[clang] 8effc8d - Reland [clang-format] Add OneLineFormatOffRegex option (#137577)
Owen Pan via cfe-commits
cfe-commits at lists.llvm.org
Wed Apr 30 19:59:08 PDT 2025
Author: Owen Pan
Date: 2025-04-30T19:58:59-07:00
New Revision: 8effc8da292bfacb823a7e3c4134296da481fedc
URL: https://github.com/llvm/llvm-project/commit/8effc8da292bfacb823a7e3c4134296da481fedc
DIFF: https://github.com/llvm/llvm-project/commit/8effc8da292bfacb823a7e3c4134296da481fedc.diff
LOG: Reland [clang-format] Add OneLineFormatOffRegex option (#137577)
Added:
Modified:
clang/docs/ClangFormatStyleOptions.rst
clang/docs/ReleaseNotes.rst
clang/include/clang/Format/Format.h
clang/lib/Format/Format.cpp
clang/lib/Format/FormatTokenLexer.cpp
clang/lib/Format/FormatTokenLexer.h
clang/unittests/Format/ConfigParseTest.cpp
clang/unittests/Format/FormatTest.cpp
Removed:
################################################################################
diff --git a/clang/docs/ClangFormatStyleOptions.rst b/clang/docs/ClangFormatStyleOptions.rst
index 3f8a5f49313b2..b47291599649d 100644
--- a/clang/docs/ClangFormatStyleOptions.rst
+++ b/clang/docs/ClangFormatStyleOptions.rst
@@ -5195,6 +5195,29 @@ the configuration (without a prefix: ``Auto``).
Add a space in front of an Objective-C protocol list, i.e. use
``Foo <Protocol>`` instead of ``Foo<Protocol>``.
+.. _OneLineFormatOffRegex:
+
+**OneLineFormatOffRegex** (``String``) :versionbadge:`clang-format 21` :ref:`¶ <OneLineFormatOffRegex>`
+ A regular expression that describes markers for turning formatting off for
+ one line. If it matches a comment that is the only token of a line,
+ clang-format skips the comment and the next line. Otherwise, clang-format
+ skips lines containing a matched token.
+
+ .. code-block:: c++
+
+ // OneLineFormatOffRegex: ^(// NOLINT|logger$)
+ // results in the output below:
+ int a;
+ int b ; // NOLINT
+ int c;
+ // NOLINTNEXTLINE
+ int d ;
+ int e;
+ s = "// NOLINT";
+ logger() ;
+ logger2();
+ my_logger();
+
.. _PPIndentWidth:
**PPIndentWidth** (``Integer``) :versionbadge:`clang-format 13` :ref:`¶ <PPIndentWidth>`
diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
index 728ceb900bcd4..ba6e40f145655 100644
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -746,6 +746,7 @@ clang-format
top of the file.
- Add ``EnumTrailingComma`` option for inserting/removing commas at the end of
``enum`` enumerator lists.
+- Add ``OneLineFormatOffRegex`` option for turning formatting off for one line.
libclang
--------
diff --git a/clang/include/clang/Format/Format.h b/clang/include/clang/Format/Format.h
index f6ceef08b46da..7fe41d800ccb3 100644
--- a/clang/include/clang/Format/Format.h
+++ b/clang/include/clang/Format/Format.h
@@ -3654,6 +3654,27 @@ struct FormatStyle {
/// \version 3.7
bool ObjCSpaceBeforeProtocolList;
+ /// A regular expression that describes markers for turning formatting off for
+ /// one line. If it matches a comment that is the only token of a line,
+ /// clang-format skips the comment and the next line. Otherwise, clang-format
+ /// skips lines containing a matched token.
+ /// \code
+ /// // OneLineFormatOffRegex: ^(// NOLINT|logger$)
+ /// // results in the output below:
+ /// int a;
+ /// int b ; // NOLINT
+ /// int c;
+ /// // NOLINTNEXTLINE
+ /// int d ;
+ /// int e;
+ /// s = "// NOLINT";
+ /// logger() ;
+ /// logger2();
+ /// my_logger();
+ /// \endcode
+ /// \version 21
+ std::string OneLineFormatOffRegex;
+
/// Different ways to try to fit all constructor initializers on a line.
enum PackConstructorInitializersStyle : int8_t {
/// Always put each constructor initializer on its own line.
@@ -5399,6 +5420,7 @@ struct FormatStyle {
ObjCPropertyAttributeOrder == R.ObjCPropertyAttributeOrder &&
ObjCSpaceAfterProperty == R.ObjCSpaceAfterProperty &&
ObjCSpaceBeforeProtocolList == R.ObjCSpaceBeforeProtocolList &&
+ OneLineFormatOffRegex == R.OneLineFormatOffRegex &&
PackConstructorInitializers == R.PackConstructorInitializers &&
PenaltyBreakAssignment == R.PenaltyBreakAssignment &&
PenaltyBreakBeforeFirstCallParameter ==
diff --git a/clang/lib/Format/Format.cpp b/clang/lib/Format/Format.cpp
index 5a1c3f556b331..2f4b64ef4f5fe 100644
--- a/clang/lib/Format/Format.cpp
+++ b/clang/lib/Format/Format.cpp
@@ -1100,6 +1100,7 @@ template <> struct MappingTraits<FormatStyle> {
IO.mapOptional("ObjCSpaceAfterProperty", Style.ObjCSpaceAfterProperty);
IO.mapOptional("ObjCSpaceBeforeProtocolList",
Style.ObjCSpaceBeforeProtocolList);
+ IO.mapOptional("OneLineFormatOffRegex", Style.OneLineFormatOffRegex);
IO.mapOptional("PackConstructorInitializers",
Style.PackConstructorInitializers);
IO.mapOptional("PenaltyBreakAssignment", Style.PenaltyBreakAssignment);
diff --git a/clang/lib/Format/FormatTokenLexer.cpp b/clang/lib/Format/FormatTokenLexer.cpp
index a4c94ac411fe0..944d30ba7874f 100644
--- a/clang/lib/Format/FormatTokenLexer.cpp
+++ b/clang/lib/Format/FormatTokenLexer.cpp
@@ -32,7 +32,8 @@ FormatTokenLexer::FormatTokenLexer(
LangOpts(getFormattingLangOpts(Style)), SourceMgr(SourceMgr), ID(ID),
Style(Style), IdentTable(IdentTable), Keywords(IdentTable),
Encoding(Encoding), Allocator(Allocator), FirstInLineIndex(0),
- FormattingDisabled(false), MacroBlockBeginRegex(Style.MacroBlockBegin),
+ FormattingDisabled(false), FormatOffRegex(Style.OneLineFormatOffRegex),
+ MacroBlockBeginRegex(Style.MacroBlockBegin),
MacroBlockEndRegex(Style.MacroBlockEnd) {
Lex.reset(new Lexer(ID, SourceMgr.getBufferOrFake(ID), SourceMgr, LangOpts));
Lex->SetKeepWhitespaceMode(true);
@@ -83,8 +84,42 @@ FormatTokenLexer::FormatTokenLexer(
ArrayRef<FormatToken *> FormatTokenLexer::lex() {
assert(Tokens.empty());
assert(FirstInLineIndex == 0);
+ enum { FO_None, FO_CurrentLine, FO_NextLine } FormatOff = FO_None;
do {
Tokens.push_back(getNextToken());
+ auto &Tok = *Tokens.back();
+ const auto NewlinesBefore = Tok.NewlinesBefore;
+ switch (FormatOff) {
+ case FO_CurrentLine:
+ if (NewlinesBefore == 0)
+ Tok.Finalized = true;
+ else
+ FormatOff = FO_None;
+ break;
+ case FO_NextLine:
+ if (NewlinesBefore > 1) {
+ FormatOff = FO_None;
+ } else {
+ Tok.Finalized = true;
+ FormatOff = FO_CurrentLine;
+ }
+ break;
+ default:
+ if (!FormattingDisabled && FormatOffRegex.match(Tok.TokenText)) {
+ if (Tok.is(tok::comment) &&
+ (NewlinesBefore > 0 || Tokens.size() == 1)) {
+ Tok.Finalized = true;
+ FormatOff = FO_NextLine;
+ } else {
+ for (auto *Token : reverse(Tokens)) {
+ Token->Finalized = true;
+ if (Token->NewlinesBefore > 0)
+ break;
+ }
+ FormatOff = FO_CurrentLine;
+ }
+ }
+ }
if (Style.isJavaScript()) {
tryParseJSRegexLiteral();
handleTemplateStrings();
diff --git a/clang/lib/Format/FormatTokenLexer.h b/clang/lib/Format/FormatTokenLexer.h
index 3f001bc69415d..105847b126e20 100644
--- a/clang/lib/Format/FormatTokenLexer.h
+++ b/clang/lib/Format/FormatTokenLexer.h
@@ -134,6 +134,7 @@ class FormatTokenLexer {
VariableTemplates;
bool FormattingDisabled;
+ llvm::Regex FormatOffRegex; // For one line.
llvm::Regex MacroBlockBeginRegex;
llvm::Regex MacroBlockEndRegex;
diff --git a/clang/unittests/Format/ConfigParseTest.cpp b/clang/unittests/Format/ConfigParseTest.cpp
index 2b08b794792e9..f7ab5546c7193 100644
--- a/clang/unittests/Format/ConfigParseTest.cpp
+++ b/clang/unittests/Format/ConfigParseTest.cpp
@@ -295,6 +295,7 @@ TEST(ConfigParseTest, ParsesConfiguration) {
FormatStyle Style = {};
Style.Language = FormatStyle::LK_Cpp;
CHECK_PARSE("CommentPragmas: '// abc$'", CommentPragmas, "// abc$");
+ CHECK_PARSE("OneLineFormatOffRegex: // ab$", OneLineFormatOffRegex, "// ab$");
Style.QualifierAlignment = FormatStyle::QAS_Right;
CHECK_PARSE("QualifierAlignment: Leave", QualifierAlignment,
diff --git a/clang/unittests/Format/FormatTest.cpp b/clang/unittests/Format/FormatTest.cpp
index 333d40d481025..c4fcc5506d152 100644
--- a/clang/unittests/Format/FormatTest.cpp
+++ b/clang/unittests/Format/FormatTest.cpp
@@ -24954,6 +24954,105 @@ TEST_F(FormatTest, DisableRegions) {
"// clang-format on");
}
+TEST_F(FormatTest, OneLineFormatOffRegex) {
+ auto Style = getLLVMStyle();
+ Style.OneLineFormatOffRegex = "// format off$";
+
+ verifyFormat(" // format off\n"
+ " int i ;\n"
+ "int j;",
+ " // format off\n"
+ " int i ;\n"
+ " int j ;",
+ Style);
+ verifyFormat("// format off?\n"
+ "int i;",
+ " // format off?\n"
+ " int i ;",
+ Style);
+ verifyFormat("f(\"// format off\");", " f(\"// format off\") ;", Style);
+
+ verifyFormat("int i;\n"
+ " // format off\n"
+ " int j ;\n"
+ "int k;",
+ " int i ;\n"
+ " // format off\n"
+ " int j ;\n"
+ " int k ;",
+ Style);
+
+ verifyFormat(" // format off\n"
+ "\n"
+ "int i;",
+ " // format off\n"
+ " \n"
+ " int i ;",
+ Style);
+
+ verifyFormat("int i;\n"
+ " int j ; // format off\n"
+ "int k;",
+ " int i ;\n"
+ " int j ; // format off\n"
+ " int k ;",
+ Style);
+
+ verifyFormat("// clang-format off\n"
+ " int i ;\n"
+ " int j ; // format off\n"
+ " int k ;\n"
+ "// clang-format on\n"
+ "f();",
+ " // clang-format off\n"
+ " int i ;\n"
+ " int j ; // format off\n"
+ " int k ;\n"
+ " // clang-format on\n"
+ " f() ;",
+ Style);
+
+ Style.OneLineFormatOffRegex = "^/\\* format off \\*/";
+ verifyFormat("int i;\n"
+ " /* format off */ int j ;\n"
+ "int k;",
+ " int i ;\n"
+ " /* format off */ int j ;\n"
+ " int k ;",
+ Style);
+ verifyFormat("f(\"/* format off */\");", " f(\"/* format off */\") ;", Style);
+
+ Style.AlignEscapedNewlines = FormatStyle::ENAS_DontAlign;
+ verifyFormat("#define A \\\n"
+ " do { \\\n"
+ " /* format off */\\\n"
+ " f() ; \\\n"
+ " g(); \\\n"
+ " } while (0)",
+ "# define A\\\n"
+ " do{ \\\n"
+ " /* format off */\\\n"
+ " f() ; \\\n"
+ " g() ;\\\n"
+ " } while (0 )",
+ Style);
+
+ Style.ColumnLimit = 50;
+ Style.OneLineFormatOffRegex = "^LogErrorPrint$";
+ verifyFormat(" myproject::LogErrorPrint(logger, \"Don't split me!\");\n"
+ "myproject::MyLogErrorPrinter(myLogger,\n"
+ " \"Split me!\");",
+ " myproject::LogErrorPrint(logger, \"Don't split me!\");\n"
+ " myproject::MyLogErrorPrinter(myLogger, \"Split me!\");",
+ Style);
+
+ Style.OneLineFormatOffRegex = "//(< clang-format off| NO_TRANSLATION)$";
+ verifyNoChange(
+ " int i ; //< clang-format off\n"
+ " msg = sprintf(\"Long string with placeholders.\"); // NO_TRANSLATION",
+ Style);
+}
+
TEST_F(FormatTest, DoNotCrashOnInvalidInput) {
format("? ) =");
verifyNoCrash("#define a\\\n /**/}");
More information about the cfe-commits
mailing list