Add support for retrieving the Doxygen comment associated with a given
declaration in the AST.
The new ASTContext::getCommentForDecl function searches for a comment
that is attached to the given declaration, and returns that comment,
which may be composed of several comment blocks.
Comments are always available in an AST. However, to avoid harming
performance, we don't actually parse the comments. Rather, we keep the
source ranges of all of the comments within a large, sorted vector,
then lazily extract comments via a binary search in that vector only
when needed (which never occurs in a "normal" compile).
Comments are written to a precompiled header/AST file as a blob of
source ranges. That blob is only lazily loaded when one requests a
comment for a declaration (this never occurs in a "normal" compile).
The indexer testbed now supports comment extraction. When the
-point-at location points to a declaration with a Doxygen-style
comment, the indexer testbed prints the associated comment
block(s). See test/Index/comments.c for an example.
Some notes:
- We don't actually attempt to parse the comment blocks themselves,
beyond identifying them as Doxygen comment blocks to associate them
with a declaration.
- We won't find comment blocks that aren't adjacent to the
declaration, because we start our search based on the location of
the declaration.
- We don't go through the necessary hops to find, for example,
whether some redeclaration of a declaration has comments when our
current declaration does not. Similarly, we don't attempt to
associate a \param Foo marker in a function body comment with the
parameter named Foo (although that is certainly possible).
- Verification of my "no performance impact" claims is still "to be
done".
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@74704 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/include/clang/AST/ASTContext.h b/include/clang/AST/ASTContext.h
index 450b28a..041a0f3 100644
--- a/include/clang/AST/ASTContext.h
+++ b/include/clang/AST/ASTContext.h
@@ -141,10 +141,19 @@
/// this ASTContext object.
LangOptions LangOpts;
+ /// \brief Whether we have already loaded comment source ranges from an
+ /// external source.
+ bool LoadedExternalComments;
+
/// MallocAlloc/BumpAlloc - The allocator objects used to create AST objects.
bool FreeMemory;
llvm::MallocAllocator MallocAlloc;
llvm::BumpPtrAllocator BumpAlloc;
+
+ /// \brief Mapping from declarations to their comments, once we have
+ /// already looked up the comment associated with a given declaration.
+ llvm::DenseMap<const Decl *, std::string> DeclComments;
+
public:
TargetInfo &Target;
IdentifierTable &Idents;
@@ -154,6 +163,10 @@
llvm::OwningPtr<ExternalASTSource> ExternalSource;
clang::PrintingPolicy PrintingPolicy;
+ /// \brief Source ranges for all of the comments in the source file,
+ /// sorted in order of appearance in the translation unit.
+ std::vector<SourceRange> Comments;
+
SourceManager& getSourceManager() { return SourceMgr; }
const SourceManager& getSourceManager() const { return SourceMgr; }
void *Allocate(unsigned Size, unsigned Align = 8) {
@@ -178,7 +191,8 @@
TranslationUnitDecl *getTranslationUnitDecl() const { return TUDecl; }
-
+ const char *getCommentForDecl(const Decl *D);
+
// Builtin Types.
QualType VoidTy;
QualType BoolTy;
diff --git a/include/clang/AST/ExternalASTSource.h b/include/clang/AST/ExternalASTSource.h
index 4219bd5..6f862a5 100644
--- a/include/clang/AST/ExternalASTSource.h
+++ b/include/clang/AST/ExternalASTSource.h
@@ -18,6 +18,7 @@
#include "clang/AST/Type.h"
#include "llvm/ADT/SmallVector.h"
#include <cassert>
+#include <vector>
namespace clang {
class ASTConsumer;
@@ -57,6 +58,14 @@
virtual ~ExternalASTSource();
+ /// \brief Reads the source ranges that correspond to comments from
+ /// an external AST source.
+ ///
+ /// \param Comments the contents of this vector will be
+ /// replaced with the sorted set of source ranges corresponding to
+ /// comments in the source code.
+ virtual void ReadComments(std::vector<SourceRange> &Comments) = 0;
+
/// \brief Resolve a type ID into a type, potentially building a new
/// type.
virtual QualType GetType(uint32_t ID) = 0;
diff --git a/include/clang/Frontend/PCHBitCodes.h b/include/clang/Frontend/PCHBitCodes.h
index 6322272..80aa248 100644
--- a/include/clang/Frontend/PCHBitCodes.h
+++ b/include/clang/Frontend/PCHBitCodes.h
@@ -218,7 +218,11 @@
/// \brief Record code for the original file that was used to
/// generate the precompiled header.
- ORIGINAL_FILE_NAME = 20
+ ORIGINAL_FILE_NAME = 20,
+
+ /// \brief Record code for the sorted array of source ranges where
+ /// comments were encountered in the source code.
+ COMMENT_RANGES = 21
};
/// \brief Record types used within a source manager block.
diff --git a/include/clang/Frontend/PCHReader.h b/include/clang/Frontend/PCHReader.h
index b3ed364..8291f46 100644
--- a/include/clang/Frontend/PCHReader.h
+++ b/include/clang/Frontend/PCHReader.h
@@ -279,6 +279,12 @@
/// been loaded.
llvm::SmallVector<Selector, 16> SelectorsLoaded;
+ /// \brief A sorted array of source ranges containing comments.
+ SourceRange *Comments;
+
+ /// \brief The number of source ranges in the Comments array.
+ unsigned NumComments;
+
/// \brief The set of external definitions stored in the the PCH
/// file.
llvm::SmallVector<uint64_t, 16> ExternalDefinitions;
@@ -452,6 +458,14 @@
/// build prior to including the precompiled header.
const std::string &getSuggestedPredefines() { return SuggestedPredefines; }
+ /// \brief Reads the source ranges that correspond to comments from
+ /// an external AST source.
+ ///
+ /// \param Comments the contents of this vector will be
+ /// replaced with the sorted set of source ranges corresponding to
+ /// comments in the source code.
+ virtual void ReadComments(std::vector<SourceRange> &Comments);
+
/// \brief Resolve a type ID into a type, potentially building a new
/// type.
virtual QualType GetType(pch::TypeID ID);
diff --git a/include/clang/Frontend/PCHWriter.h b/include/clang/Frontend/PCHWriter.h
index 5cb9395..c663442 100644
--- a/include/clang/Frontend/PCHWriter.h
+++ b/include/clang/Frontend/PCHWriter.h
@@ -166,6 +166,7 @@
void WriteSourceManagerBlock(SourceManager &SourceMgr,
const Preprocessor &PP);
void WritePreprocessor(const Preprocessor &PP);
+ void WriteComments(ASTContext &Context);
void WriteType(const Type *T);
void WriteTypesBlock(ASTContext &Context);
uint64_t WriteDeclContextLexicalBlock(ASTContext &Context, DeclContext *DC);
diff --git a/include/clang/Lex/Preprocessor.h b/include/clang/Lex/Preprocessor.h
index f229881..6d5ed72 100644
--- a/include/clang/Lex/Preprocessor.h
+++ b/include/clang/Lex/Preprocessor.h
@@ -26,6 +26,7 @@
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/OwningPtr.h"
#include "llvm/Support/Allocator.h"
+#include <vector>
namespace clang {
@@ -35,6 +36,7 @@
class HeaderSearch;
class PragmaNamespace;
class PragmaHandler;
+class CommentHandler;
class ScratchBuffer;
class TargetInfo;
class PPCallbacks;
@@ -109,6 +111,10 @@
/// with this preprocessor.
PragmaNamespace *PragmaHandlers;
+ /// \brief Tracks all of the comment handlers that the client registered
+ /// with this preprocessor.
+ std::vector<CommentHandler *> CommentHandlers;
+
/// CurLexer - This is the current top of the stack that we're lexing from if
/// not expanding a macro and we are lexing directly from source code.
/// Only one of CurLexer, CurPTHLexer, or CurTokenLexer will be non-null.
@@ -301,6 +307,14 @@
/// to remove a handler that has not been registered.
void RemovePragmaHandler(const char *Namespace, PragmaHandler *Handler);
+ /// \brief Add the specified comment handler to the preprocessor.
+ void AddCommentHandler(CommentHandler *Handler);
+
+ /// \brief Remove the specified comment handler.
+ ///
+ /// It is an error to remove a handler that has not been registered.
+ void RemoveCommentHandler(CommentHandler *Handler);
+
/// EnterMainSourceFile - Enter the specified FileID as the main source file,
/// which implicitly adds the builtin defines etc.
void EnterMainSourceFile();
@@ -791,6 +805,7 @@
void HandlePragmaSystemHeader(Token &SysHeaderTok);
void HandlePragmaDependency(Token &DependencyTok);
void HandlePragmaComment(Token &CommentTok);
+ void HandleComment(SourceRange Comment);
};
/// PreprocessorFactory - A generic factory interface for lazily creating
@@ -801,6 +816,15 @@
virtual Preprocessor* CreatePreprocessor() = 0;
};
+/// \brief Abstract base class that describes a handler that will receive
+/// source ranges for each of the comments encountered in the source file.
+class CommentHandler {
+public:
+ virtual ~CommentHandler();
+
+ virtual void HandleComment(Preprocessor &PP, SourceRange Comment) = 0;
+};
+
} // end namespace clang
#endif
diff --git a/include/clang/Parse/Action.h b/include/clang/Parse/Action.h
index 24262a1..d969562 100644
--- a/include/clang/Parse/Action.h
+++ b/include/clang/Parse/Action.h
@@ -152,6 +152,10 @@
/// an empty string if not. This is used for pretty crash reporting.
virtual std::string getDeclName(DeclPtrTy D) { return ""; }
+ /// \brief Invoked for each comment in the source code, providing the source
+ /// range that contains the comment.
+ virtual void ActOnComment(SourceRange Comment) { }
+
//===--------------------------------------------------------------------===//
// Declaration Tracking Callbacks.
//===--------------------------------------------------------------------===//
diff --git a/include/clang/Parse/Parser.h b/include/clang/Parse/Parser.h
index e5f62ec..e238054 100644
--- a/include/clang/Parse/Parser.h
+++ b/include/clang/Parse/Parser.h
@@ -82,7 +82,8 @@
llvm::OwningPtr<PragmaHandler> PackHandler;
llvm::OwningPtr<PragmaHandler> UnusedHandler;
llvm::OwningPtr<PragmaHandler> WeakHandler;
-
+ llvm::OwningPtr<clang::CommentHandler> CommentHandler;
+
/// Whether the '>' token acts as an operator or not. This will be
/// true except when we are parsing an expression within a C++
/// template argument list, where the '>' closes the template
diff --git a/lib/AST/ASTContext.cpp b/lib/AST/ASTContext.cpp
index 7f945db..2877cc3 100644
--- a/lib/AST/ASTContext.cpp
+++ b/lib/AST/ASTContext.cpp
@@ -37,7 +37,8 @@
bool FreeMem, unsigned size_reserve) :
GlobalNestedNameSpecifier(0), CFConstantStringTypeDecl(0),
ObjCFastEnumerationStateTypeDecl(0), SourceMgr(SM), LangOpts(LOpts),
- FreeMemory(FreeMem), Target(t), Idents(idents), Selectors(sels),
+ LoadedExternalComments(false), FreeMemory(FreeMem), Target(t),
+ Idents(idents), Selectors(sels),
BuiltinInfo(builtins), ExternalSource(0), PrintingPolicy(LOpts) {
if (size_reserve > 0) Types.reserve(size_reserve);
InitBuiltinTypes();
@@ -202,6 +203,207 @@
InitBuiltinType(NullPtrTy, BuiltinType::NullPtr);
}
+namespace {
+ class BeforeInTranslationUnit
+ : std::binary_function<SourceRange, SourceRange, bool> {
+ SourceManager *SourceMgr;
+
+ public:
+ explicit BeforeInTranslationUnit(SourceManager *SM) : SourceMgr(SM) { }
+
+ bool operator()(SourceRange X, SourceRange Y) {
+ return SourceMgr->isBeforeInTranslationUnit(X.getBegin(), Y.getBegin());
+ }
+ };
+}
+
+/// \brief Determine whether the given comment is a Doxygen-style comment.
+///
+/// \param Start the start of the comment text.
+///
+/// \param End the end of the comment text.
+///
+/// \param Member whether we want to check whether this is a member comment
+/// (which requires a < after the Doxygen-comment delimiter). Otherwise,
+/// we only return true when we find a non-member comment.
+static bool
+isDoxygenComment(SourceManager &SourceMgr, SourceRange Comment,
+ bool Member = false) {
+ const char *BufferStart
+ = SourceMgr.getBufferData(SourceMgr.getFileID(Comment.getBegin())).first;
+ const char *Start = BufferStart + SourceMgr.getFileOffset(Comment.getBegin());
+ const char* End = BufferStart + SourceMgr.getFileOffset(Comment.getEnd());
+
+ if (End - Start < 4)
+ return false;
+
+ assert(Start[0] == '/' && "Not a comment?");
+ if (Start[1] == '*' && !(Start[2] == '!' || Start[2] == '*'))
+ return false;
+ if (Start[1] == '/' && !(Start[2] == '!' || Start[2] == '/'))
+ return false;
+
+ return (Start[3] == '<') == Member;
+}
+
+/// \brief Retrieve the comment associated with the given declaration, if
+/// it has one.
+const char *ASTContext::getCommentForDecl(const Decl *D) {
+ if (!D)
+ return 0;
+
+ // Check whether we have cached a comment string for this declaration
+ // already.
+ llvm::DenseMap<const Decl *, std::string>::iterator Pos
+ = DeclComments.find(D);
+ if (Pos != DeclComments.end())
+ return Pos->second.c_str();
+
+ // If we have an external AST source and have not yet loaded comments from
+ // that source, do so now.
+ if (ExternalSource && !LoadedExternalComments) {
+ std::vector<SourceRange> LoadedComments;
+ ExternalSource->ReadComments(LoadedComments);
+
+ if (!LoadedComments.empty())
+ Comments.insert(Comments.begin(), LoadedComments.begin(),
+ LoadedComments.end());
+
+ LoadedExternalComments = true;
+ }
+
+ // If there are no comments anywhere, we won't find anything.
+ if (Comments.empty())
+ return 0;
+
+ // If the declaration doesn't map directly to a location in a file, we
+ // can't find the comment.
+ SourceLocation DeclStartLoc = D->getLocStart();
+ if (DeclStartLoc.isInvalid() || !DeclStartLoc.isFileID())
+ return 0;
+
+ // Find the comment that occurs just before this declaration.
+ std::vector<SourceRange>::iterator LastComment
+ = std::lower_bound(Comments.begin(), Comments.end(),
+ SourceRange(DeclStartLoc),
+ BeforeInTranslationUnit(&SourceMgr));
+
+ // Decompose the location for the start of the declaration and find the
+ // beginning of the file buffer.
+ std::pair<FileID, unsigned> DeclStartDecomp
+ = SourceMgr.getDecomposedLoc(DeclStartLoc);
+ const char *FileBufferStart
+ = SourceMgr.getBufferData(DeclStartDecomp.first).first;
+
+ // First check whether we have a comment for a member.
+ if (LastComment != Comments.end() &&
+ !isa<TagDecl>(D) && !isa<NamespaceDecl>(D) &&
+ isDoxygenComment(SourceMgr, *LastComment, true)) {
+ std::pair<FileID, unsigned> LastCommentEndDecomp
+ = SourceMgr.getDecomposedLoc(LastComment->getEnd());
+ if (DeclStartDecomp.first == LastCommentEndDecomp.first &&
+ SourceMgr.getLineNumber(DeclStartDecomp.first, DeclStartDecomp.second)
+ == SourceMgr.getLineNumber(LastCommentEndDecomp.first,
+ LastCommentEndDecomp.second)) {
+ // The Doxygen member comment comes after the declaration starts and
+ // is on the same line and in the same file as the declaration. This
+ // is the comment we want.
+ std::string &Result = DeclComments[D];
+ Result.append(FileBufferStart +
+ SourceMgr.getFileOffset(LastComment->getBegin()),
+ FileBufferStart + LastCommentEndDecomp.second + 1);
+ return Result.c_str();
+ }
+ }
+
+ if (LastComment == Comments.begin())
+ return 0;
+ --LastComment;
+
+ // Decompose the end of the comment.
+ std::pair<FileID, unsigned> LastCommentEndDecomp
+ = SourceMgr.getDecomposedLoc(LastComment->getEnd());
+
+ // If the comment and the declaration aren't in the same file, then they
+ // aren't related.
+ if (DeclStartDecomp.first != LastCommentEndDecomp.first)
+ return 0;
+
+ // Check that we actually have a Doxygen comment.
+ if (!isDoxygenComment(SourceMgr, *LastComment))
+ return 0;
+
+ // Compute the starting line for the declaration and for the end of the
+ // comment (this is expensive).
+ unsigned DeclStartLine
+ = SourceMgr.getLineNumber(DeclStartDecomp.first, DeclStartDecomp.second);
+ unsigned CommentEndLine
+ = SourceMgr.getLineNumber(LastCommentEndDecomp.first,
+ LastCommentEndDecomp.second);
+
+ // If the comment does not end on the line prior to the declaration, then
+ // the comment is not associated with the declaration at all.
+ if (CommentEndLine + 1 != DeclStartLine)
+ return 0;
+
+ // We have a comment, but there may be more comments on the previous lines.
+ // Keep looking so long as the comments are still Doxygen comments and are
+ // still adjacent.
+ unsigned ExpectedLine
+ = SourceMgr.getSpellingLineNumber(LastComment->getBegin()) - 1;
+ std::vector<SourceRange>::iterator FirstComment = LastComment;
+ while (FirstComment != Comments.begin()) {
+ // Look at the previous comment
+ --FirstComment;
+ std::pair<FileID, unsigned> Decomp
+ = SourceMgr.getDecomposedLoc(FirstComment->getEnd());
+
+ // If this previous comment is in a different file, we're done.
+ if (Decomp.first != DeclStartDecomp.first) {
+ ++FirstComment;
+ break;
+ }
+
+ // If this comment is not a Doxygen comment, we're done.
+ if (!isDoxygenComment(SourceMgr, *FirstComment)) {
+ ++FirstComment;
+ break;
+ }
+
+ // If the line number is not what we expected, we're done.
+ unsigned Line = SourceMgr.getLineNumber(Decomp.first, Decomp.second);
+ if (Line != ExpectedLine) {
+ ++FirstComment;
+ break;
+ }
+
+ // Set the next expected line number.
+ ExpectedLine
+ = SourceMgr.getSpellingLineNumber(FirstComment->getBegin()) - 1;
+ }
+
+ // The iterator range [FirstComment, LastComment] contains all of the
+ // BCPL comments that, together, are associated with this declaration.
+ // Form a single comment block string for this declaration that concatenates
+ // all of these comments.
+ std::string &Result = DeclComments[D];
+ while (FirstComment != LastComment) {
+ std::pair<FileID, unsigned> DecompStart
+ = SourceMgr.getDecomposedLoc(FirstComment->getBegin());
+ std::pair<FileID, unsigned> DecompEnd
+ = SourceMgr.getDecomposedLoc(FirstComment->getEnd());
+ Result.append(FileBufferStart + DecompStart.second,
+ FileBufferStart + DecompEnd.second + 1);
+ ++FirstComment;
+ }
+
+ // Append the last comment line.
+ Result.append(FileBufferStart +
+ SourceMgr.getFileOffset(LastComment->getBegin()),
+ FileBufferStart + LastCommentEndDecomp.second + 1);
+ return Result.c_str();
+}
+
//===----------------------------------------------------------------------===//
// Type Sizing and Analysis
//===----------------------------------------------------------------------===//
diff --git a/lib/Frontend/PCHReader.cpp b/lib/Frontend/PCHReader.cpp
index 765655b..95b1661 100644
--- a/lib/Frontend/PCHReader.cpp
+++ b/lib/Frontend/PCHReader.cpp
@@ -344,7 +344,8 @@
IdentifierOffsets(0),
MethodPoolLookupTable(0), MethodPoolLookupTableData(0),
TotalSelectorsInMethodPool(0), SelectorOffsets(0),
- TotalNumSelectors(0), NumStatHits(0), NumStatMisses(0),
+ TotalNumSelectors(0), Comments(0), NumComments(0),
+ NumStatHits(0), NumStatMisses(0),
NumSLocEntriesRead(0), NumStatementsRead(0),
NumMacrosRead(0), NumMethodPoolSelectorsRead(0), NumMethodPoolMisses(0),
NumLexicalDeclContextsRead(0), NumVisibleDeclContextsRead(0) { }
@@ -1350,6 +1351,11 @@
case pch::ORIGINAL_FILE_NAME:
OriginalFileName.assign(BlobStart, BlobLen);
break;
+
+ case pch::COMMENT_RANGES:
+ Comments = (SourceRange *)BlobStart;
+ NumComments = BlobLen / sizeof(SourceRange);
+ break;
}
}
Error("premature end of bitstream in PCH file");
@@ -1664,6 +1670,12 @@
return false;
}
+void PCHReader::ReadComments(std::vector<SourceRange> &Comments) {
+ Comments.resize(NumComments);
+ std::copy(this->Comments, this->Comments + NumComments,
+ Comments.begin());
+}
+
/// \brief Read and return the type at the given offset.
///
/// This routine actually reads the record corresponding to the type
diff --git a/lib/Frontend/PCHWriter.cpp b/lib/Frontend/PCHWriter.cpp
index 41a5025..3bfc9e8 100644
--- a/lib/Frontend/PCHWriter.cpp
+++ b/lib/Frontend/PCHWriter.cpp
@@ -374,7 +374,8 @@
RECORD(STAT_CACHE);
RECORD(EXT_VECTOR_DECLS);
RECORD(OBJC_CATEGORY_IMPLEMENTATIONS);
-
+ RECORD(COMMENT_RANGES);
+
// SourceManager Block.
BLOCK(SOURCE_MANAGER_BLOCK);
RECORD(SM_SLOC_FILE_ENTRY);
@@ -989,6 +990,24 @@
Stream.ExitBlock();
}
+void PCHWriter::WriteComments(ASTContext &Context) {
+ using namespace llvm;
+
+ if (Context.Comments.empty())
+ return;
+
+ BitCodeAbbrev *CommentAbbrev = new BitCodeAbbrev();
+ CommentAbbrev->Add(BitCodeAbbrevOp(pch::COMMENT_RANGES));
+ CommentAbbrev->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Blob));
+ unsigned CommentCode = Stream.EmitAbbrev(CommentAbbrev);
+
+ RecordData Record;
+ Record.push_back(pch::COMMENT_RANGES);
+ Stream.EmitRecordWithBlob(CommentCode, Record,
+ (const char*)&Context.Comments[0],
+ Context.Comments.size() * sizeof(SourceRange));
+}
+
//===----------------------------------------------------------------------===//
// Type Serialization
//===----------------------------------------------------------------------===//
@@ -1746,7 +1765,8 @@
WriteStatCache(*StatCalls);
WriteSourceManagerBlock(Context.getSourceManager(), PP);
WritePreprocessor(PP);
-
+ WriteComments(Context);
+
// Keep writing types and declarations until all types and
// declarations have been written.
do {
diff --git a/lib/Lex/Lexer.cpp b/lib/Lex/Lexer.cpp
index 81ea606..6f1043a 100644
--- a/lib/Lex/Lexer.cpp
+++ b/lib/Lex/Lexer.cpp
@@ -903,7 +903,10 @@
} while (C != '\n' && C != '\r');
// Found but did not consume the newline.
-
+ if (PP)
+ PP->HandleComment(SourceRange(getSourceLocation(BufferPtr),
+ getSourceLocation(CurPtr)));
+
// If we are returning comments as tokens, return this comment as a token.
if (inKeepCommentMode())
return SaveBCPLComment(Result, CurPtr);
@@ -1146,6 +1149,10 @@
C = *CurPtr++;
}
+ if (PP)
+ PP->HandleComment(SourceRange(getSourceLocation(BufferPtr),
+ getSourceLocation(CurPtr)));
+
// If we are returning comments as tokens, return this comment as a token.
if (inKeepCommentMode()) {
FormTokenWithChars(Result, CurPtr, tok::comment);
diff --git a/lib/Lex/Pragma.cpp b/lib/Lex/Pragma.cpp
index 3227d1c..bb0b71e 100644
--- a/lib/Lex/Pragma.cpp
+++ b/lib/Lex/Pragma.cpp
@@ -19,6 +19,7 @@
#include "clang/Lex/LexDiagnostic.h"
#include "clang/Basic/FileManager.h"
#include "clang/Basic/SourceManager.h"
+#include <algorithm>
using namespace clang;
// Out-of-line destructor to provide a home for the class.
diff --git a/lib/Lex/Preprocessor.cpp b/lib/Lex/Preprocessor.cpp
index 24ee5d8..9f0c15f 100644
--- a/lib/Lex/Preprocessor.cpp
+++ b/lib/Lex/Preprocessor.cpp
@@ -476,3 +476,26 @@
if (II.isExtensionToken() && !DisableMacroExpansion)
Diag(Identifier, diag::ext_token_used);
}
+
+void Preprocessor::AddCommentHandler(CommentHandler *Handler) {
+ assert(Handler && "NULL comment handler");
+ assert(std::find(CommentHandlers.begin(), CommentHandlers.end(), Handler) ==
+ CommentHandlers.end() && "Comment handler already registered");
+ CommentHandlers.push_back(Handler);
+}
+
+void Preprocessor::RemoveCommentHandler(CommentHandler *Handler) {
+ std::vector<CommentHandler *>::iterator Pos
+ = std::find(CommentHandlers.begin(), CommentHandlers.end(), Handler);
+ assert(Pos != CommentHandlers.end() && "Comment handler not registered");
+ CommentHandlers.erase(Pos);
+}
+
+void Preprocessor::HandleComment(SourceRange Comment) {
+ for (std::vector<CommentHandler *>::iterator H = CommentHandlers.begin(),
+ HEnd = CommentHandlers.end();
+ H != HEnd; ++H)
+ (*H)->HandleComment(*this, Comment);
+}
+
+CommentHandler::~CommentHandler() { }
diff --git a/lib/Parse/Parser.cpp b/lib/Parse/Parser.cpp
index 29d1d87..9771cf7 100644
--- a/lib/Parse/Parser.cpp
+++ b/lib/Parse/Parser.cpp
@@ -20,6 +20,19 @@
#include "ParsePragma.h"
using namespace clang;
+/// \brief A comment handler that passes comments found by the preprocessor
+/// to the parser action.
+class ActionCommentHandler : public CommentHandler {
+ Action &Actions;
+
+public:
+ explicit ActionCommentHandler(Action &Actions) : Actions(Actions) { }
+
+ virtual void HandleComment(Preprocessor &PP, SourceRange Comment) {
+ Actions.ActOnComment(Comment);
+ }
+};
+
Parser::Parser(Preprocessor &pp, Action &actions)
: CrashInfo(*this), PP(pp), Actions(actions), Diags(PP.getDiagnostics()),
GreaterThanIsOperator(true) {
@@ -43,6 +56,9 @@
WeakHandler.reset(new
PragmaWeakHandler(&PP.getIdentifierTable().get("weak"), actions));
PP.AddPragmaHandler(0, WeakHandler.get());
+
+ CommentHandler.reset(new ActionCommentHandler(actions));
+ PP.AddCommentHandler(CommentHandler.get());
}
/// If a crash happens while the parser is active, print out a line indicating
@@ -294,6 +310,7 @@
UnusedHandler.reset();
PP.RemovePragmaHandler(0, WeakHandler.get());
WeakHandler.reset();
+ PP.RemoveCommentHandler(CommentHandler.get());
}
/// Initialize - Warm up the parser.
diff --git a/lib/Sema/Sema.cpp b/lib/Sema/Sema.cpp
index c37f66d..d1e8e21 100644
--- a/lib/Sema/Sema.cpp
+++ b/lib/Sema/Sema.cpp
@@ -345,3 +345,7 @@
= SemaRef.ActiveTemplateInstantiations.back();
}
}
+
+void Sema::ActOnComment(SourceRange Comment) {
+ Context.Comments.push_back(Comment);
+}
diff --git a/lib/Sema/Sema.h b/lib/Sema/Sema.h
index a6c34e8..72fb0a8 100644
--- a/lib/Sema/Sema.h
+++ b/lib/Sema/Sema.h
@@ -362,6 +362,8 @@
return CurBlock ? CurBlock->SwitchStack : FunctionSwitchStack;
}
+ virtual void ActOnComment(SourceRange Comment);
+
//===--------------------------------------------------------------------===//
// Type Analysis / Processing: SemaType.cpp.
//
diff --git a/test/CMakeLists.txt b/test/CMakeLists.txt
index 7521a1d..d5e2327 100644
--- a/test/CMakeLists.txt
+++ b/test/CMakeLists.txt
@@ -8,6 +8,7 @@
"Driver"
"FixIt"
"Frontend"
+ "Index"
"Lexer"
"Misc"
"PCH"
@@ -54,6 +55,6 @@
--clang-cc=${LLVM_TOOLS_PATH}/${CMAKE_CFG_INTDIR}/clang-cc
${all_testdirs}
WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}
- DEPENDS clang clang-cc
+ DEPENDS clang clang-cc index-test
COMMENT "Running Clang regression tests")
endif()
diff --git a/test/Index/comments.c b/test/Index/comments.c
new file mode 100644
index 0000000..7ad8fd7
--- /dev/null
+++ b/test/Index/comments.c
@@ -0,0 +1,30 @@
+// RUN: clang-cc -emit-pch -o %t.ast %s &&
+// RUN: index-test %t.ast -point-at %s:22:6 | grep "starts here" &&
+// RUN: index-test %t.ast -point-at %s:22:6 | grep "block comment" &&
+// RUN: index-test %t.ast -point-at %s:28:6 | grep "BCPL" &&
+// RUN: index-test %t.ast -point-at %s:28:6 | grep "But" &&
+// RUN: index-test %t.ast -point-at %s:28:6 | grep "NOT" | count 0 &&
+// RUN: index-test %t.ast -point-at %s:30:6 | grep "member"
+
+
+
+
+
+
+//! It all starts here.
+/*! It's a little odd to continue line this,
+ *
+ * but we need more multi-line comments. */
+/// This comment comes before my other comments
+/** This is a block comment that is associated with the function f. It
+ * runs for three lines.
+ */
+void f(int, int);
+
+// NOT IN THE COMMENT
+/// This is a BCPL comment that is associated with the function g.
+/// It has only two lines.
+/** But there are other blocks that are part of the comment, too. */
+void g(int);
+
+void h(int); ///< This is a member comment.
\ No newline at end of file
diff --git a/tools/index-test/index-test.cpp b/tools/index-test/index-test.cpp
index 9b10c55..552b7b0 100644
--- a/tools/index-test/index-test.cpp
+++ b/tools/index-test/index-test.cpp
@@ -130,6 +130,9 @@
OS << ND->getNameAsString();
OS << "\n";
+ if (const char *Comment = AST->getASTContext().getCommentForDecl(Point.D))
+ OS << "Comment associated with this declaration:\n" << Comment << "\n";
+
if (Point.Node) {
OS << "Statement node at point: " << Point.Node->getStmtClassName()
<< " ";