Implement AST classes for comments, a real parser for Doxygen comments and a

very simple semantic analysis that just builds the AST; minor changes for lexer
to pick up source locations I didn't think about before.

Comments AST is modelled along the ideas of HTML AST: block and inline content.

* Block content is a paragraph or a command that has a paragraph as an argument
  or verbatim command.
* Inline content is placed within some block.  Inline content includes plain
  text, inline commands and HTML as tag soup.



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@159790 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 112 insertions, 0 deletions

diff --git a/include/clang/AST/CommentParser.h b/include/clang/AST/CommentParser.h
new file mode 100644
index 0000000000..53c58662bf
--- /dev/null
+++ b/include/clang/AST/CommentParser.h
@@ -0,0 +1,112 @@
+//===--- CommentParser.h - Doxygen comment parser ---------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+//  This file defines the Doxygen comment parser.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_AST_COMMENT_PARSER_H
+#define LLVM_CLANG_AST_COMMENT_PARSER_H
+
+#include "clang/AST/CommentLexer.h"
+#include "clang/AST/Comment.h"
+#include "clang/AST/CommentSema.h"
+#include "llvm/Support/Allocator.h"
+
+namespace clang {
+namespace comments {
+
+/// Doxygen comment parser.
+class Parser {
+  Lexer &L;
+
+  Sema &S;
+
+  llvm::BumpPtrAllocator &Allocator;
+
+  template<typename T>
+  ArrayRef<T> copyArray(ArrayRef<T> Source) {
+    size_t Size = Source.size();
+    if (Size != 0) {
+      T *Mem = new (Allocator) T[Size];
+      std::copy(Source.begin(), Source.end(), Mem);
+      return llvm::makeArrayRef(Mem, Size);
+    } else
+      return llvm::makeArrayRef(static_cast<T *>(NULL), 0);
+  }
+
+  /// Current lookahead token.  We can safely assume that all tokens are from
+  /// a single source file.
+  Token Tok;
+
+  /// A stack of additional lookahead tokens.
+  SmallVector<Token, 8> MoreLATokens;
+
+  SourceLocation consumeToken() {
+    SourceLocation Loc = Tok.getLocation();
+    if (MoreLATokens.empty())
+      L.lex(Tok);
+    else {
+      Tok = MoreLATokens.back();
+      MoreLATokens.pop_back();
+    }
+    return Loc;
+  }
+
+  void putBack(const Token &OldTok) {
+    MoreLATokens.push_back(Tok);
+    Tok = OldTok;
+  }
+
+  void putBack(ArrayRef<Token> Toks) {
+    if (Toks.empty())
+      return;
+
+    MoreLATokens.push_back(Tok);
+    for (const Token *I = &Toks.back(),
+         *B = &Toks.front() + 1;
+         I != B; --I) {
+      MoreLATokens.push_back(*I);
+    }
+
+    Tok = Toks[0];
+  }
+
+public:
+  Parser(Lexer &L, Sema &S, llvm::BumpPtrAllocator &Allocator);
+
+  /// Parse arguments for \\param command.
+  ParamCommandComment *parseParamCommandArgs(
+                                    ParamCommandComment *PC,
+                                    TextTokenRetokenizer &Retokenizer);
+
+  BlockCommandComment *parseBlockCommandArgs(
+                                    BlockCommandComment *BC,
+                                    TextTokenRetokenizer &Retokenizer,
+                                    unsigned NumArgs);
+
+  BlockCommandComment *parseBlockCommand();
+  InlineCommandComment *parseInlineCommand();
+
+  HTMLOpenTagComment *parseHTMLOpenTag();
+  HTMLCloseTagComment *parseHTMLCloseTag();
+
+  BlockContentComment *parseParagraphOrBlockCommand();
+
+  VerbatimBlockComment *parseVerbatimBlock();
+  VerbatimLineComment *parseVerbatimLine();
+  BlockContentComment *parseBlockContent();
+  FullComment *parseFullComment();
+};
+
+} // end namespace comments
+} // end namespace clang
+
+#endif
+