span><span class="p">,</span> <span class="o">*</span><span class="n">z</span><span class="p">,</span> <span class="n">carryin</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">carryout</span><span class="p">;</span>
+<span class="n">z</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">=</span> <span class="n">__builtin_addc</span><span class="p">(</span><span class="n">x</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">y</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">carryin</span><span class="p">,</span> <span class="o">&</span><span class="n">carryout</span><span class="p">);</span>
+<span class="n">carryin</span> <span class="o">=</span> <span class="n">carryout</span><span class="p">;</span>
+<span class="n">z</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">=</span> <span class="n">__builtin_addc</span><span class="p">(</span><span class="n">x</span><span class="p">[</span><span class="mi">1</span><span class="p">],</span> <span class="n">y</span><span class="p">[</span><span class="mi">1</span><span class="p">],</span> <span class="n">carryin</span><span class="p">,</span> <span class="o">&</span><span class="n">carryout</span><span class="p">);</span>
+<span class="n">carryin</span> <span class="o">=</span> <span class="n">carryout</span><span class="p">;</span>
+<span class="n">z</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">__builtin_addc</span><span class="p">(</span><span class="n">x</span><span class="p">[</span><span class="mi">2</span><span class="p">],</span> <span class="n">y</span><span class="p">[</span><span class="mi">2</span><span class="p">],</span> <span class="n">carryin</span><span class="p">,</span> <span class="o">&</span><span class="n">carryout</span><span class="p">);</span>
+<span class="n">carryin</span> <span class="o">=</span> <span class="n">carryout</span><span class="p">;</span>
+<span class="n">z</span><span class="p">[</span><span class="mi">3</span><span class="p">]</span> <span class="o">=</span> <span class="n">__builtin_addc</span><span class="p">(</span><span class="n">x</span><span class="p">[</span><span class="mi">3</span><span class="p">],</span> <span class="n">y</span><span class="p">[</span><span class="mi">3</span><span class="p">],</span> <span class="n">carryin</span><span class="p">,</span> <span class="o">&</span><span class="n">carryout</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>The complete list of builtins are:</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="kt">unsigned</span> <span class="kt">char</span> <span class="nf">__builtin_addcb</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="kt">char</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">char</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">char</span> <span class="n">carryin</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">char</span> <span class="o">*</span><span class="n">carryout</span><span class="p">);</span>
+<span class="kt">unsigned</span> <span class="kt">short</span> <span class="nf">__builtin_addcs</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="kt">short</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">short</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">short</span> <span class="n">carryin</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">short</span> <span class="o">*</span><span class="n">carryout</span><span class="p">);</span>
+<span class="kt">unsigned</span> <span class="nf">__builtin_addc</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="n">carryin</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="o">*</span><span class="n">carryout</span><span class="p">);</span>
+<span class="kt">unsigned</span> <span class="kt">long</span> <span class="nf">__builtin_addcl</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">carryin</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="o">*</span><span class="n">carryout</span><span class="p">);</span>
+<span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="nf">__builtin_addcll</span><span class="p">(</span><span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">carryin</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="o">*</span><span class="n">carryout</span><span class="p">);</span>
+<span class="kt">unsigned</span> <span class="kt">char</span> <span class="nf">__builtin_subcb</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="kt">char</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">char</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">char</span> <span class="n">carryin</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">char</span> <span class="o">*</span><span class="n">carryout</span><span class="p">);</span>
+<span class="kt">unsigned</span> <span class="kt">short</span> <span class="nf">__builtin_subcs</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="kt">short</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">short</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">short</span> <span class="n">carryin</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">short</span> <span class="o">*</span><span class="n">carryout</span><span class="p">);</span>
+<span class="kt">unsigned</span> <span class="nf">__builtin_subc</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="n">carryin</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="o">*</span><span class="n">carryout</span><span class="p">);</span>
+<span class="kt">unsigned</span> <span class="kt">long</span> <span class="nf">__builtin_subcl</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">carryin</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="o">*</span><span class="n">carryout</span><span class="p">);</span>
+<span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="nf">__builtin_subcll</span><span class="p">(</span><span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">carryin</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="o">*</span><span class="n">carryout</span><span class="p">);</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="checked-arithmetic-builtins">
+<h3>Checked Arithmetic Builtins<a class="headerlink" href="#checked-arithmetic-builtins" title="Permalink to this headline">¶</a></h3>
+<p>Clang provides a set of builtins that implement checked arithmetic for security
+critical applications in a manner that is fast and easily expressable in C. As
+an example of their usage:</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="kt">errorcode_t</span> <span class="nf">security_critical_application</span><span class="p">(...)</span> <span class="p">{</span>
+ <span class="kt">unsigned</span> <span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">,</span> <span class="n">result</span><span class="p">;</span>
+ <span class="p">...</span>
+ <span class="k">if</span> <span class="p">(</span><span class="n">__builtin_mul_overflow</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">,</span> <span class="o">&</span><span class="n">result</span><span class="p">))</span>
+ <span class="k">return</span> <span class="n">kErrorCodeHackers</span><span class="p">;</span>
+ <span class="p">...</span>
+ <span class="n">use_multiply</span><span class="p">(</span><span class="n">result</span><span class="p">);</span>
+ <span class="p">...</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>Clang provides the following checked arithmetic builtins:</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="kt">bool</span> <span class="nf">__builtin_add_overflow</span> <span class="p">(</span><span class="n">type1</span> <span class="n">x</span><span class="p">,</span> <span class="n">type2</span> <span class="n">y</span><span class="p">,</span> <span class="n">type3</span> <span class="o">*</span><span class="n">sum</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_sub_overflow</span> <span class="p">(</span><span class="n">type1</span> <span class="n">x</span><span class="p">,</span> <span class="n">type2</span> <span class="n">y</span><span class="p">,</span> <span class="n">type3</span> <span class="o">*</span><span class="n">diff</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_mul_overflow</span> <span class="p">(</span><span class="n">type1</span> <span class="n">x</span><span class="p">,</span> <span class="n">type2</span> <span class="n">y</span><span class="p">,</span> <span class="n">type3</span> <span class="o">*</span><span class="n">prod</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_uadd_overflow</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="o">*</span><span class="n">sum</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_uaddl_overflow</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="o">*</span><span class="n">sum</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_uaddll_overflow</span><span class="p">(</span><span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="o">*</span><span class="n">sum</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_usub_overflow</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="o">*</span><span class="n">diff</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_usubl_overflow</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="o">*</span><span class="n">diff</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_usubll_overflow</span><span class="p">(</span><span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="o">*</span><span class="n">diff</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_umul_overflow</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="o">*</span><span class="n">prod</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_umull_overflow</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="o">*</span><span class="n">prod</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_umulll_overflow</span><span class="p">(</span><span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">unsigned</span> <span class="kt">long</span> <span class="kt">long</span> <span class="o">*</span><span class="n">prod</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_sadd_overflow</span> <span class="p">(</span><span class="kt">int</span> <span class="n">x</span><span class="p">,</span> <span class="kt">int</span> <span class="n">y</span><span class="p">,</span> <span class="kt">int</span> <span class="o">*</span><span class="n">sum</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_saddl_overflow</span> <span class="p">(</span><span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">long</span> <span class="o">*</span><span class="n">sum</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_saddll_overflow</span><span class="p">(</span><span class="kt">long</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">long</span> <span class="kt">long</span> <span class="o">*</span><span class="n">sum</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_ssub_overflow</span> <span class="p">(</span><span class="kt">int</span> <span class="n">x</span><span class="p">,</span> <span class="kt">int</span> <span class="n">y</span><span class="p">,</span> <span class="kt">int</span> <span class="o">*</span><span class="n">diff</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_ssubl_overflow</span> <span class="p">(</span><span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">long</span> <span class="o">*</span><span class="n">diff</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_ssubll_overflow</span><span class="p">(</span><span class="kt">long</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">long</span> <span class="kt">long</span> <span class="o">*</span><span class="n">diff</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_smul_overflow</span> <span class="p">(</span><span class="kt">int</span> <span class="n">x</span><span class="p">,</span> <span class="kt">int</span> <span class="n">y</span><span class="p">,</span> <span class="kt">int</span> <span class="o">*</span><span class="n">prod</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_smull_overflow</span> <span class="p">(</span><span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">long</span> <span class="o">*</span><span class="n">prod</span><span class="p">);</span>
+<span class="kt">bool</span> <span class="nf">__builtin_smulll_overflow</span><span class="p">(</span><span class="kt">long</span> <span class="kt">long</span> <span class="n">x</span><span class="p">,</span> <span class="kt">long</span> <span class="kt">long</span> <span class="n">y</span><span class="p">,</span> <span class="kt">long</span> <span class="kt">long</span> <span class="o">*</span><span class="n">prod</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>Each builtin performs the specified mathematical operation on the
+first two arguments and stores the result in the third argument. If
+possible, the result will be equal to mathematically-correct result
+and the builtin will return 0. Otherwise, the builtin will return
+1 and the result will be equal to the unique value that is equivalent
+to the mathematically-correct result modulo two raised to the <em>k</em>
+power, where <em>k</em> is the number of bits in the result type. The
+behavior of these builtins is well-defined for all argument values.</p>
+<p>The first three builtins work generically for operands of any integer type,
+including boolean types. The operands need not have the same type as each
+other, or as the result. The other builtins may implicitly promote or
+convert their operands before performing the operation.</p>
+<p>Query for this feature with <tt class="docutils literal"><span class="pre">__has_builtin(__builtin_add_overflow)</span></tt>, etc.</p>
+</div>
+<div class="section" id="c11-atomic-builtins">
+<span id="langext-c11-atomic"></span><h3>__c11_atomic builtins<a class="headerlink" href="#c11-atomic-builtins" title="Permalink to this headline">¶</a></h3>
+<p>Clang provides a set of builtins which are intended to be used to implement
+C11’s <tt class="docutils literal"><span class="pre"><stdatomic.h></span></tt> header. These builtins provide the semantics of the
+<tt class="docutils literal"><span class="pre">_explicit</span></tt> form of the corresponding C11 operation, and are named with a
+<tt class="docutils literal"><span class="pre">__c11_</span></tt> prefix. The supported operations, and the differences from
+the corresponding C11 operations, are:</p>
+<ul class="simple">
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_init</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_thread_fence</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_signal_fence</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_is_lock_free</span></tt> (The argument is the size of the
+<tt class="docutils literal"><span class="pre">_Atomic(...)</span></tt> object, instead of its address)</li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_store</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_load</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_exchange</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_compare_exchange_strong</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_compare_exchange_weak</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_fetch_add</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_fetch_sub</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_fetch_and</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_fetch_or</span></tt></li>
+<li><tt class="docutils literal"><span class="pre">__c11_atomic_fetch_xor</span></tt></li>
+</ul>
+<p>The macros <tt class="docutils literal"><span class="pre">__ATOMIC_RELAXED</span></tt>, <tt class="docutils literal"><span class="pre">__ATOMIC_CONSUME</span></tt>, <tt class="docutils literal"><span class="pre">__ATOMIC_ACQUIRE</span></tt>,
+<tt class="docutils literal"><span class="pre">__ATOMIC_RELEASE</span></tt>, <tt class="docutils literal"><span class="pre">__ATOMIC_ACQ_REL</span></tt>, and <tt class="docutils literal"><span class="pre">__ATOMIC_SEQ_CST</span></tt> are
+provided, with values corresponding to the enumerators of C11’s
+<tt class="docutils literal"><span class="pre">memory_order</span></tt> enumeration.</p>
+<p>(Note that Clang additionally provides GCC-compatible <tt class="docutils literal"><span class="pre">__atomic_*</span></tt>
+builtins)</p>
+</div>
+<div class="section" id="low-level-arm-exclusive-memory-builtins">
+<h3>Low-level ARM exclusive memory builtins<a class="headerlink" href="#low-level-arm-exclusive-memory-builtins" title="Permalink to this headline">¶</a></h3>
+<p>Clang provides overloaded builtins giving direct access to the three key ARM
+instructions for implementing atomic operations.</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="n">T</span> <span class="nf">__builtin_arm_ldrex</span><span class="p">(</span><span class="k">const</span> <span class="k">volatile</span> <span class="n">T</span> <span class="o">*</span><span class="n">addr</span><span class="p">);</span>
+<span class="n">T</span> <span class="nf">__builtin_arm_ldaex</span><span class="p">(</span><span class="k">const</span> <span class="k">volatile</span> <span class="n">T</span> <span class="o">*</span><span class="n">addr</span><span class="p">);</span>
+<span class="kt">int</span> <span class="nf">__builtin_arm_strex</span><span class="p">(</span><span class="n">T</span> <span class="n">val</span><span class="p">,</span> <span class="k">volatile</span> <span class="n">T</span> <span class="o">*</span><span class="n">addr</span><span class="p">);</span>
+<span class="kt">int</span> <span class="nf">__builtin_arm_stlex</span><span class="p">(</span><span class="n">T</span> <span class="n">val</span><span class="p">,</span> <span class="k">volatile</span> <span class="n">T</span> <span class="o">*</span><span class="n">addr</span><span class="p">);</span>
+<span class="kt">void</span> <span class="nf">__builtin_arm_clrex</span><span class="p">(</span><span class="kt">void</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>The types <tt class="docutils literal"><span class="pre">T</span></tt> currently supported are:</p>
+<ul class="simple">
+<li>Integer types with width at most 64 bits (or 128 bits on AArch64).</li>
+<li>Floating-point types</li>
+<li>Pointer types.</li>
+</ul>
+<p>Note that the compiler does not guarantee it will not insert stores which clear
+the exclusive monitor in between an <tt class="docutils literal"><span class="pre">ldrex</span></tt> type operation and its paired
+<tt class="docutils literal"><span class="pre">strex</span></tt>. In practice this is only usually a risk when the extra store is on
+the same cache line as the variable being modified and Clang will only insert
+stack stores on its own, so it is best not to use these operations on variables
+with automatic storage duration.</p>
+<p>Also, loads and stores may be implicit in code written between the <tt class="docutils literal"><span class="pre">ldrex</span></tt> and
+<tt class="docutils literal"><span class="pre">strex</span></tt>. Clang will not necessarily mitigate the effects of these either, so
+care should be exercised.</p>
+<p>For these reasons the higher level atomic primitives should be preferred where
+possible.</p>
+</div>
+<div class="section" id="non-temporal-load-store-builtins">
+<h3>Non-temporal load/store builtins<a class="headerlink" href="#non-temporal-load-store-builtins" title="Permalink to this headline">¶</a></h3>
+<p>Clang provides overloaded builtins allowing generation of non-temporal memory
+accesses.</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="n">T</span> <span class="nf">__builtin_nontemporal_load</span><span class="p">(</span><span class="n">T</span> <span class="o">*</span><span class="n">addr</span><span class="p">);</span>
+<span class="kt">void</span> <span class="nf">__builtin_nontemporal_store</span><span class="p">(</span><span class="n">T</span> <span class="n">value</span><span class="p">,</span> <span class="n">T</span> <span class="o">*</span><span class="n">addr</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>The types <tt class="docutils literal"><span class="pre">T</span></tt> currently supported are:</p>
+<ul class="simple">
+<li>Integer types.</li>
+<li>Floating-point types.</li>
+<li>Vector types.</li>
+</ul>
+<p>Note that the compiler does not guarantee that non-temporal loads or stores
+will be used.</p>
+</div>
+</div>
+<div class="section" id="non-standard-c-11-attributes">
+<h2><a class="toc-backref" href="#id15">Non-standard C++11 Attributes</a><a class="headerlink" href="#non-standard-c-11-attributes" title="Permalink to this headline">¶</a></h2>
+<p>Clang’s non-standard C++11 attributes live in the <tt class="docutils literal"><span class="pre">clang</span></tt> attribute
+namespace.</p>
+<p>Clang supports GCC’s <tt class="docutils literal"><span class="pre">gnu</span></tt> attribute namespace. All GCC attributes which
+are accepted with the <tt class="docutils literal"><span class="pre">__attribute__((foo))</span></tt> syntax are also accepted as
+<tt class="docutils literal"><span class="pre">[[gnu::foo]]</span></tt>. This only extends to attributes which are specified by GCC
+(see the list of <a class="reference external" href="http://gcc.gnu.org/onlinedocs/gcc/Function-Attributes.html">GCC function attributes</a>, <a class="reference external" href="http://gcc.gnu.org/onlinedocs/gcc/Variable-Attributes.html">GCC variable
+attributes</a>, and
+<a class="reference external" href="http://gcc.gnu.org/onlinedocs/gcc/Type-Attributes.html">GCC type attributes</a>). As with the GCC
+implementation, these attributes must appertain to the <em>declarator-id</em> in a
+declaration, which means they must go either at the start of the declaration or
+immediately after the name being declared.</p>
+<p>For example, this applies the GNU <tt class="docutils literal"><span class="pre">unused</span></tt> attribute to <tt class="docutils literal"><span class="pre">a</span></tt> and <tt class="docutils literal"><span class="pre">f</span></tt>, and
+also applies the GNU <tt class="docutils literal"><span class="pre">noreturn</span></tt> attribute to <tt class="docutils literal"><span class="pre">f</span></tt>.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="p">[[</span><span class="n">gnu</span><span class="o">::</span><span class="n">unused</span><span class="p">]]</span> <span class="kt">int</span> <span class="n">a</span><span class="p">,</span> <span class="n">f</span> <span class="p">[[</span><span class="n">gnu</span><span class="o">::</span><span class="n">noreturn</span><span class="p">]]</span> <span class="p">();</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="target-specific-extensions">
+<h2><a class="toc-backref" href="#id16">Target-Specific Extensions</a><a class="headerlink" href="#target-specific-extensions" title="Permalink to this headline">¶</a></h2>
+<p>Clang supports some language features conditionally on some targets.</p>
+<div class="section" id="arm-aarch64-language-extensions">
+<h3>ARM/AArch64 Language Extensions<a class="headerlink" href="#arm-aarch64-language-extensions" title="Permalink to this headline">¶</a></h3>
+<div class="section" id="memory-barrier-intrinsics">
+<h4>Memory Barrier Intrinsics<a class="headerlink" href="#memory-barrier-intrinsics" title="Permalink to this headline">¶</a></h4>
+<p>Clang implements the <tt class="docutils literal"><span class="pre">__dmb</span></tt>, <tt class="docutils literal"><span class="pre">__dsb</span></tt> and <tt class="docutils literal"><span class="pre">__isb</span></tt> intrinsics as defined
+in the <a class="reference external" href="http://infocenter.arm.com/help/topic/com.arm.doc.ihi0053c/IHI0053C_acle_2_0.pdf">ARM C Language Extensions Release 2.0</a>.
+Note that these intrinsics are implemented as motion barriers that block
+reordering of memory accesses and side effect instructions. Other instructions
+like simple arithmatic may be reordered around the intrinsic. If you expect to
+have no reordering at all, use inline assembly instead.</p>
+</div>
+</div>
+<div class="section" id="x86-x86-64-language-extensions">
+<h3>X86/X86-64 Language Extensions<a class="headerlink" href="#x86-x86-64-language-extensions" title="Permalink to this headline">¶</a></h3>
+<p>The X86 backend has these language extensions:</p>
+<div class="section" id="memory-references-off-the-gs-segment">
+<h4>Memory references off the GS segment<a class="headerlink" href="#memory-references-off-the-gs-segment" title="Permalink to this headline">¶</a></h4>
+<p>Annotating a pointer with address space #256 causes it to be code generated
+relative to the X86 GS segment register, and address space #257 causes it to be
+relative to the X86 FS segment. Note that this is a very very low-level
+feature that should only be used if you know what you’re doing (for example in
+an OS kernel).</p>
+<p>Here is an example:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#define GS_RELATIVE __attribute__((address_space(256)))</span>
+<span class="kt">int</span> <span class="nf">foo</span><span class="p">(</span><span class="kt">int</span> <span class="n">GS_RELATIVE</span> <span class="o">*</span><span class="n">P</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">return</span> <span class="o">*</span><span class="n">P</span><span class="p">;</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>Which compiles to (on X86-32):</p>
+<div class="highlight-gas"><div class="highlight"><pre><span class="nl">_foo:</span>
+ <span class="nf">movl</span> <span class="mi">4</span><span class="p">(</span><span class="nv">%esp</span><span class="p">),</span> <span class="nv">%eax</span>
+ <span class="nf">movl</span> <span class="nv">%gs</span><span class="p">:(</span><span class="nv">%eax</span><span class="p">),</span> <span class="nv">%eax</span>
+ <span class="nf">ret</span>
+</pre></div>
+</div>
+</div>
+</div>
+</div>
+<div class="section" id="extensions-for-static-analysis">
+<h2><a class="toc-backref" href="#id17">Extensions for Static Analysis</a><a class="headerlink" href="#extensions-for-static-analysis" title="Permalink to this headline">¶</a></h2>
+<p>Clang supports additional attributes that are useful for documenting program
+invariants and rules for static analysis tools, such as the <a class="reference external" href="http://clang-analyzer.llvm.org/">Clang Static
+Analyzer</a>. These attributes are documented
+in the analyzer’s <a class="reference external" href="http://clang-analyzer.llvm.org/annotations.html">list of source-level annotations</a>.</p>
+</div>
+<div class="section" id="extensions-for-dynamic-analysis">
+<h2><a class="toc-backref" href="#id18">Extensions for Dynamic Analysis</a><a class="headerlink" href="#extensions-for-dynamic-analysis" title="Permalink to this headline">¶</a></h2>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(address_sanitizer)</span></tt> to check if the code is being built
+with <a class="reference internal" href="AddressSanitizer.html"><em>AddressSanitizer</em></a>.</p>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(thread_sanitizer)</span></tt> to check if the code is being built
+with <a class="reference internal" href="ThreadSanitizer.html"><em>ThreadSanitizer</em></a>.</p>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(memory_sanitizer)</span></tt> to check if the code is being built
+with <a class="reference internal" href="MemorySanitizer.html"><em>MemorySanitizer</em></a>.</p>
+<p>Use <tt class="docutils literal"><span class="pre">__has_feature(safe_stack)</span></tt> to check if the code is being built
+with <a class="reference internal" href="SafeStack.html"><em>SafeStack</em></a>.</p>
+</div>
+<div class="section" id="extensions-for-selectively-disabling-optimization">
+<h2><a class="toc-backref" href="#id19">Extensions for selectively disabling optimization</a><a class="headerlink" href="#extensions-for-selectively-disabling-optimization" title="Permalink to this headline">¶</a></h2>
+<p>Clang provides a mechanism for selectively disabling optimizations in functions
+and methods.</p>
+<p>To disable optimizations in a single function definition, the GNU-style or C++11
+non-standard attribute <tt class="docutils literal"><span class="pre">optnone</span></tt> can be used.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="c1">// The following functions will not be optimized.</span>
+<span class="c1">// GNU-style attribute</span>
+<span class="n">__attribute__</span><span class="p">((</span><span class="n">optnone</span><span class="p">))</span> <span class="kt">int</span> <span class="n">foo</span><span class="p">()</span> <span class="p">{</span>
+ <span class="c1">// ... code</span>
+<span class="p">}</span>
+<span class="c1">// C++11 attribute</span>
+<span class="p">[[</span><span class="n">clang</span><span class="o">::</span><span class="n">optnone</span><span class="p">]]</span> <span class="kt">int</span> <span class="n">bar</span><span class="p">()</span> <span class="p">{</span>
+ <span class="c1">// ... code</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>To facilitate disabling optimization for a range of function definitions, a
+range-based pragma is provided. Its syntax is <tt class="docutils literal"><span class="pre">#pragma</span> <span class="pre">clang</span> <span class="pre">optimize</span></tt>
+followed by <tt class="docutils literal"><span class="pre">off</span></tt> or <tt class="docutils literal"><span class="pre">on</span></tt>.</p>
+<p>All function definitions in the region between an <tt class="docutils literal"><span class="pre">off</span></tt> and the following
+<tt class="docutils literal"><span class="pre">on</span></tt> will be decorated with the <tt class="docutils literal"><span class="pre">optnone</span></tt> attribute unless doing so would
+conflict with explicit attributes already present on the function (e.g. the
+ones that control inlining).</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#pragma clang optimize off</span>
+<span class="c1">// This function will be decorated with optnone.</span>
+<span class="kt">int</span> <span class="nf">foo</span><span class="p">()</span> <span class="p">{</span>
+ <span class="c1">// ... code</span>
+<span class="p">}</span>
+
+<span class="c1">// optnone conflicts with always_inline, so bar() will not be decorated.</span>
+<span class="n">__attribute__</span><span class="p">((</span><span class="n">always_inline</span><span class="p">))</span> <span class="kt">int</span> <span class="n">bar</span><span class="p">()</span> <span class="p">{</span>
+ <span class="c1">// ... code</span>
+<span class="p">}</span>
+<span class="cp">#pragma clang optimize on</span>
+</pre></div>
+</div>
+<p>If no <tt class="docutils literal"><span class="pre">on</span></tt> is found to close an <tt class="docutils literal"><span class="pre">off</span></tt> region, the end of the region is the
+end of the compilation unit.</p>
+<p>Note that a stray <tt class="docutils literal"><span class="pre">#pragma</span> <span class="pre">clang</span> <span class="pre">optimize</span> <span class="pre">on</span></tt> does not selectively enable
+additional optimizations when compiling at low optimization levels. This feature
+can only be used to selectively disable optimizations.</p>
+<p>The pragma has an effect on functions only at the point of their definition; for
+function templates, this means that the state of the pragma at the point of an
+instantiation is not necessarily relevant. Consider the following example:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">template</span><span class="o"><</span><span class="k">typename</span> <span class="n">T</span><span class="o">></span> <span class="n">T</span> <span class="n">twice</span><span class="p">(</span><span class="n">T</span> <span class="n">t</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">return</span> <span class="mi">2</span> <span class="o">*</span> <span class="n">t</span><span class="p">;</span>
+<span class="p">}</span>
+
+<span class="cp">#pragma clang optimize off</span>
+<span class="k">template</span><span class="o"><</span><span class="k">typename</span> <span class="n">T</span><span class="o">></span> <span class="n">T</span> <span class="n">thrice</span><span class="p">(</span><span class="n">T</span> <span class="n">t</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">return</span> <span class="mi">3</span> <span class="o">*</span> <span class="n">t</span><span class="p">;</span>
+<span class="p">}</span>
+
+<span class="kt">int</span> <span class="n">container</span><span class="p">(</span><span class="kt">int</span> <span class="n">a</span><span class="p">,</span> <span class="kt">int</span> <span class="n">b</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">return</span> <span class="n">twice</span><span class="p">(</span><span class="n">a</span><span class="p">)</span> <span class="o">+</span> <span class="n">thrice</span><span class="p">(</span><span class="n">b</span><span class="p">);</span>
+<span class="p">}</span>
+<span class="cp">#pragma clang optimize on</span>
+</pre></div>
+</div>
+<p>In this example, the definition of the template function <tt class="docutils literal"><span class="pre">twice</span></tt> is outside
+the pragma region, whereas the definition of <tt class="docutils literal"><span class="pre">thrice</span></tt> is inside the region.
+The <tt class="docutils literal"><span class="pre">container</span></tt> function is also in the region and will not be optimized, but
+it causes the instantiation of <tt class="docutils literal"><span class="pre">twice</span></tt> and <tt class="docutils literal"><span class="pre">thrice</span></tt> with an <tt class="docutils literal"><span class="pre">int</span></tt> type; of
+these two instantiations, <tt class="docutils literal"><span class="pre">twice</span></tt> will be optimized (because its definition
+was outside the region) and <tt class="docutils literal"><span class="pre">thrice</span></tt> will not be optimized.</p>
+</div>
+<div class="section" id="extensions-for-loop-hint-optimizations">
+<h2><a class="toc-backref" href="#id20">Extensions for loop hint optimizations</a><a class="headerlink" href="#extensions-for-loop-hint-optimizations" title="Permalink to this headline">¶</a></h2>
+<p>The <tt class="docutils literal"><span class="pre">#pragma</span> <span class="pre">clang</span> <span class="pre">loop</span></tt> directive is used to specify hints for optimizing the
+subsequent for, while, do-while, or c++11 range-based for loop. The directive
+provides options for vectorization, interleaving, and unrolling. Loop hints can
+be specified before any loop and will be ignored if the optimization is not safe
+to apply.</p>
+<div class="section" id="vectorization-and-interleaving">
+<h3>Vectorization and Interleaving<a class="headerlink" href="#vectorization-and-interleaving" title="Permalink to this headline">¶</a></h3>
+<p>A vectorized loop performs multiple iterations of the original loop
+in parallel using vector instructions. The instruction set of the target
+processor determines which vector instructions are available and their vector
+widths. This restricts the types of loops that can be vectorized. The vectorizer
+automatically determines if the loop is safe and profitable to vectorize. A
+vector instruction cost model is used to select the vector width.</p>
+<p>Interleaving multiple loop iterations allows modern processors to further
+improve instruction-level parallelism (ILP) using advanced hardware features,
+such as multiple execution units and out-of-order execution. The vectorizer uses
+a cost model that depends on the register pressure and generated code size to
+select the interleaving count.</p>
+<p>Vectorization is enabled by <tt class="docutils literal"><span class="pre">vectorize(enable)</span></tt> and interleaving is enabled
+by <tt class="docutils literal"><span class="pre">interleave(enable)</span></tt>. This is useful when compiling with <tt class="docutils literal"><span class="pre">-Os</span></tt> to
+manually enable vectorization or interleaving.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#pragma clang loop vectorize(enable)</span>
+<span class="cp">#pragma clang loop interleave(enable)</span>
+<span class="k">for</span><span class="p">(...)</span> <span class="p">{</span>
+ <span class="p">...</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>The vector width is specified by <tt class="docutils literal"><span class="pre">vectorize_width(_value_)</span></tt> and the interleave
+count is specified by <tt class="docutils literal"><span class="pre">interleave_count(_value_)</span></tt>, where
+_value_ is a positive integer. This is useful for specifying the optimal
+width/count of the set of target architectures supported by your application.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#pragma clang loop vectorize_width(2)</span>
+<span class="cp">#pragma clang loop interleave_count(2)</span>
+<span class="k">for</span><span class="p">(...)</span> <span class="p">{</span>
+ <span class="p">...</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>Specifying a width/count of 1 disables the optimization, and is equivalent to
+<tt class="docutils literal"><span class="pre">vectorize(disable)</span></tt> or <tt class="docutils literal"><span class="pre">interleave(disable)</span></tt>.</p>
+</div>
+<div class="section" id="loop-unrolling">
+<h3>Loop Unrolling<a class="headerlink" href="#loop-unrolling" title="Permalink to this headline">¶</a></h3>
+<p>Unrolling a loop reduces the loop control overhead and exposes more
+opportunities for ILP. Loops can be fully or partially unrolled. Full unrolling
+eliminates the loop and replaces it with an enumerated sequence of loop
+iterations. Full unrolling is only possible if the loop trip count is known at
+compile time. Partial unrolling replicates the loop body within the loop and
+reduces the trip count.</p>
+<p>If <tt class="docutils literal"><span class="pre">unroll(enable)</span></tt> is specified the unroller will attempt to fully unroll the
+loop if the trip count is known at compile time. If the fully unrolled code size
+is greater than an internal limit the loop will be partially unrolled up to this
+limit. If the trip count is not known at compile time the loop will be partially
+unrolled with a heuristically chosen unroll factor.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#pragma clang loop unroll(enable)</span>
+<span class="k">for</span><span class="p">(...)</span> <span class="p">{</span>
+ <span class="p">...</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>If <tt class="docutils literal"><span class="pre">unroll(full)</span></tt> is specified the unroller will attempt to fully unroll the
+loop if the trip count is known at compile time identically to
+<tt class="docutils literal"><span class="pre">unroll(enable)</span></tt>. However, with <tt class="docutils literal"><span class="pre">unroll(full)</span></tt> the loop will not be unrolled
+if the loop count is not known at compile time.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#pragma clang loop unroll(full)</span>
+<span class="k">for</span><span class="p">(...)</span> <span class="p">{</span>
+ <span class="p">...</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>The unroll count can be specified explicitly with <tt class="docutils literal"><span class="pre">unroll_count(_value_)</span></tt> where
+_value_ is a positive integer. If this value is greater than the trip count the
+loop will be fully unrolled. Otherwise the loop is partially unrolled subject
+to the same code size limit as with <tt class="docutils literal"><span class="pre">unroll(enable)</span></tt>.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#pragma clang loop unroll_count(8)</span>
+<span class="k">for</span><span class="p">(...)</span> <span class="p">{</span>
+ <span class="p">...</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>Unrolling of a loop can be prevented by specifying <tt class="docutils literal"><span class="pre">unroll(disable)</span></tt>.</p>
+</div>
+<div class="section" id="additional-information">
+<h3>Additional Information<a class="headerlink" href="#additional-information" title="Permalink to this headline">¶</a></h3>
+<p>For convenience multiple loop hints can be specified on a single line.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#pragma clang loop vectorize_width(4) interleave_count(8)</span>
+<span class="k">for</span><span class="p">(...)</span> <span class="p">{</span>
+ <span class="p">...</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>If an optimization cannot be applied any hints that apply to it will be ignored.
+For example, the hint <tt class="docutils literal"><span class="pre">vectorize_width(4)</span></tt> is ignored if the loop is not
+proven safe to vectorize. To identify and diagnose optimization issues use
+<cite>-Rpass</cite>, <cite>-Rpass-missed</cite>, and <cite>-Rpass-analysis</cite> command line options. See the
+user guide for details.</p>
+</div>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="UsersManual.html">Clang Compiler User’s Manual</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ObjectiveCLiterals.html">Objective-C Literals</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/LeakSanitizer.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/LeakSanitizer.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/LeakSanitizer.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/LeakSanitizer.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,100 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>LeakSanitizer — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="SanitizerCoverage" href="SanitizerCoverage.html" />
+ <link rel="prev" title="DataFlowSanitizer Design Document" href="DataFlowSanitizerDesign.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>LeakSanitizer</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="DataFlowSanitizerDesign.html">DataFlowSanitizer Design Document</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="SanitizerCoverage.html">SanitizerCoverage</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="leaksanitizer">
+<h1>LeakSanitizer<a class="headerlink" href="#leaksanitizer" title="Permalink to this headline">¶</a></h1>
+<div class="contents local topic" id="contents">
+<ul class="simple">
+<li><a class="reference internal" href="#introduction" id="id1">Introduction</a></li>
+<li><a class="reference internal" href="#current-status" id="id2">Current status</a></li>
+<li><a class="reference internal" href="#more-information" id="id3">More Information</a></li>
+</ul>
+</div>
+<div class="section" id="introduction">
+<h2><a class="toc-backref" href="#id1">Introduction</a><a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>LeakSanitizer is a run-time memory leak detector. It can be combined with
+<a class="reference internal" href="AddressSanitizer.html"><em>AddressSanitizer</em></a> to get both memory error and leak detection.
+LeakSanitizer does not introduce any additional slowdown when used in this mode.
+The LeakSanitizer runtime can also be linked in separately to get leak detection
+only, at a minimal performance cost.</p>
+</div>
+<div class="section" id="current-status">
+<h2><a class="toc-backref" href="#id2">Current status</a><a class="headerlink" href="#current-status" title="Permalink to this headline">¶</a></h2>
+<p>LeakSanitizer is turned on by default, but it is only supported on x86_64
+Linux.</p>
+<p>The combined mode has been tested on fairly large software projects. The
+stand-alone mode has received much less testing.</p>
+<p>There are plans to support LeakSanitizer in <a class="reference internal" href="MemorySanitizer.html"><em>MemorySanitizer</em></a> builds.</p>
+</div>
+<div class="section" id="more-information">
+<h2><a class="toc-backref" href="#id3">More Information</a><a class="headerlink" href="#more-information" title="Permalink to this headline">¶</a></h2>
+<p><a class="reference external" href="https://github.com/google/sanitizers/wiki/AddressSanitizerLeakSanitizer">https://github.com/google/sanitizers/wiki/AddressSanitizerLeakSanitizer</a></p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="DataFlowSanitizerDesign.html">DataFlowSanitizer Design Document</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="SanitizerCoverage.html">SanitizerCoverage</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/LibASTMatchers.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/LibASTMatchers.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/LibASTMatchers.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/LibASTMatchers.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,167 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Matching the Clang AST — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="How To Setup Clang Tooling For LLVM" href="HowToSetupToolingForLLVM.html" />
+ <link rel="prev" title="Tutorial for building tools using LibTooling and LibASTMatchers" href="LibASTMatchersTutorial.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Matching the Clang AST</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="LibASTMatchersTutorial.html">Tutorial for building tools using LibTooling and LibASTMatchers</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="HowToSetupToolingForLLVM.html">How To Setup Clang Tooling For LLVM</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="matching-the-clang-ast">
+<h1>Matching the Clang AST<a class="headerlink" href="#matching-the-clang-ast" title="Permalink to this headline">¶</a></h1>
+<p>This document explains how to use Clang’s LibASTMatchers to match interesting
+nodes of the AST and execute code that uses the matched nodes. Combined with
+<a class="reference internal" href="LibTooling.html"><em>LibTooling</em></a>, LibASTMatchers helps to write code-to-code transformation
+tools or query tools.</p>
+<p>We assume basic knowledge about the Clang AST. See the <a class="reference internal" href="IntroductionToTheClangAST.html"><em>Introduction
+to the Clang AST</em></a> if you want to learn more
+about how the AST is structured.</p>
+<div class="section" id="introduction">
+<h2>Introduction<a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>LibASTMatchers provides a domain specific language to create predicates on
+Clang’s AST. This DSL is written in and can be used from C++, allowing users
+to write a single program to both match AST nodes and access the node’s C++
+interface to extract attributes, source locations, or any other information
+provided on the AST level.</p>
+<p>AST matchers are predicates on nodes in the AST. Matchers are created by
+calling creator functions that allow building up a tree of matchers, where
+inner matchers are used to make the match more specific.</p>
+<p>For example, to create a matcher that matches all class or union declarations
+in the AST of a translation unit, you can call <a class="reference external" href="LibASTMatchersReference.html#recordDecl0Anchor">recordDecl()</a>. To narrow the match down,
+for example to find all class or union declarations with the name “<tt class="docutils literal"><span class="pre">Foo</span></tt>”,
+insert a <a class="reference external" href="LibASTMatchersReference.html#hasName0Anchor">hasName</a> matcher: the
+call <tt class="docutils literal"><span class="pre">recordDecl(hasName("Foo"))</span></tt> returns a matcher that matches classes or
+unions that are named “<tt class="docutils literal"><span class="pre">Foo</span></tt>”, in any namespace. By default, matchers that
+accept multiple inner matchers use an implicit <a class="reference external" href="LibASTMatchersReference.html#allOf0Anchor">allOf()</a>. This allows further narrowing
+down the match, for example to match all classes that are derived from
+“<tt class="docutils literal"><span class="pre">Bar</span></tt>”: <tt class="docutils literal"><span class="pre">recordDecl(hasName("Foo"),</span> <span class="pre">isDerivedFrom("Bar"))</span></tt>.</p>
+</div>
+<div class="section" id="how-to-create-a-matcher">
+<h2>How to create a matcher<a class="headerlink" href="#how-to-create-a-matcher" title="Permalink to this headline">¶</a></h2>
+<p>With more than a thousand classes in the Clang AST, one can quickly get lost
+when trying to figure out how to create a matcher for a specific pattern. This
+section will teach you how to use a rigorous step-by-step pattern to build the
+matcher you are interested in. Note that there will always be matchers missing
+for some part of the AST. See the section about <a class="reference internal" href="#astmatchers-writing"><em>how to write your own
+AST matchers</em></a> later in this document.</p>
+<p>The precondition to using the matchers is to understand how the AST for what you
+want to match looks like. The
+<a class="reference internal" href="IntroductionToTheClangAST.html"><em>Introduction to the Clang AST</em></a> teaches you
+how to dump a translation unit’s AST into a human readable format.</p>
+<p>In general, the strategy to create the right matchers is:</p>
+<ol class="arabic simple">
+<li>Find the outermost class in Clang’s AST you want to match.</li>
+<li>Look at the <a class="reference external" href="LibASTMatchersReference.html">AST Matcher Reference</a> for
+matchers that either match the node you’re interested in or narrow down
+attributes on the node.</li>
+<li>Create your outer match expression. Verify that it works as expected.</li>
+<li>Examine the matchers for what the next inner node you want to match is.</li>
+<li>Repeat until the matcher is finished.</li>
+</ol>
+</div>
+<div class="section" id="binding-nodes-in-match-expressions">
+<span id="astmatchers-bind"></span><h2>Binding nodes in match expressions<a class="headerlink" href="#binding-nodes-in-match-expressions" title="Permalink to this headline">¶</a></h2>
+<p>Matcher expressions allow you to specify which parts of the AST are interesting
+for a certain task. Often you will want to then do something with the nodes
+that were matched, like building source code transformations.</p>
+<p>To that end, matchers that match specific AST nodes (so called node matchers)
+are bindable; for example, <tt class="docutils literal"><span class="pre">recordDecl(hasName("MyClass")).bind("id")</span></tt> will
+bind the matched <tt class="docutils literal"><span class="pre">recordDecl</span></tt> node to the string “<tt class="docutils literal"><span class="pre">id</span></tt>”, to be later
+retrieved in the <a class="reference external" href="http://clang.llvm.org/doxygen/classclang_1_1ast__matchers_1_1MatchFinder_1_1MatchCallback.html">match callback</a>.</p>
+</div>
+<div class="section" id="writing-your-own-matchers">
+<h2>Writing your own matchers<a class="headerlink" href="#writing-your-own-matchers" title="Permalink to this headline">¶</a></h2>
+<p>There are multiple different ways to define a matcher, depending on its type
+and flexibility.</p>
+<div class="section" id="variadicdyncastallofmatcher-base-derived">
+<h3><tt class="docutils literal"><span class="pre">VariadicDynCastAllOfMatcher<Base,</span> <span class="pre">Derived></span></tt><a class="headerlink" href="#variadicdyncastallofmatcher-base-derived" title="Permalink to this headline">¶</a></h3>
+<p>Those match all nodes of type <em>Base</em> if they can be dynamically casted to
+<em>Derived</em>. The names of those matchers are nouns, which closely resemble
+<em>Derived</em>. <tt class="docutils literal"><span class="pre">VariadicDynCastAllOfMatchers</span></tt> are the backbone of the matcher
+hierarchy. Most often, your match expression will start with one of them, and
+you can <a class="reference internal" href="#astmatchers-bind"><em>bind</em></a> the node they represent to ids for later
+processing.</p>
+<p><tt class="docutils literal"><span class="pre">VariadicDynCastAllOfMatchers</span></tt> are callable classes that model variadic
+template functions in C++03. They take an aribtrary number of
+<tt class="docutils literal"><span class="pre">Matcher<Derived></span></tt> and return a <tt class="docutils literal"><span class="pre">Matcher<Base></span></tt>.</p>
+</div>
+<div class="section" id="ast-matcher-p-type-name-paramtype-param">
+<h3><tt class="docutils literal"><span class="pre">AST_MATCHER_P(Type,</span> <span class="pre">Name,</span> <span class="pre">ParamType,</span> <span class="pre">Param)</span></tt><a class="headerlink" href="#ast-matcher-p-type-name-paramtype-param" title="Permalink to this headline">¶</a></h3>
+<p>Most matcher definitions use the matcher creation macros. Those define both
+the matcher of type <tt class="docutils literal"><span class="pre">Matcher<Type></span></tt> itself, and a matcher-creation function
+named <em>Name</em> that takes a parameter of type <em>ParamType</em> and returns the
+corresponding matcher.</p>
+<p>There are multiple matcher definition macros that deal with polymorphic return
+values and different parameter counts. See <a class="reference external" href="http://clang.llvm.org/doxygen/ASTMatchersMacros_8h.html">ASTMatchersMacros.h</a>.</p>
+</div>
+<div class="section" id="matcher-creation-functions">
+<span id="astmatchers-writing"></span><h3>Matcher creation functions<a class="headerlink" href="#matcher-creation-functions" title="Permalink to this headline">¶</a></h3>
+<p>Matchers are generated by nesting calls to matcher creation functions. Most of
+the time those functions are either created by using
+<tt class="docutils literal"><span class="pre">VariadicDynCastAllOfMatcher</span></tt> or the matcher creation macros (see below).
+The free-standing functions are an indication that this matcher is just a
+combination of other matchers, as is for example the case with <a class="reference external" href="LibASTMatchersReference.html#callee1Anchor">callee</a>.</p>
+</div>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="LibASTMatchersTutorial.html">Tutorial for building tools using LibTooling and LibASTMatchers</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="HowToSetupToolingForLLVM.html">How To Setup Clang Tooling For LLVM</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/LibASTMatchersTutorial.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/LibASTMatchersTutorial.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/LibASTMatchersTutorial.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/LibASTMatchersTutorial.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,542 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Tutorial for building tools using LibTooling and LibASTMatchers — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Matching the Clang AST" href="LibASTMatchers.html" />
+ <link rel="prev" title="How to write RecursiveASTVisitor based ASTFrontendActions." href="RAVFrontendAction.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Tutorial for building tools using LibTooling and LibASTMatchers</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="RAVFrontendAction.html">How to write RecursiveASTVisitor based ASTFrontendActions.</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="LibASTMatchers.html">Matching the Clang AST</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="tutorial-for-building-tools-using-libtooling-and-libastmatchers">
+<h1>Tutorial for building tools using LibTooling and LibASTMatchers<a class="headerlink" href="#tutorial-for-building-tools-using-libtooling-and-libastmatchers" title="Permalink to this headline">¶</a></h1>
+<p>This document is intended to show how to build a useful source-to-source
+translation tool based on Clang’s <a class="reference external" href="LibTooling.html">LibTooling</a>. It is
+explicitly aimed at people who are new to Clang, so all you should need
+is a working knowledge of C++ and the command line.</p>
+<p>In order to work on the compiler, you need some basic knowledge of the
+abstract syntax tree (AST). To this end, the reader is incouraged to
+skim the <a class="reference internal" href="IntroductionToTheClangAST.html"><em>Introduction to the Clang
+AST</em></a></p>
+<div class="section" id="step-0-obtaining-clang">
+<h2>Step 0: Obtaining Clang<a class="headerlink" href="#step-0-obtaining-clang" title="Permalink to this headline">¶</a></h2>
+<p>As Clang is part of the LLVM project, you’ll need to download LLVM’s
+source code first. Both Clang and LLVM are maintained as Subversion
+repositories, but we’ll be accessing them through the git mirror. For
+further information, see the <a class="reference external" href="http://llvm.org/docs/GettingStarted.html">getting started
+guide</a>.</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="go">mkdir ~/clang-llvm && cd ~/clang-llvm</span>
+<span class="go">git clone http://llvm.org/git/llvm.git</span>
+<span class="go">cd llvm/tools</span>
+<span class="go">git clone http://llvm.org/git/clang.git</span>
+<span class="go">cd clang/tools</span>
+<span class="go">git clone http://llvm.org/git/clang-tools-extra.git extra</span>
+</pre></div>
+</div>
+<p>Next you need to obtain the CMake build system and Ninja build tool. You
+may already have CMake installed, but current binary versions of CMake
+aren’t built with Ninja support.</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="go">cd ~/clang-llvm</span>
+<span class="go">git clone https://github.com/martine/ninja.git</span>
+<span class="go">cd ninja</span>
+<span class="go">git checkout release</span>
+<span class="go">./bootstrap.py</span>
+<span class="go">sudo cp ninja /usr/bin/</span>
+
+<span class="go">cd ~/clang-llvm</span>
+<span class="go">git clone git://cmake.org/stage/cmake.git</span>
+<span class="go">cd cmake</span>
+<span class="go">git checkout next</span>
+<span class="go">./bootstrap</span>
+<span class="go">make</span>
+<span class="go">sudo make install</span>
+</pre></div>
+</div>
+<p>Okay. Now we’ll build Clang!</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="go">cd ~/clang-llvm</span>
+<span class="go">mkdir build && cd build</span>
+<span class="go">cmake -G Ninja ../llvm -DLLVM_BUILD_TESTS=ON # Enable tests; default is off.</span>
+<span class="go">ninja</span>
+<span class="go">ninja check # Test LLVM only.</span>
+<span class="go">ninja clang-test # Test Clang only.</span>
+<span class="go">ninja install</span>
+</pre></div>
+</div>
+<p>And we’re live.</p>
+<p>All of the tests should pass, though there is a (very) small chance that
+you can catch LLVM and Clang out of sync. Running <tt class="docutils literal"><span class="pre">'git</span> <span class="pre">svn</span> <span class="pre">rebase'</span></tt>
+in both the llvm and clang directories should fix any problems.</p>
+<p>Finally, we want to set Clang as its own compiler.</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="go">cd ~/clang-llvm/build</span>
+<span class="go">ccmake ../llvm</span>
+</pre></div>
+</div>
+<p>The second command will bring up a GUI for configuring Clang. You need
+to set the entry for <tt class="docutils literal"><span class="pre">CMAKE_CXX_COMPILER</span></tt>. Press <tt class="docutils literal"><span class="pre">'t'</span></tt> to turn on
+advanced mode. Scroll down to <tt class="docutils literal"><span class="pre">CMAKE_CXX_COMPILER</span></tt>, and set it to
+<tt class="docutils literal"><span class="pre">/usr/bin/clang++</span></tt>, or wherever you installed it. Press <tt class="docutils literal"><span class="pre">'c'</span></tt> to
+configure, then <tt class="docutils literal"><span class="pre">'g'</span></tt> to generate CMake’s files.</p>
+<p>Finally, run ninja one last time, and you’re done.</p>
+</div>
+<div class="section" id="step-1-create-a-clangtool">
+<h2>Step 1: Create a ClangTool<a class="headerlink" href="#step-1-create-a-clangtool" title="Permalink to this headline">¶</a></h2>
+<p>Now that we have enough background knowledge, it’s time to create the
+simplest productive ClangTool in existence: a syntax checker. While this
+already exists as <tt class="docutils literal"><span class="pre">clang-check</span></tt>, it’s important to understand what’s
+going on.</p>
+<p>First, we’ll need to create a new directory for our tool and tell CMake
+that it exists. As this is not going to be a core clang tool, it will
+live in the <tt class="docutils literal"><span class="pre">tools/extra</span></tt> repository.</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="go">cd ~/clang-llvm/llvm/tools/clang</span>
+<span class="go">mkdir tools/extra/loop-convert</span>
+<span class="go">echo 'add_subdirectory(loop-convert)' >> tools/extra/CMakeLists.txt</span>
+<span class="go">vim tools/extra/loop-convert/CMakeLists.txt</span>
+</pre></div>
+</div>
+<p>CMakeLists.txt should have the following contents:</p>
+<div class="highlight-python"><div class="highlight"><pre>set(LLVM_LINK_COMPONENTS support)
+
+add_clang_executable(loop-convert
+ LoopConvert.cpp
+ )
+target_link_libraries(loop-convert
+ clangTooling
+ clangBasic
+ clangASTMatchers
+ )
+</pre></div>
+</div>
+<p>With that done, Ninja will be able to compile our tool. Let’s give it
+something to compile! Put the following into
+<tt class="docutils literal"><span class="pre">tools/extra/loop-convert/LoopConvert.cpp</span></tt>. A detailed explanation of
+why the different parts are needed can be found in the <a class="reference external" href="LibTooling.html">LibTooling
+documentation</a>.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="c1">// Declares clang::SyntaxOnlyAction.</span>
+<span class="cp">#include "clang/Frontend/FrontendActions.h"</span>
+<span class="cp">#include "clang/Tooling/CommonOptionsParser.h"</span>
+<span class="cp">#include "clang/Tooling/Tooling.h"</span>
+<span class="c1">// Declares llvm::cl::extrahelp.</span>
+<span class="cp">#include "llvm/Support/CommandLine.h"</span>
+
+<span class="k">using</span> <span class="k">namespace</span> <span class="n">clang</span><span class="o">::</span><span class="n">tooling</span><span class="p">;</span>
+<span class="k">using</span> <span class="k">namespace</span> <span class="n">llvm</span><span class="p">;</span>
+
+<span class="c1">// Apply a custom category to all command-line options so that they are the</span>
+<span class="c1">// only ones displayed.</span>
+<span class="k">static</span> <span class="n">llvm</span><span class="o">::</span><span class="n">cl</span><span class="o">::</span><span class="n">OptionCategory</span> <span class="n">MyToolCategory</span><span class="p">(</span><span class="s">"my-tool options"</span><span class="p">);</span>
+
+<span class="c1">// CommonOptionsParser declares HelpMessage with a description of the common</span>
+<span class="c1">// command-line options related to the compilation database and input files.</span>
+<span class="c1">// It's nice to have this help message in all tools.</span>
+<span class="k">static</span> <span class="n">cl</span><span class="o">::</span><span class="n">extrahelp</span> <span class="n">CommonHelp</span><span class="p">(</span><span class="n">CommonOptionsParser</span><span class="o">::</span><span class="n">HelpMessage</span><span class="p">);</span>
+
+<span class="c1">// A help message for this specific tool can be added afterwards.</span>
+<span class="k">static</span> <span class="n">cl</span><span class="o">::</span><span class="n">extrahelp</span> <span class="n">MoreHelp</span><span class="p">(</span><span class="s">"</span><span class="se">\n</span><span class="s">More help text..."</span><span class="p">);</span>
+
+<span class="kt">int</span> <span class="nf">main</span><span class="p">(</span><span class="kt">int</span> <span class="n">argc</span><span class="p">,</span> <span class="k">const</span> <span class="kt">char</span> <span class="o">**</span><span class="n">argv</span><span class="p">)</span> <span class="p">{</span>
+ <span class="n">CommonOptionsParser</span> <span class="n">OptionsParser</span><span class="p">(</span><span class="n">argc</span><span class="p">,</span> <span class="n">argv</span><span class="p">,</span> <span class="n">MyToolCategory</span><span class="p">);</span>
+ <span class="n">ClangTool</span> <span class="n">Tool</span><span class="p">(</span><span class="n">OptionsParser</span><span class="p">.</span><span class="n">getCompilations</span><span class="p">(),</span>
+ <span class="n">OptionsParser</span><span class="p">.</span><span class="n">getSourcePathList</span><span class="p">());</span>
+ <span class="k">return</span> <span class="n">Tool</span><span class="p">.</span><span class="n">run</span><span class="p">(</span><span class="n">newFrontendActionFactory</span><span class="o"><</span><span class="n">clang</span><span class="o">::</span><span class="n">SyntaxOnlyAction</span><span class="o">></span><span class="p">().</span><span class="n">get</span><span class="p">());</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>And that’s it! You can compile our new tool by running ninja from the
+<tt class="docutils literal"><span class="pre">build</span></tt> directory.</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="go">cd ~/clang-llvm/build</span>
+<span class="go">ninja</span>
+</pre></div>
+</div>
+<p>You should now be able to run the syntax checker, which is located in
+<tt class="docutils literal"><span class="pre">~/clang-llvm/build/bin</span></tt>, on any source file. Try it!</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="go">echo "int main() { return 0; }" > test.cpp</span>
+<span class="go">bin/loop-convert test.cpp --</span>
+</pre></div>
+</div>
+<p>Note the two dashes after we specify the source file. The additional
+options for the compiler are passed after the dashes rather than loading
+them from a compilation database - there just aren’t any options needed
+right now.</p>
+</div>
+<div class="section" id="intermezzo-learn-ast-matcher-basics">
+<h2>Intermezzo: Learn AST matcher basics<a class="headerlink" href="#intermezzo-learn-ast-matcher-basics" title="Permalink to this headline">¶</a></h2>
+<p>Clang recently introduced the <a class="reference internal" href="LibASTMatchers.html"><em>ASTMatcher
+library</em></a> to provide a simple, powerful, and
+concise way to describe specific patterns in the AST. Implemented as a
+DSL powered by macros and templates (see
+<a class="reference external" href="../doxygen/ASTMatchers_8h_source.html">ASTMatchers.h</a> if you’re
+curious), matchers offer the feel of algebraic data types common to
+functional programming languages.</p>
+<p>For example, suppose you wanted to examine only binary operators. There
+is a matcher to do exactly that, conveniently named <tt class="docutils literal"><span class="pre">binaryOperator</span></tt>.
+I’ll give you one guess what this matcher does:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">binaryOperator</span><span class="p">(</span><span class="n">hasOperatorName</span><span class="p">(</span><span class="s">"+"</span><span class="p">),</span> <span class="n">hasLHS</span><span class="p">(</span><span class="n">integerLiteral</span><span class="p">(</span><span class="n">equals</span><span class="p">(</span><span class="mi">0</span><span class="p">))))</span>
+</pre></div>
+</div>
+<p>Shockingly, it will match against addition expressions whose left hand
+side is exactly the literal 0. It will not match against other forms of
+0, such as <tt class="docutils literal"><span class="pre">'\0'</span></tt> or <tt class="docutils literal"><span class="pre">NULL</span></tt>, but it will match against macros that
+expand to 0. The matcher will also not match against calls to the
+overloaded operator <tt class="docutils literal"><span class="pre">'+'</span></tt>, as there is a separate <tt class="docutils literal"><span class="pre">operatorCallExpr</span></tt>
+matcher to handle overloaded operators.</p>
+<p>There are AST matchers to match all the different nodes of the AST,
+narrowing matchers to only match AST nodes fulfilling specific criteria,
+and traversal matchers to get from one kind of AST node to another. For
+a complete list of AST matchers, take a look at the <a class="reference external" href="LibASTMatchersReference.html">AST Matcher
+References</a></p>
+<p>All matcher that are nouns describe entities in the AST and can be
+bound, so that they can be referred to whenever a match is found. To do
+so, simply call the method <tt class="docutils literal"><span class="pre">bind</span></tt> on these matchers, e.g.:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">variable</span><span class="p">(</span><span class="n">hasType</span><span class="p">(</span><span class="n">isInteger</span><span class="p">())).</span><span class="n">bind</span><span class="p">(</span><span class="s">"intvar"</span><span class="p">)</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="step-2-using-ast-matchers">
+<h2>Step 2: Using AST matchers<a class="headerlink" href="#step-2-using-ast-matchers" title="Permalink to this headline">¶</a></h2>
+<p>Okay, on to using matchers for real. Let’s start by defining a matcher
+which will capture all <tt class="docutils literal"><span class="pre">for</span></tt> statements that define a new variable
+initialized to zero. Let’s start with matching all <tt class="docutils literal"><span class="pre">for</span></tt> loops:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">forStmt</span><span class="p">()</span>
+</pre></div>
+</div>
+<p>Next, we want to specify that a single variable is declared in the first
+portion of the loop, so we can extend the matcher to</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">forStmt</span><span class="p">(</span><span class="n">hasLoopInit</span><span class="p">(</span><span class="n">declStmt</span><span class="p">(</span><span class="n">hasSingleDecl</span><span class="p">(</span><span class="n">varDecl</span><span class="p">()))))</span>
+</pre></div>
+</div>
+<p>Finally, we can add the condition that the variable is initialized to
+zero.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">forStmt</span><span class="p">(</span><span class="n">hasLoopInit</span><span class="p">(</span><span class="n">declStmt</span><span class="p">(</span><span class="n">hasSingleDecl</span><span class="p">(</span><span class="n">varDecl</span><span class="p">(</span>
+ <span class="n">hasInitializer</span><span class="p">(</span><span class="n">integerLiteral</span><span class="p">(</span><span class="n">equals</span><span class="p">(</span><span class="mi">0</span><span class="p">))))))))</span>
+</pre></div>
+</div>
+<p>It is fairly easy to read and understand the matcher definition (“match
+loops whose init portion declares a single variable which is initialized
+to the integer literal 0”), but deciding that every piece is necessary
+is more difficult. Note that this matcher will not match loops whose
+variables are initialized to <tt class="docutils literal"><span class="pre">'\0'</span></tt>, <tt class="docutils literal"><span class="pre">0.0</span></tt>, <tt class="docutils literal"><span class="pre">NULL</span></tt>, or any form of
+zero besides the integer 0.</p>
+<p>The last step is giving the matcher a name and binding the <tt class="docutils literal"><span class="pre">ForStmt</span></tt>
+as we will want to do something with it:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">StatementMatcher</span> <span class="n">LoopMatcher</span> <span class="o">=</span>
+ <span class="n">forStmt</span><span class="p">(</span><span class="n">hasLoopInit</span><span class="p">(</span><span class="n">declStmt</span><span class="p">(</span><span class="n">hasSingleDecl</span><span class="p">(</span><span class="n">varDecl</span><span class="p">(</span>
+ <span class="n">hasInitializer</span><span class="p">(</span><span class="n">integerLiteral</span><span class="p">(</span><span class="n">equals</span><span class="p">(</span><span class="mi">0</span><span class="p">)))))))).</span><span class="n">bind</span><span class="p">(</span><span class="s">"forLoop"</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>Once you have defined your matchers, you will need to add a little more
+scaffolding in order to run them. Matchers are paired with a
+<tt class="docutils literal"><span class="pre">MatchCallback</span></tt> and registered with a <tt class="docutils literal"><span class="pre">MatchFinder</span></tt> object, then run
+from a <tt class="docutils literal"><span class="pre">ClangTool</span></tt>. More code!</p>
+<p>Add the following to <tt class="docutils literal"><span class="pre">LoopConvert.cpp</span></tt>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#include "clang/ASTMatchers/ASTMatchers.h"</span>
+<span class="cp">#include "clang/ASTMatchers/ASTMatchFinder.h"</span>
+
+<span class="k">using</span> <span class="k">namespace</span> <span class="n">clang</span><span class="p">;</span>
+<span class="k">using</span> <span class="k">namespace</span> <span class="n">clang</span><span class="o">::</span><span class="n">ast_matchers</span><span class="p">;</span>
+
+<span class="n">StatementMatcher</span> <span class="n">LoopMatcher</span> <span class="o">=</span>
+ <span class="n">forStmt</span><span class="p">(</span><span class="n">hasLoopInit</span><span class="p">(</span><span class="n">declStmt</span><span class="p">(</span><span class="n">hasSingleDecl</span><span class="p">(</span><span class="n">varDecl</span><span class="p">(</span>
+ <span class="n">hasInitializer</span><span class="p">(</span><span class="n">integerLiteral</span><span class="p">(</span><span class="n">equals</span><span class="p">(</span><span class="mi">0</span><span class="p">)))))))).</span><span class="n">bind</span><span class="p">(</span><span class="s">"forLoop"</span><span class="p">);</span>
+
+<span class="k">class</span> <span class="nc">LoopPrinter</span> <span class="o">:</span> <span class="k">public</span> <span class="n">MatchFinder</span><span class="o">::</span><span class="n">MatchCallback</span> <span class="p">{</span>
+<span class="k">public</span> <span class="o">:</span>
+ <span class="k">virtual</span> <span class="kt">void</span> <span class="n">run</span><span class="p">(</span><span class="k">const</span> <span class="n">MatchFinder</span><span class="o">::</span><span class="n">MatchResult</span> <span class="o">&</span><span class="n">Result</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">if</span> <span class="p">(</span><span class="k">const</span> <span class="n">ForStmt</span> <span class="o">*</span><span class="n">FS</span> <span class="o">=</span> <span class="n">Result</span><span class="p">.</span><span class="n">Nodes</span><span class="p">.</span><span class="n">getNodeAs</span><span class="o"><</span><span class="n">clang</span><span class="o">::</span><span class="n">ForStmt</span><span class="o">></span><span class="p">(</span><span class="s">"forLoop"</span><span class="p">))</span>
+ <span class="n">FS</span><span class="o">-></span><span class="n">dump</span><span class="p">();</span>
+ <span class="p">}</span>
+<span class="p">};</span>
+</pre></div>
+</div>
+<p>And change <tt class="docutils literal"><span class="pre">main()</span></tt> to:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">int</span> <span class="nf">main</span><span class="p">(</span><span class="kt">int</span> <span class="n">argc</span><span class="p">,</span> <span class="k">const</span> <span class="kt">char</span> <span class="o">**</span><span class="n">argv</span><span class="p">)</span> <span class="p">{</span>
+ <span class="n">CommonOptionsParser</span> <span class="n">OptionsParser</span><span class="p">(</span><span class="n">argc</span><span class="p">,</span> <span class="n">argv</span><span class="p">,</span> <span class="n">MyToolCategory</span><span class="p">);</span>
+ <span class="n">ClangTool</span> <span class="n">Tool</span><span class="p">(</span><span class="n">OptionsParser</span><span class="p">.</span><span class="n">getCompilations</span><span class="p">(),</span>
+ <span class="n">OptionsParser</span><span class="p">.</span><span class="n">getSourcePathList</span><span class="p">());</span>
+
+ <span class="n">LoopPrinter</span> <span class="n">Printer</span><span class="p">;</span>
+ <span class="n">MatchFinder</span> <span class="n">Finder</span><span class="p">;</span>
+ <span class="n">Finder</span><span class="p">.</span><span class="n">addMatcher</span><span class="p">(</span><span class="n">LoopMatcher</span><span class="p">,</span> <span class="o">&</span><span class="n">Printer</span><span class="p">);</span>
+
+ <span class="k">return</span> <span class="n">Tool</span><span class="p">.</span><span class="n">run</span><span class="p">(</span><span class="n">newFrontendActionFactory</span><span class="p">(</span><span class="o">&</span><span class="n">Finder</span><span class="p">).</span><span class="n">get</span><span class="p">());</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>Now, you should be able to recompile and run the code to discover for
+loops. Create a new file with a few examples, and test out our new
+handiwork:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="go">cd ~/clang-llvm/llvm/llvm_build/</span>
+<span class="go">ninja loop-convert</span>
+<span class="go">vim ~/test-files/simple-loops.cc</span>
+<span class="go">bin/loop-convert ~/test-files/simple-loops.cc</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="step-3-5-more-complicated-matchers">
+<h2>Step 3.5: More Complicated Matchers<a class="headerlink" href="#step-3-5-more-complicated-matchers" title="Permalink to this headline">¶</a></h2>
+<p>Our simple matcher is capable of discovering for loops, but we would
+still need to filter out many more ourselves. We can do a good portion
+of the remaining work with some cleverly chosen matchers, but first we
+need to decide exactly which properties we want to allow.</p>
+<p>How can we characterize for loops over arrays which would be eligible
+for translation to range-based syntax? Range based loops over arrays of
+size <tt class="docutils literal"><span class="pre">N</span></tt> that:</p>
+<ul class="simple">
+<li>start at index <tt class="docutils literal"><span class="pre">0</span></tt></li>
+<li>iterate consecutively</li>
+<li>end at index <tt class="docutils literal"><span class="pre">N-1</span></tt></li>
+</ul>
+<p>We already check for (1), so all we need to add is a check to the loop’s
+condition to ensure that the loop’s index variable is compared against
+<tt class="docutils literal"><span class="pre">N</span></tt> and another check to ensure that the increment step just
+increments this same variable. The matcher for (2) is straightforward:
+require a pre- or post-increment of the same variable declared in the
+init portion.</p>
+<p>Unfortunately, such a matcher is impossible to write. Matchers contain
+no logic for comparing two arbitrary AST nodes and determining whether
+or not they are equal, so the best we can do is matching more than we
+would like to allow, and punting extra comparisons to the callback.</p>
+<p>In any case, we can start building this sub-matcher. We can require that
+the increment step be a unary increment like this:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">hasIncrement</span><span class="p">(</span><span class="n">unaryOperator</span><span class="p">(</span><span class="n">hasOperatorName</span><span class="p">(</span><span class="s">"++"</span><span class="p">)))</span>
+</pre></div>
+</div>
+<p>Specifying what is incremented introduces another quirk of Clang’s AST:
+Usages of variables are represented as <tt class="docutils literal"><span class="pre">DeclRefExpr</span></tt>‘s (“declaration
+reference expressions”) because they are expressions which refer to
+variable declarations. To find a <tt class="docutils literal"><span class="pre">unaryOperator</span></tt> that refers to a
+specific declaration, we can simply add a second condition to it:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">hasIncrement</span><span class="p">(</span><span class="n">unaryOperator</span><span class="p">(</span>
+ <span class="n">hasOperatorName</span><span class="p">(</span><span class="s">"++"</span><span class="p">),</span>
+ <span class="n">hasUnaryOperand</span><span class="p">(</span><span class="n">declRefExpr</span><span class="p">())))</span>
+</pre></div>
+</div>
+<p>Furthermore, we can restrict our matcher to only match if the
+incremented variable is an integer:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">hasIncrement</span><span class="p">(</span><span class="n">unaryOperator</span><span class="p">(</span>
+ <span class="n">hasOperatorName</span><span class="p">(</span><span class="s">"++"</span><span class="p">),</span>
+ <span class="n">hasUnaryOperand</span><span class="p">(</span><span class="n">declRefExpr</span><span class="p">(</span><span class="n">to</span><span class="p">(</span><span class="n">varDecl</span><span class="p">(</span><span class="n">hasType</span><span class="p">(</span><span class="n">isInteger</span><span class="p">())))))))</span>
+</pre></div>
+</div>
+<p>And the last step will be to attach an identifier to this variable, so
+that we can retrieve it in the callback:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">hasIncrement</span><span class="p">(</span><span class="n">unaryOperator</span><span class="p">(</span>
+ <span class="n">hasOperatorName</span><span class="p">(</span><span class="s">"++"</span><span class="p">),</span>
+ <span class="n">hasUnaryOperand</span><span class="p">(</span><span class="n">declRefExpr</span><span class="p">(</span><span class="n">to</span><span class="p">(</span>
+ <span class="n">varDecl</span><span class="p">(</span><span class="n">hasType</span><span class="p">(</span><span class="n">isInteger</span><span class="p">())).</span><span class="n">bind</span><span class="p">(</span><span class="s">"incrementVariable"</span><span class="p">))))))</span>
+</pre></div>
+</div>
+<p>We can add this code to the definition of <tt class="docutils literal"><span class="pre">LoopMatcher</span></tt> and make sure
+that our program, outfitted with the new matcher, only prints out loops
+that declare a single variable initialized to zero and have an increment
+step consisting of a unary increment of some variable.</p>
+<p>Now, we just need to add a matcher to check if the condition part of the
+<tt class="docutils literal"><span class="pre">for</span></tt> loop compares a variable against the size of the array. There is
+only one problem - we don’t know which array we’re iterating over
+without looking at the body of the loop! We are again restricted to
+approximating the result we want with matchers, filling in the details
+in the callback. So we start with:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">hasCondition</span><span class="p">(</span><span class="n">binaryOperator</span><span class="p">(</span><span class="n">hasOperatorName</span><span class="p">(</span><span class="s">"<"</span><span class="p">))</span>
+</pre></div>
+</div>
+<p>It makes sense to ensure that the left-hand side is a reference to a
+variable, and that the right-hand side has integer type.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">hasCondition</span><span class="p">(</span><span class="n">binaryOperator</span><span class="p">(</span>
+ <span class="n">hasOperatorName</span><span class="p">(</span><span class="s">"<"</span><span class="p">),</span>
+ <span class="n">hasLHS</span><span class="p">(</span><span class="n">declRefExpr</span><span class="p">(</span><span class="n">to</span><span class="p">(</span><span class="n">varDecl</span><span class="p">(</span><span class="n">hasType</span><span class="p">(</span><span class="n">isInteger</span><span class="p">()))))),</span>
+ <span class="n">hasRHS</span><span class="p">(</span><span class="n">expr</span><span class="p">(</span><span class="n">hasType</span><span class="p">(</span><span class="n">isInteger</span><span class="p">())))))</span>
+</pre></div>
+</div>
+<p>Why? Because it doesn’t work. Of the three loops provided in
+<tt class="docutils literal"><span class="pre">test-files/simple.cpp</span></tt>, zero of them have a matching condition. A
+quick look at the AST dump of the first for loop, produced by the
+previous iteration of loop-convert, shows us the answer:</p>
+<div class="highlight-python"><div class="highlight"><pre>(ForStmt 0x173b240
+ (DeclStmt 0x173afc8
+ 0x173af50 "int i =
+ (IntegerLiteral 0x173afa8 'int' 0)")
+ <<>>
+ (BinaryOperator 0x173b060 '_Bool' '<'
+ (ImplicitCastExpr 0x173b030 'int'
+ (DeclRefExpr 0x173afe0 'int' lvalue Var 0x173af50 'i' 'int'))
+ (ImplicitCastExpr 0x173b048 'int'
+ (DeclRefExpr 0x173b008 'const int' lvalue Var 0x170fa80 'N' 'const int')))
+ (UnaryOperator 0x173b0b0 'int' lvalue prefix '++'
+ (DeclRefExpr 0x173b088 'int' lvalue Var 0x173af50 'i' 'int'))
+ (CompoundStatement ...
+</pre></div>
+</div>
+<p>We already know that the declaration and increments both match, or this
+loop wouldn’t have been dumped. The culprit lies in the implicit cast
+applied to the first operand (i.e. the LHS) of the less-than operator,
+an L-value to R-value conversion applied to the expression referencing
+<tt class="docutils literal"><span class="pre">i</span></tt>. Thankfully, the matcher library offers a solution to this problem
+in the form of <tt class="docutils literal"><span class="pre">ignoringParenImpCasts</span></tt>, which instructs the matcher to
+ignore implicit casts and parentheses before continuing to match.
+Adjusting the condition operator will restore the desired match.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">hasCondition</span><span class="p">(</span><span class="n">binaryOperator</span><span class="p">(</span>
+ <span class="n">hasOperatorName</span><span class="p">(</span><span class="s">"<"</span><span class="p">),</span>
+ <span class="n">hasLHS</span><span class="p">(</span><span class="n">ignoringParenImpCasts</span><span class="p">(</span><span class="n">declRefExpr</span><span class="p">(</span>
+ <span class="n">to</span><span class="p">(</span><span class="n">varDecl</span><span class="p">(</span><span class="n">hasType</span><span class="p">(</span><span class="n">isInteger</span><span class="p">())))))),</span>
+ <span class="n">hasRHS</span><span class="p">(</span><span class="n">expr</span><span class="p">(</span><span class="n">hasType</span><span class="p">(</span><span class="n">isInteger</span><span class="p">())))))</span>
+</pre></div>
+</div>
+<p>After adding binds to the expressions we wished to capture and
+extracting the identifier strings into variables, we have array-step-2
+completed.</p>
+</div>
+<div class="section" id="step-4-retrieving-matched-nodes">
+<h2>Step 4: Retrieving Matched Nodes<a class="headerlink" href="#step-4-retrieving-matched-nodes" title="Permalink to this headline">¶</a></h2>
+<p>So far, the matcher callback isn’t very interesting: it just dumps the
+loop’s AST. At some point, we will need to make changes to the input
+source code. Next, we’ll work on using the nodes we bound in the
+previous step.</p>
+<p>The <tt class="docutils literal"><span class="pre">MatchFinder::run()</span></tt> callback takes a
+<tt class="docutils literal"><span class="pre">MatchFinder::MatchResult&</span></tt> as its parameter. We’re most interested in
+its <tt class="docutils literal"><span class="pre">Context</span></tt> and <tt class="docutils literal"><span class="pre">Nodes</span></tt> members. Clang uses the <tt class="docutils literal"><span class="pre">ASTContext</span></tt>
+class to represent contextual information about the AST, as the name
+implies, though the most functionally important detail is that several
+operations require an <tt class="docutils literal"><span class="pre">ASTContext*</span></tt> parameter. More immediately useful
+is the set of matched nodes, and how we retrieve them.</p>
+<p>Since we bind three variables (identified by ConditionVarName,
+InitVarName, and IncrementVarName), we can obtain the matched nodes by
+using the <tt class="docutils literal"><span class="pre">getNodeAs()</span></tt> member function.</p>
+<p>In <tt class="docutils literal"><span class="pre">LoopConvert.cpp</span></tt> add</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#include "clang/AST/ASTContext.h"</span>
+</pre></div>
+</div>
+<p>Change <tt class="docutils literal"><span class="pre">LoopMatcher</span></tt> to</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">StatementMatcher</span> <span class="n">LoopMatcher</span> <span class="o">=</span>
+ <span class="n">forStmt</span><span class="p">(</span><span class="n">hasLoopInit</span><span class="p">(</span><span class="n">declStmt</span><span class="p">(</span>
+ <span class="n">hasSingleDecl</span><span class="p">(</span><span class="n">varDecl</span><span class="p">(</span><span class="n">hasInitializer</span><span class="p">(</span><span class="n">integerLiteral</span><span class="p">(</span><span class="n">equals</span><span class="p">(</span><span class="mi">0</span><span class="p">))))</span>
+ <span class="p">.</span><span class="n">bind</span><span class="p">(</span><span class="s">"initVarName"</span><span class="p">)))),</span>
+ <span class="n">hasIncrement</span><span class="p">(</span><span class="n">unaryOperator</span><span class="p">(</span>
+ <span class="n">hasOperatorName</span><span class="p">(</span><span class="s">"++"</span><span class="p">),</span>
+ <span class="n">hasUnaryOperand</span><span class="p">(</span><span class="n">declRefExpr</span><span class="p">(</span>
+ <span class="n">to</span><span class="p">(</span><span class="n">varDecl</span><span class="p">(</span><span class="n">hasType</span><span class="p">(</span><span class="n">isInteger</span><span class="p">())).</span><span class="n">bind</span><span class="p">(</span><span class="s">"incVarName"</span><span class="p">)))))),</span>
+ <span class="n">hasCondition</span><span class="p">(</span><span class="n">binaryOperator</span><span class="p">(</span>
+ <span class="n">hasOperatorName</span><span class="p">(</span><span class="s">"<"</span><span class="p">),</span>
+ <span class="n">hasLHS</span><span class="p">(</span><span class="n">ignoringParenImpCasts</span><span class="p">(</span><span class="n">declRefExpr</span><span class="p">(</span>
+ <span class="n">to</span><span class="p">(</span><span class="n">varDecl</span><span class="p">(</span><span class="n">hasType</span><span class="p">(</span><span class="n">isInteger</span><span class="p">())).</span><span class="n">bind</span><span class="p">(</span><span class="s">"condVarName"</span><span class="p">))))),</span>
+ <span class="n">hasRHS</span><span class="p">(</span><span class="n">expr</span><span class="p">(</span><span class="n">hasType</span><span class="p">(</span><span class="n">isInteger</span><span class="p">())))))).</span><span class="n">bind</span><span class="p">(</span><span class="s">"forLoop"</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>And change <tt class="docutils literal"><span class="pre">LoopPrinter::run</span></tt> to</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="kt">void</span> <span class="n">LoopPrinter</span><span class="o">::</span><span class="n">run</span><span class="p">(</span><span class="k">const</span> <span class="n">MatchFinder</span><span class="o">::</span><span class="n">MatchResult</span> <span class="o">&</span><span class="n">Result</span><span class="p">)</span> <span class="p">{</span>
+ <span class="n">ASTContext</span> <span class="o">*</span><span class="n">Context</span> <span class="o">=</span> <span class="n">Result</span><span class="p">.</span><span class="n">Context</span><span class="p">;</span>
+ <span class="k">const</span> <span class="n">ForStmt</span> <span class="o">*</span><span class="n">FS</span> <span class="o">=</span> <span class="n">Result</span><span class="p">.</span><span class="n">Nodes</span><span class="p">.</span><span class="n">getStmtAs</span><span class="o"><</span><span class="n">ForStmt</span><span class="o">></span><span class="p">(</span><span class="s">"forLoop"</span><span class="p">);</span>
+ <span class="c1">// We do not want to convert header files!</span>
+ <span class="k">if</span> <span class="p">(</span><span class="o">!</span><span class="n">FS</span> <span class="o">||</span> <span class="o">!</span><span class="n">Context</span><span class="o">-></span><span class="n">getSourceManager</span><span class="p">().</span><span class="n">isFromMainFile</span><span class="p">(</span><span class="n">FS</span><span class="o">-></span><span class="n">getForLoc</span><span class="p">()))</span>
+ <span class="k">return</span><span class="p">;</span>
+ <span class="k">const</span> <span class="n">VarDecl</span> <span class="o">*</span><span class="n">IncVar</span> <span class="o">=</span> <span class="n">Result</span><span class="p">.</span><span class="n">Nodes</span><span class="p">.</span><span class="n">getNodeAs</span><span class="o"><</span><span class="n">VarDecl</span><span class="o">></span><span class="p">(</span><span class="s">"incVarName"</span><span class="p">);</span>
+ <span class="k">const</span> <span class="n">VarDecl</span> <span class="o">*</span><span class="n">CondVar</span> <span class="o">=</span> <span class="n">Result</span><span class="p">.</span><span class="n">Nodes</span><span class="p">.</span><span class="n">getNodeAs</span><span class="o"><</span><span class="n">VarDecl</span><span class="o">></span><span class="p">(</span><span class="s">"condVarName"</span><span class="p">);</span>
+ <span class="k">const</span> <span class="n">VarDecl</span> <span class="o">*</span><span class="n">InitVar</span> <span class="o">=</span> <span class="n">Result</span><span class="p">.</span><span class="n">Nodes</span><span class="p">.</span><span class="n">getNodeAs</span><span class="o"><</span><span class="n">VarDecl</span><span class="o">></span><span class="p">(</span><span class="s">"initVarName"</span><span class="p">);</span>
+
+ <span class="k">if</span> <span class="p">(</span><span class="o">!</span><span class="n">areSameVariable</span><span class="p">(</span><span class="n">IncVar</span><span class="p">,</span> <span class="n">CondVar</span><span class="p">)</span> <span class="o">||</span> <span class="o">!</span><span class="n">areSameVariable</span><span class="p">(</span><span class="n">IncVar</span><span class="p">,</span> <span class="n">InitVar</span><span class="p">))</span>
+ <span class="k">return</span><span class="p">;</span>
+ <span class="n">llvm</span><span class="o">::</span><span class="n">outs</span><span class="p">()</span> <span class="o"><<</span> <span class="s">"Potential array-based loop discovered.</span><span class="se">\n</span><span class="s">"</span><span class="p">;</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>Clang associates a <tt class="docutils literal"><span class="pre">VarDecl</span></tt> with each variable to represent the variable’s
+declaration. Since the “canonical” form of each declaration is unique by
+address, all we need to do is make sure neither <tt class="docutils literal"><span class="pre">ValueDecl</span></tt> (base class of
+<tt class="docutils literal"><span class="pre">VarDecl</span></tt>) is <tt class="docutils literal"><span class="pre">NULL</span></tt> and compare the canonical Decls.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">static</span> <span class="kt">bool</span> <span class="nf">areSameVariable</span><span class="p">(</span><span class="k">const</span> <span class="n">ValueDecl</span> <span class="o">*</span><span class="n">First</span><span class="p">,</span> <span class="k">const</span> <span class="n">ValueDecl</span> <span class="o">*</span><span class="n">Second</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">return</span> <span class="n">First</span> <span class="o">&&</span> <span class="n">Second</span> <span class="o">&&</span>
+ <span class="n">First</span><span class="o">-></span><span class="n">getCanonicalDecl</span><span class="p">()</span> <span class="o">==</span> <span class="n">Second</span><span class="o">-></span><span class="n">getCanonicalDecl</span><span class="p">();</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>If execution reaches the end of <tt class="docutils literal"><span class="pre">LoopPrinter::run()</span></tt>, we know that the
+loop shell that looks like</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">for</span> <span class="p">(</span><span class="kt">int</span> <span class="n">i</span><span class="o">=</span> <span class="mi">0</span><span class="p">;</span> <span class="n">i</span> <span class="o"><</span> <span class="n">expr</span><span class="p">();</span> <span class="o">++</span><span class="n">i</span><span class="p">)</span> <span class="p">{</span> <span class="p">...</span> <span class="p">}</span>
+</pre></div>
+</div>
+<p>For now, we will just print a message explaining that we found a loop.
+The next section will deal with recursively traversing the AST to
+discover all changes needed.</p>
+<p>As a side note, it’s not as trivial to test if two expressions are the same,
+though Clang has already done the hard work for us by providing a way to
+canonicalize expressions:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">static</span> <span class="kt">bool</span> <span class="nf">areSameExpr</span><span class="p">(</span><span class="n">ASTContext</span> <span class="o">*</span><span class="n">Context</span><span class="p">,</span> <span class="k">const</span> <span class="n">Expr</span> <span class="o">*</span><span class="n">First</span><span class="p">,</span>
+ <span class="k">const</span> <span class="n">Expr</span> <span class="o">*</span><span class="n">Second</span><span class="p">)</span> <span class="p">{</span>
+ <span class="k">if</span> <span class="p">(</span><span class="o">!</span><span class="n">First</span> <span class="o">||</span> <span class="o">!</span><span class="n">Second</span><span class="p">)</span>
+ <span class="k">return</span> <span class="nb">false</span><span class="p">;</span>
+ <span class="n">llvm</span><span class="o">::</span><span class="n">FoldingSetNodeID</span> <span class="n">FirstID</span><span class="p">,</span> <span class="n">SecondID</span><span class="p">;</span>
+ <span class="n">First</span><span class="o">-></span><span class="n">Profile</span><span class="p">(</span><span class="n">FirstID</span><span class="p">,</span> <span class="o">*</span><span class="n">Context</span><span class="p">,</span> <span class="nb">true</span><span class="p">);</span>
+ <span class="n">Second</span><span class="o">-></span><span class="n">Profile</span><span class="p">(</span><span class="n">SecondID</span><span class="p">,</span> <span class="o">*</span><span class="n">Context</span><span class="p">,</span> <span class="nb">true</span><span class="p">);</span>
+ <span class="k">return</span> <span class="n">FirstID</span> <span class="o">==</span> <span class="n">SecondID</span><span class="p">;</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>This code relies on the comparison between two
+<tt class="docutils literal"><span class="pre">llvm::FoldingSetNodeIDs</span></tt>. As the documentation for
+<tt class="docutils literal"><span class="pre">Stmt::Profile()</span></tt> indicates, the <tt class="docutils literal"><span class="pre">Profile()</span></tt> member function builds
+a description of a node in the AST, based on its properties, along with
+those of its children. <tt class="docutils literal"><span class="pre">FoldingSetNodeID</span></tt> then serves as a hash we can
+use to compare expressions. We will need <tt class="docutils literal"><span class="pre">areSameExpr</span></tt> later. Before
+you run the new code on the additional loops added to
+test-files/simple.cpp, try to figure out which ones will be considered
+potentially convertible.</p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="RAVFrontendAction.html">How to write RecursiveASTVisitor based ASTFrontendActions.</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="LibASTMatchers.html">Matching the Clang AST</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/LibFormat.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/LibFormat.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/LibFormat.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/LibFormat.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,112 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>LibFormat — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Clang Plugins" href="ClangPlugins.html" />
+ <link rel="prev" title="LibTooling" href="LibTooling.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>LibFormat</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="LibTooling.html">LibTooling</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ClangPlugins.html">Clang Plugins</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="libformat">
+<h1>LibFormat<a class="headerlink" href="#libformat" title="Permalink to this headline">¶</a></h1>
+<p>LibFormat is a library that implements automatic source code formatting based
+on Clang. This documents describes the LibFormat interface and design as well
+as some basic style discussions.</p>
+<p>If you just want to use <cite>clang-format</cite> as a tool or integrated into an editor,
+checkout <a class="reference internal" href="ClangFormat.html"><em>ClangFormat</em></a>.</p>
+<div class="section" id="design">
+<h2>Design<a class="headerlink" href="#design" title="Permalink to this headline">¶</a></h2>
+<p>FIXME: Write up design.</p>
+</div>
+<div class="section" id="interface">
+<h2>Interface<a class="headerlink" href="#interface" title="Permalink to this headline">¶</a></h2>
+<p>The core routine of LibFormat is <tt class="docutils literal"><span class="pre">reformat()</span></tt>:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="n">tooling</span><span class="o">::</span><span class="n">Replacements</span> <span class="n">reformat</span><span class="p">(</span><span class="k">const</span> <span class="n">FormatStyle</span> <span class="o">&</span><span class="n">Style</span><span class="p">,</span> <span class="n">Lexer</span> <span class="o">&</span><span class="n">Lex</span><span class="p">,</span>
+ <span class="n">SourceManager</span> <span class="o">&</span><span class="n">SourceMgr</span><span class="p">,</span>
+ <span class="n">std</span><span class="o">::</span><span class="n">vector</span><span class="o"><</span><span class="n">CharSourceRange</span><span class="o">></span> <span class="n">Ranges</span><span class="p">);</span>
+</pre></div>
+</div>
+<p>This reads a token stream out of the lexer <tt class="docutils literal"><span class="pre">Lex</span></tt> and reformats all the code
+ranges in <tt class="docutils literal"><span class="pre">Ranges</span></tt>. The <tt class="docutils literal"><span class="pre">FormatStyle</span></tt> controls basic decisions made during
+formatting. A list of options can be found under <a class="reference internal" href="#style-options"><em>Style Options</em></a>.</p>
+</div>
+<div class="section" id="style-options">
+<span id="id1"></span><h2>Style Options<a class="headerlink" href="#style-options" title="Permalink to this headline">¶</a></h2>
+<p>The style options describe specific formatting options that can be used in
+order to make <cite>ClangFormat</cite> comply with different style guides. Currently,
+two style guides are hard-coded:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="c1">/// \brief Returns a format style complying with the LLVM coding standards:</span>
+<span class="c1">/// http://llvm.org/docs/CodingStandards.html.</span>
+<span class="n">FormatStyle</span> <span class="nf">getLLVMStyle</span><span class="p">();</span>
+
+<span class="c1">/// \brief Returns a format style complying with Google's C++ style guide:</span>
+<span class="c1">/// http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml.</span>
+<span class="n">FormatStyle</span> <span class="nf">getGoogleStyle</span><span class="p">();</span>
+</pre></div>
+</div>
+<p>These options are also exposed in the <a class="reference internal" href="ClangFormat.html"><em>standalone tools</em></a>
+through the <cite>-style</cite> option.</p>
+<p>In the future, we plan on making this configurable.</p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="LibTooling.html">LibTooling</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="ClangPlugins.html">Clang Plugins</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/LibTooling.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/LibTooling.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/LibTooling.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/LibTooling.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,242 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>LibTooling — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="LibFormat" href="LibFormat.html" />
+ <link rel="prev" title="Introduction to the Clang AST" href="IntroductionToTheClangAST.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>LibTooling</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="IntroductionToTheClangAST.html">Introduction to the Clang AST</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="LibFormat.html">LibFormat</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="libtooling">
+<h1>LibTooling<a class="headerlink" href="#libtooling" title="Permalink to this headline">¶</a></h1>
+<p>LibTooling is a library to support writing standalone tools based on Clang.
+This document will provide a basic walkthrough of how to write a tool using
+LibTooling.</p>
+<p>For the information on how to setup Clang Tooling for LLVM see
+<a class="reference internal" href="HowToSetupToolingForLLVM.html"><em>How To Setup Clang Tooling For LLVM</em></a></p>
+<div class="section" id="introduction">
+<h2>Introduction<a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>Tools built with LibTooling, like Clang Plugins, run <tt class="docutils literal"><span class="pre">FrontendActions</span></tt> over
+code.</p>
+<p>In this tutorial, we’ll demonstrate the different ways of running Clang’s
+<tt class="docutils literal"><span class="pre">SyntaxOnlyAction</span></tt>, which runs a quick syntax check, over a bunch of code.</p>
+</div>
+<div class="section" id="parsing-a-code-snippet-in-memory">
+<h2>Parsing a code snippet in memory<a class="headerlink" href="#parsing-a-code-snippet-in-memory" title="Permalink to this headline">¶</a></h2>
+<p>If you ever wanted to run a <tt class="docutils literal"><span class="pre">FrontendAction</span></tt> over some sample code, for
+example to unit test parts of the Clang AST, <tt class="docutils literal"><span class="pre">runToolOnCode</span></tt> is what you
+looked for. Let me give you an example:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#include "clang/Tooling/Tooling.h"</span>
+
+<span class="n">TEST</span><span class="p">(</span><span class="n">runToolOnCode</span><span class="p">,</span> <span class="n">CanSyntaxCheckCode</span><span class="p">)</span> <span class="p">{</span>
+ <span class="c1">// runToolOnCode returns whether the action was correctly run over the</span>
+ <span class="c1">// given code.</span>
+ <span class="n">EXPECT_TRUE</span><span class="p">(</span><span class="n">runToolOnCode</span><span class="p">(</span><span class="k">new</span> <span class="n">clang</span><span class="o">::</span><span class="n">SyntaxOnlyAction</span><span class="p">,</span> <span class="s">"class X {};"</span><span class="p">));</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="writing-a-standalone-tool">
+<h2>Writing a standalone tool<a class="headerlink" href="#writing-a-standalone-tool" title="Permalink to this headline">¶</a></h2>
+<p>Once you unit tested your <tt class="docutils literal"><span class="pre">FrontendAction</span></tt> to the point where it cannot
+possibly break, it’s time to create a standalone tool. For a standalone tool
+to run clang, it first needs to figure out what command line arguments to use
+for a specified file. To that end we create a <tt class="docutils literal"><span class="pre">CompilationDatabase</span></tt>. There
+are different ways to create a compilation database, and we need to support all
+of them depending on command-line options. There’s the <tt class="docutils literal"><span class="pre">CommonOptionsParser</span></tt>
+class that takes the responsibility to parse command-line parameters related to
+compilation databases and inputs, so that all tools share the implementation.</p>
+<div class="section" id="parsing-common-tools-options">
+<h3>Parsing common tools options<a class="headerlink" href="#parsing-common-tools-options" title="Permalink to this headline">¶</a></h3>
+<p><tt class="docutils literal"><span class="pre">CompilationDatabase</span></tt> can be read from a build directory or the command line.
+Using <tt class="docutils literal"><span class="pre">CommonOptionsParser</span></tt> allows for explicit specification of a compile
+command line, specification of build path using the <tt class="docutils literal"><span class="pre">-p</span></tt> command-line option,
+and automatic location of the compilation database using source files paths.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#include "clang/Tooling/CommonOptionsParser.h"</span>
+<span class="cp">#include "llvm/Support/CommandLine.h"</span>
+
+<span class="k">using</span> <span class="k">namespace</span> <span class="n">clang</span><span class="o">::</span><span class="n">tooling</span><span class="p">;</span>
+
+<span class="c1">// Apply a custom category to all command-line options so that they are the</span>
+<span class="c1">// only ones displayed.</span>
+<span class="k">static</span> <span class="n">llvm</span><span class="o">::</span><span class="n">cl</span><span class="o">::</span><span class="n">OptionCategory</span> <span class="n">MyToolCategory</span><span class="p">(</span><span class="s">"my-tool options"</span><span class="p">);</span>
+
+<span class="kt">int</span> <span class="nf">main</span><span class="p">(</span><span class="kt">int</span> <span class="n">argc</span><span class="p">,</span> <span class="k">const</span> <span class="kt">char</span> <span class="o">**</span><span class="n">argv</span><span class="p">)</span> <span class="p">{</span>
+ <span class="c1">// CommonOptionsParser constructor will parse arguments and create a</span>
+ <span class="c1">// CompilationDatabase. In case of error it will terminate the program.</span>
+ <span class="n">CommonOptionsParser</span> <span class="n">OptionsParser</span><span class="p">(</span><span class="n">argc</span><span class="p">,</span> <span class="n">argv</span><span class="p">,</span> <span class="n">MyToolCategory</span><span class="p">);</span>
+
+ <span class="c1">// Use OptionsParser.getCompilations() and OptionsParser.getSourcePathList()</span>
+ <span class="c1">// to retrieve CompilationDatabase and the list of input file paths.</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="creating-and-running-a-clangtool">
+<h3>Creating and running a ClangTool<a class="headerlink" href="#creating-and-running-a-clangtool" title="Permalink to this headline">¶</a></h3>
+<p>Once we have a <tt class="docutils literal"><span class="pre">CompilationDatabase</span></tt>, we can create a <tt class="docutils literal"><span class="pre">ClangTool</span></tt> and run
+our <tt class="docutils literal"><span class="pre">FrontendAction</span></tt> over some code. For example, to run the
+<tt class="docutils literal"><span class="pre">SyntaxOnlyAction</span></tt> over the files “a.cc” and “b.cc” one would write:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="c1">// A clang tool can run over a number of sources in the same process...</span>
+<span class="n">std</span><span class="o">::</span><span class="n">vector</span><span class="o"><</span><span class="n">std</span><span class="o">::</span><span class="n">string</span><span class="o">></span> <span class="n">Sources</span><span class="p">;</span>
+<span class="n">Sources</span><span class="p">.</span><span class="n">push_back</span><span class="p">(</span><span class="s">"a.cc"</span><span class="p">);</span>
+<span class="n">Sources</span><span class="p">.</span><span class="n">push_back</span><span class="p">(</span><span class="s">"b.cc"</span><span class="p">);</span>
+
+<span class="c1">// We hand the CompilationDatabase we created and the sources to run over into</span>
+<span class="c1">// the tool constructor.</span>
+<span class="n">ClangTool</span> <span class="nf">Tool</span><span class="p">(</span><span class="n">OptionsParser</span><span class="p">.</span><span class="n">getCompilations</span><span class="p">(),</span> <span class="n">Sources</span><span class="p">);</span>
+
+<span class="c1">// The ClangTool needs a new FrontendAction for each translation unit we run</span>
+<span class="c1">// on. Thus, it takes a FrontendActionFactory as parameter. To create a</span>
+<span class="c1">// FrontendActionFactory from a given FrontendAction type, we call</span>
+<span class="c1">// newFrontendActionFactory<clang::SyntaxOnlyAction>().</span>
+<span class="kt">int</span> <span class="n">result</span> <span class="o">=</span> <span class="n">Tool</span><span class="p">.</span><span class="n">run</span><span class="p">(</span><span class="n">newFrontendActionFactory</span><span class="o"><</span><span class="n">clang</span><span class="o">::</span><span class="n">SyntaxOnlyAction</span><span class="o">></span><span class="p">().</span><span class="n">get</span><span class="p">());</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="putting-it-together-the-first-tool">
+<h3>Putting it together — the first tool<a class="headerlink" href="#putting-it-together-the-first-tool" title="Permalink to this headline">¶</a></h3>
+<p>Now we combine the two previous steps into our first real tool. A more advanced
+version of this example tool is also checked into the clang tree at
+<tt class="docutils literal"><span class="pre">tools/clang-check/ClangCheck.cpp</span></tt>.</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="c1">// Declares clang::SyntaxOnlyAction.</span>
+<span class="cp">#include "clang/Frontend/FrontendActions.h"</span>
+<span class="cp">#include "clang/Tooling/CommonOptionsParser.h"</span>
+<span class="cp">#include "clang/Tooling/Tooling.h"</span>
+<span class="c1">// Declares llvm::cl::extrahelp.</span>
+<span class="cp">#include "llvm/Support/CommandLine.h"</span>
+
+<span class="k">using</span> <span class="k">namespace</span> <span class="n">clang</span><span class="o">::</span><span class="n">tooling</span><span class="p">;</span>
+<span class="k">using</span> <span class="k">namespace</span> <span class="n">llvm</span><span class="p">;</span>
+
+<span class="c1">// Apply a custom category to all command-line options so that they are the</span>
+<span class="c1">// only ones displayed.</span>
+<span class="k">static</span> <span class="n">cl</span><span class="o">::</span><span class="n">OptionCategory</span> <span class="n">MyToolCategory</span><span class="p">(</span><span class="s">"my-tool options"</span><span class="p">);</span>
+
+<span class="c1">// CommonOptionsParser declares HelpMessage with a description of the common</span>
+<span class="c1">// command-line options related to the compilation database and input files.</span>
+<span class="c1">// It's nice to have this help message in all tools.</span>
+<span class="k">static</span> <span class="n">cl</span><span class="o">::</span><span class="n">extrahelp</span> <span class="n">CommonHelp</span><span class="p">(</span><span class="n">CommonOptionsParser</span><span class="o">::</span><span class="n">HelpMessage</span><span class="p">);</span>
+
+<span class="c1">// A help message for this specific tool can be added afterwards.</span>
+<span class="k">static</span> <span class="n">cl</span><span class="o">::</span><span class="n">extrahelp</span> <span class="n">MoreHelp</span><span class="p">(</span><span class="s">"</span><span class="se">\n</span><span class="s">More help text..."</span><span class="p">);</span>
+
+<span class="kt">int</span> <span class="nf">main</span><span class="p">(</span><span class="kt">int</span> <span class="n">argc</span><span class="p">,</span> <span class="k">const</span> <span class="kt">char</span> <span class="o">**</span><span class="n">argv</span><span class="p">)</span> <span class="p">{</span>
+ <span class="n">CommonOptionsParser</span> <span class="n">OptionsParser</span><span class="p">(</span><span class="n">argc</span><span class="p">,</span> <span class="n">argv</span><span class="p">,</span> <span class="n">MyToolCategory</span><span class="p">);</span>
+ <span class="n">ClangTool</span> <span class="n">Tool</span><span class="p">(</span><span class="n">OptionsParser</span><span class="p">.</span><span class="n">getCompilations</span><span class="p">(),</span>
+ <span class="n">OptionsParser</span><span class="p">.</span><span class="n">getSourcePathList</span><span class="p">());</span>
+ <span class="k">return</span> <span class="n">Tool</span><span class="p">.</span><span class="n">run</span><span class="p">(</span><span class="n">newFrontendActionFactory</span><span class="o"><</span><span class="n">clang</span><span class="o">::</span><span class="n">SyntaxOnlyAction</span><span class="o">></span><span class="p">().</span><span class="n">get</span><span class="p">());</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="running-the-tool-on-some-code">
+<h3>Running the tool on some code<a class="headerlink" href="#running-the-tool-on-some-code" title="Permalink to this headline">¶</a></h3>
+<p>When you check out and build clang, clang-check is already built and available
+to you in bin/clang-check inside your build directory.</p>
+<p>You can run clang-check on a file in the llvm repository by specifying all the
+needed parameters after a “<tt class="docutils literal"><span class="pre">--</span></tt>” separator:</p>
+<div class="highlight-bash"><div class="highlight"><pre><span class="nv">$ </span><span class="nb">cd</span> /path/to/source/llvm
+<span class="nv">$ </span><span class="nb">export </span><span class="nv">BD</span><span class="o">=</span>/path/to/build/llvm
+<span class="nv">$ $BD</span>/bin/clang-check tools/clang/tools/clang-check/ClangCheck.cpp -- <span class="se">\</span>
+ clang++ -D__STDC_CONSTANT_MACROS -D__STDC_LIMIT_MACROS <span class="se">\</span>
+ -Itools/clang/include -I<span class="nv">$BD</span>/include -Iinclude <span class="se">\</span>
+ -Itools/clang/lib/Headers -c
+</pre></div>
+</div>
+<p>As an alternative, you can also configure cmake to output a compile command
+database into its build directory:</p>
+<div class="highlight-bash"><div class="highlight"><pre><span class="c"># Alternatively to calling cmake, use ccmake, toggle to advanced mode and</span>
+<span class="c"># set the parameter CMAKE_EXPORT_COMPILE_COMMANDS from the UI.</span>
+<span class="nv">$ </span>cmake -DCMAKE_EXPORT_COMPILE_COMMANDS<span class="o">=</span>ON .
+</pre></div>
+</div>
+<p>This creates a file called <tt class="docutils literal"><span class="pre">compile_commands.json</span></tt> in the build directory.
+Now you can run <strong class="program">clang-check</strong> over files in the project by specifying
+the build path as first argument and some source files as further positional
+arguments:</p>
+<div class="highlight-bash"><div class="highlight"><pre><span class="nv">$ </span><span class="nb">cd</span> /path/to/source/llvm
+<span class="nv">$ </span><span class="nb">export </span><span class="nv">BD</span><span class="o">=</span>/path/to/build/llvm
+<span class="nv">$ $BD</span>/bin/clang-check -p <span class="nv">$BD</span> tools/clang/tools/clang-check/ClangCheck.cpp
+</pre></div>
+</div>
+</div>
+<div class="section" id="builtin-includes">
+<span id="libtooling-builtin-includes"></span><h3>Builtin includes<a class="headerlink" href="#builtin-includes" title="Permalink to this headline">¶</a></h3>
+<p>Clang tools need their builtin headers and search for them the same way Clang
+does. Thus, the default location to look for builtin headers is in a path
+<tt class="docutils literal"><span class="pre">$(dirname</span> <span class="pre">/path/to/tool)/../lib/clang/3.3/include</span></tt> relative to the tool
+binary. This works out-of-the-box for tools running from llvm’s toplevel
+binary directory after building clang-headers, or if the tool is running from
+the binary directory of a clang install next to the clang binary.</p>
+<p>Tips: if your tool fails to find <tt class="docutils literal"><span class="pre">stddef.h</span></tt> or similar headers, call the tool
+with <tt class="docutils literal"><span class="pre">-v</span></tt> and look at the search paths it looks through.</p>
+</div>
+<div class="section" id="linking">
+<h3>Linking<a class="headerlink" href="#linking" title="Permalink to this headline">¶</a></h3>
+<p>For a list of libraries to link, look at one of the tools’ Makefiles (for
+example <a class="reference external" href="http://llvm.org/viewvc/llvm-project/cfe/trunk/tools/clang-check/Makefile?view=markup">clang-check/Makefile</a>).</p>
+</div>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="IntroductionToTheClangAST.html">Introduction to the Clang AST</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="LibFormat.html">LibFormat</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/MSVCCompatibility.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/MSVCCompatibility.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/MSVCCompatibility.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/MSVCCompatibility.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,190 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>MSVC compatibility — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Clang âmanâ pages" href="CommandGuide/index.html" />
+ <link rel="prev" title="Modules" href="Modules.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>MSVC compatibility</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="Modules.html">Modules</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="CommandGuide/index.html">Clang “man” pages</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <style type="text/css">
+ .none { background-color: #FFCCCC }
+ .partial { background-color: #FFFF99 }
+ .good { background-color: #CCFF99 }
+</style><div class="section" id="msvc-compatibility">
+<h1>MSVC compatibility<a class="headerlink" href="#msvc-compatibility" title="Permalink to this headline">¶</a></h1>
+<p>When Clang compiles C++ code for Windows, it attempts to be compatible with
+MSVC. There are multiple dimensions to compatibility.</p>
+<p>First, Clang attempts to be ABI-compatible, meaning that Clang-compiled code
+should be able to link against MSVC-compiled code successfully. However, C++
+ABIs are particularly large and complicated, and Clang’s support for MSVC’s C++
+ABI is a work in progress. If you don’t require MSVC ABI compatibility or don’t
+want to use Microsoft’s C and C++ runtimes, the mingw32 toolchain might be a
+better fit for your project.</p>
+<p>Second, Clang implements many MSVC language extensions, such as
+<tt class="docutils literal"><span class="pre">__declspec(dllexport)</span></tt> and a handful of pragmas. These are typically
+controlled by <tt class="docutils literal"><span class="pre">-fms-extensions</span></tt>.</p>
+<p>Third, MSVC accepts some C++ code that Clang will typically diagnose as
+invalid. When these constructs are present in widely included system headers,
+Clang attempts to recover and continue compiling the user’s program. Most
+parsing and semantic compatibility tweaks are controlled by
+<tt class="docutils literal"><span class="pre">-fms-compatibility</span></tt> and <tt class="docutils literal"><span class="pre">-fdelayed-template-parsing</span></tt>, and they are a work
+in progress.</p>
+<p>Finally, there is <a class="reference internal" href="UsersManual.html#clang-cl"><em>clang-cl</em></a>, a driver program for clang that attempts to
+be compatible with MSVC’s cl.exe.</p>
+<div class="section" id="abi-features">
+<h2>ABI features<a class="headerlink" href="#abi-features" title="Permalink to this headline">¶</a></h2>
+<p>The status of major ABI-impacting C++ features:</p>
+<ul class="simple">
+<li>Record layout: <span class="good">Complete</span>. We’ve tested this with a fuzzer and have
+fixed all known bugs.</li>
+<li>Class inheritance: <span class="good">Mostly complete</span>. This covers all of the standard
+OO features you would expect: virtual method inheritance, multiple
+inheritance, and virtual inheritance. Every so often we uncover a bug where
+our tables are incompatible, but this is pretty well in hand. This feature
+has also been fuzz tested.</li>
+<li>Name mangling: <span class="good">Ongoing</span>. Every new C++ feature generally needs its own
+mangling. For example, member pointer template arguments have an interesting
+and distinct mangling. Fortunately, incorrect manglings usually do not result
+in runtime errors. Non-inline functions with incorrect manglings usually
+result in link errors, which are relatively easy to diagnose. Incorrect
+manglings for inline functions and templates result in multiple copies in the
+final image. The C++ standard requires that those addresses be equal, but few
+programs rely on this.</li>
+<li>Member pointers: <span class="good">Mostly complete</span>. Standard C++ member pointers are
+fully implemented and should be ABI compatible. Both <a class="reference external" href="http://msdn.microsoft.com/en-us/library/83cch5a6.aspx">#pragma
+pointers_to_members</a> and the <a class="reference external" href="http://msdn.microsoft.com/en-us/library/yad46a6z.aspx">/vm</a> flags are supported. However, MSVC
+supports an extension to allow creating a <a class="reference external" href="http://llvm.org/PR15713">pointer to a member of a virtual
+base class</a>. Clang does not yet support this.</li>
+</ul>
+<ul class="simple">
+<li>Debug info: <span class="partial">Minimal</span>. Clang emits both CodeView line tables
+(similar to what MSVC emits when given the <tt class="docutils literal"><span class="pre">/Z7</span></tt> flag) and DWARF debug
+information into the object file.
+Microsoft’s link.exe will transform the CodeView line tables into a PDB,
+enabling stack traces in all modern Windows debuggers. Clang does not emit
+any CodeView-compatible type info or description of variable layout.
+Binaries linked with either binutils’ ld or LLVM’s lld should be usable with
+GDB however sophisticated C++ expressions are likely to fail.</li>
+<li>RTTI: <span class="good">Complete</span>. Generation of RTTI data structures has been
+finished, along with support for the <tt class="docutils literal"><span class="pre">/GR</span></tt> flag.</li>
+<li>Exceptions and SEH: <span class="partial">Partial</span>.
+C++ exceptions (<tt class="docutils literal"><span class="pre">try</span></tt> / <tt class="docutils literal"><span class="pre">catch</span></tt> / <tt class="docutils literal"><span class="pre">throw</span></tt>) and
+structured exceptions (<tt class="docutils literal"><span class="pre">__try</span></tt> / <tt class="docutils literal"><span class="pre">__except</span></tt> / <tt class="docutils literal"><span class="pre">__finally</span></tt>) mostly
+work on x64. 32-bit exception handling support is being worked on. LLVM does
+not model asynchronous exceptions, so it is currently impossible to catch an
+asynchronous exception generated in the same frame as the catching <tt class="docutils literal"><span class="pre">__try</span></tt>.
+C++ exception specifications are ignored, but this is <a class="reference external" href="https://msdn.microsoft.com/en-us/library/wfa0edys.aspx">consistent with Visual
+C++</a>.</li>
+</ul>
+<ul class="simple">
+<li>Thread-safe initialization of local statics: <span class="good">Complete</span>. MSVC 2015
+added support for thread-safe initialization of such variables by taking an
+ABI break.
+We are ABI compatible with both the MSVC 2013 and 2015 ABI for static local
+variables.</li>
+<li>Lambdas: <span class="good">Mostly complete</span>. Clang is compatible with Microsoft’s
+implementation of lambdas except for providing overloads for conversion to
+function pointer for different calling conventions. However, Microsoft’s
+extension is non-conforming.</li>
+</ul>
+</div>
+<div class="section" id="template-instantiation-and-name-lookup">
+<h2>Template instantiation and name lookup<a class="headerlink" href="#template-instantiation-and-name-lookup" title="Permalink to this headline">¶</a></h2>
+<p>MSVC allows many invalid constructs in class templates that Clang has
+historically rejected. In order to parse widely distributed headers for
+libraries such as the Active Template Library (ATL) and Windows Runtime Library
+(WRL), some template rules have been relaxed or extended in Clang on Windows.</p>
+<p>The first major semantic difference is that MSVC appears to defer all parsing
+an analysis of inline method bodies in class templates until instantiation
+time. By default on Windows, Clang attempts to follow suit. This behavior is
+controlled by the <tt class="docutils literal"><span class="pre">-fdelayed-template-parsing</span></tt> flag. While Clang delays
+parsing of method bodies, it still parses the bodies <em>before</em> template argument
+substitution, which is not what MSVC does. The following compatibility tweaks
+are necessary to parse the template in those cases.</p>
+<p>MSVC allows some name lookup into dependent base classes. Even on other
+platforms, this has been a <a class="reference external" href="http://clang.llvm.org/compatibility.html#dep_lookup">frequently asked question</a> for Clang users. A
+dependent base class is a base class that depends on the value of a template
+parameter. Clang cannot see any of the names inside dependent bases while it
+is parsing your template, so the user is sometimes required to use the
+<tt class="docutils literal"><span class="pre">typename</span></tt> keyword to assist the parser. On Windows, Clang attempts to
+follow the normal lookup rules, but if lookup fails, it will assume that the
+user intended to find the name in a dependent base. While parsing the
+following program, Clang will recover as if the user had written the
+commented-out code:</p>
+<div class="highlight-c++"><div class="highlight"><pre><span class="k">template</span> <span class="o"><</span><span class="k">typename</span> <span class="n">T</span><span class="o">></span>
+<span class="k">struct</span> <span class="n">Foo</span> <span class="o">:</span> <span class="n">T</span> <span class="p">{</span>
+ <span class="kt">void</span> <span class="n">f</span><span class="p">()</span> <span class="p">{</span>
+ <span class="cm">/*typename*/</span> <span class="n">T</span><span class="o">::</span><span class="n">UnknownType</span> <span class="n">x</span> <span class="o">=</span> <span class="cm">/*this->*/</span><span class="n">unknownMember</span><span class="p">;</span>
+ <span class="p">}</span>
+<span class="p">};</span>
+</pre></div>
+</div>
+<p>After recovery, Clang warns the user that this code is non-standard and issues
+a hint suggesting how to fix the problem.</p>
+<p>As of this writing, Clang is able to compile a simple ATL hello world
+application. There are still issues parsing WRL headers for modern Windows 8
+apps, but they should be addressed soon.</p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="Modules.html">Modules</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="CommandGuide/index.html">Clang “man” pages</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/MemorySanitizer.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/MemorySanitizer.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/MemorySanitizer.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/MemorySanitizer.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,274 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>MemorySanitizer — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="UndefinedBehaviorSanitizer" href="UndefinedBehaviorSanitizer.html" />
+ <link rel="prev" title="ThreadSanitizer" href="ThreadSanitizer.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>MemorySanitizer</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="ThreadSanitizer.html">ThreadSanitizer</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="UndefinedBehaviorSanitizer.html">UndefinedBehaviorSanitizer</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="memorysanitizer">
+<h1>MemorySanitizer<a class="headerlink" href="#memorysanitizer" title="Permalink to this headline">¶</a></h1>
+<div class="contents local topic" id="contents">
+<ul class="simple">
+<li><a class="reference internal" href="#introduction" id="id1">Introduction</a></li>
+<li><a class="reference internal" href="#how-to-build" id="id2">How to build</a></li>
+<li><a class="reference internal" href="#usage" id="id3">Usage</a><ul>
+<li><a class="reference internal" href="#has-feature-memory-sanitizer" id="id4"><tt class="docutils literal"><span class="pre">__has_feature(memory_sanitizer)</span></tt></a></li>
+<li><a class="reference internal" href="#attribute-no-sanitize-memory" id="id5"><tt class="docutils literal"><span class="pre">__attribute__((no_sanitize_memory))</span></tt></a></li>
+<li><a class="reference internal" href="#blacklist" id="id6">Blacklist</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#report-symbolization" id="id7">Report symbolization</a></li>
+<li><a class="reference internal" href="#origin-tracking" id="id8">Origin Tracking</a></li>
+<li><a class="reference internal" href="#use-after-destruction-detection" id="id9">Use-after-destruction detection</a></li>
+<li><a class="reference internal" href="#handling-external-code" id="id10">Handling external code</a></li>
+<li><a class="reference internal" href="#supported-platforms" id="id11">Supported Platforms</a></li>
+<li><a class="reference internal" href="#limitations" id="id12">Limitations</a></li>
+<li><a class="reference internal" href="#current-status" id="id13">Current Status</a></li>
+<li><a class="reference internal" href="#more-information" id="id14">More Information</a></li>
+</ul>
+</div>
+<div class="section" id="introduction">
+<h2><a class="toc-backref" href="#id1">Introduction</a><a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>MemorySanitizer is a detector of uninitialized reads. It consists of a
+compiler instrumentation module and a run-time library.</p>
+<p>Typical slowdown introduced by MemorySanitizer is <strong>3x</strong>.</p>
+</div>
+<div class="section" id="how-to-build">
+<h2><a class="toc-backref" href="#id2">How to build</a><a class="headerlink" href="#how-to-build" title="Permalink to this headline">¶</a></h2>
+<p>Build LLVM/Clang with <a class="reference external" href="http://llvm.org/docs/CMake.html">CMake</a>.</p>
+</div>
+<div class="section" id="usage">
+<h2><a class="toc-backref" href="#id3">Usage</a><a class="headerlink" href="#usage" title="Permalink to this headline">¶</a></h2>
+<p>Simply compile and link your program with <tt class="docutils literal"><span class="pre">-fsanitize=memory</span></tt> flag.
+The MemorySanitizer run-time library should be linked to the final
+executable, so make sure to use <tt class="docutils literal"><span class="pre">clang</span></tt> (not <tt class="docutils literal"><span class="pre">ld</span></tt>) for the final
+link step. When linking shared libraries, the MemorySanitizer run-time
+is not linked, so <tt class="docutils literal"><span class="pre">-Wl,-z,defs</span></tt> may cause link errors (don’t use it
+with MemorySanitizer). To get a reasonable performance add <tt class="docutils literal"><span class="pre">-O1</span></tt> or
+higher. To get meaninful stack traces in error messages add
+<tt class="docutils literal"><span class="pre">-fno-omit-frame-pointer</span></tt>. To get perfect stack traces you may need
+to disable inlining (just use <tt class="docutils literal"><span class="pre">-O1</span></tt>) and tail call elimination
+(<tt class="docutils literal"><span class="pre">-fno-optimize-sibling-calls</span></tt>).</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">%</span> cat umr.cc
+<span class="gp">#</span>include <stdio.h>
+
+<span class="go">int main(int argc, char** argv) {</span>
+<span class="go"> int* a = new int[10];</span>
+<span class="go"> a[5] = 0;</span>
+<span class="go"> if (a[argc])</span>
+<span class="go"> printf("xx\n");</span>
+<span class="go"> return 0;</span>
+<span class="go">}</span>
+
+<span class="gp">%</span> clang -fsanitize<span class="o">=</span>memory -fno-omit-frame-pointer -g -O2 umr.cc
+</pre></div>
+</div>
+<p>If a bug is detected, the program will print an error message to
+stderr and exit with a non-zero exit code.</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">%</span> ./a.out
+<span class="go">WARNING: MemorySanitizer: use-of-uninitialized-value</span>
+<span class="go"> #0 0x7f45944b418a in main umr.cc:6</span>
+<span class="go"> #1 0x7f45938b676c in __libc_start_main libc-start.c:226</span>
+</pre></div>
+</div>
+<p>By default, MemorySanitizer exits on the first detected error. If you
+find the error report hard to understand, try enabling
+<a class="reference internal" href="#msan-origins"><em>origin tracking</em></a>.</p>
+<div class="section" id="has-feature-memory-sanitizer">
+<h3><a class="toc-backref" href="#id4"><tt class="docutils literal"><span class="pre">__has_feature(memory_sanitizer)</span></tt></a><a class="headerlink" href="#has-feature-memory-sanitizer" title="Permalink to this headline">¶</a></h3>
+<p>In some cases one may need to execute different code depending on
+whether MemorySanitizer is enabled. <a class="reference internal" href="LanguageExtensions.html#langext-has-feature-has-extension"><em>__has_feature</em></a> can be used for this purpose.</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="cp">#if defined(__has_feature)</span>
+<span class="cp"># if __has_feature(memory_sanitizer)</span>
+<span class="c1">// code that builds only under MemorySanitizer</span>
+<span class="cp"># endif</span>
+<span class="cp">#endif</span>
+</pre></div>
+</div>
+</div>
+<div class="section" id="attribute-no-sanitize-memory">
+<h3><a class="toc-backref" href="#id5"><tt class="docutils literal"><span class="pre">__attribute__((no_sanitize_memory))</span></tt></a><a class="headerlink" href="#attribute-no-sanitize-memory" title="Permalink to this headline">¶</a></h3>
+<p>Some code should not be checked by MemorySanitizer. One may use the function
+attribute <cite>no_sanitize_memory</cite> to disable uninitialized checks in a particular
+function. MemorySanitizer may still instrument such functions to avoid false
+positives. This attribute may not be supported by other compilers, so we
+suggest to use it together with <tt class="docutils literal"><span class="pre">__has_feature(memory_sanitizer)</span></tt>.</p>
+</div>
+<div class="section" id="blacklist">
+<h3><a class="toc-backref" href="#id6">Blacklist</a><a class="headerlink" href="#blacklist" title="Permalink to this headline">¶</a></h3>
+<p>MemorySanitizer supports <tt class="docutils literal"><span class="pre">src</span></tt> and <tt class="docutils literal"><span class="pre">fun</span></tt> entity types in
+<a class="reference internal" href="SanitizerSpecialCaseList.html"><em>Sanitizer special case list</em></a>, that can be used to relax MemorySanitizer
+checks for certain source files and functions. All “Use of uninitialized value”
+warnings will be suppressed and all values loaded from memory will be
+considered fully initialized.</p>
+</div>
+</div>
+<div class="section" id="report-symbolization">
+<h2><a class="toc-backref" href="#id7">Report symbolization</a><a class="headerlink" href="#report-symbolization" title="Permalink to this headline">¶</a></h2>
+<p>MemorySanitizer uses an external symbolizer to print files and line numbers in
+reports. Make sure that <tt class="docutils literal"><span class="pre">llvm-symbolizer</span></tt> binary is in <tt class="docutils literal"><span class="pre">PATH</span></tt>,
+or set environment variable <tt class="docutils literal"><span class="pre">MSAN_SYMBOLIZER_PATH</span></tt> to point to it.</p>
+</div>
+<div class="section" id="origin-tracking">
+<span id="msan-origins"></span><h2><a class="toc-backref" href="#id8">Origin Tracking</a><a class="headerlink" href="#origin-tracking" title="Permalink to this headline">¶</a></h2>
+<p>MemorySanitizer can track origins of uninitialized values, similar to
+Valgrind’s –track-origins option. This feature is enabled by
+<tt class="docutils literal"><span class="pre">-fsanitize-memory-track-origins=2</span></tt> (or simply
+<tt class="docutils literal"><span class="pre">-fsanitize-memory-track-origins</span></tt>) Clang option. With the code from
+the example above,</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">%</span> cat umr2.cc
+<span class="gp">#</span>include <stdio.h>
+
+<span class="go">int main(int argc, char** argv) {</span>
+<span class="go"> int* a = new int[10];</span>
+<span class="go"> a[5] = 0;</span>
+<span class="go"> volatile int b = a[argc];</span>
+<span class="go"> if (b)</span>
+<span class="go"> printf("xx\n");</span>
+<span class="go"> return 0;</span>
+<span class="go">}</span>
+
+<span class="gp">%</span> clang -fsanitize<span class="o">=</span>memory -fsanitize-memory-track-origins<span class="o">=</span>2 -fno-omit-frame-pointer -g -O2 umr2.cc
+<span class="gp">%</span> ./a.out
+<span class="go">WARNING: MemorySanitizer: use-of-uninitialized-value</span>
+<span class="go"> #0 0x7f7893912f0b in main umr2.cc:7</span>
+<span class="go"> #1 0x7f789249b76c in __libc_start_main libc-start.c:226</span>
+
+<span class="go"> Uninitialized value was stored to memory at</span>
+<span class="go"> #0 0x7f78938b5c25 in __msan_chain_origin msan.cc:484</span>
+<span class="go"> #1 0x7f7893912ecd in main umr2.cc:6</span>
+
+<span class="go"> Uninitialized value was created by a heap allocation</span>
+<span class="go"> #0 0x7f7893901cbd in operator new[](unsigned long) msan_new_delete.cc:44</span>
+<span class="go"> #1 0x7f7893912e06 in main umr2.cc:4</span>
+</pre></div>
+</div>
+<p>By default, MemorySanitizer collects both allocation points and all
+intermediate stores the uninitialized value went through. Origin
+tracking has proved to be very useful for debugging MemorySanitizer
+reports. It slows down program execution by a factor of 1.5x-2x on top
+of the usual MemorySanitizer slowdown and increases memory overhead.</p>
+<p>Clang option <tt class="docutils literal"><span class="pre">-fsanitize-memory-track-origins=1</span></tt> enables a slightly
+faster mode when MemorySanitizer collects only allocation points but
+not intermediate stores.</p>
+</div>
+<div class="section" id="use-after-destruction-detection">
+<h2><a class="toc-backref" href="#id9">Use-after-destruction detection</a><a class="headerlink" href="#use-after-destruction-detection" title="Permalink to this headline">¶</a></h2>
+<p>You can enable experimental use-after-destruction detection in MemorySanitizer.
+After invocation of the destructor, the object will be considered no longer
+readable, and using underlying memory will lead to error reports in runtime.</p>
+<p>This feature is still experimental, in order to enable it at runtime you need
+to:</p>
+<ol class="arabic simple">
+<li>Pass addition Clang option <tt class="docutils literal"><span class="pre">-fsanitize-memory-use-after-dtor</span></tt> during
+compilation.</li>
+<li>Set environment variable <cite>MSAN_OPTIONS=poison_in_dtor=1</cite> before running
+the program.</li>
+</ol>
+</div>
+<div class="section" id="handling-external-code">
+<h2><a class="toc-backref" href="#id10">Handling external code</a><a class="headerlink" href="#handling-external-code" title="Permalink to this headline">¶</a></h2>
+<p>MemorySanitizer requires that all program code is instrumented. This
+also includes any libraries that the program depends on, even libc.
+Failing to achieve this may result in false reports.</p>
+<p>Full MemorySanitizer instrumentation is very difficult to achieve. To
+make it easier, MemorySanitizer runtime library includes 70+
+interceptors for the most common libc functions. They make it possible
+to run MemorySanitizer-instrumented programs linked with
+uninstrumented libc. For example, the authors were able to bootstrap
+MemorySanitizer-instrumented Clang compiler by linking it with
+self-built instrumented libc++ (as a replacement for libstdc++).</p>
+</div>
+<div class="section" id="supported-platforms">
+<h2><a class="toc-backref" href="#id11">Supported Platforms</a><a class="headerlink" href="#supported-platforms" title="Permalink to this headline">¶</a></h2>
+<p>MemorySanitizer is supported on Linux x86_64/MIPS64/AArch64.</p>
+</div>
+<div class="section" id="limitations">
+<h2><a class="toc-backref" href="#id12">Limitations</a><a class="headerlink" href="#limitations" title="Permalink to this headline">¶</a></h2>
+<ul class="simple">
+<li>MemorySanitizer uses 2x more real memory than a native run, 3x with
+origin tracking.</li>
+<li>MemorySanitizer maps (but not reserves) 64 Terabytes of virtual
+address space. This means that tools like <tt class="docutils literal"><span class="pre">ulimit</span></tt> may not work as
+usually expected.</li>
+<li>Static linking is not supported.</li>
+<li>Older versions of MSan (LLVM 3.7 and older) didn’t work with
+non-position-independent executables, and could fail on some Linux
+kernel versions with disabled ASLR. Refer to documentation for older versions
+for more details.</li>
+</ul>
+</div>
+<div class="section" id="current-status">
+<h2><a class="toc-backref" href="#id13">Current Status</a><a class="headerlink" href="#current-status" title="Permalink to this headline">¶</a></h2>
+<p>MemorySanitizer is known to work on large real-world programs
+(like Clang/LLVM itself) that can be recompiled from source, including all
+dependent libraries.</p>
+</div>
+<div class="section" id="more-information">
+<h2><a class="toc-backref" href="#id14">More Information</a><a class="headerlink" href="#more-information" title="Permalink to this headline">¶</a></h2>
+<p><a class="reference external" href="https://github.com/google/sanitizers/wiki/MemorySanitizer">https://github.com/google/sanitizers/wiki/MemorySanitizer</a></p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="ThreadSanitizer.html">ThreadSanitizer</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="UndefinedBehaviorSanitizer.html">UndefinedBehaviorSanitizer</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/Modules.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/Modules.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/Modules.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/Modules.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,891 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Modules — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="MSVC compatibility" href="MSVCCompatibility.html" />
+ <link rel="prev" title="SafeStack" href="SafeStack.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Modules</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="SafeStack.html">SafeStack</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="MSVCCompatibility.html">MSVC compatibility</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="modules">
+<h1>Modules<a class="headerlink" href="#modules" title="Permalink to this headline">¶</a></h1>
+<div class="contents local topic" id="contents">
+<ul class="simple">
+<li><a class="reference internal" href="#introduction" id="id9">Introduction</a><ul>
+<li><a class="reference internal" href="#problems-with-the-current-model" id="id10">Problems with the current model</a></li>
+<li><a class="reference internal" href="#semantic-import" id="id11">Semantic import</a></li>
+<li><a class="reference internal" href="#problems-modules-do-not-solve" id="id12">Problems modules do not solve</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#using-modules" id="id13">Using Modules</a><ul>
+<li><a class="reference internal" href="#objective-c-import-declaration" id="id14">Objective-C Import declaration</a></li>
+<li><a class="reference internal" href="#includes-as-imports" id="id15">Includes as imports</a></li>
+<li><a class="reference internal" href="#module-maps" id="id16">Module maps</a></li>
+<li><a class="reference internal" href="#compilation-model" id="id17">Compilation model</a></li>
+<li><a class="reference internal" href="#command-line-parameters" id="id18">Command-line parameters</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#module-semantics" id="id19">Module Semantics</a><ul>
+<li><a class="reference internal" href="#macros" id="id20">Macros</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#module-map-language" id="id21">Module Map Language</a><ul>
+<li><a class="reference internal" href="#lexical-structure" id="id22">Lexical structure</a></li>
+<li><a class="reference internal" href="#module-map-file" id="id23">Module map file</a></li>
+<li><a class="reference internal" href="#module-declaration" id="id24">Module declaration</a><ul>
+<li><a class="reference internal" href="#requires-declaration" id="id25">Requires declaration</a></li>
+<li><a class="reference internal" href="#header-declaration" id="id26">Header declaration</a></li>
+<li><a class="reference internal" href="#umbrella-directory-declaration" id="id27">Umbrella directory declaration</a></li>
+<li><a class="reference internal" href="#submodule-declaration" id="id28">Submodule declaration</a></li>
+<li><a class="reference internal" href="#export-declaration" id="id29">Export declaration</a></li>
+<li><a class="reference internal" href="#use-declaration" id="id30">Use declaration</a></li>
+<li><a class="reference internal" href="#link-declaration" id="id31">Link declaration</a></li>
+<li><a class="reference internal" href="#configuration-macros-declaration" id="id32">Configuration macros declaration</a></li>
+<li><a class="reference internal" href="#conflict-declarations" id="id33">Conflict declarations</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#attributes" id="id34">Attributes</a></li>
+<li><a class="reference internal" href="#private-module-map-files" id="id35">Private Module Map Files</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#modularizing-a-platform" id="id36">Modularizing a Platform</a></li>
+<li><a class="reference internal" href="#future-directions" id="id37">Future Directions</a></li>
+<li><a class="reference internal" href="#where-to-learn-more-about-modules" id="id38">Where To Learn More About Modules</a></li>
+</ul>
+</div>
+<div class="section" id="introduction">
+<h2><a class="toc-backref" href="#id9">Introduction</a><a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>Most software is built using a number of software libraries, including libraries supplied by the platform, internal libraries built as part of the software itself to provide structure, and third-party libraries. For each library, one needs to access both its interface (API) and its implementation. In the C family of languages, the interface to a library is accessed by including the appropriate header files(s):</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="cp">#include <SomeLib.h></span>
+</pre></div>
+</div>
+<p>The implementation is handled separately by linking against the appropriate library. For example, by passing <tt class="docutils literal"><span class="pre">-lSomeLib</span></tt> to the linker.</p>
+<p>Modules provide an alternative, simpler way to use software libraries that provides better compile-time scalability and eliminates many of the problems inherent to using the C preprocessor to access the API of a library.</p>
+<div class="section" id="problems-with-the-current-model">
+<h3><a class="toc-backref" href="#id10">Problems with the current model</a><a class="headerlink" href="#problems-with-the-current-model" title="Permalink to this headline">¶</a></h3>
+<p>The <tt class="docutils literal"><span class="pre">#include</span></tt> mechanism provided by the C preprocessor is a very poor way to access the API of a library, for a number of reasons:</p>
+<ul class="simple">
+<li><strong>Compile-time scalability</strong>: Each time a header is included, the
+compiler must preprocess and parse the text in that header and every
+header it includes, transitively. This process must be repeated for
+every translation unit in the application, which involves a huge
+amount of redundant work. In a project with <em>N</em> translation units
+and <em>M</em> headers included in each translation unit, the compiler is
+performing <em>M x N</em> work even though most of the <em>M</em> headers are
+shared among multiple translation units. C++ is particularly bad,
+because the compilation model for templates forces a huge amount of
+code into headers.</li>
+<li><strong>Fragility</strong>: <tt class="docutils literal"><span class="pre">#include</span></tt> directives are treated as textual
+inclusion by the preprocessor, and are therefore subject to any
+active macro definitions at the time of inclusion. If any of the
+active macro definitions happens to collide with a name in the
+library, it can break the library API or cause compilation failures
+in the library header itself. For an extreme example,
+<tt class="docutils literal"><span class="pre">#define</span> <span class="pre">std</span> <span class="pre">"The</span> <span class="pre">C++</span> <span class="pre">Standard"</span></tt> and then include a standard
+library header: the result is a horrific cascade of failures in the
+C++ Standard Library’s implementation. More subtle real-world
+problems occur when the headers for two different libraries interact
+due to macro collisions, and users are forced to reorder
+<tt class="docutils literal"><span class="pre">#include</span></tt> directives or introduce <tt class="docutils literal"><span class="pre">#undef</span></tt> directives to break
+the (unintended) dependency.</li>
+<li><strong>Conventional workarounds</strong>: C programmers have
+adopted a number of conventions to work around the fragility of the
+C preprocessor model. Include guards, for example, are required for
+the vast majority of headers to ensure that multiple inclusion
+doesn’t break the compile. Macro names are written with
+<tt class="docutils literal"><span class="pre">LONG_PREFIXED_UPPERCASE_IDENTIFIERS</span></tt> to avoid collisions, and some
+library/framework developers even use <tt class="docutils literal"><span class="pre">__underscored</span></tt> names
+in headers to avoid collisions with “normal” names that (by
+convention) shouldn’t even be macros. These conventions are a
+barrier to entry for developers coming from non-C languages, are
+boilerplate for more experienced developers, and make our headers
+far uglier than they should be.</li>
+<li><strong>Tool confusion</strong>: In a C-based language, it is hard to build tools
+that work well with software libraries, because the boundaries of
+the libraries are not clear. Which headers belong to a particular
+library, and in what order should those headers be included to
+guarantee that they compile correctly? Are the headers C, C++,
+Objective-C++, or one of the variants of these languages? What
+declarations in those headers are actually meant to be part of the
+API, and what declarations are present only because they had to be
+written as part of the header file?</li>
+</ul>
+</div>
+<div class="section" id="semantic-import">
+<h3><a class="toc-backref" href="#id11">Semantic import</a><a class="headerlink" href="#semantic-import" title="Permalink to this headline">¶</a></h3>
+<p>Modules improve access to the API of software libraries by replacing the textual preprocessor inclusion model with a more robust, more efficient semantic model. From the user’s perspective, the code looks only slightly different, because one uses an <tt class="docutils literal"><span class="pre">import</span></tt> declaration rather than a <tt class="docutils literal"><span class="pre">#include</span></tt> preprocessor directive:</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="n">import</span> <span class="n">std</span><span class="p">.</span><span class="n">io</span><span class="p">;</span> <span class="c1">// pseudo-code; see below for syntax discussion</span>
+</pre></div>
+</div>
+<p>However, this module import behaves quite differently from the corresponding <tt class="docutils literal"><span class="pre">#include</span> <span class="pre"><stdio.h></span></tt>: when the compiler sees the module import above, it loads a binary representation of the <tt class="docutils literal"><span class="pre">std.io</span></tt> module and makes its API available to the application directly. Preprocessor definitions that precede the import declaration have no impact on the API provided by <tt class="docutils literal"><span class="pre">std.io</span></tt>, because the module itself was compiled as a separate, standalone module. Additionally, any linker flags required to use the <tt class="docutils literal"><span class="pre">std.io</span></tt> module will automatically be provided when the module is imported <a class="footnote-reference" href="#id5" id="id1">[1]</a>
+This semantic import model addresses many of the problems of the preprocessor inclusion model:</p>
+<ul class="simple">
+<li><strong>Compile-time scalability</strong>: The <tt class="docutils literal"><span class="pre">std.io</span></tt> module is only compiled once, and importing the module into a translation unit is a constant-time operation (independent of module system). Thus, the API of each software library is only parsed once, reducing the <em>M x N</em> compilation problem to an <em>M + N</em> problem.</li>
+<li><strong>Fragility</strong>: Each module is parsed as a standalone entity, so it has a consistent preprocessor environment. This completely eliminates the need for <tt class="docutils literal"><span class="pre">__underscored</span></tt> names and similarly defensive tricks. Moreover, the current preprocessor definitions when an import declaration is encountered are ignored, so one software library can not affect how another software library is compiled, eliminating include-order dependencies.</li>
+<li><strong>Tool confusion</strong>: Modules describe the API of software libraries, and tools can reason about and present a module as a representation of that API. Because modules can only be built standalone, tools can rely on the module definition to ensure that they get the complete API for the library. Moreover, modules can specify which languages they work with, so, e.g., one can not accidentally attempt to load a C++ module into a C program.</li>
+</ul>
+</div>
+<div class="section" id="problems-modules-do-not-solve">
+<h3><a class="toc-backref" href="#id12">Problems modules do not solve</a><a class="headerlink" href="#problems-modules-do-not-solve" title="Permalink to this headline">¶</a></h3>
+<p>Many programming languages have a module or package system, and because of the variety of features provided by these languages it is important to define what modules do <em>not</em> do. In particular, all of the following are considered out-of-scope for modules:</p>
+<ul class="simple">
+<li><strong>Rewrite the world’s code</strong>: It is not realistic to require applications or software libraries to make drastic or non-backward-compatible changes, nor is it feasible to completely eliminate headers. Modules must interoperate with existing software libraries and allow a gradual transition.</li>
+<li><strong>Versioning</strong>: Modules have no notion of version information. Programmers must still rely on the existing versioning mechanisms of the underlying language (if any exist) to version software libraries.</li>
+<li><strong>Namespaces</strong>: Unlike in some languages, modules do not imply any notion of namespaces. Thus, a struct declared in one module will still conflict with a struct of the same name declared in a different module, just as they would if declared in two different headers. This aspect is important for backward compatibility, because (for example) the mangled names of entities in software libraries must not change when introducing modules.</li>
+<li><strong>Binary distribution of modules</strong>: Headers (particularly C++ headers) expose the full complexity of the language. Maintaining a stable binary module format across architectures, compiler versions, and compiler vendors is technically infeasible.</li>
+</ul>
+</div>
+</div>
+<div class="section" id="using-modules">
+<h2><a class="toc-backref" href="#id13">Using Modules</a><a class="headerlink" href="#using-modules" title="Permalink to this headline">¶</a></h2>
+<p>To enable modules, pass the command-line flag <tt class="docutils literal"><span class="pre">-fmodules</span></tt>. This will make any modules-enabled software libraries available as modules as well as introducing any modules-specific syntax. Additional <a class="reference internal" href="#command-line-parameters">command-line parameters</a> are described in a separate section later.</p>
+<div class="section" id="objective-c-import-declaration">
+<h3><a class="toc-backref" href="#id14">Objective-C Import declaration</a><a class="headerlink" href="#objective-c-import-declaration" title="Permalink to this headline">¶</a></h3>
+<p>Objective-C provides syntax for importing a module via an <em>@import declaration</em>, which imports the named module:</p>
+<div class="highlight-python"><div class="highlight"><pre>@import std;
+</pre></div>
+</div>
+<p>The <tt class="docutils literal"><span class="pre">@import</span></tt> declaration above imports the entire contents of the <tt class="docutils literal"><span class="pre">std</span></tt> module (which would contain, e.g., the entire C or C++ standard library) and make its API available within the current translation unit. To import only part of a module, one may use dot syntax to specific a particular submodule, e.g.,</p>
+<div class="highlight-python"><div class="highlight"><pre>@import std.io;
+</pre></div>
+</div>
+<p>Redundant import declarations are ignored, and one is free to import modules at any point within the translation unit, so long as the import declaration is at global scope.</p>
+<p>At present, there is no C or C++ syntax for import declarations. Clang
+will track the modules proposal in the C++ committee. See the section
+<a class="reference internal" href="#includes-as-imports">Includes as imports</a> to see how modules get imported today.</p>
+</div>
+<div class="section" id="includes-as-imports">
+<h3><a class="toc-backref" href="#id15">Includes as imports</a><a class="headerlink" href="#includes-as-imports" title="Permalink to this headline">¶</a></h3>
+<p>The primary user-level feature of modules is the import operation, which provides access to the API of software libraries. However, today’s programs make extensive use of <tt class="docutils literal"><span class="pre">#include</span></tt>, and it is unrealistic to assume that all of this code will change overnight. Instead, modules automatically translate <tt class="docutils literal"><span class="pre">#include</span></tt> directives into the corresponding module import. For example, the include directive</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="cp">#include <stdio.h></span>
+</pre></div>
+</div>
+<p>will be automatically mapped to an import of the module <tt class="docutils literal"><span class="pre">std.io</span></tt>. Even with specific <tt class="docutils literal"><span class="pre">import</span></tt> syntax in the language, this particular feature is important for both adoption and backward compatibility: automatic translation of <tt class="docutils literal"><span class="pre">#include</span></tt> to <tt class="docutils literal"><span class="pre">import</span></tt> allows an application to get the benefits of modules (for all modules-enabled libraries) without any changes to the application itself. Thus, users can easily use modules with one compiler while falling back to the preprocessor-inclusion mechanism with other compilers.</p>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">The automatic mapping of <tt class="docutils literal"><span class="pre">#include</span></tt> to <tt class="docutils literal"><span class="pre">import</span></tt> also solves an implementation problem: importing a module with a definition of some entity (say, a <tt class="docutils literal"><span class="pre">struct</span> <span class="pre">Point</span></tt>) and then parsing a header containing another definition of <tt class="docutils literal"><span class="pre">struct</span> <span class="pre">Point</span></tt> would cause a redefinition error, even if it is the same <tt class="docutils literal"><span class="pre">struct</span> <span class="pre">Point</span></tt>. By mapping <tt class="docutils literal"><span class="pre">#include</span></tt> to <tt class="docutils literal"><span class="pre">import</span></tt>, the compiler can guarantee that it always sees just the already-parsed definition from the module.</p>
+</div>
+<p>While building a module, <tt class="docutils literal"><span class="pre">#include_next</span></tt> is also supported, with one caveat.
+The usual behavior of <tt class="docutils literal"><span class="pre">#include_next</span></tt> is to search for the specified filename
+in the list of include paths, starting from the path <em>after</em> the one
+in which the current file was found.
+Because files listed in module maps are not found through include paths, a
+different strategy is used for <tt class="docutils literal"><span class="pre">#include_next</span></tt> directives in such files: the
+list of include paths is searched for the specified header name, to find the
+first include path that would refer to the current file. <tt class="docutils literal"><span class="pre">#include_next</span></tt> is
+interpreted as if the current file had been found in that path.
+If this search finds a file named by a module map, the <tt class="docutils literal"><span class="pre">#include_next</span></tt>
+directive is translated into an import, just like for a <tt class="docutils literal"><span class="pre">#include</span></tt>
+directive.``</p>
+</div>
+<div class="section" id="module-maps">
+<h3><a class="toc-backref" href="#id16">Module maps</a><a class="headerlink" href="#module-maps" title="Permalink to this headline">¶</a></h3>
+<p>The crucial link between modules and headers is described by a <em>module map</em>, which describes how a collection of existing headers maps on to the (logical) structure of a module. For example, one could imagine a module <tt class="docutils literal"><span class="pre">std</span></tt> covering the C standard library. Each of the C standard library headers (<tt class="docutils literal"><span class="pre"><stdio.h></span></tt>, <tt class="docutils literal"><span class="pre"><stdlib.h></span></tt>, <tt class="docutils literal"><span class="pre"><math.h></span></tt>, etc.) would contribute to the <tt class="docutils literal"><span class="pre">std</span></tt> module, by placing their respective APIs into the corresponding submodule (<tt class="docutils literal"><span class="pre">std.io</span></tt>, <tt class="docutils literal"><span class="pre">std.lib</span></tt>, <tt class="docutils literal"><span class="pre">std.math</span></tt>, etc.). Having a list of the heade
rs that are part of the <tt class="docutils literal"><span class="pre">std</span></tt> module allows the compiler to build the <tt class="docutils literal"><span class="pre">std</span></tt> module as a standalone entity, and having the mapping from header names to (sub)modules allows the automatic translation of <tt class="docutils literal"><span class="pre">#include</span></tt> directives to module imports.</p>
+<p>Module maps are specified as separate files (each named <tt class="docutils literal"><span class="pre">module.modulemap</span></tt>) alongside the headers they describe, which allows them to be added to existing software libraries without having to change the library headers themselves (in most cases <a class="footnote-reference" href="#id6" id="id2">[2]</a>). The actual <a class="reference internal" href="#module-map-language">Module map language</a> is described in a later section.</p>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">To actually see any benefits from modules, one first has to introduce module maps for the underlying C standard library and the libraries and headers on which it depends. The section <a class="reference internal" href="#modularizing-a-platform">Modularizing a Platform</a> describes the steps one must take to write these module maps.</p>
+</div>
+<p>One can use module maps without modules to check the integrity of the use of header files. To do this, use the <tt class="docutils literal"><span class="pre">-fimplicit-module-maps</span></tt> option instead of the <tt class="docutils literal"><span class="pre">-fmodules</span></tt> option, or use <tt class="docutils literal"><span class="pre">-fmodule-map-file=</span></tt> option to explicitly specify the module map files to load.</p>
+</div>
+<div class="section" id="compilation-model">
+<h3><a class="toc-backref" href="#id17">Compilation model</a><a class="headerlink" href="#compilation-model" title="Permalink to this headline">¶</a></h3>
+<p>The binary representation of modules is automatically generated by the compiler on an as-needed basis. When a module is imported (e.g., by an <tt class="docutils literal"><span class="pre">#include</span></tt> of one of the module’s headers), the compiler will spawn a second instance of itself <a class="footnote-reference" href="#id7" id="id3">[3]</a>, with a fresh preprocessing context <a class="footnote-reference" href="#id8" id="id4">[4]</a>, to parse just the headers in that module. The resulting Abstract Syntax Tree (AST) is then persisted into the binary representation of the module that is then loaded into translation unit where the module import was encountered.</p>
+<p>The binary representation of modules is persisted in the <em>module cache</em>. Imports of a module will first query the module cache and, if a binary representation of the required module is already available, will load that representation directly. Thus, a module’s headers will only be parsed once per language configuration, rather than once per translation unit that uses the module.</p>
+<p>Modules maintain references to each of the headers that were part of the module build. If any of those headers changes, or if any of the modules on which a module depends change, then the module will be (automatically) recompiled. The process should never require any user intervention.</p>
+</div>
+<div class="section" id="command-line-parameters">
+<h3><a class="toc-backref" href="#id18">Command-line parameters</a><a class="headerlink" href="#command-line-parameters" title="Permalink to this headline">¶</a></h3>
+<dl class="docutils">
+<dt><tt class="docutils literal"><span class="pre">-fmodules</span></tt></dt>
+<dd>Enable the modules feature.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fimplicit-module-maps</span></tt></dt>
+<dd>Enable implicit search for module map files named <tt class="docutils literal"><span class="pre">module.modulemap</span></tt> and similar. This option is implied by <tt class="docutils literal"><span class="pre">-fmodules</span></tt>. If this is disabled with <tt class="docutils literal"><span class="pre">-fno-implicit-module-maps</span></tt>, module map files will only be loaded if they are explicitly specified via <tt class="docutils literal"><span class="pre">-fmodule-map-file</span></tt> or transitively used by another module map file.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fmodules-cache-path=<directory></span></tt></dt>
+<dd>Specify the path to the modules cache. If not provided, Clang will select a system-appropriate default.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fno-autolink</span></tt></dt>
+<dd>Disable automatic linking against the libraries associated with imported modules.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fmodules-ignore-macro=macroname</span></tt></dt>
+<dd>Instruct modules to ignore the named macro when selecting an appropriate module variant. Use this for macros defined on the command line that don’t affect how modules are built, to improve sharing of compiled module files.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fmodules-prune-interval=seconds</span></tt></dt>
+<dd>Specify the minimum delay (in seconds) between attempts to prune the module cache. Module cache pruning attempts to clear out old, unused module files so that the module cache itself does not grow without bound. The default delay is large (604,800 seconds, or 7 days) because this is an expensive operation. Set this value to 0 to turn off pruning.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fmodules-prune-after=seconds</span></tt></dt>
+<dd>Specify the minimum time (in seconds) for which a file in the module cache must be unused (according to access time) before module pruning will remove it. The default delay is large (2,678,400 seconds, or 31 days) to avoid excessive module rebuilding.</dd>
+<dt><tt class="docutils literal"><span class="pre">-module-file-info</span> <span class="pre"><module</span> <span class="pre">file</span> <span class="pre">name></span></tt></dt>
+<dd>Debugging aid that prints information about a given module file (with a <tt class="docutils literal"><span class="pre">.pcm</span></tt> extension), including the language and preprocessor options that particular module variant was built with.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fmodules-decluse</span></tt></dt>
+<dd>Enable checking of module <tt class="docutils literal"><span class="pre">use</span></tt> declarations.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fmodule-name=module-id</span></tt></dt>
+<dd>Consider a source file as a part of the given module.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fmodule-map-file=<file></span></tt></dt>
+<dd>Load the given module map file if a header from its directory or one of its subdirectories is loaded.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fmodules-search-all</span></tt></dt>
+<dd>If a symbol is not found, search modules referenced in the current module maps but not imported for symbols, so the error message can reference the module by name. Note that if the global module index has not been built before, this might take some time as it needs to build all the modules. Note that this option doesn’t apply in module builds, to avoid the recursion.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fno-implicit-modules</span></tt></dt>
+<dd>All modules used by the build must be specified with <tt class="docutils literal"><span class="pre">-fmodule-file</span></tt>.</dd>
+<dt><tt class="docutils literal"><span class="pre">-fmodule-file=<file></span></tt></dt>
+<dd>Load the given precompiled module file.</dd>
+</dl>
+</div>
+</div>
+<div class="section" id="module-semantics">
+<h2><a class="toc-backref" href="#id19">Module Semantics</a><a class="headerlink" href="#module-semantics" title="Permalink to this headline">¶</a></h2>
+<p>Modules are modeled as if each submodule were a separate translation unit, and a module import makes names from the other translation unit visible. Each submodule starts with a new preprocessor state and an empty translation unit.</p>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">This behavior is currently only approximated when building a module with submodules. Entities within a submodule that has already been built are visible when building later submodules in that module. This can lead to fragile modules that depend on the build order used for the submodules of the module, and should not be relied upon. This behavior is subject to change.</p>
+</div>
+<p>As an example, in C, this implies that if two structs are defined in different submodules with the same name, those two types are distinct types (but may be <em>compatible</em> types if their definitions match). In C++, two structs defined with the same name in different submodules are the <em>same</em> type, and must be equivalent under C++’s One Definition Rule.</p>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">Clang currently only performs minimal checking for violations of the One Definition Rule.</p>
+</div>
+<p>If any submodule of a module is imported into any part of a program, the entire top-level module is considered to be part of the program. As a consequence of this, Clang may diagnose conflicts between an entity declared in an unimported submodule and an entity declared in the current translation unit, and Clang may inline or devirtualize based on knowledge from unimported submodules.</p>
+<div class="section" id="macros">
+<h3><a class="toc-backref" href="#id20">Macros</a><a class="headerlink" href="#macros" title="Permalink to this headline">¶</a></h3>
+<p>The C and C++ preprocessor assumes that the input text is a single linear buffer, but with modules this is not the case. It is possible to import two modules that have conflicting definitions for a macro (or where one <tt class="docutils literal"><span class="pre">#define</span></tt>s a macro and the other <tt class="docutils literal"><span class="pre">#undef</span></tt>ines it). The rules for handling macro definitions in the presence of modules are as follows:</p>
+<ul class="simple">
+<li>Each definition and undefinition of a macro is considered to be a distinct entity.</li>
+<li>Such entities are <em>visible</em> if they are from the current submodule or translation unit, or if they were exported from a submodule that has been imported.</li>
+<li>A <tt class="docutils literal"><span class="pre">#define</span> <span class="pre">X</span></tt> or <tt class="docutils literal"><span class="pre">#undef</span> <span class="pre">X</span></tt> directive <em>overrides</em> all definitions of <tt class="docutils literal"><span class="pre">X</span></tt> that are visible at the point of the directive.</li>
+<li>A <tt class="docutils literal"><span class="pre">#define</span></tt> or <tt class="docutils literal"><span class="pre">#undef</span></tt> directive is <em>active</em> if it is visible and no visible directive overrides it.</li>
+<li>A set of macro directives is <em>consistent</em> if it consists of only <tt class="docutils literal"><span class="pre">#undef</span></tt> directives, or if all <tt class="docutils literal"><span class="pre">#define</span></tt> directives in the set define the macro name to the same sequence of tokens (following the usual rules for macro redefinitions).</li>
+<li>If a macro name is used and the set of active directives is not consistent, the program is ill-formed. Otherwise, the (unique) meaning of the macro name is used.</li>
+</ul>
+<p>For example, suppose:</p>
+<ul class="simple">
+<li><tt class="docutils literal"><span class="pre"><stdio.h></span></tt> defines a macro <tt class="docutils literal"><span class="pre">getc</span></tt> (and exports its <tt class="docutils literal"><span class="pre">#define</span></tt>)</li>
+<li><tt class="docutils literal"><span class="pre"><cstdio></span></tt> imports the <tt class="docutils literal"><span class="pre"><stdio.h></span></tt> module and undefines the macro (and exports its <tt class="docutils literal"><span class="pre">#undef</span></tt>)</li>
+</ul>
+<p>The <tt class="docutils literal"><span class="pre">#undef</span></tt> overrides the <tt class="docutils literal"><span class="pre">#define</span></tt>, and a source file that imports both modules <em>in any order</em> will not see <tt class="docutils literal"><span class="pre">getc</span></tt> defined as a macro.</p>
+</div>
+</div>
+<div class="section" id="module-map-language">
+<h2><a class="toc-backref" href="#id21">Module Map Language</a><a class="headerlink" href="#module-map-language" title="Permalink to this headline">¶</a></h2>
+<div class="admonition warning">
+<p class="first admonition-title">Warning</p>
+<p class="last">The module map language is not currently guaranteed to be stable between major revisions of Clang.</p>
+</div>
+<p>The module map language describes the mapping from header files to the
+logical structure of modules. To enable support for using a library as
+a module, one must write a <tt class="docutils literal"><span class="pre">module.modulemap</span></tt> file for that library. The
+<tt class="docutils literal"><span class="pre">module.modulemap</span></tt> file is placed alongside the header files themselves,
+and is written in the module map language described below.</p>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">For compatibility with previous releases, if a module map file named
+<tt class="docutils literal"><span class="pre">module.modulemap</span></tt> is not found, Clang will also search for a file named
+<tt class="docutils literal"><span class="pre">module.map</span></tt>. This behavior is deprecated and we plan to eventually
+remove it.</p>
+</div>
+<p>As an example, the module map file for the C standard library might look a bit like this:</p>
+<div class="highlight-python"><div class="highlight"><pre>module std [system] [extern_c] {
+ module assert {
+ textual header "assert.h"
+ header "bits/assert-decls.h"
+ export *
+ }
+
+ module complex {
+ header "complex.h"
+ export *
+ }
+
+ module ctype {
+ header "ctype.h"
+ export *
+ }
+
+ module errno {
+ header "errno.h"
+ header "sys/errno.h"
+ export *
+ }
+
+ module fenv {
+ header "fenv.h"
+ export *
+ }
+
+ // ...more headers follow...
+}
+</pre></div>
+</div>
+<p>Here, the top-level module <tt class="docutils literal"><span class="pre">std</span></tt> encompasses the whole C standard library. It has a number of submodules containing different parts of the standard library: <tt class="docutils literal"><span class="pre">complex</span></tt> for complex numbers, <tt class="docutils literal"><span class="pre">ctype</span></tt> for character types, etc. Each submodule lists one of more headers that provide the contents for that submodule. Finally, the <tt class="docutils literal"><span class="pre">export</span> <span class="pre">*</span></tt> command specifies that anything included by that submodule will be automatically re-exported.</p>
+<div class="section" id="lexical-structure">
+<h3><a class="toc-backref" href="#id22">Lexical structure</a><a class="headerlink" href="#lexical-structure" title="Permalink to this headline">¶</a></h3>
+<p>Module map files use a simplified form of the C99 lexer, with the same rules for identifiers, tokens, string literals, <tt class="docutils literal"><span class="pre">/*</span> <span class="pre">*/</span></tt> and <tt class="docutils literal"><span class="pre">//</span></tt> comments. The module map language has the following reserved words; all other C identifiers are valid identifiers.</p>
+<pre class="literal-block">
+<tt class="docutils literal"><span class="pre">config_macros</span></tt> <tt class="docutils literal"><span class="pre">export</span></tt> <tt class="docutils literal"><span class="pre">private</span></tt>
+<tt class="docutils literal"><span class="pre">conflict</span></tt> <tt class="docutils literal"><span class="pre">framework</span></tt> <tt class="docutils literal"><span class="pre">requires</span></tt>
+<tt class="docutils literal"><span class="pre">exclude</span></tt> <tt class="docutils literal"><span class="pre">header</span></tt> <tt class="docutils literal"><span class="pre">textual</span></tt>
+<tt class="docutils literal"><span class="pre">explicit</span></tt> <tt class="docutils literal"><span class="pre">link</span></tt> <tt class="docutils literal"><span class="pre">umbrella</span></tt>
+<tt class="docutils literal"><span class="pre">extern</span></tt> <tt class="docutils literal"><span class="pre">module</span></tt> <tt class="docutils literal"><span class="pre">use</span></tt>
+</pre>
+</div>
+<div class="section" id="module-map-file">
+<h3><a class="toc-backref" href="#id23">Module map file</a><a class="headerlink" href="#module-map-file" title="Permalink to this headline">¶</a></h3>
+<p>A module map file consists of a series of module declarations:</p>
+<pre class="literal-block">
+<em>module-map-file</em>:
+ <em>module-declaration*</em>
+</pre>
+<p>Within a module map file, modules are referred to by a <em>module-id</em>, which uses periods to separate each part of a module’s name:</p>
+<pre class="literal-block">
+<em>module-id</em>:
+ <em>identifier</em> ('.' <em>identifier</em>)*
+</pre>
+</div>
+<div class="section" id="module-declaration">
+<h3><a class="toc-backref" href="#id24">Module declaration</a><a class="headerlink" href="#module-declaration" title="Permalink to this headline">¶</a></h3>
+<p>A module declaration describes a module, including the headers that contribute to that module, its submodules, and other aspects of the module.</p>
+<pre class="literal-block">
+<em>module-declaration</em>:
+ <tt class="docutils literal"><span class="pre">explicit</span></tt><sub>opt</sub> <tt class="docutils literal"><span class="pre">framework</span></tt><sub>opt</sub> <tt class="docutils literal"><span class="pre">module</span></tt> <em>module-id</em> <em>attributes</em><sub>opt</sub> '{' <em>module-member*</em> '}'
+ <tt class="docutils literal"><span class="pre">extern</span></tt> <tt class="docutils literal"><span class="pre">module</span></tt> <em>module-id</em> <em>string-literal</em>
+</pre>
+<p>The <em>module-id</em> should consist of only a single <em>identifier</em>, which provides the name of the module being defined. Each module shall have a single definition.</p>
+<p>The <tt class="docutils literal"><span class="pre">explicit</span></tt> qualifier can only be applied to a submodule, i.e., a module that is nested within another module. The contents of explicit submodules are only made available when the submodule itself was explicitly named in an import declaration or was re-exported from an imported module.</p>
+<p>The <tt class="docutils literal"><span class="pre">framework</span></tt> qualifier specifies that this module corresponds to a Darwin-style framework. A Darwin-style framework (used primarily on Mac OS X and iOS) is contained entirely in directory <tt class="docutils literal"><span class="pre">Name.framework</span></tt>, where <tt class="docutils literal"><span class="pre">Name</span></tt> is the name of the framework (and, therefore, the name of the module). That directory has the following layout:</p>
+<div class="highlight-python"><div class="highlight"><pre>Name.framework/
+ Modules/module.modulemap Module map for the framework
+ Headers/ Subdirectory containing framework headers
+ Frameworks/ Subdirectory containing embedded frameworks
+ Resources/ Subdirectory containing additional resources
+ Name Symbolic link to the shared library for the framework
+</pre></div>
+</div>
+<p>The <tt class="docutils literal"><span class="pre">system</span></tt> attribute specifies that the module is a system module. When a system module is rebuilt, all of the module’s headers will be considered system headers, which suppresses warnings. This is equivalent to placing <tt class="docutils literal"><span class="pre">#pragma</span> <span class="pre">GCC</span> <span class="pre">system_header</span></tt> in each of the module’s headers. The form of attributes is described in the section <a class="reference internal" href="#attributes">Attributes</a>, below.</p>
+<p>The <tt class="docutils literal"><span class="pre">extern_c</span></tt> attribute specifies that the module contains C code that can be used from within C++. When such a module is built for use in C++ code, all of the module’s headers will be treated as if they were contained within an implicit <tt class="docutils literal"><span class="pre">extern</span> <span class="pre">"C"</span></tt> block. An import for a module with this attribute can appear within an <tt class="docutils literal"><span class="pre">extern</span> <span class="pre">"C"</span></tt> block. No other restrictions are lifted, however: the module currently cannot be imported within an <tt class="docutils literal"><span class="pre">extern</span> <span class="pre">"C"</span></tt> block in a namespace.</p>
+<p>Modules can have a number of different kinds of members, each of which is described below:</p>
+<pre class="literal-block">
+<em>module-member</em>:
+ <em>requires-declaration</em>
+ <em>header-declaration</em>
+ <em>umbrella-dir-declaration</em>
+ <em>submodule-declaration</em>
+ <em>export-declaration</em>
+ <em>use-declaration</em>
+ <em>link-declaration</em>
+ <em>config-macros-declaration</em>
+ <em>conflict-declaration</em>
+</pre>
+<p>An extern module references a module defined by the <em>module-id</em> in a file given by the <em>string-literal</em>. The file can be referenced either by an absolute path or by a path relative to the current map file.</p>
+<div class="section" id="requires-declaration">
+<h4><a class="toc-backref" href="#id25">Requires declaration</a><a class="headerlink" href="#requires-declaration" title="Permalink to this headline">¶</a></h4>
+<p>A <em>requires-declaration</em> specifies the requirements that an importing translation unit must satisfy to use the module.</p>
+<pre class="literal-block">
+<em>requires-declaration</em>:
+ <tt class="docutils literal"><span class="pre">requires</span></tt> <em>feature-list</em>
+
+<em>feature-list</em>:
+ <em>feature</em> (',' <em>feature</em>)*
+
+<em>feature</em>:
+ <tt class="docutils literal"><span class="pre">!</span></tt><sub>opt</sub> <em>identifier</em>
+</pre>
+<p>The requirements clause allows specific modules or submodules to specify that they are only accessible with certain language dialects or on certain platforms. The feature list is a set of identifiers, defined below. If any of the features is not available in a given translation unit, that translation unit shall not import the module. The optional <tt class="docutils literal"><span class="pre">!</span></tt> indicates that a feature is incompatible with the module.</p>
+<p>The following features are defined:</p>
+<dl class="docutils">
+<dt>altivec</dt>
+<dd>The target supports AltiVec.</dd>
+<dt>blocks</dt>
+<dd>The “blocks” language feature is available.</dd>
+<dt>cplusplus</dt>
+<dd>C++ support is available.</dd>
+<dt>cplusplus11</dt>
+<dd>C++11 support is available.</dd>
+<dt>objc</dt>
+<dd>Objective-C support is available.</dd>
+<dt>objc_arc</dt>
+<dd>Objective-C Automatic Reference Counting (ARC) is available</dd>
+<dt>opencl</dt>
+<dd>OpenCL is available</dd>
+<dt>tls</dt>
+<dd>Thread local storage is available.</dd>
+<dt><em>target feature</em></dt>
+<dd>A specific target feature (e.g., <tt class="docutils literal"><span class="pre">sse4</span></tt>, <tt class="docutils literal"><span class="pre">avx</span></tt>, <tt class="docutils literal"><span class="pre">neon</span></tt>) is available.</dd>
+</dl>
+<p><strong>Example:</strong> The <tt class="docutils literal"><span class="pre">std</span></tt> module can be extended to also include C++ and C++11 headers using a <em>requires-declaration</em>:</p>
+<div class="highlight-python"><div class="highlight"><pre>module std {
+ // C standard library...
+
+ module vector {
+ requires cplusplus
+ header "vector"
+ }
+
+ module type_traits {
+ requires cplusplus11
+ header "type_traits"
+ }
+ }
+</pre></div>
+</div>
+</div>
+<div class="section" id="header-declaration">
+<h4><a class="toc-backref" href="#id26">Header declaration</a><a class="headerlink" href="#header-declaration" title="Permalink to this headline">¶</a></h4>
+<p>A header declaration specifies that a particular header is associated with the enclosing module.</p>
+<pre class="literal-block">
+<em>header-declaration</em>:
+ <tt class="docutils literal"><span class="pre">private</span></tt><sub>opt</sub> <tt class="docutils literal"><span class="pre">textual</span></tt><sub>opt</sub> <tt class="docutils literal"><span class="pre">header</span></tt> <em>string-literal</em>
+ <tt class="docutils literal"><span class="pre">umbrella</span></tt> <tt class="docutils literal"><span class="pre">header</span></tt> <em>string-literal</em>
+ <tt class="docutils literal"><span class="pre">exclude</span></tt> <tt class="docutils literal"><span class="pre">header</span></tt> <em>string-literal</em>
+</pre>
+<p>A header declaration that does not contain <tt class="docutils literal"><span class="pre">exclude</span></tt> nor <tt class="docutils literal"><span class="pre">textual</span></tt> specifies a header that contributes to the enclosing module. Specifically, when the module is built, the named header will be parsed and its declarations will be (logically) placed into the enclosing submodule.</p>
+<p>A header with the <tt class="docutils literal"><span class="pre">umbrella</span></tt> specifier is called an umbrella header. An umbrella header includes all of the headers within its directory (and any subdirectories), and is typically used (in the <tt class="docutils literal"><span class="pre">#include</span></tt> world) to easily access the full API provided by a particular library. With modules, an umbrella header is a convenient shortcut that eliminates the need to write out <tt class="docutils literal"><span class="pre">header</span></tt> declarations for every library header. A given directory can only contain a single umbrella header.</p>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">Any headers not included by the umbrella header should have
+explicit <tt class="docutils literal"><span class="pre">header</span></tt> declarations. Use the
+<tt class="docutils literal"><span class="pre">-Wincomplete-umbrella</span></tt> warning option to ask Clang to complain
+about headers not covered by the umbrella header or the module map.</p>
+</div>
+<p>A header with the <tt class="docutils literal"><span class="pre">private</span></tt> specifier may not be included from outside the module itself.</p>
+<p>A header with the <tt class="docutils literal"><span class="pre">textual</span></tt> specifier will not be compiled when the module is
+built, and will be textually included if it is named by a <tt class="docutils literal"><span class="pre">#include</span></tt>
+directive. However, it is considered to be part of the module for the purpose
+of checking <em>use-declaration</em>s, and must still be a lexically-valid header
+file. In the future, we intend to pre-tokenize such headers and include the
+token sequence within the prebuilt module representation.</p>
+<p>A header with the <tt class="docutils literal"><span class="pre">exclude</span></tt> specifier is excluded from the module. It will not be included when the module is built, nor will it be considered to be part of the module, even if an <tt class="docutils literal"><span class="pre">umbrella</span></tt> header or directory would otherwise make it part of the module.</p>
+<p><strong>Example:</strong> The C header <tt class="docutils literal"><span class="pre">assert.h</span></tt> is an excellent candidate for a textual header, because it is meant to be included multiple times (possibly with different <tt class="docutils literal"><span class="pre">NDEBUG</span></tt> settings). However, declarations within it should typically be split into a separate modular header.</p>
+<div class="highlight-python"><div class="highlight"><pre>module std [system] {
+ textual header "assert.h"
+}
+</pre></div>
+</div>
+<p>A given header shall not be referenced by more than one <em>header-declaration</em>.</p>
+</div>
+<div class="section" id="umbrella-directory-declaration">
+<h4><a class="toc-backref" href="#id27">Umbrella directory declaration</a><a class="headerlink" href="#umbrella-directory-declaration" title="Permalink to this headline">¶</a></h4>
+<p>An umbrella directory declaration specifies that all of the headers in the specified directory should be included within the module.</p>
+<pre class="literal-block">
+<em>umbrella-dir-declaration</em>:
+ <tt class="docutils literal"><span class="pre">umbrella</span></tt> <em>string-literal</em>
+</pre>
+<p>The <em>string-literal</em> refers to a directory. When the module is built, all of the header files in that directory (and its subdirectories) are included in the module.</p>
+<p>An <em>umbrella-dir-declaration</em> shall not refer to the same directory as the location of an umbrella <em>header-declaration</em>. In other words, only a single kind of umbrella can be specified for a given directory.</p>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">Umbrella directories are useful for libraries that have a large number of headers but do not have an umbrella header.</p>
+</div>
+</div>
+<div class="section" id="submodule-declaration">
+<h4><a class="toc-backref" href="#id28">Submodule declaration</a><a class="headerlink" href="#submodule-declaration" title="Permalink to this headline">¶</a></h4>
+<p>Submodule declarations describe modules that are nested within their enclosing module.</p>
+<pre class="literal-block">
+<em>submodule-declaration</em>:
+ <em>module-declaration</em>
+ <em>inferred-submodule-declaration</em>
+</pre>
+<p>A <em>submodule-declaration</em> that is a <em>module-declaration</em> is a nested module. If the <em>module-declaration</em> has a <tt class="docutils literal"><span class="pre">framework</span></tt> specifier, the enclosing module shall have a <tt class="docutils literal"><span class="pre">framework</span></tt> specifier; the submodule’s contents shall be contained within the subdirectory <tt class="docutils literal"><span class="pre">Frameworks/SubName.framework</span></tt>, where <tt class="docutils literal"><span class="pre">SubName</span></tt> is the name of the submodule.</p>
+<p>A <em>submodule-declaration</em> that is an <em>inferred-submodule-declaration</em> describes a set of submodules that correspond to any headers that are part of the module but are not explicitly described by a <em>header-declaration</em>.</p>
+<pre class="literal-block">
+<em>inferred-submodule-declaration</em>:
+ <tt class="docutils literal"><span class="pre">explicit</span></tt><sub>opt</sub> <tt class="docutils literal"><span class="pre">framework</span></tt><sub>opt</sub> <tt class="docutils literal"><span class="pre">module</span></tt> '*' <em>attributes</em><sub>opt</sub> '{' <em>inferred-submodule-member*</em> '}'
+
+<em>inferred-submodule-member</em>:
+ <tt class="docutils literal"><span class="pre">export</span></tt> '*'
+</pre>
+<p>A module containing an <em>inferred-submodule-declaration</em> shall have either an umbrella header or an umbrella directory. The headers to which the <em>inferred-submodule-declaration</em> applies are exactly those headers included by the umbrella header (transitively) or included in the module because they reside within the umbrella directory (or its subdirectories).</p>
+<p>For each header included by the umbrella header or in the umbrella directory that is not named by a <em>header-declaration</em>, a module declaration is implicitly generated from the <em>inferred-submodule-declaration</em>. The module will:</p>
+<ul class="simple">
+<li>Have the same name as the header (without the file extension)</li>
+<li>Have the <tt class="docutils literal"><span class="pre">explicit</span></tt> specifier, if the <em>inferred-submodule-declaration</em> has the <tt class="docutils literal"><span class="pre">explicit</span></tt> specifier</li>
+<li>Have the <tt class="docutils literal"><span class="pre">framework</span></tt> specifier, if the
+<em>inferred-submodule-declaration</em> has the <tt class="docutils literal"><span class="pre">framework</span></tt> specifier</li>
+<li>Have the attributes specified by the <em>inferred-submodule-declaration</em></li>
+<li>Contain a single <em>header-declaration</em> naming that header</li>
+<li>Contain a single <em>export-declaration</em> <tt class="docutils literal"><span class="pre">export</span> <span class="pre">*</span></tt>, if the <em>inferred-submodule-declaration</em> contains the <em>inferred-submodule-member</em> <tt class="docutils literal"><span class="pre">export</span> <span class="pre">*</span></tt></li>
+</ul>
+<p><strong>Example:</strong> If the subdirectory “MyLib” contains the headers <tt class="docutils literal"><span class="pre">A.h</span></tt> and <tt class="docutils literal"><span class="pre">B.h</span></tt>, then the following module map:</p>
+<div class="highlight-python"><div class="highlight"><pre>module MyLib {
+ umbrella "MyLib"
+ explicit module * {
+ export *
+ }
+}
+</pre></div>
+</div>
+<p>is equivalent to the (more verbose) module map:</p>
+<div class="highlight-python"><div class="highlight"><pre>module MyLib {
+ explicit module A {
+ header "A.h"
+ export *
+ }
+
+ explicit module B {
+ header "B.h"
+ export *
+ }
+}
+</pre></div>
+</div>
+</div>
+<div class="section" id="export-declaration">
+<h4><a class="toc-backref" href="#id29">Export declaration</a><a class="headerlink" href="#export-declaration" title="Permalink to this headline">¶</a></h4>
+<p>An <em>export-declaration</em> specifies which imported modules will automatically be re-exported as part of a given module’s API.</p>
+<pre class="literal-block">
+<em>export-declaration</em>:
+ <tt class="docutils literal"><span class="pre">export</span></tt> <em>wildcard-module-id</em>
+
+<em>wildcard-module-id</em>:
+ <em>identifier</em>
+ '*'
+ <em>identifier</em> '.' <em>wildcard-module-id</em>
+</pre>
+<p>The <em>export-declaration</em> names a module or a set of modules that will be re-exported to any translation unit that imports the enclosing module. Each imported module that matches the <em>wildcard-module-id</em> up to, but not including, the first <tt class="docutils literal"><span class="pre">*</span></tt> will be re-exported.</p>
+<p><strong>Example:</strong> In the following example, importing <tt class="docutils literal"><span class="pre">MyLib.Derived</span></tt> also provides the API for <tt class="docutils literal"><span class="pre">MyLib.Base</span></tt>:</p>
+<div class="highlight-python"><div class="highlight"><pre>module MyLib {
+ module Base {
+ header "Base.h"
+ }
+
+ module Derived {
+ header "Derived.h"
+ export Base
+ }
+}
+</pre></div>
+</div>
+<p>Note that, if <tt class="docutils literal"><span class="pre">Derived.h</span></tt> includes <tt class="docutils literal"><span class="pre">Base.h</span></tt>, one can simply use a wildcard export to re-export everything <tt class="docutils literal"><span class="pre">Derived.h</span></tt> includes:</p>
+<div class="highlight-python"><div class="highlight"><pre>module MyLib {
+ module Base {
+ header "Base.h"
+ }
+
+ module Derived {
+ header "Derived.h"
+ export *
+ }
+}
+</pre></div>
+</div>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">The wildcard export syntax <tt class="docutils literal"><span class="pre">export</span> <span class="pre">*</span></tt> re-exports all of the
+modules that were imported in the actual header file. Because
+<tt class="docutils literal"><span class="pre">#include</span></tt> directives are automatically mapped to module imports,
+<tt class="docutils literal"><span class="pre">export</span> <span class="pre">*</span></tt> provides the same transitive-inclusion behavior
+provided by the C preprocessor, e.g., importing a given module
+implicitly imports all of the modules on which it depends.
+Therefore, liberal use of <tt class="docutils literal"><span class="pre">export</span> <span class="pre">*</span></tt> provides excellent backward
+compatibility for programs that rely on transitive inclusion (i.e.,
+all of them).</p>
+</div>
+</div>
+<div class="section" id="use-declaration">
+<h4><a class="toc-backref" href="#id30">Use declaration</a><a class="headerlink" href="#use-declaration" title="Permalink to this headline">¶</a></h4>
+<p>A <em>use-declaration</em> specifies another module that the current top-level module
+intends to use. When the option <em>-fmodules-decluse</em> is specified, a module can
+only use other modules that are explicitly specified in this way.</p>
+<pre class="literal-block">
+<em>use-declaration</em>:
+ <tt class="docutils literal"><span class="pre">use</span></tt> <em>module-id</em>
+</pre>
+<p><strong>Example:</strong> In the following example, use of A from C is not declared, so will trigger a warning.</p>
+<div class="highlight-python"><div class="highlight"><pre>module A {
+ header "a.h"
+}
+
+module B {
+ header "b.h"
+}
+
+module C {
+ header "c.h"
+ use B
+}
+</pre></div>
+</div>
+<p>When compiling a source file that implements a module, use the option
+<tt class="docutils literal"><span class="pre">-fmodule-name=module-id</span></tt> to indicate that the source file is logically part
+of that module.</p>
+<p>The compiler at present only applies restrictions to the module directly being built.</p>
+</div>
+<div class="section" id="link-declaration">
+<h4><a class="toc-backref" href="#id31">Link declaration</a><a class="headerlink" href="#link-declaration" title="Permalink to this headline">¶</a></h4>
+<p>A <em>link-declaration</em> specifies a library or framework against which a program should be linked if the enclosing module is imported in any translation unit in that program.</p>
+<pre class="literal-block">
+<em>link-declaration</em>:
+ <tt class="docutils literal"><span class="pre">link</span></tt> <tt class="docutils literal"><span class="pre">framework</span></tt><sub>opt</sub> <em>string-literal</em>
+</pre>
+<p>The <em>string-literal</em> specifies the name of the library or framework against which the program should be linked. For example, specifying “clangBasic” would instruct the linker to link with <tt class="docutils literal"><span class="pre">-lclangBasic</span></tt> for a Unix-style linker.</p>
+<p>A <em>link-declaration</em> with the <tt class="docutils literal"><span class="pre">framework</span></tt> specifies that the linker should link against the named framework, e.g., with <tt class="docutils literal"><span class="pre">-framework</span> <span class="pre">MyFramework</span></tt>.</p>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">Automatic linking with the <tt class="docutils literal"><span class="pre">link</span></tt> directive is not yet widely
+implemented, because it requires support from both the object file
+format and the linker. The notion is similar to Microsoft Visual
+Studio’s <tt class="docutils literal"><span class="pre">#pragma</span> <span class="pre">comment(lib...)</span></tt>.</p>
+</div>
+</div>
+<div class="section" id="configuration-macros-declaration">
+<h4><a class="toc-backref" href="#id32">Configuration macros declaration</a><a class="headerlink" href="#configuration-macros-declaration" title="Permalink to this headline">¶</a></h4>
+<p>The <em>config-macros-declaration</em> specifies the set of configuration macros that have an effect on the API of the enclosing module.</p>
+<pre class="literal-block">
+<em>config-macros-declaration</em>:
+ <tt class="docutils literal"><span class="pre">config_macros</span></tt> <em>attributes</em><sub>opt</sub> <em>config-macro-list</em><sub>opt</sub>
+
+<em>config-macro-list</em>:
+ <em>identifier</em> (',' <em>identifier</em>)*
+</pre>
+<p>Each <em>identifier</em> in the <em>config-macro-list</em> specifies the name of a macro. The compiler is required to maintain different variants of the given module for differing definitions of any of the named macros.</p>
+<p>A <em>config-macros-declaration</em> shall only be present on a top-level module, i.e., a module that is not nested within an enclosing module.</p>
+<p>The <tt class="docutils literal"><span class="pre">exhaustive</span></tt> attribute specifies that the list of macros in the <em>config-macros-declaration</em> is exhaustive, meaning that no other macro definition is intended to have an effect on the API of that module.</p>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">The <tt class="docutils literal"><span class="pre">exhaustive</span></tt> attribute implies that any macro definitions
+for macros not listed as configuration macros should be ignored
+completely when building the module. As an optimization, the
+compiler could reduce the number of unique module variants by not
+considering these non-configuration macros. This optimization is not
+yet implemented in Clang.</p>
+</div>
+<p>A translation unit shall not import the same module under different definitions of the configuration macros.</p>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">Clang implements a weak form of this requirement: the definitions
+used for configuration macros are fixed based on the definitions
+provided by the command line. If an import occurs and the definition
+of any configuration macro has changed, the compiler will produce a
+warning (under the control of <tt class="docutils literal"><span class="pre">-Wconfig-macros</span></tt>).</p>
+</div>
+<p><strong>Example:</strong> A logging library might provide different API (e.g., in the form of different definitions for a logging macro) based on the <tt class="docutils literal"><span class="pre">NDEBUG</span></tt> macro setting:</p>
+<div class="highlight-python"><div class="highlight"><pre>module MyLogger {
+ umbrella header "MyLogger.h"
+ config_macros [exhaustive] NDEBUG
+}
+</pre></div>
+</div>
+</div>
+<div class="section" id="conflict-declarations">
+<h4><a class="toc-backref" href="#id33">Conflict declarations</a><a class="headerlink" href="#conflict-declarations" title="Permalink to this headline">¶</a></h4>
+<p>A <em>conflict-declaration</em> describes a case where the presence of two different modules in the same translation unit is likely to cause a problem. For example, two modules may provide similar-but-incompatible functionality.</p>
+<pre class="literal-block">
+<em>conflict-declaration</em>:
+ <tt class="docutils literal"><span class="pre">conflict</span></tt> <em>module-id</em> ',' <em>string-literal</em>
+</pre>
+<p>The <em>module-id</em> of the <em>conflict-declaration</em> specifies the module with which the enclosing module conflicts. The specified module shall not have been imported in the translation unit when the enclosing module is imported.</p>
+<p>The <em>string-literal</em> provides a message to be provided as part of the compiler diagnostic when two modules conflict.</p>
+<div class="admonition note">
+<p class="first admonition-title">Note</p>
+<p class="last">Clang emits a warning (under the control of <tt class="docutils literal"><span class="pre">-Wmodule-conflict</span></tt>)
+when a module conflict is discovered.</p>
+</div>
+<p><strong>Example:</strong></p>
+<div class="highlight-python"><div class="highlight"><pre>module Conflicts {
+ explicit module A {
+ header "conflict_a.h"
+ conflict B, "we just don't like B"
+ }
+
+ module B {
+ header "conflict_b.h"
+ }
+}
+</pre></div>
+</div>
+</div>
+</div>
+<div class="section" id="attributes">
+<h3><a class="toc-backref" href="#id34">Attributes</a><a class="headerlink" href="#attributes" title="Permalink to this headline">¶</a></h3>
+<p>Attributes are used in a number of places in the grammar to describe specific behavior of other declarations. The format of attributes is fairly simple.</p>
+<pre class="literal-block">
+<em>attributes</em>:
+ <em>attribute</em> <em>attributes</em><sub>opt</sub>
+
+<em>attribute</em>:
+ '[' <em>identifier</em> ']'
+</pre>
+<p>Any <em>identifier</em> can be used as an attribute, and each declaration specifies what attributes can be applied to it.</p>
+</div>
+<div class="section" id="private-module-map-files">
+<h3><a class="toc-backref" href="#id35">Private Module Map Files</a><a class="headerlink" href="#private-module-map-files" title="Permalink to this headline">¶</a></h3>
+<p>Module map files are typically named <tt class="docutils literal"><span class="pre">module.modulemap</span></tt> and live
+either alongside the headers they describe or in a parent directory of
+the headers they describe. These module maps typically describe all of
+the API for the library.</p>
+<p>However, in some cases, the presence or absence of particular headers
+is used to distinguish between the “public” and “private” APIs of a
+particular library. For example, a library may contain the headers
+<tt class="docutils literal"><span class="pre">Foo.h</span></tt> and <tt class="docutils literal"><span class="pre">Foo_Private.h</span></tt>, providing public and private APIs,
+respectively. Additionally, <tt class="docutils literal"><span class="pre">Foo_Private.h</span></tt> may only be available on
+some versions of library, and absent in others. One cannot easily
+express this with a single module map file in the library:</p>
+<div class="highlight-python"><div class="highlight"><pre>module Foo {
+ header "Foo.h"
+
+ explicit module Private {
+ header "Foo_Private.h"
+ }
+}
+</pre></div>
+</div>
+<p>because the header <tt class="docutils literal"><span class="pre">Foo_Private.h</span></tt> won’t always be available. The
+module map file could be customized based on whether
+<tt class="docutils literal"><span class="pre">Foo_Private.h</span></tt> is available or not, but doing so requires custom
+build machinery.</p>
+<p>Private module map files, which are named <tt class="docutils literal"><span class="pre">module.private.modulemap</span></tt>
+(or, for backward compatibility, <tt class="docutils literal"><span class="pre">module_private.map</span></tt>), allow one to
+augment the primary module map file with an additional submodule. For
+example, we would split the module map file above into two module map
+files:</p>
+<div class="highlight-c"><div class="highlight"><pre><span class="cm">/* module.modulemap */</span>
+<span class="n">module</span> <span class="n">Foo</span> <span class="p">{</span>
+ <span class="n">header</span> <span class="s">"Foo.h"</span>
+<span class="p">}</span>
+
+<span class="cm">/* module.private.modulemap */</span>
+<span class="n">explicit</span> <span class="n">module</span> <span class="n">Foo</span><span class="p">.</span><span class="n">Private</span> <span class="p">{</span>
+ <span class="n">header</span> <span class="s">"Foo_Private.h"</span>
+<span class="p">}</span>
+</pre></div>
+</div>
+<p>When a <tt class="docutils literal"><span class="pre">module.private.modulemap</span></tt> file is found alongside a
+<tt class="docutils literal"><span class="pre">module.modulemap</span></tt> file, it is loaded after the <tt class="docutils literal"><span class="pre">module.modulemap</span></tt>
+file. In our example library, the <tt class="docutils literal"><span class="pre">module.private.modulemap</span></tt> file
+would be available when <tt class="docutils literal"><span class="pre">Foo_Private.h</span></tt> is available, making it
+easier to split a library’s public and private APIs along header
+boundaries.</p>
+</div>
+</div>
+<div class="section" id="modularizing-a-platform">
+<h2><a class="toc-backref" href="#id36">Modularizing a Platform</a><a class="headerlink" href="#modularizing-a-platform" title="Permalink to this headline">¶</a></h2>
+<p>To get any benefit out of modules, one needs to introduce module maps for software libraries starting at the bottom of the stack. This typically means introducing a module map covering the operating system’s headers and the C standard library headers (in <tt class="docutils literal"><span class="pre">/usr/include</span></tt>, for a Unix system).</p>
+<p>The module maps will be written using the <a class="reference internal" href="#module-map-language">module map language</a>, which provides the tools necessary to describe the mapping between headers and modules. Because the set of headers differs from one system to the next, the module map will likely have to be somewhat customized for, e.g., a particular distribution and version of the operating system. Moreover, the system headers themselves may require some modification, if they exhibit any anti-patterns that break modules. Such common patterns are described below.</p>
+<dl class="docutils">
+<dt><strong>Macro-guarded copy-and-pasted definitions</strong></dt>
+<dd><p class="first">System headers vend core types such as <tt class="docutils literal"><span class="pre">size_t</span></tt> for users. These types are often needed in a number of system headers, and are almost trivial to write. Hence, it is fairly common to see a definition such as the following copy-and-pasted throughout the headers:</p>
+<div class="highlight-python"><div class="highlight"><pre>#ifndef _SIZE_T
+#define _SIZE_T
+typedef __SIZE_TYPE__ size_t;
+#endif
+</pre></div>
+</div>
+<p class="last">Unfortunately, when modules compiles all of the C library headers together into a single module, only the first actual type definition of <tt class="docutils literal"><span class="pre">size_t</span></tt> will be visible, and then only in the submodule corresponding to the lucky first header. Any other headers that have copy-and-pasted versions of this pattern will <em>not</em> have a definition of <tt class="docutils literal"><span class="pre">size_t</span></tt>. Importing the submodule corresponding to one of those headers will therefore not yield <tt class="docutils literal"><span class="pre">size_t</span></tt> as part of the API, because it wasn’t there when the header was parsed. The fix for this problem is either to pull the copied declarations into a common header that gets included everywhere <tt class="docutils literal"><span class="pre">size_t</span></tt> is part of the API, or to eliminate the <tt class="docutils literal"><span class="pre">#ifndef</sp
an></tt> and redefine the <tt class="docutils literal"><span class="pre">size_t</span></tt> type. The latter works for C++ headers and C11, but will cause an error for non-modules C90/C99, where redefinition of <tt class="docutils literal"><span class="pre">typedefs</span></tt> is not permitted.</p>
+</dd>
+<dt><strong>Conflicting definitions</strong></dt>
+<dd>Different system headers may provide conflicting definitions for various macros, functions, or types. These conflicting definitions don’t tend to cause problems in a pre-modules world unless someone happens to include both headers in one translation unit. Since the fix is often simply “don’t do that”, such problems persist. Modules requires that the conflicting definitions be eliminated or that they be placed in separate modules (the former is generally the better answer).</dd>
+<dt><strong>Missing includes</strong></dt>
+<dd>Headers are often missing <tt class="docutils literal"><span class="pre">#include</span></tt> directives for headers that they actually depend on. As with the problem of conflicting definitions, this only affects unlucky users who don’t happen to include headers in the right order. With modules, the headers of a particular module will be parsed in isolation, so the module may fail to build if there are missing includes.</dd>
+<dt><strong>Headers that vend multiple APIs at different times</strong></dt>
+<dd>Some systems have headers that contain a number of different kinds of API definitions, only some of which are made available with a given include. For example, the header may vend <tt class="docutils literal"><span class="pre">size_t</span></tt> only when the macro <tt class="docutils literal"><span class="pre">__need_size_t</span></tt> is defined before that header is included, and also vend <tt class="docutils literal"><span class="pre">wchar_t</span></tt> only when the macro <tt class="docutils literal"><span class="pre">__need_wchar_t</span></tt> is defined. Such headers are often included many times in a single translation unit, and will have no include guards. There is no sane way to map this header to a submodule. One can either eliminate the header (e.g., by splitting it into separate headers, one per actual API) or simply <tt class="docutils literal"><span class="pre">exclude</span></tt> it in the module map.</dd>
+</dl>
+<p>To detect and help address some of these problems, the <tt class="docutils literal"><span class="pre">clang-tools-extra</span></tt> repository contains a <tt class="docutils literal"><span class="pre">modularize</span></tt> tool that parses a set of given headers and attempts to detect these problems and produce a report. See the tool’s in-source documentation for information on how to check your system or library headers.</p>
+</div>
+<div class="section" id="future-directions">
+<h2><a class="toc-backref" href="#id37">Future Directions</a><a class="headerlink" href="#future-directions" title="Permalink to this headline">¶</a></h2>
+<p>Modules support is under active development, and there are many opportunities remaining to improve it. Here are a few ideas:</p>
+<dl class="docutils">
+<dt><strong>Detect unused module imports</strong></dt>
+<dd>Unlike with <tt class="docutils literal"><span class="pre">#include</span></tt> directives, it should be fairly simple to track whether a directly-imported module has ever been used. By doing so, Clang can emit <tt class="docutils literal"><span class="pre">unused</span> <span class="pre">import</span></tt> or <tt class="docutils literal"><span class="pre">unused</span> <span class="pre">#include</span></tt> diagnostics, including Fix-Its to remove the useless imports/includes.</dd>
+<dt><strong>Fix-Its for missing imports</strong></dt>
+<dd>It’s fairly common for one to make use of some API while writing code, only to get a compiler error about “unknown type” or “no function named” because the corresponding header has not been included. Clang can detect such cases and auto-import the required module, but should provide a Fix-It to add the import.</dd>
+<dt><strong>Improve modularize</strong></dt>
+<dd>The modularize tool is both extremely important (for deployment) and extremely crude. It needs better UI, better detection of problems (especially for C++), and perhaps an assistant mode to help write module maps for you.</dd>
+</dl>
+</div>
+<div class="section" id="where-to-learn-more-about-modules">
+<h2><a class="toc-backref" href="#id38">Where To Learn More About Modules</a><a class="headerlink" href="#where-to-learn-more-about-modules" title="Permalink to this headline">¶</a></h2>
+<p>The Clang source code provides additional information about modules:</p>
+<dl class="docutils">
+<dt><tt class="docutils literal"><span class="pre">clang/lib/Headers/module.modulemap</span></tt></dt>
+<dd>Module map for Clang’s compiler-specific header files.</dd>
+<dt><tt class="docutils literal"><span class="pre">clang/test/Modules/</span></tt></dt>
+<dd>Tests specifically related to modules functionality.</dd>
+<dt><tt class="docutils literal"><span class="pre">clang/include/clang/Basic/Module.h</span></tt></dt>
+<dd>The <tt class="docutils literal"><span class="pre">Module</span></tt> class in this header describes a module, and is used throughout the compiler to implement modules.</dd>
+<dt><tt class="docutils literal"><span class="pre">clang/include/clang/Lex/ModuleMap.h</span></tt></dt>
+<dd>The <tt class="docutils literal"><span class="pre">ModuleMap</span></tt> class in this header describes the full module map, consisting of all of the module map files that have been parsed, and providing facilities for looking up module maps and mapping between modules and headers (in both directions).</dd>
+<dt><a class="reference external" href="PCHInternals.html">PCHInternals</a></dt>
+<dd>Information about the serialized AST format used for precompiled headers and modules. The actual implementation is in the <tt class="docutils literal"><span class="pre">clangSerialization</span></tt> library.</dd>
+</dl>
+<table class="docutils footnote" frame="void" id="id5" rules="none">
+<colgroup><col class="label" /><col /></colgroup>
+<tbody valign="top">
+<tr><td class="label"><a class="fn-backref" href="#id1">[1]</a></td><td>Automatic linking against the libraries of modules requires specific linker support, which is not widely available.</td></tr>
+</tbody>
+</table>
+<table class="docutils footnote" frame="void" id="id6" rules="none">
+<colgroup><col class="label" /><col /></colgroup>
+<tbody valign="top">
+<tr><td class="label"><a class="fn-backref" href="#id2">[2]</a></td><td>There are certain anti-patterns that occur in headers, particularly system headers, that cause problems for modules. The section <a class="reference internal" href="#modularizing-a-platform">Modularizing a Platform</a> describes some of them.</td></tr>
+</tbody>
+</table>
+<table class="docutils footnote" frame="void" id="id7" rules="none">
+<colgroup><col class="label" /><col /></colgroup>
+<tbody valign="top">
+<tr><td class="label"><a class="fn-backref" href="#id3">[3]</a></td><td>The second instance is actually a new thread within the current process, not a separate process. However, the original compiler instance is blocked on the execution of this thread.</td></tr>
+</tbody>
+</table>
+<table class="docutils footnote" frame="void" id="id8" rules="none">
+<colgroup><col class="label" /><col /></colgroup>
+<tbody valign="top">
+<tr><td class="label"><a class="fn-backref" href="#id4">[4]</a></td><td>The preprocessing context in which the modules are parsed is actually dependent on the command-line options provided to the compiler, including the language dialect and any <tt class="docutils literal"><span class="pre">-D</span></tt> options. However, the compiled modules for different command-line options are kept distinct, and any preprocessor directives that occur within the translation unit are ignored. See the section on the <a class="reference internal" href="#configuration-macros-declaration">Configuration macros declaration</a> for more information.</td></tr>
+</tbody>
+</table>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="SafeStack.html">SafeStack</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="MSVCCompatibility.html">MSVC compatibility</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/ObjectiveCLiterals.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/ObjectiveCLiterals.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/ObjectiveCLiterals.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/ObjectiveCLiterals.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,598 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Objective-C Literals — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="up" title="Clang Language Extensions" href="LanguageExtensions.html" />
+ <link rel="next" title="Language Specification for Blocks" href="BlockLanguageSpec.html" />
+ <link rel="prev" title="Clang Language Extensions" href="LanguageExtensions.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Objective-C Literals</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="LanguageExtensions.html">Clang Language Extensions</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="BlockLanguageSpec.html">Language Specification for Blocks</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="objective-c-literals">
+<h1>Objective-C Literals<a class="headerlink" href="#objective-c-literals" title="Permalink to this headline">¶</a></h1>
+<div class="section" id="introduction">
+<h2>Introduction<a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
+<p>Three new features were introduced into clang at the same time:
+<em>NSNumber Literals</em> provide a syntax for creating <tt class="docutils literal"><span class="pre">NSNumber</span></tt> from
+scalar literal expressions; <em>Collection Literals</em> provide a short-hand
+for creating arrays and dictionaries; <em>Object Subscripting</em> provides a
+way to use subscripting with Objective-C objects. Users of Apple
+compiler releases can use these features starting with the Apple LLVM
+Compiler 4.0. Users of open-source LLVM.org compiler releases can use
+these features starting with clang v3.1.</p>
+<p>These language additions simplify common Objective-C programming
+patterns, make programs more concise, and improve the safety of
+container creation.</p>
+<p>This document describes how the features are implemented in clang, and
+how to use them in your own programs.</p>
+</div>
+<div class="section" id="nsnumber-literals">
+<h2>NSNumber Literals<a class="headerlink" href="#nsnumber-literals" title="Permalink to this headline">¶</a></h2>
+<p>The framework class <tt class="docutils literal"><span class="pre">NSNumber</span></tt> is used to wrap scalar values inside
+objects: signed and unsigned integers (<tt class="docutils literal"><span class="pre">char</span></tt>, <tt class="docutils literal"><span class="pre">short</span></tt>, <tt class="docutils literal"><span class="pre">int</span></tt>,
+<tt class="docutils literal"><span class="pre">long</span></tt>, <tt class="docutils literal"><span class="pre">long</span> <span class="pre">long</span></tt>), floating point numbers (<tt class="docutils literal"><span class="pre">float</span></tt>,
+<tt class="docutils literal"><span class="pre">double</span></tt>), and boolean values (<tt class="docutils literal"><span class="pre">BOOL</span></tt>, C++ <tt class="docutils literal"><span class="pre">bool</span></tt>). Scalar values
+wrapped in objects are also known as <em>boxed</em> values.</p>
+<p>In Objective-C, any character, numeric or boolean literal prefixed with
+the <tt class="docutils literal"><span class="pre">'@'</span></tt> character will evaluate to a pointer to an <tt class="docutils literal"><span class="pre">NSNumber</span></tt>
+object initialized with that value. C’s type suffixes may be used to
+control the size of numeric literals.</p>
+<div class="section" id="examples">
+<h3>Examples<a class="headerlink" href="#examples" title="Permalink to this headline">¶</a></h3>
+<p>The following program illustrates the rules for <tt class="docutils literal"><span class="pre">NSNumber</span></tt> literals:</p>
+<div class="highlight-objc"><div class="highlight"><pre>void main(int argc, const char *argv[]) {
+ // character literals.
+ NSNumber *theLetterZ = @'Z'; // equivalent to [NSNumber numberWithChar:'Z']
+
+ // integral literals.
+ NSNumber *fortyTwo = @42; // equivalent to [NSNumber numberWithInt:42]
+ NSNumber *fortyTwoUnsigned = @42U; // equivalent to [NSNumber numberWithUnsignedInt:42U]
+ NSNumber *fortyTwoLong = @42L; // equivalent to [NSNumber numberWithLong:42L]
+ NSNumber *fortyTwoLongLong = @42LL; // equivalent to [NSNumber numberWithLongLong:42LL]
+
+ // floating point literals.
+ NSNumber *piFloat = @3.141592654F; // equivalent to [NSNumber numberWithFloat:3.141592654F]
+ NSNumber *piDouble = @3.1415926535; // equivalent to [NSNumber numberWithDouble:3.1415926535]
+
+ // BOOL literals.
+ NSNumber *yesNumber = @YES; // equivalent to [NSNumber numberWithBool:YES]
+ NSNumber *noNumber = @NO; // equivalent to [NSNumber numberWithBool:NO]
+
+#ifdef __cplusplus
+ NSNumber *trueNumber = @true; // equivalent to [NSNumber numberWithBool:(BOOL)true]
+ NSNumber *falseNumber = @false; // equivalent to [NSNumber numberWithBool:(BOOL)false]
+#endif
+}
+</pre></div>
+</div>
+</div>
+<div class="section" id="discussion">
+<h3>Discussion<a class="headerlink" href="#discussion" title="Permalink to this headline">¶</a></h3>
+<p>NSNumber literals only support literal scalar values after the <tt class="docutils literal"><span class="pre">'@'</span></tt>.
+Consequently, <tt class="docutils literal"><span class="pre">@INT_MAX</span></tt> works, but <tt class="docutils literal"><span class="pre">@INT_MIN</span></tt> does not, because
+they are defined like this:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="cp">#define INT_MAX 2147483647 </span><span class="cm">/* max value for an int */</span><span class="cp"></span>
+<span class="cp">#define INT_MIN (-2147483647-1) </span><span class="cm">/* min value for an int */</span><span class="cp"></span>
+</pre></div>
+</div>
+<p>The definition of <tt class="docutils literal"><span class="pre">INT_MIN</span></tt> is not a simple literal, but a
+parenthesized expression. Parenthesized expressions are supported using
+the <a class="reference external" href="#objc_boxed_expressions">boxed expression</a> syntax, which is
+described in the next section.</p>
+<p>Because <tt class="docutils literal"><span class="pre">NSNumber</span></tt> does not currently support wrapping <tt class="docutils literal"><span class="pre">long</span> <span class="pre">double</span></tt>
+values, the use of a <tt class="docutils literal"><span class="pre">long</span> <span class="pre">double</span> <span class="pre">NSNumber</span></tt> literal (e.g.
+<tt class="docutils literal"><span class="pre">@123.23L</span></tt>) will be rejected by the compiler.</p>
+<p>Previously, the <tt class="docutils literal"><span class="pre">BOOL</span></tt> type was simply a typedef for <tt class="docutils literal"><span class="pre">signed</span> <span class="pre">char</span></tt>,
+and <tt class="docutils literal"><span class="pre">YES</span></tt> and <tt class="docutils literal"><span class="pre">NO</span></tt> were macros that expand to <tt class="docutils literal"><span class="pre">(BOOL)1</span></tt> and
+<tt class="docutils literal"><span class="pre">(BOOL)0</span></tt> respectively. To support <tt class="docutils literal"><span class="pre">@YES</span></tt> and <tt class="docutils literal"><span class="pre">@NO</span></tt> expressions,
+these macros are now defined using new language keywords in
+<tt class="docutils literal"><span class="pre"><objc/objc.h></span></tt>:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="cp">#if __has_feature(objc_bool)</span>
+<span class="cp">#define YES __objc_yes</span>
+<span class="cp">#define NO __objc_no</span>
+<span class="cp">#else</span>
+<span class="cp">#define YES ((BOOL)1)</span>
+<span class="cp">#define NO ((BOOL)0)</span>
+<span class="cp">#endif</span>
+</pre></div>
+</div>
+<p>The compiler implicitly converts <tt class="docutils literal"><span class="pre">__objc_yes</span></tt> and <tt class="docutils literal"><span class="pre">__objc_no</span></tt> to
+<tt class="docutils literal"><span class="pre">(BOOL)1</span></tt> and <tt class="docutils literal"><span class="pre">(BOOL)0</span></tt>. The keywords are used to disambiguate
+<tt class="docutils literal"><span class="pre">BOOL</span></tt> and integer literals.</p>
+<p>Objective-C++ also supports <tt class="docutils literal"><span class="pre">@true</span></tt> and <tt class="docutils literal"><span class="pre">@false</span></tt> expressions, which
+are equivalent to <tt class="docutils literal"><span class="pre">@YES</span></tt> and <tt class="docutils literal"><span class="pre">@NO</span></tt>.</p>
+</div>
+</div>
+<div class="section" id="boxed-expressions">
+<h2>Boxed Expressions<a class="headerlink" href="#boxed-expressions" title="Permalink to this headline">¶</a></h2>
+<p>Objective-C provides a new syntax for boxing C expressions:</p>
+<div class="highlight-objc"><div class="highlight"><pre>@( <expression> )
+</pre></div>
+</div>
+<p>Expressions of scalar (numeric, enumerated, BOOL), C string pointer
+and some C structures (via NSValue) are supported:</p>
+<div class="highlight-objc"><div class="highlight"><pre>// numbers.
+NSNumber *smallestInt = @(-INT_MAX - 1); // [NSNumber numberWithInt:(-INT_MAX - 1)]
+NSNumber *piOverTwo = @(M_PI / 2); // [NSNumber numberWithDouble:(M_PI / 2)]
+
+// enumerated types.
+typedef enum { Red, Green, Blue } Color;
+NSNumber *favoriteColor = @(Green); // [NSNumber numberWithInt:((int)Green)]
+
+// strings.
+NSString *path = @(getenv("PATH")); // [NSString stringWithUTF8String:(getenv("PATH"))]
+NSArray *pathComponents = [path componentsSeparatedByString:@":"];
+
+// structs.
+NSValue *center = @(view.center); // Point p = view.center;
+ // [NSValue valueWithBytes:&p objCType:@encode(Point)];
+NSValue *frame = @(view.frame); // Rect r = view.frame;
+ // [NSValue valueWithBytes:&r objCType:@encode(Rect)];
+</pre></div>
+</div>
+<div class="section" id="boxed-enums">
+<h3>Boxed Enums<a class="headerlink" href="#boxed-enums" title="Permalink to this headline">¶</a></h3>
+<p>Cocoa frameworks frequently define constant values using <em>enums.</em>
+Although enum values are integral, they may not be used directly as
+boxed literals (this avoids conflicts with future <tt class="docutils literal"><span class="pre">'@'</span></tt>-prefixed
+Objective-C keywords). Instead, an enum value must be placed inside a
+boxed expression. The following example demonstrates configuring an
+<tt class="docutils literal"><span class="pre">AVAudioRecorder</span></tt> using a dictionary that contains a boxed enumeration
+value:</p>
+<div class="highlight-objc"><div class="highlight"><pre>enum {
+ AVAudioQualityMin = 0,
+ AVAudioQualityLow = 0x20,
+ AVAudioQualityMedium = 0x40,
+ AVAudioQualityHigh = 0x60,
+ AVAudioQualityMax = 0x7F
+};
+
+- (AVAudioRecorder *)recordToFile:(NSURL *)fileURL {
+ NSDictionary *settings = @{ AVEncoderAudioQualityKey : @(AVAudioQualityMax) };
+ return [[AVAudioRecorder alloc] initWithURL:fileURL settings:settings error:NULL];
+}
+</pre></div>
+</div>
+<p>The expression <tt class="docutils literal"><span class="pre">@(AVAudioQualityMax)</span></tt> converts <tt class="docutils literal"><span class="pre">AVAudioQualityMax</span></tt>
+to an integer type, and boxes the value accordingly. If the enum has a
+<a class="reference internal" href="LanguageExtensions.html#objc-fixed-enum"><em>fixed underlying type</em></a> as in:</p>
+<div class="highlight-objc"><div class="highlight"><pre>typedef enum : unsigned char { Red, Green, Blue } Color;
+NSNumber *red = @(Red), *green = @(Green), *blue = @(Blue); // => [NSNumber numberWithUnsignedChar:]
+</pre></div>
+</div>
+<p>then the fixed underlying type will be used to select the correct
+<tt class="docutils literal"><span class="pre">NSNumber</span></tt> creation method.</p>
+<p>Boxing a value of enum type will result in a <tt class="docutils literal"><span class="pre">NSNumber</span></tt> pointer with a
+creation method according to the underlying type of the enum, which can
+be a <a class="reference internal" href="LanguageExtensions.html#objc-fixed-enum"><em>fixed underlying type</em></a>
+or a compiler-defined integer type capable of representing the values of
+all the members of the enumeration:</p>
+<div class="highlight-objc"><div class="highlight"><pre>typedef enum : unsigned char { Red, Green, Blue } Color;
+Color col = Red;
+NSNumber *nsCol = @(col); // => [NSNumber numberWithUnsignedChar:]
+</pre></div>
+</div>
+</div>
+<div class="section" id="boxed-c-strings">
+<h3>Boxed C Strings<a class="headerlink" href="#boxed-c-strings" title="Permalink to this headline">¶</a></h3>
+<p>A C string literal prefixed by the <tt class="docutils literal"><span class="pre">'@'</span></tt> token denotes an <tt class="docutils literal"><span class="pre">NSString</span></tt>
+literal in the same way a numeric literal prefixed by the <tt class="docutils literal"><span class="pre">'@'</span></tt> token
+denotes an <tt class="docutils literal"><span class="pre">NSNumber</span></tt> literal. When the type of the parenthesized
+expression is <tt class="docutils literal"><span class="pre">(char</span> <span class="pre">*)</span></tt> or <tt class="docutils literal"><span class="pre">(const</span> <span class="pre">char</span> <span class="pre">*)</span></tt>, the result of the
+boxed expression is a pointer to an <tt class="docutils literal"><span class="pre">NSString</span></tt> object containing
+equivalent character data, which is assumed to be ‘\0’-terminated and
+UTF-8 encoded. The following example converts C-style command line
+arguments into <tt class="docutils literal"><span class="pre">NSString</span></tt> objects.</p>
+<div class="highlight-objc"><div class="highlight"><pre>// Partition command line arguments into positional and option arguments.
+NSMutableArray *args = [NSMutableArray new];
+NSMutableDictionary *options = [NSMutableDictionary new];
+while (--argc) {
+ const char *arg = *++argv;
+ if (strncmp(arg, "--", 2) == 0) {
+ options[@(arg + 2)] = @(*++argv); // --key value
+ } else {
+ [args addObject:@(arg)]; // positional argument
+ }
+}
+</pre></div>
+</div>
+<p>As with all C pointers, character pointer expressions can involve
+arbitrary pointer arithmetic, therefore programmers must ensure that the
+character data is valid. Passing <tt class="docutils literal"><span class="pre">NULL</span></tt> as the character pointer will
+raise an exception at runtime. When possible, the compiler will reject
+<tt class="docutils literal"><span class="pre">NULL</span></tt> character pointers used in boxed expressions.</p>
+</div>
+<div class="section" id="boxed-c-structures">
+<h3>Boxed C Structures<a class="headerlink" href="#boxed-c-structures" title="Permalink to this headline">¶</a></h3>
+<p>Boxed expressions support construction of NSValue objects.
+It said that C structures can be used, the only requirement is:
+structure should be marked with <tt class="docutils literal"><span class="pre">objc_boxable</span></tt> attribute.
+To support older version of frameworks and/or third-party libraries
+you may need to add the attribute via <tt class="docutils literal"><span class="pre">typedef</span></tt>.</p>
+<div class="highlight-objc"><div class="highlight"><pre>struct __attribute__((objc_boxable)) Point {
+ // ...
+};
+
+typedef struct __attribute__((objc_boxable)) _Size {
+ // ...
+} Size;
+
+typedef struct _Rect {
+ // ...
+} Rect;
+
+struct Point p;
+NSValue *point = @(p); // ok
+Size s;
+NSValue *size = @(s); // ok
+
+Rect r;
+NSValue *bad_rect = @(r); // error
+
+typedef struct __attribute__((objc_boxable)) _Rect Rect;
+
+NSValue *good_rect = @(r); // ok
+</pre></div>
+</div>
+</div>
+</div>
+<div class="section" id="container-literals">
+<h2>Container Literals<a class="headerlink" href="#container-literals" title="Permalink to this headline">¶</a></h2>
+<p>Objective-C now supports a new expression syntax for creating immutable
+array and dictionary container objects.</p>
+<div class="section" id="id1">
+<h3>Examples<a class="headerlink" href="#id1" title="Permalink to this headline">¶</a></h3>
+<p>Immutable array expression:</p>
+<div class="highlight-objc"><div class="highlight"><pre>NSArray *array = @[ @"Hello", NSApp, [NSNumber numberWithInt:42] ];
+</pre></div>
+</div>
+<p>This creates an <tt class="docutils literal"><span class="pre">NSArray</span></tt> with 3 elements. The comma-separated
+sub-expressions of an array literal can be any Objective-C object
+pointer typed expression.</p>
+<p>Immutable dictionary expression:</p>
+<div class="highlight-objc"><div class="highlight"><pre>NSDictionary *dictionary = @{
+ @"name" : NSUserName(),
+ @"date" : [NSDate date],
+ @"processInfo" : [NSProcessInfo processInfo]
+};
+</pre></div>
+</div>
+<p>This creates an <tt class="docutils literal"><span class="pre">NSDictionary</span></tt> with 3 key/value pairs. Value
+sub-expressions of a dictionary literal must be Objective-C object
+pointer typed, as in array literals. Key sub-expressions must be of an
+Objective-C object pointer type that implements the
+<tt class="docutils literal"><span class="pre"><NSCopying></span></tt> protocol.</p>
+</div>
+<div class="section" id="id2">
+<h3>Discussion<a class="headerlink" href="#id2" title="Permalink to this headline">¶</a></h3>
+<p>Neither keys nor values can have the value <tt class="docutils literal"><span class="pre">nil</span></tt> in containers. If the
+compiler can prove that a key or value is <tt class="docutils literal"><span class="pre">nil</span></tt> at compile time, then
+a warning will be emitted. Otherwise, a runtime error will occur.</p>
+<p>Using array and dictionary literals is safer than the variadic creation
+forms commonly in use today. Array literal expressions expand to calls
+to <tt class="docutils literal"><span class="pre">+[NSArray</span> <span class="pre">arrayWithObjects:count:]</span></tt>, which validates that all
+objects are non-<tt class="docutils literal"><span class="pre">nil</span></tt>. The variadic form,
+<tt class="docutils literal"><span class="pre">+[NSArray</span> <span class="pre">arrayWithObjects:]</span></tt> uses <tt class="docutils literal"><span class="pre">nil</span></tt> as an argument list
+terminator, which can lead to malformed array objects. Dictionary
+literals are similarly created with
+<tt class="docutils literal"><span class="pre">+[NSDictionary</span> <span class="pre">dictionaryWithObjects:forKeys:count:]</span></tt> which validates
+all objects and keys, unlike
+<tt class="docutils literal"><span class="pre">+[NSDictionary</span> <span class="pre">dictionaryWithObjectsAndKeys:]</span></tt> which also uses a
+<tt class="docutils literal"><span class="pre">nil</span></tt> parameter as an argument list terminator.</p>
+</div>
+</div>
+<div class="section" id="object-subscripting">
+<h2>Object Subscripting<a class="headerlink" href="#object-subscripting" title="Permalink to this headline">¶</a></h2>
+<p>Objective-C object pointer values can now be used with C’s subscripting
+operator.</p>
+<div class="section" id="id3">
+<h3>Examples<a class="headerlink" href="#id3" title="Permalink to this headline">¶</a></h3>
+<p>The following code demonstrates the use of object subscripting syntax
+with <tt class="docutils literal"><span class="pre">NSMutableArray</span></tt> and <tt class="docutils literal"><span class="pre">NSMutableDictionary</span></tt> objects:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="n">NSMutableArray</span> <span class="o">*</span><span class="n">array</span> <span class="o">=</span> <span class="p">...;</span>
+<span class="n">NSUInteger</span> <span class="n">idx</span> <span class="o">=</span> <span class="p">...;</span>
+<span class="kt">id</span> <span class="n">newObject</span> <span class="o">=</span> <span class="p">...;</span>
+<span class="kt">id</span> <span class="n">oldObject</span> <span class="o">=</span> <span class="n">array</span><span class="p">[</span><span class="n">idx</span><span class="p">];</span>
+<span class="n">array</span><span class="p">[</span><span class="n">idx</span><span class="p">]</span> <span class="o">=</span> <span class="n">newObject</span><span class="p">;</span> <span class="c1">// replace oldObject with newObject</span>
+
+<span class="n">NSMutableDictionary</span> <span class="o">*</span><span class="n">dictionary</span> <span class="o">=</span> <span class="p">...;</span>
+<span class="n">NSString</span> <span class="o">*</span><span class="n">key</span> <span class="o">=</span> <span class="p">...;</span>
+<span class="n">oldObject</span> <span class="o">=</span> <span class="n">dictionary</span><span class="p">[</span><span class="n">key</span><span class="p">];</span>
+<span class="n">dictionary</span><span class="p">[</span><span class="n">key</span><span class="p">]</span> <span class="o">=</span> <span class="n">newObject</span><span class="p">;</span> <span class="c1">// replace oldObject with newObject</span>
+</pre></div>
+</div>
+<p>The next section explains how subscripting expressions map to accessor
+methods.</p>
+</div>
+<div class="section" id="subscripting-methods">
+<h3>Subscripting Methods<a class="headerlink" href="#subscripting-methods" title="Permalink to this headline">¶</a></h3>
+<p>Objective-C supports two kinds of subscript expressions: <em>array-style</em>
+subscript expressions use integer typed subscripts; <em>dictionary-style</em>
+subscript expressions use Objective-C object pointer typed subscripts.
+Each type of subscript expression is mapped to a message send using a
+predefined selector. The advantage of this design is flexibility: class
+designers are free to introduce subscripting by declaring methods or by
+adopting protocols. Moreover, because the method names are selected by
+the type of the subscript, an object can be subscripted using both array
+and dictionary styles.</p>
+<div class="section" id="array-style-subscripting">
+<h4>Array-Style Subscripting<a class="headerlink" href="#array-style-subscripting" title="Permalink to this headline">¶</a></h4>
+<p>When the subscript operand has an integral type, the expression is
+rewritten to use one of two different selectors, depending on whether
+the element is being read or written. When an expression reads an
+element using an integral index, as in the following example:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="n">NSUInteger</span> <span class="n">idx</span> <span class="o">=</span> <span class="p">...;</span>
+<span class="kt">id</span> <span class="n">value</span> <span class="o">=</span> <span class="n">object</span><span class="p">[</span><span class="n">idx</span><span class="p">];</span>
+</pre></div>
+</div>
+<p>it is translated into a call to <tt class="docutils literal"><span class="pre">objectAtIndexedSubscript:</span></tt></p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="kt">id</span> <span class="n">value</span> <span class="o">=</span> <span class="p">[</span><span class="n">object</span> <span class="n">objectAtIndexedSubscript</span><span class="o">:</span><span class="n">idx</span><span class="p">];</span>
+</pre></div>
+</div>
+<p>When an expression writes an element using an integral index:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="n">object</span><span class="p">[</span><span class="n">idx</span><span class="p">]</span> <span class="o">=</span> <span class="n">newValue</span><span class="p">;</span>
+</pre></div>
+</div>
+<p>it is translated to a call to <tt class="docutils literal"><span class="pre">setObject:atIndexedSubscript:</span></tt></p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="p">[</span><span class="n">object</span> <span class="n">setObject</span><span class="o">:</span><span class="n">newValue</span> <span class="n">atIndexedSubscript</span><span class="o">:</span><span class="n">idx</span><span class="p">];</span>
+</pre></div>
+</div>
+<p>These message sends are then type-checked and performed just like
+explicit message sends. The method used for objectAtIndexedSubscript:
+must be declared with an argument of integral type and a return value of
+some Objective-C object pointer type. The method used for
+setObject:atIndexedSubscript: must be declared with its first argument
+having some Objective-C pointer type and its second argument having
+integral type.</p>
+<p>The meaning of indexes is left up to the declaring class. The compiler
+will coerce the index to the appropriate argument type of the method it
+uses for type-checking. For an instance of <tt class="docutils literal"><span class="pre">NSArray</span></tt>, reading an
+element using an index outside the range <tt class="docutils literal"><span class="pre">[0,</span> <span class="pre">array.count)</span></tt> will raise
+an exception. For an instance of <tt class="docutils literal"><span class="pre">NSMutableArray</span></tt>, assigning to an
+element using an index within this range will replace that element, but
+assigning to an element using an index outside this range will raise an
+exception; no syntax is provided for inserting, appending, or removing
+elements for mutable arrays.</p>
+<p>A class need not declare both methods in order to take advantage of this
+language feature. For example, the class <tt class="docutils literal"><span class="pre">NSArray</span></tt> declares only
+<tt class="docutils literal"><span class="pre">objectAtIndexedSubscript:</span></tt>, so that assignments to elements will fail
+to type-check; moreover, its subclass <tt class="docutils literal"><span class="pre">NSMutableArray</span></tt> declares
+<tt class="docutils literal"><span class="pre">setObject:atIndexedSubscript:</span></tt>.</p>
+</div>
+<div class="section" id="dictionary-style-subscripting">
+<h4>Dictionary-Style Subscripting<a class="headerlink" href="#dictionary-style-subscripting" title="Permalink to this headline">¶</a></h4>
+<p>When the subscript operand has an Objective-C object pointer type, the
+expression is rewritten to use one of two different selectors, depending
+on whether the element is being read from or written to. When an
+expression reads an element using an Objective-C object pointer
+subscript operand, as in the following example:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="kt">id</span> <span class="n">key</span> <span class="o">=</span> <span class="p">...;</span>
+<span class="kt">id</span> <span class="n">value</span> <span class="o">=</span> <span class="n">object</span><span class="p">[</span><span class="n">key</span><span class="p">];</span>
+</pre></div>
+</div>
+<p>it is translated into a call to the <tt class="docutils literal"><span class="pre">objectForKeyedSubscript:</span></tt> method:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="kt">id</span> <span class="n">value</span> <span class="o">=</span> <span class="p">[</span><span class="n">object</span> <span class="n">objectForKeyedSubscript</span><span class="o">:</span><span class="n">key</span><span class="p">];</span>
+</pre></div>
+</div>
+<p>When an expression writes an element using an Objective-C object pointer
+subscript:</p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="n">object</span><span class="p">[</span><span class="n">key</span><span class="p">]</span> <span class="o">=</span> <span class="n">newValue</span><span class="p">;</span>
+</pre></div>
+</div>
+<p>it is translated to a call to <tt class="docutils literal"><span class="pre">setObject:forKeyedSubscript:</span></tt></p>
+<div class="highlight-objc"><div class="highlight"><pre><span class="p">[</span><span class="n">object</span> <span class="n">setObject</span><span class="o">:</span><span class="n">newValue</span> <span class="n">forKeyedSubscript</span><span class="o">:</span><span class="n">key</span><span class="p">];</span>
+</pre></div>
+</div>
+<p>The behavior of <tt class="docutils literal"><span class="pre">setObject:forKeyedSubscript:</span></tt> is class-specific; but
+in general it should replace an existing value if one is already
+associated with a key, otherwise it should add a new value for the key.
+No syntax is provided for removing elements from mutable dictionaries.</p>
+</div>
+</div>
+<div class="section" id="id4">
+<h3>Discussion<a class="headerlink" href="#id4" title="Permalink to this headline">¶</a></h3>
+<p>An Objective-C subscript expression occurs when the base operand of the
+C subscript operator has an Objective-C object pointer type. Since this
+potentially collides with pointer arithmetic on the value, these
+expressions are only supported under the modern Objective-C runtime,
+which categorically forbids such arithmetic.</p>
+<p>Currently, only subscripts of integral or Objective-C object pointer
+type are supported. In C++, a class type can be used if it has a single
+conversion function to an integral or Objective-C pointer type, in which
+case that conversion is applied and analysis continues as appropriate.
+Otherwise, the expression is ill-formed.</p>
+<p>An Objective-C object subscript expression is always an l-value. If the
+expression appears on the left-hand side of a simple assignment operator
+(=), the element is written as described below. If the expression
+appears on the left-hand side of a compound assignment operator (e.g.
++=), the program is ill-formed, because the result of reading an element
+is always an Objective-C object pointer and no binary operators are
+legal on such pointers. If the expression appears in any other position,
+the element is read as described below. It is an error to take the
+address of a subscript expression, or (in C++) to bind a reference to
+it.</p>
+<p>Programs can use object subscripting with Objective-C object pointers of
+type <tt class="docutils literal"><span class="pre">id</span></tt>. Normal dynamic message send rules apply; the compiler must
+see <em>some</em> declaration of the subscripting methods, and will pick the
+declaration seen first.</p>
+</div>
+</div>
+<div class="section" id="caveats">
+<h2>Caveats<a class="headerlink" href="#caveats" title="Permalink to this headline">¶</a></h2>
+<p>Objects created using the literal or boxed expression syntax are not
+guaranteed to be uniqued by the runtime, but nor are they guaranteed to
+be newly-allocated. As such, the result of performing direct comparisons
+against the location of an object literal (using <tt class="docutils literal"><span class="pre">==</span></tt>, <tt class="docutils literal"><span class="pre">!=</span></tt>, <tt class="docutils literal"><span class="pre"><</span></tt>,
+<tt class="docutils literal"><span class="pre"><=</span></tt>, <tt class="docutils literal"><span class="pre">></span></tt>, or <tt class="docutils literal"><span class="pre">>=</span></tt>) is not well-defined. This is usually a simple
+mistake in code that intended to call the <tt class="docutils literal"><span class="pre">isEqual:</span></tt> method (or the
+<tt class="docutils literal"><span class="pre">compare:</span></tt> method).</p>
+<p>This caveat applies to compile-time string literals as well.
+Historically, string literals (using the <tt class="docutils literal"><span class="pre">@"..."</span></tt> syntax) have been
+uniqued across translation units during linking. This is an
+implementation detail of the compiler and should not be relied upon. If
+you are using such code, please use global string constants instead
+(<tt class="docutils literal"><span class="pre">NSString</span> <span class="pre">*</span> <span class="pre">const</span> <span class="pre">MyConst</span> <span class="pre">=</span> <span class="pre">@"..."</span></tt>) or use <tt class="docutils literal"><span class="pre">isEqual:</span></tt>.</p>
+</div>
+<div class="section" id="grammar-additions">
+<h2>Grammar Additions<a class="headerlink" href="#grammar-additions" title="Permalink to this headline">¶</a></h2>
+<p>To support the new syntax described above, the Objective-C
+<tt class="docutils literal"><span class="pre">@</span></tt>-expression grammar has the following new productions:</p>
+<div class="highlight-python"><div class="highlight"><pre>objc-at-expression : '@' (string-literal | encode-literal | selector-literal | protocol-literal | object-literal)
+ ;
+
+object-literal : ('+' | '-')? numeric-constant
+ | character-constant
+ | boolean-constant
+ | array-literal
+ | dictionary-literal
+ ;
+
+boolean-constant : '__objc_yes' | '__objc_no' | 'true' | 'false' /* boolean keywords. */
+ ;
+
+array-literal : '[' assignment-expression-list ']'
+ ;
+
+assignment-expression-list : assignment-expression (',' assignment-expression-list)?
+ | /* empty */
+ ;
+
+dictionary-literal : '{' key-value-list '}'
+ ;
+
+key-value-list : key-value-pair (',' key-value-list)?
+ | /* empty */
+ ;
+
+key-value-pair : assignment-expression ':' assignment-expression
+ ;
+</pre></div>
+</div>
+<p>Note: <tt class="docutils literal"><span class="pre">@true</span></tt> and <tt class="docutils literal"><span class="pre">@false</span></tt> are only supported in Objective-C++.</p>
+</div>
+<div class="section" id="availability-checks">
+<h2>Availability Checks<a class="headerlink" href="#availability-checks" title="Permalink to this headline">¶</a></h2>
+<p>Programs test for the new features by using clang’s __has_feature
+checks. Here are examples of their use:</p>
+<div class="highlight-objc"><div class="highlight"><pre>#if __has_feature(objc_array_literals)
+ // new way.
+ NSArray *elements = @[ @"H", @"He", @"O", @"C" ];
+#else
+ // old way (equivalent).
+ id objects[] = { @"H", @"He", @"O", @"C" };
+ NSArray *elements = [NSArray arrayWithObjects:objects count:4];
+#endif
+
+#if __has_feature(objc_dictionary_literals)
+ // new way.
+ NSDictionary *masses = @{ @"H" : @1.0078, @"He" : @4.0026, @"O" : @15.9990, @"C" : @12.0096 };
+#else
+ // old way (equivalent).
+ id keys[] = { @"H", @"He", @"O", @"C" };
+ id values[] = { [NSNumber numberWithDouble:1.0078], [NSNumber numberWithDouble:4.0026],
+ [NSNumber numberWithDouble:15.9990], [NSNumber numberWithDouble:12.0096] };
+ NSDictionary *masses = [NSDictionary dictionaryWithObjects:objects forKeys:keys count:4];
+#endif
+
+#if __has_feature(objc_subscripting)
+ NSUInteger i, count = elements.count;
+ for (i = 0; i < count; ++i) {
+ NSString *element = elements[i];
+ NSNumber *mass = masses[element];
+ NSLog(@"the mass of %@ is %@", element, mass);
+ }
+#else
+ NSUInteger i, count = [elements count];
+ for (i = 0; i < count; ++i) {
+ NSString *element = [elements objectAtIndex:i];
+ NSNumber *mass = [masses objectForKey:element];
+ NSLog(@"the mass of %@ is %@", element, mass);
+ }
+#endif
+
+#if __has_attribute(objc_boxable)
+ typedef struct __attribute__((objc_boxable)) _Rect Rect;
+#endif
+
+#if __has_feature(objc_boxed_nsvalue_expressions)
+ CABasicAnimation animation = [CABasicAnimation animationWithKeyPath:@"position"];
+ animation.fromValue = @(layer.position);
+ animation.toValue = @(newPosition);
+ [layer addAnimation:animation forKey:@"move"];
+#else
+ CABasicAnimation animation = [CABasicAnimation animationWithKeyPath:@"position"];
+ animation.fromValue = [NSValue valueWithCGPoint:layer.position];
+ animation.toValue = [NSValue valueWithCGPoint:newPosition];
+ [layer addAnimation:animation forKey:@"move"];
+#endif
+</pre></div>
+</div>
+<p>Code can use also <tt class="docutils literal"><span class="pre">__has_feature(objc_bool)</span></tt> to check for the
+availability of numeric literals support. This checks for the new
+<tt class="docutils literal"><span class="pre">__objc_yes</span> <span class="pre">/</span> <span class="pre">__objc_no</span></tt> keywords, which enable the use of
+<tt class="docutils literal"><span class="pre">@YES</span> <span class="pre">/</span> <span class="pre">@NO</span></tt> literals.</p>
+<p>To check whether boxed expressions are supported, use
+<tt class="docutils literal"><span class="pre">__has_feature(objc_boxed_expressions)</span></tt> feature macro.</p>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="LanguageExtensions.html">Clang Language Extensions</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="BlockLanguageSpec.html">Language Specification for Blocks</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/PCHInternals.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/PCHInternals.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/PCHInternals.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/PCHInternals.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,585 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Precompiled Header and Modules Internals — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="prev" title="Pretokenized Headers (PTH)" href="PTHInternals.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Precompiled Header and Modules Internals</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="PTHInternals.html">Pretokenized Headers (PTH)</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="precompiled-header-and-modules-internals">
+<h1>Precompiled Header and Modules Internals<a class="headerlink" href="#precompiled-header-and-modules-internals" title="Permalink to this headline">¶</a></h1>
+<div class="contents local topic" id="contents">
+<ul class="simple">
+<li><a class="reference internal" href="#using-precompiled-headers-with-clang" id="id1">Using Precompiled Headers with <tt class="docutils literal"><span class="pre">clang</span></tt></a></li>
+<li><a class="reference internal" href="#design-philosophy" id="id2">Design Philosophy</a></li>
+<li><a class="reference internal" href="#ast-file-contents" id="id3">AST File Contents</a><ul>
+<li><a class="reference internal" href="#metadata-block" id="id4">Metadata Block</a></li>
+<li><a class="reference internal" href="#source-manager-block" id="id5">Source Manager Block</a></li>
+<li><a class="reference internal" href="#preprocessor-block" id="id6">Preprocessor Block</a></li>
+<li><a class="reference internal" href="#types-block" id="id7">Types Block</a></li>
+<li><a class="reference internal" href="#declarations-block" id="id8">Declarations Block</a></li>
+<li><a class="reference internal" href="#statements-and-expressions" id="id9">Statements and Expressions</a></li>
+<li><a class="reference internal" href="#pchinternals-ident-table" id="id10">Identifier Table Block</a></li>
+<li><a class="reference internal" href="#method-pool-block" id="id11">Method Pool Block</a></li>
+</ul>
+</li>
+<li><a class="reference internal" href="#ast-reader-integration-points" id="id12">AST Reader Integration Points</a></li>
+<li><a class="reference internal" href="#chained-precompiled-headers" id="id13">Chained precompiled headers</a></li>
+<li><a class="reference internal" href="#modules" id="id14">Modules</a></li>
+</ul>
+</div>
+<p>This document describes the design and implementation of Clang’s precompiled
+headers (PCH) and modules. If you are interested in the end-user view, please
+see the <a class="reference internal" href="UsersManual.html#usersmanual-precompiled-headers"><em>User’s Manual</em></a>.</p>
+<div class="section" id="using-precompiled-headers-with-clang">
+<h2><a class="toc-backref" href="#id1">Using Precompiled Headers with <tt class="docutils literal"><span class="pre">clang</span></tt></a><a class="headerlink" href="#using-precompiled-headers-with-clang" title="Permalink to this headline">¶</a></h2>
+<p>The Clang compiler frontend, <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span></tt>, supports two command line options
+for generating and using PCH files.</p>
+<p>To generate PCH files using <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span></tt>, use the option <em class="xref std std-option">-emit-pch</em>:</p>
+<div class="highlight-bash"><div class="highlight"><pre><span class="nv">$ </span>clang -cc1 test.h -emit-pch -o test.h.pch
+</pre></div>
+</div>
+<p>This option is transparently used by <tt class="docutils literal"><span class="pre">clang</span></tt> when generating PCH files. The
+resulting PCH file contains the serialized form of the compiler’s internal
+representation after it has completed parsing and semantic analysis. The PCH
+file can then be used as a prefix header with the <em class="xref std std-option">-include-pch</em>
+option:</p>
+<div class="highlight-bash"><div class="highlight"><pre><span class="nv">$ </span>clang -cc1 -include-pch test.h.pch test.c -o test.s
+</pre></div>
+</div>
+</div>
+<div class="section" id="design-philosophy">
+<h2><a class="toc-backref" href="#id2">Design Philosophy</a><a class="headerlink" href="#design-philosophy" title="Permalink to this headline">¶</a></h2>
+<p>Precompiled headers are meant to improve overall compile times for projects, so
+the design of precompiled headers is entirely driven by performance concerns.
+The use case for precompiled headers is relatively simple: when there is a
+common set of headers that is included in nearly every source file in the
+project, we <em>precompile</em> that bundle of headers into a single precompiled
+header (PCH file). Then, when compiling the source files in the project, we
+load the PCH file first (as a prefix header), which acts as a stand-in for that
+bundle of headers.</p>
+<p>A precompiled header implementation improves performance when:</p>
+<ul class="simple">
+<li>Loading the PCH file is significantly faster than re-parsing the bundle of
+headers stored within the PCH file. Thus, a precompiled header design
+attempts to minimize the cost of reading the PCH file. Ideally, this cost
+should not vary with the size of the precompiled header file.</li>
+<li>The cost of generating the PCH file initially is not so large that it
+counters the per-source-file performance improvement due to eliminating the
+need to parse the bundled headers in the first place. This is particularly
+important on multi-core systems, because PCH file generation serializes the
+build when all compilations require the PCH file to be up-to-date.</li>
+</ul>
+<p>Modules, as implemented in Clang, use the same mechanisms as precompiled
+headers to save a serialized AST file (one per module) and use those AST
+modules. From an implementation standpoint, modules are a generalization of
+precompiled headers, lifting a number of restrictions placed on precompiled
+headers. In particular, there can only be one precompiled header and it must
+be included at the beginning of the translation unit. The extensions to the
+AST file format required for modules are discussed in the section on
+<a class="reference internal" href="#pchinternals-modules"><em>modules</em></a>.</p>
+<p>Clang’s AST files are designed with a compact on-disk representation, which
+minimizes both creation time and the time required to initially load the AST
+file. The AST file itself contains a serialized representation of Clang’s
+abstract syntax trees and supporting data structures, stored using the same
+compressed bitstream as <a class="reference external" href="http://llvm.org/docs/BitCodeFormat.html">LLVM’s bitcode file format</a>.</p>
+<p>Clang’s AST files are loaded “lazily” from disk. When an AST file is initially
+loaded, Clang reads only a small amount of data from the AST file to establish
+where certain important data structures are stored. The amount of data read in
+this initial load is independent of the size of the AST file, such that a
+larger AST file does not lead to longer AST load times. The actual header data
+in the AST file — macros, functions, variables, types, etc. — is loaded
+only when it is referenced from the user’s code, at which point only that
+entity (and those entities it depends on) are deserialized from the AST file.
+With this approach, the cost of using an AST file for a translation unit is
+proportional to the amount of code actually used from the AST file, rather than
+being proportional to the size of the AST file itself.</p>
+<p>When given the <em class="xref std std-option">-print-stats</em> option, Clang produces statistics
+describing how much of the AST file was actually loaded from disk. For a
+simple “Hello, World!” program that includes the Apple <tt class="docutils literal"><span class="pre">Cocoa.h</span></tt> header
+(which is built as a precompiled header), this option illustrates how little of
+the actual precompiled header is required:</p>
+<div class="highlight-none"><div class="highlight"><pre>*** AST File Statistics:
+ 895/39981 source location entries read (2.238563%)
+ 19/15315 types read (0.124061%)
+ 20/82685 declarations read (0.024188%)
+ 154/58070 identifiers read (0.265197%)
+ 0/7260 selectors read (0.000000%)
+ 0/30842 statements read (0.000000%)
+ 4/8400 macros read (0.047619%)
+ 1/4995 lexical declcontexts read (0.020020%)
+ 0/4413 visible declcontexts read (0.000000%)
+ 0/7230 method pool entries read (0.000000%)
+ 0 method pool misses
+</pre></div>
+</div>
+<p>For this small program, only a tiny fraction of the source locations, types,
+declarations, identifiers, and macros were actually deserialized from the
+precompiled header. These statistics can be useful to determine whether the
+AST file implementation can be improved by making more of the implementation
+lazy.</p>
+<p>Precompiled headers can be chained. When you create a PCH while including an
+existing PCH, Clang can create the new PCH by referencing the original file and
+only writing the new data to the new file. For example, you could create a PCH
+out of all the headers that are very commonly used throughout your project, and
+then create a PCH for every single source file in the project that includes the
+code that is specific to that file, so that recompiling the file itself is very
+fast, without duplicating the data from the common headers for every file. The
+mechanisms behind chained precompiled headers are discussed in a <a class="reference internal" href="#pchinternals-chained"><em>later
+section</em></a>.</p>
+</div>
+<div class="section" id="ast-file-contents">
+<h2><a class="toc-backref" href="#id3">AST File Contents</a><a class="headerlink" href="#ast-file-contents" title="Permalink to this headline">¶</a></h2>
+<p>An AST file produced by clang is an object file container with a <tt class="docutils literal"><span class="pre">clangast</span></tt>
+(COFF) or <tt class="docutils literal"><span class="pre">__clangast</span></tt> (ELF and Mach-O) section containing the serialized AST.
+Other target-specific sections in the object file container are used to hold
+debug information for the data types defined in the AST. Tools built on top of
+libclang that do not need debug information may also produce raw AST files that
+only contain the serialized AST.</p>
+<p>The <tt class="docutils literal"><span class="pre">clangast</span></tt> section is organized into several different blocks, each of
+which contains the serialized representation of a part of Clang’s internal
+representation. Each of the blocks corresponds to either a block or a record
+within <a class="reference external" href="http://llvm.org/docs/BitCodeFormat.html">LLVM’s bitstream format</a>.
+The contents of each of these logical blocks are described below.</p>
+<img alt="_images/PCHLayout.png" src="_images/PCHLayout.png" />
+<p>The <tt class="docutils literal"><span class="pre">llvm-objdump</span></tt> utility provides a <tt class="docutils literal"><span class="pre">-raw-clang-ast</span></tt> option to extract the
+binary contents of the AST section from an object file container.</p>
+<p>The <a class="reference external" href="http://llvm.org/docs/CommandGuide/llvm-bcanalyzer.html">llvm-bcanalyzer</a>
+utility can be used to examine the actual structure of the bitstream for the AST
+section. This information can be used both to help understand the structure of
+the AST section and to isolate areas where the AST representation can still be
+optimized, e.g., through the introduction of abbreviations.</p>
+<div class="section" id="metadata-block">
+<h3><a class="toc-backref" href="#id4">Metadata Block</a><a class="headerlink" href="#metadata-block" title="Permalink to this headline">¶</a></h3>
+<p>The metadata block contains several records that provide information about how
+the AST file was built. This metadata is primarily used to validate the use of
+an AST file. For example, a precompiled header built for a 32-bit x86 target
+cannot be used when compiling for a 64-bit x86 target. The metadata block
+contains information about:</p>
+<dl class="docutils">
+<dt>Language options</dt>
+<dd>Describes the particular language dialect used to compile the AST file,
+including major options (e.g., Objective-C support) and more minor options
+(e.g., support for “<tt class="docutils literal"><span class="pre">//</span></tt>” comments). The contents of this record correspond to
+the <tt class="docutils literal"><span class="pre">LangOptions</span></tt> class.</dd>
+<dt>Target architecture</dt>
+<dd>The target triple that describes the architecture, platform, and ABI for
+which the AST file was generated, e.g., <tt class="docutils literal"><span class="pre">i386-apple-darwin9</span></tt>.</dd>
+<dt>AST version</dt>
+<dd>The major and minor version numbers of the AST file format. Changes in the
+minor version number should not affect backward compatibility, while changes
+in the major version number imply that a newer compiler cannot read an older
+precompiled header (and vice-versa).</dd>
+<dt>Original file name</dt>
+<dd>The full path of the header that was used to generate the AST file.</dd>
+<dt>Predefines buffer</dt>
+<dd>Although not explicitly stored as part of the metadata, the predefines buffer
+is used in the validation of the AST file. The predefines buffer itself
+contains code generated by the compiler to initialize the preprocessor state
+according to the current target, platform, and command-line options. For
+example, the predefines buffer will contain “<tt class="docutils literal"><span class="pre">#define</span> <span class="pre">__STDC__</span> <span class="pre">1</span></tt>” when we
+are compiling C without Microsoft extensions. The predefines buffer itself
+is stored within the <a class="reference internal" href="#pchinternals-sourcemgr"><em>Source Manager Block</em></a>, but its contents are
+verified along with the rest of the metadata.</dd>
+</dl>
+<p>A chained PCH file (that is, one that references another PCH) and a module
+(which may import other modules) have additional metadata containing the list
+of all AST files that this AST file depends on. Each of those files will be
+loaded along with this AST file.</p>
+<p>For chained precompiled headers, the language options, target architecture and
+predefines buffer data is taken from the end of the chain, since they have to
+match anyway.</p>
+</div>
+<div class="section" id="source-manager-block">
+<span id="pchinternals-sourcemgr"></span><h3><a class="toc-backref" href="#id5">Source Manager Block</a><a class="headerlink" href="#source-manager-block" title="Permalink to this headline">¶</a></h3>
+<p>The source manager block contains the serialized representation of Clang’s
+<a class="reference internal" href="InternalsManual.html#sourcemanager"><em>SourceManager</em></a> class, which handles the mapping from
+source locations (as represented in Clang’s abstract syntax tree) into actual
+column/line positions within a source file or macro instantiation. The AST
+file’s representation of the source manager also includes information about all
+of the headers that were (transitively) included when building the AST file.</p>
+<p>The bulk of the source manager block is dedicated to information about the
+various files, buffers, and macro instantiations into which a source location
+can refer. Each of these is referenced by a numeric “file ID”, which is a
+unique number (allocated starting at 1) stored in the source location. Clang
+serializes the information for each kind of file ID, along with an index that
+maps file IDs to the position within the AST file where the information about
+that file ID is stored. The data associated with a file ID is loaded only when
+required by the front end, e.g., to emit a diagnostic that includes a macro
+instantiation history inside the header itself.</p>
+<p>The source manager block also contains information about all of the headers
+that were included when building the AST file. This includes information about
+the controlling macro for the header (e.g., when the preprocessor identified
+that the contents of the header dependent on a macro like
+<tt class="docutils literal"><span class="pre">LLVM_CLANG_SOURCEMANAGER_H</span></tt>).</p>
+</div>
+<div class="section" id="preprocessor-block">
+<span id="pchinternals-preprocessor"></span><h3><a class="toc-backref" href="#id6">Preprocessor Block</a><a class="headerlink" href="#preprocessor-block" title="Permalink to this headline">¶</a></h3>
+<p>The preprocessor block contains the serialized representation of the
+preprocessor. Specifically, it contains all of the macros that have been
+defined by the end of the header used to build the AST file, along with the
+token sequences that comprise each macro. The macro definitions are only read
+from the AST file when the name of the macro first occurs in the program. This
+lazy loading of macro definitions is triggered by lookups into the
+<a class="reference internal" href="#pchinternals-ident-table"><em>identifier table</em></a>.</p>
+</div>
+<div class="section" id="types-block">
+<span id="pchinternals-types"></span><h3><a class="toc-backref" href="#id7">Types Block</a><a class="headerlink" href="#types-block" title="Permalink to this headline">¶</a></h3>
+<p>The types block contains the serialized representation of all of the types
+referenced in the translation unit. Each Clang type node (<tt class="docutils literal"><span class="pre">PointerType</span></tt>,
+<tt class="docutils literal"><span class="pre">FunctionProtoType</span></tt>, etc.) has a corresponding record type in the AST file.
+When types are deserialized from the AST file, the data within the record is
+used to reconstruct the appropriate type node using the AST context.</p>
+<p>Each type has a unique type ID, which is an integer that uniquely identifies
+that type. Type ID 0 represents the NULL type, type IDs less than
+<tt class="docutils literal"><span class="pre">NUM_PREDEF_TYPE_IDS</span></tt> represent predefined types (<tt class="docutils literal"><span class="pre">void</span></tt>, <tt class="docutils literal"><span class="pre">float</span></tt>, etc.),
+while other “user-defined” type IDs are assigned consecutively from
+<tt class="docutils literal"><span class="pre">NUM_PREDEF_TYPE_IDS</span></tt> upward as the types are encountered. The AST file has
+an associated mapping from the user-defined types block to the location within
+the types block where the serialized representation of that type resides,
+enabling lazy deserialization of types. When a type is referenced from within
+the AST file, that reference is encoded using the type ID shifted left by 3
+bits. The lower three bits are used to represent the <tt class="docutils literal"><span class="pre">const</span></tt>, <tt class="docutils literal"><span class="pre">volatile</span></tt>,
+and <tt class="docutils literal"><span class="pre">restrict</span></tt> qualifiers, as in Clang’s <a class="reference internal" href="InternalsManual.html#qualtype"><em>QualType</em></a> class.</p>
+</div>
+<div class="section" id="declarations-block">
+<span id="pchinternals-decls"></span><h3><a class="toc-backref" href="#id8">Declarations Block</a><a class="headerlink" href="#declarations-block" title="Permalink to this headline">¶</a></h3>
+<p>The declarations block contains the serialized representation of all of the
+declarations referenced in the translation unit. Each Clang declaration node
+(<tt class="docutils literal"><span class="pre">VarDecl</span></tt>, <tt class="docutils literal"><span class="pre">FunctionDecl</span></tt>, etc.) has a corresponding record type in the
+AST file. When declarations are deserialized from the AST file, the data
+within the record is used to build and populate a new instance of the
+corresponding <tt class="docutils literal"><span class="pre">Decl</span></tt> node. As with types, each declaration node has a
+numeric ID that is used to refer to that declaration within the AST file. In
+addition, a lookup table provides a mapping from that numeric ID to the offset
+within the precompiled header where that declaration is described.</p>
+<p>Declarations in Clang’s abstract syntax trees are stored hierarchically. At
+the top of the hierarchy is the translation unit (<tt class="docutils literal"><span class="pre">TranslationUnitDecl</span></tt>),
+which contains all of the declarations in the translation unit but is not
+actually written as a specific declaration node. Its child declarations (such
+as functions or struct types) may also contain other declarations inside them,
+and so on. Within Clang, each declaration is stored within a <a class="reference internal" href="InternalsManual.html#declcontext"><em>declaration
+context</em></a>, as represented by the <tt class="docutils literal"><span class="pre">DeclContext</span></tt> class.
+Declaration contexts provide the mechanism to perform name lookup within a
+given declaration (e.g., find the member named <tt class="docutils literal"><span class="pre">x</span></tt> in a structure) and
+iterate over the declarations stored within a context (e.g., iterate over all
+of the fields of a structure for structure layout).</p>
+<p>In Clang’s AST file format, deserializing a declaration that is a
+<tt class="docutils literal"><span class="pre">DeclContext</span></tt> is a separate operation from deserializing all of the
+declarations stored within that declaration context. Therefore, Clang will
+deserialize the translation unit declaration without deserializing the
+declarations within that translation unit. When required, the declarations
+stored within a declaration context will be deserialized. There are two
+representations of the declarations within a declaration context, which
+correspond to the name-lookup and iteration behavior described above:</p>
+<ul class="simple">
+<li>When the front end performs name lookup to find a name <tt class="docutils literal"><span class="pre">x</span></tt> within a given
+declaration context (for example, during semantic analysis of the expression
+<tt class="docutils literal"><span class="pre">p->x</span></tt>, where <tt class="docutils literal"><span class="pre">p</span></tt>‘s type is defined in the precompiled header), Clang
+refers to an on-disk hash table that maps from the names within that
+declaration context to the declaration IDs that represent each visible
+declaration with that name. The actual declarations will then be
+deserialized to provide the results of name lookup.</li>
+<li>When the front end performs iteration over all of the declarations within a
+declaration context, all of those declarations are immediately
+de-serialized. For large declaration contexts (e.g., the translation unit),
+this operation is expensive; however, large declaration contexts are not
+traversed in normal compilation, since such a traversal is unnecessary.
+However, it is common for the code generator and semantic analysis to
+traverse declaration contexts for structs, classes, unions, and
+enumerations, although those contexts contain relatively few declarations in
+the common case.</li>
+</ul>
+</div>
+<div class="section" id="statements-and-expressions">
+<h3><a class="toc-backref" href="#id9">Statements and Expressions</a><a class="headerlink" href="#statements-and-expressions" title="Permalink to this headline">¶</a></h3>
+<p>Statements and expressions are stored in the AST file in both the <a class="reference internal" href="#pchinternals-types"><em>types</em></a> and the <a class="reference internal" href="#pchinternals-decls"><em>declarations</em></a> blocks,
+because every statement or expression will be associated with either a type or
+declaration. The actual statement and expression records are stored
+immediately following the declaration or type that owns the statement or
+expression. For example, the statement representing the body of a function
+will be stored directly following the declaration of the function.</p>
+<p>As with types and declarations, each statement and expression kind in Clang’s
+abstract syntax tree (<tt class="docutils literal"><span class="pre">ForStmt</span></tt>, <tt class="docutils literal"><span class="pre">CallExpr</span></tt>, etc.) has a corresponding
+record type in the AST file, which contains the serialized representation of
+that statement or expression. Each substatement or subexpression within an
+expression is stored as a separate record (which keeps most records to a fixed
+size). Within the AST file, the subexpressions of an expression are stored, in
+reverse order, prior to the expression that owns those expression, using a form
+of <a class="reference external" href="http://en.wikipedia.org/wiki/Reverse_Polish_notation">Reverse Polish Notation</a>. For example, an
+expression <tt class="docutils literal"><span class="pre">3</span> <span class="pre">-</span> <span class="pre">4</span> <span class="pre">+</span> <span class="pre">5</span></tt> would be represented as follows:</p>
+<table border="1" class="docutils">
+<colgroup>
+<col width="100%" />
+</colgroup>
+<tbody valign="top">
+<tr class="row-odd"><td><tt class="docutils literal"><span class="pre">IntegerLiteral(5)</span></tt></td>
+</tr>
+<tr class="row-even"><td><tt class="docutils literal"><span class="pre">IntegerLiteral(4)</span></tt></td>
+</tr>
+<tr class="row-odd"><td><tt class="docutils literal"><span class="pre">IntegerLiteral(3)</span></tt></td>
+</tr>
+<tr class="row-even"><td><tt class="docutils literal"><span class="pre">IntegerLiteral(-)</span></tt></td>
+</tr>
+<tr class="row-odd"><td><tt class="docutils literal"><span class="pre">IntegerLiteral(+)</span></tt></td>
+</tr>
+<tr class="row-even"><td><tt class="docutils literal"><span class="pre">STOP</span></tt></td>
+</tr>
+</tbody>
+</table>
+<p>When reading this representation, Clang evaluates each expression record it
+encounters, builds the appropriate abstract syntax tree node, and then pushes
+that expression on to a stack. When a record contains <em>N</em> subexpressions —
+<tt class="docutils literal"><span class="pre">BinaryOperator</span></tt> has two of them — those expressions are popped from the
+top of the stack. The special STOP code indicates that we have reached the end
+of a serialized expression or statement; other expression or statement records
+may follow, but they are part of a different expression.</p>
+</div>
+<div class="section" id="pchinternals-ident-table">
+<span id="identifier-table-block"></span><h3><a class="toc-backref" href="#id10">Identifier Table Block</a><a class="headerlink" href="#pchinternals-ident-table" title="Permalink to this headline">¶</a></h3>
+<p>The identifier table block contains an on-disk hash table that maps each
+identifier mentioned within the AST file to the serialized representation of
+the identifier’s information (e.g, the <tt class="docutils literal"><span class="pre">IdentifierInfo</span></tt> structure). The
+serialized representation contains:</p>
+<ul class="simple">
+<li>The actual identifier string.</li>
+<li>Flags that describe whether this identifier is the name of a built-in, a
+poisoned identifier, an extension token, or a macro.</li>
+<li>If the identifier names a macro, the offset of the macro definition within
+the <a class="reference internal" href="#pchinternals-preprocessor"><em>Preprocessor Block</em></a>.</li>
+<li>If the identifier names one or more declarations visible from translation
+unit scope, the <a class="reference internal" href="#pchinternals-decls"><em>declaration IDs</em></a> of these
+declarations.</li>
+</ul>
+<p>When an AST file is loaded, the AST file reader mechanism introduces itself
+into the identifier table as an external lookup source. Thus, when the user
+program refers to an identifier that has not yet been seen, Clang will perform
+a lookup into the identifier table. If an identifier is found, its contents
+(macro definitions, flags, top-level declarations, etc.) will be deserialized,
+at which point the corresponding <tt class="docutils literal"><span class="pre">IdentifierInfo</span></tt> structure will have the
+same contents it would have after parsing the headers in the AST file.</p>
+<p>Within the AST file, the identifiers used to name declarations are represented
+with an integral value. A separate table provides a mapping from this integral
+value (the identifier ID) to the location within the on-disk hash table where
+that identifier is stored. This mapping is used when deserializing the name of
+a declaration, the identifier of a token, or any other construct in the AST
+file that refers to a name.</p>
+</div>
+<div class="section" id="method-pool-block">
+<span id="pchinternals-method-pool"></span><h3><a class="toc-backref" href="#id11">Method Pool Block</a><a class="headerlink" href="#method-pool-block" title="Permalink to this headline">¶</a></h3>
+<p>The method pool block is represented as an on-disk hash table that serves two
+purposes: it provides a mapping from the names of Objective-C selectors to the
+set of Objective-C instance and class methods that have that particular
+selector (which is required for semantic analysis in Objective-C) and also
+stores all of the selectors used by entities within the AST file. The design
+of the method pool is similar to that of the <a class="reference internal" href="#pchinternals-ident-table"><em>identifier table</em></a>: the first time a particular selector is formed
+during the compilation of the program, Clang will search in the on-disk hash
+table of selectors; if found, Clang will read the Objective-C methods
+associated with that selector into the appropriate front-end data structure
+(<tt class="docutils literal"><span class="pre">Sema::InstanceMethodPool</span></tt> and <tt class="docutils literal"><span class="pre">Sema::FactoryMethodPool</span></tt> for instance and
+class methods, respectively).</p>
+<p>As with identifiers, selectors are represented by numeric values within the AST
+file. A separate index maps these numeric selector values to the offset of the
+selector within the on-disk hash table, and will be used when de-serializing an
+Objective-C method declaration (or other Objective-C construct) that refers to
+the selector.</p>
+</div>
+</div>
+<div class="section" id="ast-reader-integration-points">
+<h2><a class="toc-backref" href="#id12">AST Reader Integration Points</a><a class="headerlink" href="#ast-reader-integration-points" title="Permalink to this headline">¶</a></h2>
+<p>The “lazy” deserialization behavior of AST files requires their integration
+into several completely different submodules of Clang. For example, lazily
+deserializing the declarations during name lookup requires that the name-lookup
+routines be able to query the AST file to find entities stored there.</p>
+<p>For each Clang data structure that requires direct interaction with the AST
+reader logic, there is an abstract class that provides the interface between
+the two modules. The <tt class="docutils literal"><span class="pre">ASTReader</span></tt> class, which handles the loading of an AST
+file, inherits from all of these abstract classes to provide lazy
+deserialization of Clang’s data structures. <tt class="docutils literal"><span class="pre">ASTReader</span></tt> implements the
+following abstract classes:</p>
+<dl class="docutils">
+<dt><tt class="docutils literal"><span class="pre">ExternalSLocEntrySource</span></tt></dt>
+<dd>This abstract interface is associated with the <tt class="docutils literal"><span class="pre">SourceManager</span></tt> class, and
+is used whenever the <a class="reference internal" href="#pchinternals-sourcemgr"><em>source manager</em></a> needs to
+load the details of a file, buffer, or macro instantiation.</dd>
+<dt><tt class="docutils literal"><span class="pre">IdentifierInfoLookup</span></tt></dt>
+<dd>This abstract interface is associated with the <tt class="docutils literal"><span class="pre">IdentifierTable</span></tt> class, and
+is used whenever the program source refers to an identifier that has not yet
+been seen. In this case, the AST reader searches for this identifier within
+its <a class="reference internal" href="#pchinternals-ident-table"><em>identifier table</em></a> to load any top-level
+declarations or macros associated with that identifier.</dd>
+<dt><tt class="docutils literal"><span class="pre">ExternalASTSource</span></tt></dt>
+<dd>This abstract interface is associated with the <tt class="docutils literal"><span class="pre">ASTContext</span></tt> class, and is
+used whenever the abstract syntax tree nodes need to loaded from the AST
+file. It provides the ability to de-serialize declarations and types
+identified by their numeric values, read the bodies of functions when
+required, and read the declarations stored within a declaration context
+(either for iteration or for name lookup).</dd>
+<dt><tt class="docutils literal"><span class="pre">ExternalSemaSource</span></tt></dt>
+<dd>This abstract interface is associated with the <tt class="docutils literal"><span class="pre">Sema</span></tt> class, and is used
+whenever semantic analysis needs to read information from the <a class="reference internal" href="#pchinternals-method-pool"><em>global
+method pool</em></a>.</dd>
+</dl>
+</div>
+<div class="section" id="chained-precompiled-headers">
+<span id="pchinternals-chained"></span><h2><a class="toc-backref" href="#id13">Chained precompiled headers</a><a class="headerlink" href="#chained-precompiled-headers" title="Permalink to this headline">¶</a></h2>
+<p>Chained precompiled headers were initially intended to improve the performance
+of IDE-centric operations such as syntax highlighting and code completion while
+a particular source file is being edited by the user. To minimize the amount
+of reparsing required after a change to the file, a form of precompiled header
+— called a precompiled <em>preamble</em> — is automatically generated by parsing
+all of the headers in the source file, up to and including the last
+<tt class="docutils literal"><span class="pre">#include</span></tt>. When only the source file changes (and none of the headers it
+depends on), reparsing of that source file can use the precompiled preamble and
+start parsing after the <tt class="docutils literal"><span class="pre">#include</span></tt>s, so parsing time is proportional to the
+size of the source file (rather than all of its includes). However, the
+compilation of that translation unit may already use a precompiled header: in
+this case, Clang will create the precompiled preamble as a chained precompiled
+header that refers to the original precompiled header. This drastically
+reduces the time needed to serialize the precompiled preamble for use in
+reparsing.</p>
+<p>Chained precompiled headers get their name because each precompiled header can
+depend on one other precompiled header, forming a chain of dependencies. A
+translation unit will then include the precompiled header that starts the chain
+(i.e., nothing depends on it). This linearity of dependencies is important for
+the semantic model of chained precompiled headers, because the most-recent
+precompiled header can provide information that overrides the information
+provided by the precompiled headers it depends on, just like a header file
+<tt class="docutils literal"><span class="pre">B.h</span></tt> that includes another header <tt class="docutils literal"><span class="pre">A.h</span></tt> can modify the state produced by
+parsing <tt class="docutils literal"><span class="pre">A.h</span></tt>, e.g., by <tt class="docutils literal"><span class="pre">#undef</span></tt>‘ing a macro defined in <tt class="docutils literal"><span class="pre">A.h</span></tt>.</p>
+<p>There are several ways in which chained precompiled headers generalize the AST
+file model:</p>
+<dl class="docutils">
+<dt>Numbering of IDs</dt>
+<dd>Many different kinds of entities — identifiers, declarations, types, etc.
+— have ID numbers that start at 1 or some other predefined constant and
+grow upward. Each precompiled header records the maximum ID number it has
+assigned in each category. Then, when a new precompiled header is generated
+that depends on (chains to) another precompiled header, it will start
+counting at the next available ID number. This way, one can determine, given
+an ID number, which AST file actually contains the entity.</dd>
+<dt>Name lookup</dt>
+<dd>When writing a chained precompiled header, Clang attempts to write only
+information that has changed from the precompiled header on which it is
+based. This changes the lookup algorithm for the various tables, such as the
+<a class="reference internal" href="#pchinternals-ident-table"><em>identifier table</em></a>: the search starts at the
+most-recent precompiled header. If no entry is found, lookup then proceeds
+to the identifier table in the precompiled header it depends on, and so one.
+Once a lookup succeeds, that result is considered definitive, overriding any
+results from earlier precompiled headers.</dd>
+<dt>Update records</dt>
+<dd>There are various ways in which a later precompiled header can modify the
+entities described in an earlier precompiled header. For example, later
+precompiled headers can add entries into the various name-lookup tables for
+the translation unit or namespaces, or add new categories to an Objective-C
+class. Each of these updates is captured in an “update record” that is
+stored in the chained precompiled header file and will be loaded along with
+the original entity.</dd>
+</dl>
+</div>
+<div class="section" id="modules">
+<span id="pchinternals-modules"></span><h2><a class="toc-backref" href="#id14">Modules</a><a class="headerlink" href="#modules" title="Permalink to this headline">¶</a></h2>
+<p>Modules generalize the chained precompiled header model yet further, from a
+linear chain of precompiled headers to an arbitrary directed acyclic graph
+(DAG) of AST files. All of the same techniques used to make chained
+precompiled headers work — ID number, name lookup, update records — are
+shared with modules. However, the DAG nature of modules introduce a number of
+additional complications to the model:</p>
+<dl class="docutils">
+<dt>Numbering of IDs</dt>
+<dd>The simple, linear numbering scheme used in chained precompiled headers falls
+apart with the module DAG, because different modules may end up with
+different numbering schemes for entities they imported from common shared
+modules. To account for this, each module file provides information about
+which modules it depends on and which ID numbers it assigned to the entities
+in those modules, as well as which ID numbers it took for its own new
+entities. The AST reader then maps these “local” ID numbers into a “global”
+ID number space for the current translation unit, providing a 1-1 mapping
+between entities (in whatever AST file they inhabit) and global ID numbers.
+If that translation unit is then serialized into an AST file, this mapping
+will be stored for use when the AST file is imported.</dd>
+<dt>Declaration merging</dt>
+<dd>It is possible for a given entity (from the language’s perspective) to be
+declared multiple times in different places. For example, two different
+headers can have the declaration of <tt class="docutils literal"><span class="pre">printf</span></tt> or could forward-declare
+<tt class="docutils literal"><span class="pre">struct</span> <span class="pre">stat</span></tt>. If each of those headers is included in a module, and some
+third party imports both of those modules, there is a potentially serious
+problem: name lookup for <tt class="docutils literal"><span class="pre">printf</span></tt> or <tt class="docutils literal"><span class="pre">struct</span> <span class="pre">stat</span></tt> will find both
+declarations, but the AST nodes are unrelated. This would result in a
+compilation error, due to an ambiguity in name lookup. Therefore, the AST
+reader performs declaration merging according to the appropriate language
+semantics, ensuring that the two disjoint declarations are merged into a
+single redeclaration chain (with a common canonical declaration), so that it
+is as if one of the headers had been included before the other.</dd>
+<dt>Name Visibility</dt>
+<dd>Modules allow certain names that occur during module creation to be “hidden”,
+so that they are not part of the public interface of the module and are not
+visible to its clients. The AST reader maintains a “visible” bit on various
+AST nodes (declarations, macros, etc.) to indicate whether that particular
+AST node is currently visible; the various name lookup mechanisms in Clang
+inspect the visible bit to determine whether that entity, which is still in
+the AST (because other, visible AST nodes may depend on it), can actually be
+found by name lookup. When a new (sub)module is imported, it may make
+existing, non-visible, already-deserialized AST nodes visible; it is the
+responsibility of the AST reader to find and update these AST nodes when it
+is notified of the import.</dd>
+</dl>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="PTHInternals.html">Pretokenized Headers (PTH)</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
Added: www-releases/trunk/3.8.0/tools/clang/docs/PTHInternals.html
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/3.8.0/tools/clang/docs/PTHInternals.html?rev=262945&view=auto
==============================================================================
--- www-releases/trunk/3.8.0/tools/clang/docs/PTHInternals.html (added)
+++ www-releases/trunk/3.8.0/tools/clang/docs/PTHInternals.html Tue Mar 8 12:28:17 2016
@@ -0,0 +1,214 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
+
+<html xmlns="http://www.w3.org/1999/xhtml">
+ <head>
+ <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
+ <title>Pretokenized Headers (PTH) — Clang 3.8 documentation</title>
+
+ <link rel="stylesheet" href="_static/haiku.css" type="text/css" />
+ <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
+
+ <script type="text/javascript">
+ var DOCUMENTATION_OPTIONS = {
+ URL_ROOT: './',
+ VERSION: '3.8',
+ COLLAPSE_INDEX: false,
+ FILE_SUFFIX: '.html',
+ HAS_SOURCE: true
+ };
+ </script>
+ <script type="text/javascript" src="_static/jquery.js"></script>
+ <script type="text/javascript" src="_static/underscore.js"></script>
+ <script type="text/javascript" src="_static/doctools.js"></script>
+ <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+ <link rel="top" title="Clang 3.8 documentation" href="index.html" />
+ <link rel="next" title="Precompiled Header and Modules Internals" href="PCHInternals.html" />
+ <link rel="prev" title="Driver Design & Internals" href="DriverInternals.html" />
+ </head>
+ <body>
+ <div class="header"><h1 class="heading"><a href="index.html">
+ <span>Clang 3.8 documentation</span></a></h1>
+ <h2 class="heading"><span>Pretokenized Headers (PTH)</span></h2>
+ </div>
+ <div class="topnav">
+
+ <p>
+ « <a href="DriverInternals.html">Driver Design & Internals</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="PCHInternals.html">Precompiled Header and Modules Internals</a> »
+ </p>
+
+ </div>
+ <div class="content">
+
+
+ <div class="section" id="pretokenized-headers-pth">
+<h1>Pretokenized Headers (PTH)<a class="headerlink" href="#pretokenized-headers-pth" title="Permalink to this headline">¶</a></h1>
+<p>This document first describes the low-level interface for using PTH and
+then briefly elaborates on its design and implementation. If you are
+interested in the end-user view, please see the <a class="reference internal" href="UsersManual.html#usersmanual-precompiled-headers"><em>User’s Manual</em></a>.</p>
+<div class="section" id="using-pretokenized-headers-with-clang-low-level-interface">
+<h2>Using Pretokenized Headers with <tt class="docutils literal"><span class="pre">clang</span></tt> (Low-level Interface)<a class="headerlink" href="#using-pretokenized-headers-with-clang-low-level-interface" title="Permalink to this headline">¶</a></h2>
+<p>The Clang compiler frontend, <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span></tt>, supports three command line
+options for generating and using PTH files.</p>
+<p>To generate PTH files using <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span></tt>, use the option <tt class="docutils literal"><span class="pre">-emit-pth</span></tt>:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> clang -cc1 test.h -emit-pth -o test.h.pth
+</pre></div>
+</div>
+<p>This option is transparently used by <tt class="docutils literal"><span class="pre">clang</span></tt> when generating PTH
+files. Similarly, PTH files can be used as prefix headers using the
+<tt class="docutils literal"><span class="pre">-include-pth</span></tt> option:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> clang -cc1 -include-pth test.h.pth test.c -o test.s
+</pre></div>
+</div>
+<p>Alternatively, Clang’s PTH files can be used as a raw “token-cache” (or
+“content” cache) of the source included by the original header file.
+This means that the contents of the PTH file are searched as substitutes
+for <em>any</em> source files that are used by <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span></tt> to process a
+source file. This is done by specifying the <tt class="docutils literal"><span class="pre">-token-cache</span></tt> option:</p>
+<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> cat test.h
+<span class="gp">#</span>include <stdio.h>
+<span class="gp">$</span> clang -cc1 -emit-pth test.h -o test.h.pth
+<span class="gp">$</span> cat test.c
+<span class="gp">#</span>include <span class="s2">"test.h"</span>
+<span class="gp">$</span> clang -cc1 test.c -o <span class="nb">test</span> -token-cache test.h.pth
+</pre></div>
+</div>
+<p>In this example the contents of <tt class="docutils literal"><span class="pre">stdio.h</span></tt> (and the files it includes)
+will be retrieved from <tt class="docutils literal"><span class="pre">test.h.pth</span></tt>, as the PTH file is being used in
+this case as a raw cache of the contents of <tt class="docutils literal"><span class="pre">test.h</span></tt>. This is a
+low-level interface used to both implement the high-level PTH interface
+as well as to provide alternative means to use PTH-style caching.</p>
+</div>
+<div class="section" id="pth-design-and-implementation">
+<h2>PTH Design and Implementation<a class="headerlink" href="#pth-design-and-implementation" title="Permalink to this headline">¶</a></h2>
+<p>Unlike GCC’s precompiled headers, which cache the full ASTs and
+preprocessor state of a header file, Clang’s pretokenized header files
+mainly cache the raw lexer <em>tokens</em> that are needed to segment the
+stream of characters in a source file into keywords, identifiers, and
+operators. Consequently, PTH serves to mainly directly speed up the
+lexing and preprocessing of a source file, while parsing and
+type-checking must be completely redone every time a PTH file is used.</p>
+<div class="section" id="basic-design-tradeoffs">
+<h3>Basic Design Tradeoffs<a class="headerlink" href="#basic-design-tradeoffs" title="Permalink to this headline">¶</a></h3>
+<p>In the long term there are plans to provide an alternate PCH
+implementation for Clang that also caches the work for parsing and type
+checking the contents of header files. The current implementation of PCH
+in Clang as pretokenized header files was motivated by the following
+factors:</p>
+<dl class="docutils">
+<dt><strong>Language independence</strong></dt>
+<dd>PTH files work with any language that
+Clang’s lexer can handle, including C, Objective-C, and (in the early
+stages) C++. This means development on language features at the
+parsing level or above (which is basically almost all interesting
+pieces) does not require PTH to be modified.</dd>
+<dt><strong>Simple design</strong></dt>
+<dd>Relatively speaking, PTH has a simple design and
+implementation, making it easy to test. Further, because the
+machinery for PTH resides at the lower-levels of the Clang library
+stack it is fairly straightforward to profile and optimize.</dd>
+</dl>
+<p>Further, compared to GCC’s PCH implementation (which is the dominate
+precompiled header file implementation that Clang can be directly
+compared against) the PTH design in Clang yields several attractive
+features:</p>
+<dl class="docutils">
+<dt><strong>Architecture independence</strong></dt>
+<dd><p class="first">In contrast to GCC’s PCH files (and
+those of several other compilers), Clang’s PTH files are architecture
+independent, requiring only a single PTH file when building a
+program for multiple architectures.</p>
+<p class="last">For example, on Mac OS X one may wish to compile a “universal binary”
+that runs on PowerPC, 32-bit Intel (i386), and 64-bit Intel
+architectures. In contrast, GCC requires a PCH file for each
+architecture, as the definitions of types in the AST are
+architecture-specific. Since a Clang PTH file essentially represents
+a lexical cache of header files, a single PTH file can be safely used
+when compiling for multiple architectures. This can also reduce
+compile times because only a single PTH file needs to be generated
+during a build instead of several.</p>
+</dd>
+<dt><strong>Reduced memory pressure</strong></dt>
+<dd>Similar to GCC, Clang reads PTH files
+via the use of memory mapping (i.e., <tt class="docutils literal"><span class="pre">mmap</span></tt>). Clang, however,
+memory maps PTH files as read-only, meaning that multiple invocations
+of <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span></tt> can share the same pages in memory from a
+memory-mapped PTH file. In comparison, GCC also memory maps its PCH
+files but also modifies those pages in memory, incurring the
+copy-on-write costs. The read-only nature of PTH can greatly reduce
+memory pressure for builds involving multiple cores, thus improving
+overall scalability.</dd>
+<dt><strong>Fast generation</strong></dt>
+<dd>PTH files can be generated in a small fraction
+of the time needed to generate GCC’s PCH files. Since PTH/PCH
+generation is a serial operation that typically blocks progress
+during a build, faster generation time leads to improved processor
+utilization with parallel builds on multicore machines.</dd>
+</dl>
+<p>Despite these strengths, PTH’s simple design suffers some algorithmic
+handicaps compared to other PCH strategies such as those used by GCC.
+While PTH can greatly speed up the processing time of a header file, the
+amount of work required to process a header file is still roughly linear
+in the size of the header file. In contrast, the amount of work done by
+GCC to process a precompiled header is (theoretically) constant (the
+ASTs for the header are literally memory mapped into the compiler). This
+means that only the pieces of the header file that are referenced by the
+source file including the header are the only ones the compiler needs to
+process during actual compilation. While GCC’s particular implementation
+of PCH mitigates some of these algorithmic strengths via the use of
+copy-on-write pages, the approach itself can fundamentally dominate at
+an algorithmic level, especially when one considers header files of
+arbitrary size.</p>
+<p>There is also a PCH implementation for Clang based on the lazy
+deserialization of ASTs. This approach theoretically has the same
+constant-time algorithmic advantages just mentioned but also retains some
+of the strengths of PTH such as reduced memory pressure (ideal for
+multi-core builds).</p>
+</div>
+<div class="section" id="internal-pth-optimizations">
+<h3>Internal PTH Optimizations<a class="headerlink" href="#internal-pth-optimizations" title="Permalink to this headline">¶</a></h3>
+<p>While the main optimization employed by PTH is to reduce lexing time of
+header files by caching pre-lexed tokens, PTH also employs several other
+optimizations to speed up the processing of header files:</p>
+<ul class="simple">
+<li><tt class="docutils literal"><span class="pre">stat</span></tt> caching: PTH files cache information obtained via calls to
+<tt class="docutils literal"><span class="pre">stat</span></tt> that <tt class="docutils literal"><span class="pre">clang</span> <span class="pre">-cc1</span></tt> uses to resolve which files are included
+by <tt class="docutils literal"><span class="pre">#include</span></tt> directives. This greatly reduces the overhead
+involved in context-switching to the kernel to resolve included
+files.</li>
+<li>Fast skipping of <tt class="docutils literal"><span class="pre">#ifdef</span></tt> ... <tt class="docutils literal"><span class="pre">#endif</span></tt> chains: PTH files
+record the basic structure of nested preprocessor blocks. When the
+condition of the preprocessor block is false, all of its tokens are
+immediately skipped instead of requiring them to be handled by
+Clang’s preprocessor.</li>
+</ul>
+</div>
+</div>
+</div>
+
+
+ </div>
+ <div class="bottomnav">
+
+ <p>
+ « <a href="DriverInternals.html">Driver Design & Internals</a>
+ ::
+ <a class="uplink" href="index.html">Contents</a>
+ ::
+ <a href="PCHInternals.html">Precompiled Header and Modules Internals</a> »
+ </p>
+
+ </div>
+
+ <div class="footer">
+ © Copyright 2007-2016, The Clang Team.
+ Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
+ </div>
+ </body>
+</html>
\ No newline at end of file
More information about the llvm-commits
mailing list