<div dir="ltr">On Tue, Apr 2, 2013 at 12:04 AM, Alexander Kornienko <span dir="ltr"><<a href="mailto:alexfh@google.com" target="_blank">alexfh@google.com</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5"><p dir="ltr"><br>
On Apr 1, 2013 6:21 PM, "Manuel Klimek" <<a href="mailto:klimek@google.com" target="_blank">klimek@google.com</a>> wrote:<br>
><br>
> On Thu, Mar 28, 2013 at 7:40 PM, Alexander Kornienko <<a href="mailto:alexfh@google.com" target="_blank">alexfh@google.com</a>> wrote:<br>
>><br>
>> Author: alexfh<br>
>> Date: Thu Mar 28 13:40:55 2013<br>
>> New Revision: 178261<br>
>><br>
>> URL: <a href="http://llvm.org/viewvc/llvm-project?rev=178261&view=rev" target="_blank">http://llvm.org/viewvc/llvm-project?rev=178261&view=rev</a><br>
>> Log:<br>
>> Fixed handling of comments before preprocessor directives.<br>
>><br>
>> Comments before preprocessor directives used to be stored with InPPDirective<br>
>> flag set, which prevented correct comment splitting in this case. Fixed by<br>
>> flushing comments before switching on InPPDirective. Added a new test and fixed<br>
>> one of the existing tests.<br>
>><br>
>><br>
>> Modified:<br>
>> cfe/trunk/lib/Format/UnwrappedLineParser.cpp<br>
>> cfe/trunk/unittests/Format/FormatTest.cpp<br>
>><br>
>> Modified: cfe/trunk/lib/Format/UnwrappedLineParser.cpp<br>
>> URL: <a href="http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Format/UnwrappedLineParser.cpp?rev=178261&r1=178260&r2=178261&view=diff" target="_blank">http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Format/UnwrappedLineParser.cpp?rev=178261&r1=178260&r2=178261&view=diff</a><br>
>> ==============================================================================<br>
>> --- cfe/trunk/lib/Format/UnwrappedLineParser.cpp (original)<br>
>> +++ cfe/trunk/lib/Format/UnwrappedLineParser.cpp Thu Mar 28 13:40:55 2013<br>
>> @@ -822,6 +822,7 @@ void UnwrappedLineParser::readToken() {<br>
>> while (!Line->InPPDirective && <a href="http://FormatTok.Tok.is" target="_blank">FormatTok.Tok.is</a>(tok::hash) &&<br>
>> ((FormatTok.NewlinesBefore > 0 && FormatTok.HasUnescapedNewline) ||<br>
>> FormatTok.IsFirst)) {<br>
><br>
><br>
> Please add a comment indicating why we need this.</p>
</div></div><p dir="ltr">By "this" you mean the flushComments call I suppose? I wanted to add a comment, but "flush comments before starting to parse the preprocessor directive" seems like rephrasing the code. Or you can suggest a better comment?</p>
</blockquote><div><br></div><div style>In general, when I ask for a comment, I ask for an answer to the question "why are we doing it", not "what are we doing" (I'm pretty good at seeing what the code does :P)</div>
<div style><br></div><div style>Since the patch seems to have problems (Daniel just discovered some regressions) we can talk later about what exactly to do here.</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div><div class="h5">
<p dir="ltr">> <br>
>><br>
>> + flushComments(FormatTok.NewlinesBefore > 0);<br>
>> // If there is an unfinished unwrapped line, we flush the preprocessor<br>
>> // directives only after that unwrapped line was finished later.<br>
>> bool SwitchToPreprocessorLines =<br>
>><br>
>> Modified: cfe/trunk/unittests/Format/FormatTest.cpp<br>
>> URL: <a href="http://llvm.org/viewvc/llvm-project/cfe/trunk/unittests/Format/FormatTest.cpp?rev=178261&r1=178260&r2=178261&view=diff" target="_blank">http://llvm.org/viewvc/llvm-project/cfe/trunk/unittests/Format/FormatTest.cpp?rev=178261&r1=178260&r2=178261&view=diff</a><br>
>> ==============================================================================<br>
>> --- cfe/trunk/unittests/Format/FormatTest.cpp (original)<br>
>> +++ cfe/trunk/unittests/Format/FormatTest.cpp Thu Mar 28 13:40:55 2013<br>
>> @@ -699,6 +699,13 @@ TEST_F(FormatTest, SplitsLongCxxComments<br>
>> "//Don't add leading\n"<br>
>> "//whitespace",<br>
>> format("//Don't add leading whitespace", getLLVMStyleWithColumns(20)));<br>
>> + EXPECT_EQ("// A comment before\n"<br>
>> + "// a macro\n"<br>
>> + "// definition\n"<br>
>> + "#define a b",<br>
>> + format("// A comment before a macro definition\n"<br>
>> + "#define a b",<br>
>> + getLLVMStyleWithColumns(20)));<br>
>> }<br>
>><br>
>> TEST_F(FormatTest, SplitsLongLinesInComments) {<br>
>> @@ -1203,13 +1210,13 @@ TEST_F(FormatTest, IndentsPPDirectiveInR<br>
>> }<br>
>><br>
>> TEST_F(FormatTest, HandlePreprocessorDirectiveContext) {<br>
>> - EXPECT_EQ("// some comment\n"<br>
>> + EXPECT_EQ("// somecomment\n"<br>
><br>
><br>
> What's the reason for this part of the change?</p>
</div></div><p dir="ltr">This is the smallest change possible to make this test pass after we started breaking long line comments. There are many alternative solutions, of course.</p></blockquote><div><br></div><div style>
Ah, I missed the column restriction. Makes sense then.</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">
<p dir="ltr">> <br>
>><br>
>> "#include \"a.h\"\n"<br>
>> "#define A( \\\n"<br>
>> " A, B)\n"<br>
>> "#include \"b.h\"\n"<br>
>> "// somecomment\n",<br>
>> - format(" // some comment\n"<br>
>> + format(" // somecomment\n"<br>
>> " #include \"a.h\"\n"<br>
>> "#define A(A,\\\n"<br>
>> " B)\n"<br>
>><br>
>><br>
>> _______________________________________________<br>
>> cfe-commits mailing list<br>
>> <a href="mailto:cfe-commits@cs.uiuc.edu" target="_blank">cfe-commits@cs.uiuc.edu</a><br>
>> <a href="http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits" target="_blank">http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits</a><br>
><br>
><br>
</p>
</div></div></blockquote></div><br></div></div>