zcov: / include/clang/Lex/Lexer.h


Files: 1 Branches Taken: 35.7% 5 / 14
Generated: 2010-02-10 01:31 Branches Executed: 57.1% 8 / 14
Line Coverage: 93.8% 60 / 64


Programs: 58 Runs 95850


       1                 : //===--- Lexer.h - C Language Family Lexer ----------------------*- C++ -*-===//
       2                 : //
       3                 : //                     The LLVM Compiler Infrastructure
       4                 : //
       5                 : // This file is distributed under the University of Illinois Open Source
       6                 : // License. See LICENSE.TXT for details.
       7                 : //
       8                 : //===----------------------------------------------------------------------===//
       9                 : //
      10                 : //  This file defines the Lexer interface.
      11                 : //
      12                 : //===----------------------------------------------------------------------===//
      13                 : 
      14                 : #ifndef LLVM_CLANG_LEXER_H
      15                 : #define LLVM_CLANG_LEXER_H
      16                 : 
      17                 : #include "clang/Lex/PreprocessorLexer.h"
      18                 : #include "clang/Basic/LangOptions.h"
      19                 : #include "llvm/ADT/SmallVector.h"
      20                 : #include <string>
      21                 : #include <vector>
      22                 : #include <cassert>
      23                 : 
      24                 : namespace clang {
      25                 : class Diagnostic;
      26                 : class SourceManager;
      27                 : class Preprocessor;
      28                 : class DiagnosticBuilder;
      29                 : 
      30                 : /// Lexer - This provides a simple interface that turns a text buffer into a
      31                 : /// stream of tokens.  This provides no support for file reading or buffering,
      32                 : /// or buffering/seeking of tokens, only forward lexing is supported.  It relies
      33                 : /// on the specified Preprocessor object to handle preprocessor directives, etc.
                        0: branch 2 not taken
                   138969: branch 3 taken
                     5618: branch 7 taken
                        0: branch 8 not taken
      34           144587: class Lexer : public PreprocessorLexer {
      35                 :   //===--------------------------------------------------------------------===//
      36                 :   // Constant configuration values for this lexer.
      37                 :   const char *BufferStart;       // Start of the buffer.
      38                 :   const char *BufferEnd;         // End of the buffer.
      39                 :   SourceLocation FileLoc;        // Location for start of file.
      40                 :   LangOptions Features;          // Features enabled by this language (cache).
      41                 :   bool Is_PragmaLexer : 1;       // True if lexer for _Pragma handling.
      42                 :   bool IsInConflictMarker : 1;   // True if in a VCS conflict marker '<<<<<<<'
      43                 :   
      44                 :   //===--------------------------------------------------------------------===//
      45                 :   // Context-specific lexing flags set by the preprocessor.
      46                 :   //
      47                 : 
      48                 :   /// ExtendedTokenMode - The lexer can optionally keep comments and whitespace
      49                 :   /// and return them as tokens.  This is used for -C and -CC modes, and
      50                 :   /// whitespace preservation can be useful for some clients that want to lex
      51                 :   /// the file in raw mode and get every character from the file.
      52                 :   ///
      53                 :   /// When this is set to 2 it returns comments and whitespace.  When set to 1
      54                 :   /// it returns comments, when it is set to 0 it returns normal tokens only.
      55                 :   unsigned char ExtendedTokenMode;
      56                 : 
      57                 :   //===--------------------------------------------------------------------===//
      58                 :   // Context that changes as the file is lexed.
      59                 :   // NOTE: any state that mutates when in raw mode must have save/restore code
      60                 :   // in Lexer::isNextPPTokenLParen.
      61                 : 
      62                 :   // BufferPtr - Current pointer into the buffer.  This is the next character
      63                 :   // to be lexed.
      64                 :   const char *BufferPtr;
      65                 : 
      66                 :   // IsAtStartOfLine - True if the next lexed token should get the "start of
      67                 :   // line" flag set on it.
      68                 :   bool IsAtStartOfLine;
      69                 : 
      70                 :   Lexer(const Lexer&);          // DO NOT IMPLEMENT
      71                 :   void operator=(const Lexer&); // DO NOT IMPLEMENT
      72                 :   friend class Preprocessor;
      73                 : 
      74                 :   void InitLexer(const char *BufStart, const char *BufPtr, const char *BufEnd);
      75                 : public:
      76                 : 
      77                 :   /// Lexer constructor - Create a new lexer object for the specified buffer
      78                 :   /// with the specified preprocessor managing the lexing process.  This lexer
      79                 :   /// assumes that the associated file buffer and Preprocessor objects will
      80                 :   /// outlive it, so it doesn't take ownership of either of them.
      81                 :   Lexer(FileID FID, const llvm::MemoryBuffer *InputBuffer, Preprocessor &PP);
      82                 : 
      83                 :   /// Lexer constructor - Create a new raw lexer object.  This object is only
      84                 :   /// suitable for calls to 'LexRawToken'.  This lexer assumes that the text
      85                 :   /// range will outlive it, so it doesn't take ownership of it.
      86                 :   Lexer(SourceLocation FileLoc, const LangOptions &Features,
      87                 :         const char *BufStart, const char *BufPtr, const char *BufEnd);
      88                 : 
      89                 :   /// Lexer constructor - Create a new raw lexer object.  This object is only
      90                 :   /// suitable for calls to 'LexRawToken'.  This lexer assumes that the text
      91                 :   /// range will outlive it, so it doesn't take ownership of it.
      92                 :   Lexer(FileID FID, const llvm::MemoryBuffer *InputBuffer,
      93                 :         const SourceManager &SM, const LangOptions &Features);
      94                 : 
      95                 :   /// Create_PragmaLexer: Lexer constructor - Create a new lexer object for
      96                 :   /// _Pragma expansion.  This has a variety of magic semantics that this method
      97                 :   /// sets up.  It returns a new'd Lexer that must be delete'd when done.
      98                 :   static Lexer *Create_PragmaLexer(SourceLocation SpellingLoc,
      99                 :                                    SourceLocation InstantiationLocStart,
     100                 :                                    SourceLocation InstantiationLocEnd,
     101                 :                                    unsigned TokLen, Preprocessor &PP);
     102                 : 
     103                 : 
     104                 :   /// getFeatures - Return the language features currently enabled.  NOTE: this
     105                 :   /// lexer modifies features as a file is parsed!
     106               76:   const LangOptions &getFeatures() const { return Features; }
     107                 : 
     108                 :   /// getFileLoc - Return the File Location for the file we are lexing out of.
     109                 :   /// The physical location encodes the location where the characters come from,
     110                 :   /// the virtual location encodes where we should *claim* the characters came
     111                 :   /// from.  Currently this is only used by _Pragma handling.
     112              944:   SourceLocation getFileLoc() const { return FileLoc; }
     113                 : 
     114                 :   /// Lex - Return the next token in the file.  If this is the end of file, it
     115                 :   /// return the tok::eof token.  Return true if an error occurred and
     116                 :   /// compilation should terminate, false if normal.  This implicitly involves
     117                 :   /// the preprocessor.
     118          2413704:   void Lex(Token &Result) {
     119                 :     // Start a new token.
     120          2413704:     Result.startToken();
     121                 : 
     122                 :     // NOTE, any changes here should also change code after calls to
     123                 :     // Preprocessor::HandleDirective
                   151794: branch 0 taken
                  2261910: branch 1 taken
     124          2413704:     if (IsAtStartOfLine) {
     125           151794:       Result.setFlag(Token::StartOfLine);
     126           151794:       IsAtStartOfLine = false;
     127                 :     }
     128                 : 
     129                 :     // Get a token.  Note that this may delete the current lexer if the end of
     130                 :     // file is reached.
     131          2413704:     LexTokenInternal(Result);
     132          2413704:   }
     133                 : 
     134                 :   /// isPragmaLexer - Returns true if this Lexer is being used to lex a pragma.
     135           270688:   bool isPragmaLexer() const { return Is_PragmaLexer; }
     136                 : 
     137                 :   /// IndirectLex - An indirect call to 'Lex' that can be invoked via
     138                 :   ///  the PreprocessorLexer interface.
     139              729:   void IndirectLex(Token &Result) { Lex(Result); }
     140                 : 
     141                 :   /// LexFromRawLexer - Lex a token from a designated raw lexer (one with no
     142                 :   /// associated preprocessor object.  Return true if the 'next character to
     143                 :   /// read' pointer points at the end of the lexer buffer, false otherwise.
     144           137981:   bool LexFromRawLexer(Token &Result) {
                        0: branch 0 not taken
                     3188: branch 1 taken
     145           137981:     assert(LexingRawMode && "Not already in raw mode!");
     146           137981:     Lex(Result);
     147                 :     // Note that lexing to the end of the buffer doesn't implicitly delete the
     148                 :     // lexer when in raw mode.
     149           137981:     return BufferPtr == BufferEnd;
     150                 :   }
     151                 : 
     152                 :   /// isKeepWhitespaceMode - Return true if the lexer should return tokens for
     153                 :   /// every character in the file, including whitespace and comments.  This
     154                 :   /// should only be used in raw mode, as the preprocessor is not prepared to
     155                 :   /// deal with the excess tokens.
     156          1913977:   bool isKeepWhitespaceMode() const {
     157          1913977:     return ExtendedTokenMode > 1;
     158                 :   }
     159                 : 
     160                 :   /// SetKeepWhitespaceMode - This method lets clients enable or disable
     161                 :   /// whitespace retention mode.
     162                0:   void SetKeepWhitespaceMode(bool Val) {
     163                 :     assert((!Val || LexingRawMode) &&
                        0: branch 0 not taken
                        0: branch 1 not taken
                        0: branch 2 not taken
                        0: branch 3 not taken
     164                0:            "Can only enable whitespace retention in raw mode");
                        0: branch 0 not taken
                        0: branch 1 not taken
     165                0:     ExtendedTokenMode = Val ? 2 : 0;
     166                0:   }
     167                 : 
     168                 :   /// inKeepCommentMode - Return true if the lexer should return comments as
     169                 :   /// tokens.
     170            87569:   bool inKeepCommentMode() const {
     171            87569:     return ExtendedTokenMode > 0;
     172                 :   }
     173                 : 
     174                 :   /// SetCommentRetentionMode - Change the comment retention mode of the lexer
     175                 :   /// to the specified mode.  This is really only useful when lexing in raw
     176                 :   /// mode, because otherwise the lexer needs to manage this.
     177           739038:   void SetCommentRetentionState(bool Mode) {
     178                 :     assert(!isKeepWhitespaceMode() &&
     179           739038:            "Can't play with comment retention state when retaining whitespace");
     180           877839:     ExtendedTokenMode = Mode ? 1 : 0;
     181           739038:   }
     182                 : 
     183                4:   const char *getBufferStart() const { return BufferStart; }
     184                 : 
     185                 :   /// ReadToEndOfLine - Read the rest of the current preprocessor line as an
     186                 :   /// uninterpreted string.  This switches the lexer out of directive mode.
     187                 :   std::string ReadToEndOfLine();
     188                 : 
     189                 : 
     190                 :   /// Diag - Forwarding function for diagnostics.  This translate a source
     191                 :   /// position in the current buffer into a SourceLocation object for rendering.
     192                 :   DiagnosticBuilder Diag(const char *Loc, unsigned DiagID) const;
     193                 : 
     194                 :   /// getSourceLocation - Return a source location identifier for the specified
     195                 :   /// offset in the current file.
     196                 :   SourceLocation getSourceLocation(const char *Loc, unsigned TokLen = 1) const;
     197                 : 
     198                 :   /// getSourceLocation - Return a source location for the next character in
     199                 :   /// the current file.
     200              552:   SourceLocation getSourceLocation() { return getSourceLocation(BufferPtr); }
     201                 : 
     202                 :   /// \brief Return the current location in the buffer.
     203               51:   const char *getBufferLocation() const { return BufferPtr; }
     204                 :   
     205                 :   /// Stringify - Convert the specified string into a C string by escaping '\'
     206                 :   /// and " characters.  This does not add surrounding ""'s to the string.
     207                 :   /// If Charify is true, this escapes the ' character instead of ".
     208                 :   static std::string Stringify(const std::string &Str, bool Charify = false);
     209                 : 
     210                 :   /// Stringify - Convert the specified string into a C string by escaping '\'
     211                 :   /// and " characters.  This does not add surrounding ""'s to the string.
     212                 :   static void Stringify(llvm::SmallVectorImpl<char> &Str);
     213                 : 
     214                 :   /// MeasureTokenLength - Relex the token at the specified location and return
     215                 :   /// its length in bytes in the input file.  If the token needs cleaning (e.g.
     216                 :   /// includes a trigraph or an escaped newline) then this count includes bytes
     217                 :   /// that are part of that.
     218                 :   static unsigned MeasureTokenLength(SourceLocation Loc,
     219                 :                                      const SourceManager &SM,
     220                 :                                      const LangOptions &LangOpts);
     221                 : 
     222                 :   //===--------------------------------------------------------------------===//
     223                 :   // Internal implementation interfaces.
     224                 : private:
     225                 : 
     226                 :   /// LexTokenInternal - Internal interface to lex a preprocessing token. Called
     227                 :   /// by Lex.
     228                 :   ///
     229                 :   void LexTokenInternal(Token &Result);
     230                 : 
     231                 :   /// FormTokenWithChars - When we lex a token, we have identified a span
     232                 :   /// starting at BufferPtr, going to TokEnd that forms the token.  This method
     233                 :   /// takes that range and assigns it to the token as its location and size.  In
     234                 :   /// addition, since tokens cannot overlap, this also updates BufferPtr to be
     235                 :   /// TokEnd.
     236                 :   void FormTokenWithChars(Token &Result, const char *TokEnd,
     237          2700239:                           tok::TokenKind Kind) {
     238          2700239:     unsigned TokLen = TokEnd-BufferPtr;
     239          2700239:     Result.setLength(TokLen);
     240          2700239:     Result.setLocation(getSourceLocation(BufferPtr, TokLen));
     241          2700239:     Result.setKind(Kind);
     242          2700239:     BufferPtr = TokEnd;
     243          2700239:   }
     244                 : 
     245                 :   /// isNextPPTokenLParen - Return 1 if the next unexpanded token will return a
     246                 :   /// tok::l_paren token, 0 if it is something else and 2 if there are no more
     247                 :   /// tokens in the buffer controlled by this lexer.
     248                 :   unsigned isNextPPTokenLParen();
     249                 : 
     250                 :   //===--------------------------------------------------------------------===//
     251                 :   // Lexer character reading interfaces.
     252                 : public:
     253                 : 
     254                 :   // This lexer is built on two interfaces for reading characters, both of which
     255                 :   // automatically provide phase 1/2 translation.  getAndAdvanceChar is used
     256                 :   // when we know that we will be reading a character from the input buffer and
     257                 :   // that this character will be part of the result token. This occurs in (f.e.)
     258                 :   // string processing, because we know we need to read until we find the
     259                 :   // closing '"' character.
     260                 :   //
     261                 :   // The second interface is the combination of getCharAndSize with
     262                 :   // ConsumeChar.  getCharAndSize reads a phase 1/2 translated character,
     263                 :   // returning it and its size.  If the lexer decides that this character is
     264                 :   // part of the current token, it calls ConsumeChar on it.  This two stage
     265                 :   // approach allows us to emit diagnostics for characters (e.g. warnings about
     266                 :   // trigraphs), knowing that they only are emitted if the character is
     267                 :   // consumed.
     268                 : 
     269                 :   /// isObviouslySimpleCharacter - Return true if the specified character is
     270                 :   /// obviously the same in translation phase 1 and translation phase 3.  This
     271                 :   /// can return false for characters that end up being the same, but it will
     272                 :   /// never return true for something that needs to be mapped.
     273          5559562:   static bool isObviouslySimpleCharacter(char C) {
     274          5559562:     return C != '?' && C != '\\';
     275                 :   }
     276                 : 
     277                 :   /// getAndAdvanceChar - Read a single 'character' from the specified buffer,
     278                 :   /// advance over it, and return it.  This is tricky in several cases.  Here we
     279                 :   /// just handle the trivial case and fall-back to the non-inlined
     280                 :   /// getCharAndSizeSlow method to handle the hard case.
     281          3126992:   inline char getAndAdvanceChar(const char *&Ptr, Token &Tok) {
     282                 :     // If this is not a trigraph and not a UCN or escaped newline, return
     283                 :     // quickly.
     284          3126992:     if (isObviouslySimpleCharacter(Ptr[0])) return *Ptr++;
     285                 : 
     286             6427:     unsigned Size = 0;
     287             6427:     char C = getCharAndSizeSlow(Ptr, Size, &Tok);
     288             6427:     Ptr += Size;
     289             6427:     return C;
     290                 :   }
     291                 : 
     292                 : private:
     293                 :   /// ConsumeChar - When a character (identified by getCharAndSize) is consumed
     294                 :   /// and added to a given token, check to see if there are diagnostics that
     295                 :   /// need to be emitted or flags that need to be set on the token.  If so, do
     296                 :   /// it.
     297          1047912:   const char *ConsumeChar(const char *Ptr, unsigned Size, Token &Tok) {
     298                 :     // Normal case, we consumed exactly one token.  Just return it.
     299          1047912:     if (Size == 1)
     300          1047901:       return Ptr+Size;
     301                 : 
     302                 :     // Otherwise, re-lex the character with a current token, allowing
     303                 :     // diagnostics to be emitted and flags to be set.
     304               11:     Size = 0;
     305               11:     getCharAndSizeSlow(Ptr, Size, &Tok);
     306               11:     return Ptr+Size;
     307                 :   }
     308                 : 
     309                 :   /// getCharAndSize - Peek a single 'character' from the specified buffer,
     310                 :   /// get its size, and return it.  This is tricky in several cases.  Here we
     311                 :   /// just handle the trivial case and fall-back to the non-inlined
     312                 :   /// getCharAndSizeSlow method to handle the hard case.
     313          1768246:   inline char getCharAndSize(const char *Ptr, unsigned &Size) {
     314                 :     // If this is not a trigraph and not a UCN or escaped newline, return
     315                 :     // quickly.
     316          1768246:     if (isObviouslySimpleCharacter(Ptr[0])) {
     317          1767985:       Size = 1;
     318          1767985:       return *Ptr;
     319                 :     }
     320                 : 
     321              261:     Size = 0;
     322              261:     return getCharAndSizeSlow(Ptr, Size);
     323                 :   }
     324                 : 
     325                 :   /// getCharAndSizeSlow - Handle the slow/uncommon case of the getCharAndSize
     326                 :   /// method.
     327                 :   char getCharAndSizeSlow(const char *Ptr, unsigned &Size, Token *Tok = 0);
     328                 : public:
     329                 : 
     330                 :   /// getCharAndSizeNoWarn - Like the getCharAndSize method, but does not ever
     331                 :   /// emit a warning.
     332                 :   static inline char getCharAndSizeNoWarn(const char *Ptr, unsigned &Size,
     333            35378:                                           const LangOptions &Features) {
     334                 :     // If this is not a trigraph and not a UCN or escaped newline, return
     335                 :     // quickly.
     336            35378:     if (isObviouslySimpleCharacter(Ptr[0])) {
     337            35045:       Size = 1;
     338            35045:       return *Ptr;
     339                 :     }
     340                 : 
     341              333:     Size = 0;
     342              333:     return getCharAndSizeSlowNoWarn(Ptr, Size, Features);
     343                 :   }
     344                 : 
     345                 :   /// getEscapedNewLineSize - Return the size of the specified escaped newline,
     346                 :   /// or 0 if it is not an escaped newline. P[-1] is known to be a "\" on entry
     347                 :   /// to this function.
     348                 :   static unsigned getEscapedNewLineSize(const char *P);
     349                 : 
     350                 :   /// SkipEscapedNewLines - If P points to an escaped newline (or a series of
     351                 :   /// them), skip over them and return the first non-escaped-newline found,
     352                 :   /// otherwise return P.
     353                 :   static const char *SkipEscapedNewLines(const char *P);
     354                 : private:
     355                 : 
     356                 :   /// getCharAndSizeSlowNoWarn - Same as getCharAndSizeSlow, but never emits a
     357                 :   /// diagnostic.
     358                 :   static char getCharAndSizeSlowNoWarn(const char *Ptr, unsigned &Size,
     359                 :                                        const LangOptions &Features);
     360                 : 
     361                 :   //===--------------------------------------------------------------------===//
     362                 :   // Other lexer functions.
     363                 : 
     364                 :   // Helper functions to lex the remainder of a token of the specific type.
     365                 :   void LexIdentifier         (Token &Result, const char *CurPtr);
     366                 :   void LexNumericConstant    (Token &Result, const char *CurPtr);
     367                 :   void LexStringLiteral      (Token &Result, const char *CurPtr,bool Wide);
     368                 :   void LexAngledStringLiteral(Token &Result, const char *CurPtr);
     369                 :   void LexCharConstant       (Token &Result, const char *CurPtr);
     370                 :   bool LexEndOfFile          (Token &Result, const char *CurPtr);
     371                 : 
     372                 :   bool SkipWhitespace        (Token &Result, const char *CurPtr);
     373                 :   bool SkipBCPLComment       (Token &Result, const char *CurPtr);
     374                 :   bool SkipBlockComment      (Token &Result, const char *CurPtr);
     375                 :   bool SaveBCPLComment       (Token &Result, const char *CurPtr);
     376                 :   
     377                 :   bool IsStartOfConflictMarker(const char *CurPtr);
     378                 :   bool HandleEndOfConflictMarker(const char *CurPtr);
     379                 : };
     380                 : 
     381                 : 
     382                 : }  // end namespace clang
     383                 : 
     384                 : #endif

Generated: 2010-02-10 01:31 by zcov