Added language detection and a test. Addressed comments.
Hi djasper, klimek,
http://llvm-reviews.chandlerc.com/D2242
CHANGE SINCE LAST DIFF
http://llvm-reviews.chandlerc.com/D2242?vs=5799&id=5812#toc
Files:
include/clang/Format/Format.h
lib/Format/Format.cpp
test/Format/language-detection.cpp
unittests/Format/FormatTest.cpp
Index: include/clang/Format/Format.h
===================================================================
--- include/clang/Format/Format.h
+++ include/clang/Format/Format.h
@@ -30,6 +30,23 @@
/// \brief The \c FormatStyle is used to configure the formatting to follow
/// specific guidelines.
struct FormatStyle {
+ /// \brief Supported languages. When stored in a configuration file, specifies
+ /// the language, that the configuration targets. When passed to the
+ /// reformat() function, enables syntax features specific for the language.
+ enum LanguageKind {
+ /// Should be used for C, C++, ObjectiveC, ObjectiveC++.
+ LK_Cpp,
+ /// Should be used for JavaScript.
+ LK_JavaScript,
+ /// When stored in a configuration file, means that configuration
+ /// may be used for all languages. Should not be passed to the reformat()
+ /// function.
+ LK_Any
+ };
+
+ /// \brief Language, this format style is targeted at.
+ LanguageKind Language;
+
/// \brief The column limit.
///
/// A column limit of \c 0 means that there is no column limit. In this case,
@@ -284,6 +301,7 @@
IndentFunctionDeclarationAfterType ==
R.IndentFunctionDeclarationAfterType &&
IndentWidth == R.IndentWidth &&
+ Language == R.Language &&
MaxEmptyLinesToKeep == R.MaxEmptyLinesToKeep &&
NamespaceIndentation == R.NamespaceIndentation &&
ObjCSpaceBeforeProtocolList == R.ObjCSpaceBeforeProtocolList &&
Index: lib/Format/Format.cpp
===================================================================
--- lib/Format/Format.cpp
+++ lib/Format/Format.cpp
@@ -34,6 +34,17 @@
namespace llvm {
namespace yaml {
template <>
+struct ScalarEnumerationTraits<clang::format::FormatStyle::LanguageKind> {
+ static void enumeration(IO &IO,
+ clang::format::FormatStyle::LanguageKind &Value) {
+ IO.enumCase(Value, "Cpp", clang::format::FormatStyle::LK_Cpp);
+ IO.enumCase(Value, "JavaScript",
+ clang::format::FormatStyle::LK_JavaScript);
+ IO.enumCase(Value, "Any", clang::format::FormatStyle::LK_Any);
+ }
+};
+
+template <>
struct ScalarEnumerationTraits<clang::format::FormatStyle::LanguageStandard> {
static void enumeration(IO &IO,
clang::format::FormatStyle::LanguageStandard &Value) {
@@ -106,6 +117,7 @@
}
}
+ IO.mapOptional("Language", Style.Language);
IO.mapOptional("AccessModifierOffset", Style.AccessModifierOffset);
IO.mapOptional("ConstructorInitializerIndentWidth",
Style.ConstructorInitializerIndentWidth);
@@ -173,6 +185,26 @@
IO.mapOptional("ContinuationIndentWidth", Style.ContinuationIndentWidth);
}
};
+
+// Allows to read vector<FormatStyle> using the 0th element as a default value.
+// Elements will be written or read starting from the 1st element.
+// When writing, the 0th element is ignored.
+// When reading, keys, that are not present in the serialized form, will be
+// copied from the 0th element of the vector.
+template <>
+struct DocumentListTraits<std::vector<clang::format::FormatStyle> > {
+ static size_t size(IO &io, std::vector<clang::format::FormatStyle> &Seq) {
+ return Seq.size() - 1;
+ }
+ static clang::format::FormatStyle &
+ element(IO &io, std::vector<clang::format::FormatStyle> &Seq, size_t Index) {
+ if (Index + 2 > Seq.size()) {
+ clang::format::FormatStyle Template = Seq[0];
+ Seq.resize(Index + 2, Template);
+ }
+ return Seq[Index + 1];
+ }
+};
}
}
@@ -188,6 +220,7 @@
FormatStyle getLLVMStyle() {
FormatStyle LLVMStyle;
+ LLVMStyle.Language= FormatStyle::LK_Cpp;
LLVMStyle.AccessModifierOffset = -2;
LLVMStyle.AlignEscapedNewlinesLeft = false;
LLVMStyle.AlignTrailingComments = true;
@@ -236,6 +269,7 @@
FormatStyle getGoogleStyle() {
FormatStyle GoogleStyle;
+ GoogleStyle.Language= FormatStyle::LK_Cpp;
GoogleStyle.AccessModifierOffset = -1;
GoogleStyle.AlignEscapedNewlinesLeft = true;
GoogleStyle.AlignTrailingComments = true;
@@ -339,9 +373,20 @@
llvm::error_code parseConfiguration(StringRef Text, FormatStyle *Style) {
if (Text.trim().empty())
return llvm::make_error_code(llvm::errc::invalid_argument);
+ std::vector<FormatStyle> Styles;
+ Styles.push_back(*Style);
llvm::yaml::Input Input(Text);
- Input >> *Style;
- return Input.error();
+ Input >> Styles;
+ for (unsigned i = 1; i < Styles.size(); ++i) {
+ if ((Styles[i].Language== Styles[0].Language) ||
+ (Styles[i].Language == FormatStyle::LK_Any)) {
+ // FIXME: Error out on multiple configurations for the same language.
+ *Style = Styles[i];
+ Style->Language = Styles[0].Language;
+ return Input.error();
+ }
+ }
+ return llvm::make_error_code(llvm::errc::not_supported);
}
std::string configurationAsText(const FormatStyle &Style) {
@@ -979,24 +1024,44 @@
private:
void tryMergePreviousTokens() {
- tryMerge_TMacro() || tryMergeJavaScriptIdentityOperators();
+ if (tryMerge_TMacro())
+ return;
+
+ if (Style.Language == FormatStyle::LK_JavaScript) {
+ static tok::TokenKind JSIdentity[] = { tok::equalequal, tok::equal };
+ static tok::TokenKind JSNotIdentity[] = { tok::exclaimequal, tok::equal };
+ static tok::TokenKind JSShiftEqual[] = { tok::greater, tok::greater,
+ tok::greaterequal };
+ // FIXME: We probably need to change token type to mimic operator with the
+ // correct priority.
+ if (tryMergeTokens(JSIdentity))
+ return;
+ if (tryMergeTokens(JSNotIdentity))
+ return;
+ if (tryMergeTokens(JSShiftEqual))
+ return;
+ }
}
- bool tryMergeJavaScriptIdentityOperators() {
- if (Tokens.size() < 2)
+ bool tryMergeTokens(ArrayRef<tok::TokenKind> Kinds) {
+ if (Tokens.size() < Kinds.size())
return false;
- FormatToken &First = *Tokens[Tokens.size() - 2];
- if (!First.isOneOf(tok::exclaimequal, tok::equalequal))
- return false;
- FormatToken &Second = *Tokens.back();
- if (!Second.is(tok::equal))
- return false;
- if (Second.WhitespaceRange.getBegin() != Second.WhitespaceRange.getEnd())
+
+ SmallVectorImpl<FormatToken *>::const_iterator First =
+ Tokens.end() - Kinds.size();
+ if (!First[0]->is(Kinds[0]))
return false;
- First.TokenText =
- StringRef(First.TokenText.data(), First.TokenText.size() + 1);
- First.ColumnWidth += 1;
- Tokens.pop_back();
+ unsigned AddLength = 0;
+ for (unsigned i = 1; i < Kinds.size(); ++i) {
+ if (!First[i]->is(Kinds[i]) || First[i]->WhitespaceRange.getBegin() !=
+ First[i]->WhitespaceRange.getEnd())
+ return false;
+ AddLength += First[i]->TokenText.size();
+ }
+ Tokens.resize(Tokens.size() - Kinds.size() + 1);
+ First[0]->TokenText = StringRef(First[0]->TokenText.data(),
+ First[0]->TokenText.size() + AddLength);
+ First[0]->ColumnWidth += AddLength;
return true;
}
@@ -1194,6 +1259,17 @@
}
};
+static const char *getLanguageName(FormatStyle::LanguageKind Language) {
+ switch (Language) {
+ case FormatStyle::LK_Cpp:
+ return "C++";
+ case FormatStyle::LK_JavaScript:
+ return "JavaScript";
+ default:
+ return "Any";
+ }
+}
+
class Formatter : public UnwrappedLineConsumer {
public:
Formatter(const FormatStyle &Style, Lexer &Lex, SourceManager &SourceMgr,
@@ -1206,6 +1282,9 @@
<< (Encoding == encoding::Encoding_UTF8 ? "UTF8"
: "unknown")
<< "\n");
+ assert(Style.Language != FormatStyle::LK_Any);
+ DEBUG(llvm::dbgs() << "Language: " << getLanguageName(Style.Language)
+ << "\n");
}
tooling::Replacements format() {
@@ -1495,11 +1574,26 @@
"parameters, e.g.:\n"
" -style=\"{BasedOnStyle: llvm, IndentWidth: 8}\"";
+static void fillLanguageByFileName(StringRef FileName, FormatStyle *Style) {
+ if (FileName.endswith_lower(".c") || FileName.endswith_lower(".h") ||
+ FileName.endswith_lower(".cpp") || FileName.endswith_lower(".hpp") ||
+ FileName.endswith_lower(".cc") || FileName.endswith_lower(".hh") ||
+ FileName.endswith_lower(".cxx") || FileName.endswith_lower(".hxx") ||
+ FileName.endswith_lower(".m") || FileName.endswith_lower(".mm")) {
+ Style->Language = FormatStyle::LK_Cpp;
+ }
+ if (FileName.endswith_lower(".js")) {
+ Style->Language = FormatStyle::LK_JavaScript;
+ }
+}
+
FormatStyle getStyle(StringRef StyleName, StringRef FileName) {
+ // FIXME: Configure fallback style from outside (add a command line option).
// Fallback style in case the rest of this function can't determine a style.
StringRef FallbackStyle = "LLVM";
FormatStyle Style;
getPredefinedStyle(FallbackStyle, &Style);
+ fillLanguageByFileName(FileName, &Style);
if (StyleName.startswith("{")) {
// Parse YAML/JSON style from the command line.
@@ -1514,9 +1608,11 @@
if (!getPredefinedStyle(StyleName, &Style))
llvm::errs() << "Invalid value for -style, using " << FallbackStyle
<< " style\n";
+ fillLanguageByFileName(FileName, &Style);
return Style;
}
+ SmallString<128> UnsuitableConfigFiles;
SmallString<128> Path(FileName);
llvm::sys::fs::make_absolute(Path);
for (StringRef Directory = Path; !Directory.empty();
@@ -1548,16 +1644,27 @@
continue;
}
if (llvm::error_code ec = parseConfiguration(Text->getBuffer(), &Style)) {
- llvm::errs() << "Error reading " << ConfigFile << ": " << ec.message()
- << "\n";
+ if (ec == llvm::errc::not_supported) {
+ if (!UnsuitableConfigFiles.empty())
+ UnsuitableConfigFiles.append(", ");
+ UnsuitableConfigFiles.append(ConfigFile);
+ } else {
+ llvm::errs() << "Error reading " << ConfigFile << ": " << ec.message()
+ << "\n";
+ }
continue;
}
DEBUG(llvm::dbgs() << "Using configuration file " << ConfigFile << "\n");
return Style;
}
}
llvm::errs() << "Can't find usable .clang-format, using " << FallbackStyle
<< " style\n";
+ if (!UnsuitableConfigFiles.empty()) {
+ llvm::errs() << "Configuration file(s) do(es) not support "
+ << getLanguageName(Style.Language) << ": "
+ << UnsuitableConfigFiles << "\n";
+ }
return Style;
}
Index: test/Format/language-detection.cpp
===================================================================
--- /dev/null
+++ test/Format/language-detection.cpp
@@ -0,0 +1,7 @@
+// RUN: grep -Ev "// *[A-Z0-9_]+:" %s > %t.js
+// RUN: grep -Ev "// *[A-Z0-9_]+:" %s > %t.cpp
+// RUN: clang-format -style=llvm %t.js | FileCheck -strict-whitespace -check-prefix=CHECK1 %s
+// RUN: clang-format -style=llvm %t.cpp | FileCheck -strict-whitespace -check-prefix=CHECK2 %s
+// CHECK1: {{^a >>>= b;$}}
+// CHECK2: {{^a >> >= b;$}}
+a >>>= b;
Index: unittests/Format/FormatTest.cpp
===================================================================
--- unittests/Format/FormatTest.cpp
+++ unittests/Format/FormatTest.cpp
@@ -6913,6 +6913,61 @@
CHECK_PARSE("Standard: C++11", Standard, FormatStyle::LS_Cpp11);
CHECK_PARSE("Standard: Auto", Standard, FormatStyle::LS_Auto);
+ Style.Language= FormatStyle::LK_Cpp;
+ CHECK_PARSE("Language: Cpp\n"
+ "IndentWidth: 12", IndentWidth, 12u);
+ EXPECT_EQ(parseConfiguration("Language: JavaScript\n"
+ "IndentWidth: 34",
+ &Style),
+ llvm::errc::not_supported);
+ EXPECT_EQ(Style.IndentWidth, 12u);
+ CHECK_PARSE("Language: Any\n"
+ "IndentWidth: 56", IndentWidth, 56u);
+ EXPECT_EQ(Style.Language, FormatStyle::LK_Cpp);
+
+ Style.Language = FormatStyle::LK_JavaScript;
+ CHECK_PARSE("Language: JavaScript\n"
+ "IndentWidth: 12", IndentWidth, 12u);
+ EXPECT_EQ(parseConfiguration("Language: Cpp\n"
+ "IndentWidth: 34",
+ &Style),
+ llvm::errc::not_supported);
+ EXPECT_EQ(Style.IndentWidth, 12u);
+ CHECK_PARSE("Language: Any\n"
+ "IndentWidth: 56", IndentWidth, 56u);
+ EXPECT_EQ(Style.Language, FormatStyle::LK_JavaScript);
+
+ CHECK_PARSE("---\n"
+ "Language: JavaScript\n"
+ "IndentWidth: 12\n"
+ "---\n"
+ "Language: Cpp\n"
+ "IndentWidth: 34\n"
+ "...\n",
+ IndentWidth, 12u);
+
+ Style.Language = FormatStyle::LK_Cpp;
+ CHECK_PARSE("---\n"
+ "Language: JavaScript\n"
+ "IndentWidth: 12\n"
+ "---\n"
+ "Language: Cpp\n"
+ "IndentWidth: 34\n"
+ "...\n",
+ IndentWidth, 34u);
+ CHECK_PARSE("---\n"
+ "Language: JavaScript\n"
+ "IndentWidth: 56\n"
+ "---\n"
+ "Language: Any\n"
+ "IndentWidth: 78\n"
+ "...\n",
+ IndentWidth, 78u);
+
+ // FIXME: Error out on multiple configs for the same language.
+
+ EXPECT_EQ(Style.Language, FormatStyle::LK_Cpp);
+
Style.UseTab = FormatStyle::UT_ForIndentation;
CHECK_PARSE("UseTab: false", UseTab, FormatStyle::UT_Never);
CHECK_PARSE("UseTab: true", UseTab, FormatStyle::UT_Always);
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
