clang  8.0.0svn
Lexer.h
Go to the documentation of this file.
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_LEX_LEXER_H
15 #define LLVM_CLANG_LEX_LEXER_H
16 
19 #include "clang/Basic/TokenKinds.h"
21 #include "clang/Lex/Token.h"
22 #include "llvm/ADT/Optional.h"
23 #include "llvm/ADT/SmallVector.h"
24 #include "llvm/ADT/StringRef.h"
25 #include <cassert>
26 #include <cstdint>
27 #include <string>
28 
29 namespace llvm {
30 
31 class MemoryBuffer;
32 
33 } // namespace llvm
34 
35 namespace clang {
36 
37 class DiagnosticBuilder;
38 class Preprocessor;
39 class SourceManager;
40 
41 /// ConflictMarkerKind - Kinds of conflict marker which the lexer might be
42 /// recovering from.
44  /// Not within a conflict marker.
46 
47  /// A normal or diff3 conflict marker, initiated by at least 7 "<"s,
48  /// separated by at least 7 "="s or "|"s, and terminated by at least 7 ">"s.
50 
51  /// A Perforce-style conflict marker, initiated by 4 ">"s,
52  /// separated by 4 "="s, and terminated by 4 "<"s.
54 };
55 
56 /// Describes the bounds (start, size) of the preamble and a flag required by
57 /// PreprocessorOptions::PrecompiledPreambleBytes.
58 /// The preamble includes the BOM, if any.
60  /// Size of the preamble in bytes.
61  unsigned Size;
62 
63  /// Whether the preamble ends at the start of a new line.
64  ///
65  /// Used to inform the lexer as to whether it's starting at the beginning of
66  /// a line after skipping the preamble.
68 
69  PreambleBounds(unsigned Size, bool PreambleEndsAtStartOfLine)
70  : Size(Size), PreambleEndsAtStartOfLine(PreambleEndsAtStartOfLine) {}
71 };
72 
73 /// Lexer - This provides a simple interface that turns a text buffer into a
74 /// stream of tokens. This provides no support for file reading or buffering,
75 /// or buffering/seeking of tokens, only forward lexing is supported. It relies
76 /// on the specified Preprocessor object to handle preprocessor directives, etc.
77 class Lexer : public PreprocessorLexer {
78  friend class Preprocessor;
79 
80  void anchor() override;
81 
82  //===--------------------------------------------------------------------===//
83  // Constant configuration values for this lexer.
84 
85  // Start of the buffer.
86  const char *BufferStart;
87 
88  // End of the buffer.
89  const char *BufferEnd;
90 
91  // Location for start of file.
92  SourceLocation FileLoc;
93 
94  // LangOpts enabled by this language (cache).
95  LangOptions LangOpts;
96 
97  // True if lexer for _Pragma handling.
98  bool Is_PragmaLexer;
99 
100  //===--------------------------------------------------------------------===//
101  // Context-specific lexing flags set by the preprocessor.
102  //
103 
104  /// ExtendedTokenMode - The lexer can optionally keep comments and whitespace
105  /// and return them as tokens. This is used for -C and -CC modes, and
106  /// whitespace preservation can be useful for some clients that want to lex
107  /// the file in raw mode and get every character from the file.
108  ///
109  /// When this is set to 2 it returns comments and whitespace. When set to 1
110  /// it returns comments, when it is set to 0 it returns normal tokens only.
111  unsigned char ExtendedTokenMode;
112 
113  //===--------------------------------------------------------------------===//
114  // Context that changes as the file is lexed.
115  // NOTE: any state that mutates when in raw mode must have save/restore code
116  // in Lexer::isNextPPTokenLParen.
117 
118  // BufferPtr - Current pointer into the buffer. This is the next character
119  // to be lexed.
120  const char *BufferPtr;
121 
122  // IsAtStartOfLine - True if the next lexed token should get the "start of
123  // line" flag set on it.
124  bool IsAtStartOfLine;
125 
126  bool IsAtPhysicalStartOfLine;
127 
128  bool HasLeadingSpace;
129 
130  bool HasLeadingEmptyMacro;
131 
132  // CurrentConflictMarkerState - The kind of conflict marker we are handling.
133  ConflictMarkerKind CurrentConflictMarkerState;
134 
135  void InitLexer(const char *BufStart, const char *BufPtr, const char *BufEnd);
136 
137 public:
138  /// Lexer constructor - Create a new lexer object for the specified buffer
139  /// with the specified preprocessor managing the lexing process. This lexer
140  /// assumes that the associated file buffer and Preprocessor objects will
141  /// outlive it, so it doesn't take ownership of either of them.
142  Lexer(FileID FID, const llvm::MemoryBuffer *InputBuffer, Preprocessor &PP);
143 
144  /// Lexer constructor - Create a new raw lexer object. This object is only
145  /// suitable for calls to 'LexFromRawLexer'. This lexer assumes that the
146  /// text range will outlive it, so it doesn't take ownership of it.
147  Lexer(SourceLocation FileLoc, const LangOptions &LangOpts,
148  const char *BufStart, const char *BufPtr, const char *BufEnd);
149 
150  /// Lexer constructor - Create a new raw lexer object. This object is only
151  /// suitable for calls to 'LexFromRawLexer'. This lexer assumes that the
152  /// text range will outlive it, so it doesn't take ownership of it.
153  Lexer(FileID FID, const llvm::MemoryBuffer *InputBuffer,
154  const SourceManager &SM, const LangOptions &LangOpts);
155 
156  Lexer(const Lexer &) = delete;
157  Lexer &operator=(const Lexer &) = delete;
158 
159  /// Create_PragmaLexer: Lexer constructor - Create a new lexer object for
160  /// _Pragma expansion. This has a variety of magic semantics that this method
161  /// sets up. It returns a new'd Lexer that must be delete'd when done.
162  static Lexer *Create_PragmaLexer(SourceLocation SpellingLoc,
163  SourceLocation ExpansionLocStart,
164  SourceLocation ExpansionLocEnd,
165  unsigned TokLen, Preprocessor &PP);
166 
167  /// getLangOpts - Return the language features currently enabled.
168  /// NOTE: this lexer modifies features as a file is parsed!
169  const LangOptions &getLangOpts() const { return LangOpts; }
170 
171  /// getFileLoc - Return the File Location for the file we are lexing out of.
172  /// The physical location encodes the location where the characters come from,
173  /// the virtual location encodes where we should *claim* the characters came
174  /// from. Currently this is only used by _Pragma handling.
175  SourceLocation getFileLoc() const { return FileLoc; }
176 
177 private:
178  /// Lex - Return the next token in the file. If this is the end of file, it
179  /// return the tok::eof token. This implicitly involves the preprocessor.
180  bool Lex(Token &Result);
181 
182 public:
183  /// isPragmaLexer - Returns true if this Lexer is being used to lex a pragma.
184  bool isPragmaLexer() const { return Is_PragmaLexer; }
185 
186 private:
187  /// IndirectLex - An indirect call to 'Lex' that can be invoked via
188  /// the PreprocessorLexer interface.
189  void IndirectLex(Token &Result) override { Lex(Result); }
190 
191 public:
192  /// LexFromRawLexer - Lex a token from a designated raw lexer (one with no
193  /// associated preprocessor object. Return true if the 'next character to
194  /// read' pointer points at the end of the lexer buffer, false otherwise.
195  bool LexFromRawLexer(Token &Result) {
196  assert(LexingRawMode && "Not already in raw mode!");
197  Lex(Result);
198  // Note that lexing to the end of the buffer doesn't implicitly delete the
199  // lexer when in raw mode.
200  return BufferPtr == BufferEnd;
201  }
202 
203  /// isKeepWhitespaceMode - Return true if the lexer should return tokens for
204  /// every character in the file, including whitespace and comments. This
205  /// should only be used in raw mode, as the preprocessor is not prepared to
206  /// deal with the excess tokens.
207  bool isKeepWhitespaceMode() const {
208  return ExtendedTokenMode > 1;
209  }
210 
211  /// SetKeepWhitespaceMode - This method lets clients enable or disable
212  /// whitespace retention mode.
213  void SetKeepWhitespaceMode(bool Val) {
214  assert((!Val || LexingRawMode || LangOpts.TraditionalCPP) &&
215  "Can only retain whitespace in raw mode or -traditional-cpp");
216  ExtendedTokenMode = Val ? 2 : 0;
217  }
218 
219  /// inKeepCommentMode - Return true if the lexer should return comments as
220  /// tokens.
221  bool inKeepCommentMode() const {
222  return ExtendedTokenMode > 0;
223  }
224 
225  /// SetCommentRetentionMode - Change the comment retention mode of the lexer
226  /// to the specified mode. This is really only useful when lexing in raw
227  /// mode, because otherwise the lexer needs to manage this.
228  void SetCommentRetentionState(bool Mode) {
229  assert(!isKeepWhitespaceMode() &&
230  "Can't play with comment retention state when retaining whitespace");
231  ExtendedTokenMode = Mode ? 1 : 0;
232  }
233 
234  /// Sets the extended token mode back to its initial value, according to the
235  /// language options and preprocessor. This controls whether the lexer
236  /// produces comment and whitespace tokens.
237  ///
238  /// This requires the lexer to have an associated preprocessor. A standalone
239  /// lexer has nothing to reset to.
240  void resetExtendedTokenMode();
241 
242  /// Gets source code buffer.
243  StringRef getBuffer() const {
244  return StringRef(BufferStart, BufferEnd - BufferStart);
245  }
246 
247  /// ReadToEndOfLine - Read the rest of the current preprocessor line as an
248  /// uninterpreted string. This switches the lexer out of directive mode.
249  void ReadToEndOfLine(SmallVectorImpl<char> *Result = nullptr);
250 
251 
252  /// Diag - Forwarding function for diagnostics. This translate a source
253  /// position in the current buffer into a SourceLocation object for rendering.
254  DiagnosticBuilder Diag(const char *Loc, unsigned DiagID) const;
255 
256  /// getSourceLocation - Return a source location identifier for the specified
257  /// offset in the current file.
258  SourceLocation getSourceLocation(const char *Loc, unsigned TokLen = 1) const;
259 
260  /// getSourceLocation - Return a source location for the next character in
261  /// the current file.
263  return getSourceLocation(BufferPtr);
264  }
265 
266  /// Return the current location in the buffer.
267  const char *getBufferLocation() const { return BufferPtr; }
268 
269  /// Stringify - Convert the specified string into a C string by i) escaping
270  /// '\\' and " characters and ii) replacing newline character(s) with "\\n".
271  /// If Charify is true, this escapes the ' character instead of ".
272  static std::string Stringify(StringRef Str, bool Charify = false);
273 
274  /// Stringify - Convert the specified string into a C string by i) escaping
275  /// '\\' and " characters and ii) replacing newline character(s) with "\\n".
276  static void Stringify(SmallVectorImpl<char> &Str);
277 
278  /// getSpelling - This method is used to get the spelling of a token into a
279  /// preallocated buffer, instead of as an std::string. The caller is required
280  /// to allocate enough space for the token, which is guaranteed to be at least
281  /// Tok.getLength() bytes long. The length of the actual result is returned.
282  ///
283  /// Note that this method may do two possible things: it may either fill in
284  /// the buffer specified with characters, or it may *change the input pointer*
285  /// to point to a constant buffer with the data already in it (avoiding a
286  /// copy). The caller is not allowed to modify the returned buffer pointer
287  /// if an internal buffer is returned.
288  static unsigned getSpelling(const Token &Tok, const char *&Buffer,
289  const SourceManager &SourceMgr,
290  const LangOptions &LangOpts,
291  bool *Invalid = nullptr);
292 
293  /// getSpelling() - Return the 'spelling' of the Tok token. The spelling of a
294  /// token is the characters used to represent the token in the source file
295  /// after trigraph expansion and escaped-newline folding. In particular, this
296  /// wants to get the true, uncanonicalized, spelling of things like digraphs
297  /// UCNs, etc.
298  static std::string getSpelling(const Token &Tok,
299  const SourceManager &SourceMgr,
300  const LangOptions &LangOpts,
301  bool *Invalid = nullptr);
302 
303  /// getSpelling - This method is used to get the spelling of the
304  /// token at the given source location. If, as is usually true, it
305  /// is not necessary to copy any data, then the returned string may
306  /// not point into the provided buffer.
307  ///
308  /// This method lexes at the expansion depth of the given
309  /// location and does not jump to the expansion or spelling
310  /// location.
311  static StringRef getSpelling(SourceLocation loc,
312  SmallVectorImpl<char> &buffer,
313  const SourceManager &SourceMgr,
314  const LangOptions &LangOpts,
315  bool *invalid = nullptr);
316 
317  /// MeasureTokenLength - Relex the token at the specified location and return
318  /// its length in bytes in the input file. If the token needs cleaning (e.g.
319  /// includes a trigraph or an escaped newline) then this count includes bytes
320  /// that are part of that.
321  static unsigned MeasureTokenLength(SourceLocation Loc,
322  const SourceManager &SM,
323  const LangOptions &LangOpts);
324 
325  /// Relex the token at the specified location.
326  /// \returns true if there was a failure, false on success.
327  static bool getRawToken(SourceLocation Loc, Token &Result,
328  const SourceManager &SM,
329  const LangOptions &LangOpts,
330  bool IgnoreWhiteSpace = false);
331 
332  /// Given a location any where in a source buffer, find the location
333  /// that corresponds to the beginning of the token in which the original
334  /// source location lands.
335  static SourceLocation GetBeginningOfToken(SourceLocation Loc,
336  const SourceManager &SM,
337  const LangOptions &LangOpts);
338 
339  /// Get the physical length (including trigraphs and escaped newlines) of the
340  /// first \p Characters characters of the token starting at TokStart.
341  static unsigned getTokenPrefixLength(SourceLocation TokStart,
342  unsigned Characters,
343  const SourceManager &SM,
344  const LangOptions &LangOpts);
345 
346  /// AdvanceToTokenCharacter - If the current SourceLocation specifies a
347  /// location at the start of a token, return a new location that specifies a
348  /// character within the token. This handles trigraphs and escaped newlines.
350  unsigned Characters,
351  const SourceManager &SM,
352  const LangOptions &LangOpts) {
353  return TokStart.getLocWithOffset(
354  getTokenPrefixLength(TokStart, Characters, SM, LangOpts));
355  }
356 
357  /// Computes the source location just past the end of the
358  /// token at this source location.
359  ///
360  /// This routine can be used to produce a source location that
361  /// points just past the end of the token referenced by \p Loc, and
362  /// is generally used when a diagnostic needs to point just after a
363  /// token where it expected something different that it received. If
364  /// the returned source location would not be meaningful (e.g., if
365  /// it points into a macro), this routine returns an invalid
366  /// source location.
367  ///
368  /// \param Offset an offset from the end of the token, where the source
369  /// location should refer to. The default offset (0) produces a source
370  /// location pointing just past the end of the token; an offset of 1 produces
371  /// a source location pointing to the last character in the token, etc.
372  static SourceLocation getLocForEndOfToken(SourceLocation Loc, unsigned Offset,
373  const SourceManager &SM,
374  const LangOptions &LangOpts);
375 
376  /// Given a token range, produce a corresponding CharSourceRange that
377  /// is not a token range. This allows the source range to be used by
378  /// components that don't have access to the lexer and thus can't find the
379  /// end of the range for themselves.
381  const SourceManager &SM,
382  const LangOptions &LangOpts) {
383  SourceLocation End = getLocForEndOfToken(Range.getEnd(), 0, SM, LangOpts);
384  return End.isInvalid() ? CharSourceRange()
385  : CharSourceRange::getCharRange(
386  Range.getBegin(), End.getLocWithOffset(-1));
387  }
389  const SourceManager &SM,
390  const LangOptions &LangOpts) {
391  return Range.isTokenRange()
392  ? getAsCharRange(Range.getAsRange(), SM, LangOpts)
393  : Range;
394  }
395 
396  /// Returns true if the given MacroID location points at the first
397  /// token of the macro expansion.
398  ///
399  /// \param MacroBegin If non-null and function returns true, it is set to
400  /// begin location of the macro.
401  static bool isAtStartOfMacroExpansion(SourceLocation loc,
402  const SourceManager &SM,
403  const LangOptions &LangOpts,
404  SourceLocation *MacroBegin = nullptr);
405 
406  /// Returns true if the given MacroID location points at the last
407  /// token of the macro expansion.
408  ///
409  /// \param MacroEnd If non-null and function returns true, it is set to
410  /// end location of the macro.
411  static bool isAtEndOfMacroExpansion(SourceLocation loc,
412  const SourceManager &SM,
413  const LangOptions &LangOpts,
414  SourceLocation *MacroEnd = nullptr);
415 
416  /// Accepts a range and returns a character range with file locations.
417  ///
418  /// Returns a null range if a part of the range resides inside a macro
419  /// expansion or the range does not reside on the same FileID.
420  ///
421  /// This function is trying to deal with macros and return a range based on
422  /// file locations. The cases where it can successfully handle macros are:
423  ///
424  /// -begin or end range lies at the start or end of a macro expansion, in
425  /// which case the location will be set to the expansion point, e.g:
426  /// \#define M 1 2
427  /// a M
428  /// If you have a range [a, 2] (where 2 came from the macro), the function
429  /// will return a range for "a M"
430  /// if you have range [a, 1], the function will fail because the range
431  /// overlaps with only a part of the macro
432  ///
433  /// -The macro is a function macro and the range can be mapped to the macro
434  /// arguments, e.g:
435  /// \#define M 1 2
436  /// \#define FM(x) x
437  /// FM(a b M)
438  /// if you have range [b, 2], the function will return the file range "b M"
439  /// inside the macro arguments.
440  /// if you have range [a, 2], the function will return the file range
441  /// "FM(a b M)" since the range includes all of the macro expansion.
442  static CharSourceRange makeFileCharRange(CharSourceRange Range,
443  const SourceManager &SM,
444  const LangOptions &LangOpts);
445 
446  /// Returns a string for the source that the range encompasses.
447  static StringRef getSourceText(CharSourceRange Range,
448  const SourceManager &SM,
449  const LangOptions &LangOpts,
450  bool *Invalid = nullptr);
451 
452  /// Retrieve the name of the immediate macro expansion.
453  ///
454  /// This routine starts from a source location, and finds the name of the macro
455  /// responsible for its immediate expansion. It looks through any intervening
456  /// macro argument expansions to compute this. It returns a StringRef which
457  /// refers to the SourceManager-owned buffer of the source where that macro
458  /// name is spelled. Thus, the result shouldn't out-live that SourceManager.
459  static StringRef getImmediateMacroName(SourceLocation Loc,
460  const SourceManager &SM,
461  const LangOptions &LangOpts);
462 
463  /// Retrieve the name of the immediate macro expansion.
464  ///
465  /// This routine starts from a source location, and finds the name of the
466  /// macro responsible for its immediate expansion. It looks through any
467  /// intervening macro argument expansions to compute this. It returns a
468  /// StringRef which refers to the SourceManager-owned buffer of the source
469  /// where that macro name is spelled. Thus, the result shouldn't out-live
470  /// that SourceManager.
471  ///
472  /// This differs from Lexer::getImmediateMacroName in that any macro argument
473  /// location will result in the topmost function macro that accepted it.
474  /// e.g.
475  /// \code
476  /// MAC1( MAC2(foo) )
477  /// \endcode
478  /// for location of 'foo' token, this function will return "MAC1" while
479  /// Lexer::getImmediateMacroName will return "MAC2".
480  static StringRef getImmediateMacroNameForDiagnostics(
481  SourceLocation Loc, const SourceManager &SM, const LangOptions &LangOpts);
482 
483  /// Compute the preamble of the given file.
484  ///
485  /// The preamble of a file contains the initial comments, include directives,
486  /// and other preprocessor directives that occur before the code in this
487  /// particular file actually begins. The preamble of the main source file is
488  /// a potential prefix header.
489  ///
490  /// \param Buffer The memory buffer containing the file's contents.
491  ///
492  /// \param MaxLines If non-zero, restrict the length of the preamble
493  /// to fewer than this number of lines.
494  ///
495  /// \returns The offset into the file where the preamble ends and the rest
496  /// of the file begins along with a boolean value indicating whether
497  /// the preamble ends at the beginning of a new line.
498  static PreambleBounds ComputePreamble(StringRef Buffer,
499  const LangOptions &LangOpts,
500  unsigned MaxLines = 0);
501 
502  /// Finds the token that comes right after the given location.
503  ///
504  /// Returns the next token, or none if the location is inside a macro.
505  static Optional<Token> findNextToken(SourceLocation Loc,
506  const SourceManager &SM,
507  const LangOptions &LangOpts);
508 
509  /// Checks that the given token is the first token that occurs after
510  /// the given location (this excludes comments and whitespace). Returns the
511  /// location immediately after the specified token. If the token is not found
512  /// or the location is inside a macro, the returned source location will be
513  /// invalid.
514  static SourceLocation findLocationAfterToken(SourceLocation loc,
515  tok::TokenKind TKind,
516  const SourceManager &SM,
517  const LangOptions &LangOpts,
518  bool SkipTrailingWhitespaceAndNewLine);
519 
520  /// Returns true if the given character could appear in an identifier.
521  static bool isIdentifierBodyChar(char c, const LangOptions &LangOpts);
522 
523  /// Checks whether new line pointed by Str is preceded by escape
524  /// sequence.
525  static bool isNewLineEscaped(const char *BufferStart, const char *Str);
526 
527  /// getCharAndSizeNoWarn - Like the getCharAndSize method, but does not ever
528  /// emit a warning.
529  static inline char getCharAndSizeNoWarn(const char *Ptr, unsigned &Size,
530  const LangOptions &LangOpts) {
531  // If this is not a trigraph and not a UCN or escaped newline, return
532  // quickly.
533  if (isObviouslySimpleCharacter(Ptr[0])) {
534  Size = 1;
535  return *Ptr;
536  }
537 
538  Size = 0;
539  return getCharAndSizeSlowNoWarn(Ptr, Size, LangOpts);
540  }
541 
542  /// Returns the leading whitespace for line that corresponds to the given
543  /// location \p Loc.
544  static StringRef getIndentationForLine(SourceLocation Loc,
545  const SourceManager &SM);
546 
547 private:
548  //===--------------------------------------------------------------------===//
549  // Internal implementation interfaces.
550 
551  /// LexTokenInternal - Internal interface to lex a preprocessing token. Called
552  /// by Lex.
553  ///
554  bool LexTokenInternal(Token &Result, bool TokAtPhysicalStartOfLine);
555 
556  bool CheckUnicodeWhitespace(Token &Result, uint32_t C, const char *CurPtr);
557 
558  /// Given that a token begins with the Unicode character \p C, figure out
559  /// what kind of token it is and dispatch to the appropriate lexing helper
560  /// function.
561  bool LexUnicode(Token &Result, uint32_t C, const char *CurPtr);
562 
563  /// FormTokenWithChars - When we lex a token, we have identified a span
564  /// starting at BufferPtr, going to TokEnd that forms the token. This method
565  /// takes that range and assigns it to the token as its location and size. In
566  /// addition, since tokens cannot overlap, this also updates BufferPtr to be
567  /// TokEnd.
568  void FormTokenWithChars(Token &Result, const char *TokEnd,
570  unsigned TokLen = TokEnd-BufferPtr;
571  Result.setLength(TokLen);
572  Result.setLocation(getSourceLocation(BufferPtr, TokLen));
573  Result.setKind(Kind);
574  BufferPtr = TokEnd;
575  }
576 
577  /// isNextPPTokenLParen - Return 1 if the next unexpanded token will return a
578  /// tok::l_paren token, 0 if it is something else and 2 if there are no more
579  /// tokens in the buffer controlled by this lexer.
580  unsigned isNextPPTokenLParen();
581 
582  //===--------------------------------------------------------------------===//
583  // Lexer character reading interfaces.
584 
585  // This lexer is built on two interfaces for reading characters, both of which
586  // automatically provide phase 1/2 translation. getAndAdvanceChar is used
587  // when we know that we will be reading a character from the input buffer and
588  // that this character will be part of the result token. This occurs in (f.e.)
589  // string processing, because we know we need to read until we find the
590  // closing '"' character.
591  //
592  // The second interface is the combination of getCharAndSize with
593  // ConsumeChar. getCharAndSize reads a phase 1/2 translated character,
594  // returning it and its size. If the lexer decides that this character is
595  // part of the current token, it calls ConsumeChar on it. This two stage
596  // approach allows us to emit diagnostics for characters (e.g. warnings about
597  // trigraphs), knowing that they only are emitted if the character is
598  // consumed.
599 
600  /// isObviouslySimpleCharacter - Return true if the specified character is
601  /// obviously the same in translation phase 1 and translation phase 3. This
602  /// can return false for characters that end up being the same, but it will
603  /// never return true for something that needs to be mapped.
604  static bool isObviouslySimpleCharacter(char C) {
605  return C != '?' && C != '\\';
606  }
607 
608  /// getAndAdvanceChar - Read a single 'character' from the specified buffer,
609  /// advance over it, and return it. This is tricky in several cases. Here we
610  /// just handle the trivial case and fall-back to the non-inlined
611  /// getCharAndSizeSlow method to handle the hard case.
612  inline char getAndAdvanceChar(const char *&Ptr, Token &Tok) {
613  // If this is not a trigraph and not a UCN or escaped newline, return
614  // quickly.
615  if (isObviouslySimpleCharacter(Ptr[0])) return *Ptr++;
616 
617  unsigned Size = 0;
618  char C = getCharAndSizeSlow(Ptr, Size, &Tok);
619  Ptr += Size;
620  return C;
621  }
622 
623  /// ConsumeChar - When a character (identified by getCharAndSize) is consumed
624  /// and added to a given token, check to see if there are diagnostics that
625  /// need to be emitted or flags that need to be set on the token. If so, do
626  /// it.
627  const char *ConsumeChar(const char *Ptr, unsigned Size, Token &Tok) {
628  // Normal case, we consumed exactly one token. Just return it.
629  if (Size == 1)
630  return Ptr+Size;
631 
632  // Otherwise, re-lex the character with a current token, allowing
633  // diagnostics to be emitted and flags to be set.
634  Size = 0;
635  getCharAndSizeSlow(Ptr, Size, &Tok);
636  return Ptr+Size;
637  }
638 
639  /// getCharAndSize - Peek a single 'character' from the specified buffer,
640  /// get its size, and return it. This is tricky in several cases. Here we
641  /// just handle the trivial case and fall-back to the non-inlined
642  /// getCharAndSizeSlow method to handle the hard case.
643  inline char getCharAndSize(const char *Ptr, unsigned &Size) {
644  // If this is not a trigraph and not a UCN or escaped newline, return
645  // quickly.
646  if (isObviouslySimpleCharacter(Ptr[0])) {
647  Size = 1;
648  return *Ptr;
649  }
650 
651  Size = 0;
652  return getCharAndSizeSlow(Ptr, Size);
653  }
654 
655  /// getCharAndSizeSlow - Handle the slow/uncommon case of the getCharAndSize
656  /// method.
657  char getCharAndSizeSlow(const char *Ptr, unsigned &Size,
658  Token *Tok = nullptr);
659 
660  /// getEscapedNewLineSize - Return the size of the specified escaped newline,
661  /// or 0 if it is not an escaped newline. P[-1] is known to be a "\" on entry
662  /// to this function.
663  static unsigned getEscapedNewLineSize(const char *P);
664 
665  /// SkipEscapedNewLines - If P points to an escaped newline (or a series of
666  /// them), skip over them and return the first non-escaped-newline found,
667  /// otherwise return P.
668  static const char *SkipEscapedNewLines(const char *P);
669 
670  /// getCharAndSizeSlowNoWarn - Same as getCharAndSizeSlow, but never emits a
671  /// diagnostic.
672  static char getCharAndSizeSlowNoWarn(const char *Ptr, unsigned &Size,
673  const LangOptions &LangOpts);
674 
675  //===--------------------------------------------------------------------===//
676  // Other lexer functions.
677 
678  void SetByteOffset(unsigned Offset, bool StartOfLine);
679 
680  void PropagateLineStartLeadingSpaceInfo(Token &Result);
681 
682  const char *LexUDSuffix(Token &Result, const char *CurPtr,
683  bool IsStringLiteral);
684 
685  // Helper functions to lex the remainder of a token of the specific type.
686  bool LexIdentifier (Token &Result, const char *CurPtr);
687  bool LexNumericConstant (Token &Result, const char *CurPtr);
688  bool LexStringLiteral (Token &Result, const char *CurPtr,
689  tok::TokenKind Kind);
690  bool LexRawStringLiteral (Token &Result, const char *CurPtr,
691  tok::TokenKind Kind);
692  bool LexAngledStringLiteral(Token &Result, const char *CurPtr);
693  bool LexCharConstant (Token &Result, const char *CurPtr,
694  tok::TokenKind Kind);
695  bool LexEndOfFile (Token &Result, const char *CurPtr);
696  bool SkipWhitespace (Token &Result, const char *CurPtr,
697  bool &TokAtPhysicalStartOfLine);
698  bool SkipLineComment (Token &Result, const char *CurPtr,
699  bool &TokAtPhysicalStartOfLine);
700  bool SkipBlockComment (Token &Result, const char *CurPtr,
701  bool &TokAtPhysicalStartOfLine);
702  bool SaveLineComment (Token &Result, const char *CurPtr);
703 
704  bool IsStartOfConflictMarker(const char *CurPtr);
705  bool HandleEndOfConflictMarker(const char *CurPtr);
706 
707  bool lexEditorPlaceholder(Token &Result, const char *CurPtr);
708 
709  bool isCodeCompletionPoint(const char *CurPtr) const;
710  void cutOffLexing() { BufferPtr = BufferEnd; }
711 
712  bool isHexaLiteral(const char *Start, const LangOptions &LangOpts);
713 
714 
715  /// Read a universal character name.
716  ///
717  /// \param CurPtr The position in the source buffer after the initial '\'.
718  /// If the UCN is syntactically well-formed (but not necessarily
719  /// valid), this parameter will be updated to point to the
720  /// character after the UCN.
721  /// \param SlashLoc The position in the source buffer of the '\'.
722  /// \param Tok The token being formed. Pass \c nullptr to suppress diagnostics
723  /// and handle token formation in the caller.
724  ///
725  /// \return The Unicode codepoint specified by the UCN, or 0 if the UCN is
726  /// invalid.
727  uint32_t tryReadUCN(const char *&CurPtr, const char *SlashLoc, Token *Tok);
728 
729  /// Try to consume a UCN as part of an identifier at the current
730  /// location.
731  /// \param CurPtr Initially points to the range of characters in the source
732  /// buffer containing the '\'. Updated to point past the end of
733  /// the UCN on success.
734  /// \param Size The number of characters occupied by the '\' (including
735  /// trigraphs and escaped newlines).
736  /// \param Result The token being produced. Marked as containing a UCN on
737  /// success.
738  /// \return \c true if a UCN was lexed and it produced an acceptable
739  /// identifier character, \c false otherwise.
740  bool tryConsumeIdentifierUCN(const char *&CurPtr, unsigned Size,
741  Token &Result);
742 
743  /// Try to consume an identifier character encoded in UTF-8.
744  /// \param CurPtr Points to the start of the (potential) UTF-8 code unit
745  /// sequence. On success, updated to point past the end of it.
746  /// \return \c true if a UTF-8 sequence mapping to an acceptable identifier
747  /// character was lexed, \c false otherwise.
748  bool tryConsumeIdentifierUTF8Char(const char *&CurPtr);
749 };
750 
751 } // namespace clang
752 
753 #endif // LLVM_CLANG_LEX_LEXER_H
Describes the bounds (start, size) of the preamble and a flag required by PreprocessorOptions::Precom...
Definition: Lexer.h:59
static DiagnosticBuilder Diag(DiagnosticsEngine *Diags, const LangOptions &Features, FullSourceLoc TokLoc, const char *TokBegin, const char *TokRangeBegin, const char *TokRangeEnd, unsigned DiagID)
Produce a diagnostic highlighting some portion of a literal.
Lexer - This provides a simple interface that turns a text buffer into a stream of tokens...
Definition: Lexer.h:77
SourceLocation getLocWithOffset(int Offset) const
Return a source location with the specified offset from this SourceLocation.
const char * getBufferLocation() const
Return the current location in the buffer.
Definition: Lexer.h:267
DominatorTree GraphTraits specialization so the DominatorTree can be iterable by generic graph iterat...
Definition: Dominators.h:30
bool LexFromRawLexer(Token &Result)
LexFromRawLexer - Lex a token from a designated raw lexer (one with no associated preprocessor object...
Definition: Lexer.h:195
StringRef P
static char getCharAndSizeNoWarn(const char *Ptr, unsigned &Size, const LangOptions &LangOpts)
getCharAndSizeNoWarn - Like the getCharAndSize method, but does not ever emit a warning.
Definition: Lexer.h:529
ConflictMarkerKind
ConflictMarkerKind - Kinds of conflict marker which the lexer might be recovering from...
Definition: Lexer.h:43
static CharSourceRange getAsCharRange(CharSourceRange Range, const SourceManager &SM, const LangOptions &LangOpts)
Definition: Lexer.h:388
Token - This structure provides full information about a lexed token.
Definition: Token.h:35
void setKind(tok::TokenKind K)
Definition: Token.h:91
SourceLocation getFileLoc() const
getFileLoc - Return the File Location for the file we are lexing out of.
Definition: Lexer.h:175
Keeps track of the various options that can be enabled, which controls the dialect of C or C++ that i...
Definition: LangOptions.h:50
bool isPragmaLexer() const
isPragmaLexer - Returns true if this Lexer is being used to lex a pragma.
Definition: Lexer.h:184
A Perforce-style conflict marker, initiated by 4 ">"s, separated by 4 "="s, and terminated by 4 "<"s...
Definition: Lexer.h:53
SourceLocation getSourceLocation() override
getSourceLocation - Return a source location for the next character in the current file...
Definition: Lexer.h:262
uint32_t Offset
Definition: CacheTokens.cpp:43
const FormatToken & Tok
StringRef getBuffer() const
Gets source code buffer.
Definition: Lexer.h:243
static SourceLocation AdvanceToTokenCharacter(SourceLocation TokStart, unsigned Characters, const SourceManager &SM, const LangOptions &LangOpts)
AdvanceToTokenCharacter - If the current SourceLocation specifies a location at the start of a token...
Definition: Lexer.h:349
bool PreambleEndsAtStartOfLine
Whether the preamble ends at the start of a new line.
Definition: Lexer.h:67
A little helper class used to produce diagnostics.
Definition: Diagnostic.h:1042
const LangOptions & getLangOpts() const
getLangOpts - Return the language features currently enabled.
Definition: Lexer.h:169
Defines the clang::LangOptions interface.
SourceLocation End
Represents a character-granular source range.
bool isKeepWhitespaceMode() const
isKeepWhitespaceMode - Return true if the lexer should return tokens for every character in the file...
Definition: Lexer.h:207
SourceLocation getEnd() const
const SourceManager & SM
Definition: Format.cpp:1475
Kind
Encodes a location in the source.
void setLength(unsigned Len)
Definition: Token.h:133
PreambleBounds(unsigned Size, bool PreambleEndsAtStartOfLine)
Definition: Lexer.h:69
TokenKind
Provides a simple uniform namespace for tokens from all C languages.
Definition: TokenKinds.h:25
bool inKeepCommentMode() const
inKeepCommentMode - Return true if the lexer should return comments as tokens.
Definition: Lexer.h:221
bool isTokenRange() const
Return true if the end of this range specifies the start of the last token.
An opaque identifier used by SourceManager which refers to a source file (MemoryBuffer) along with it...
SourceRange getAsRange() const
Dataflow Directional Tag Classes.
static CharSourceRange getAsCharRange(SourceRange Range, const SourceManager &SM, const LangOptions &LangOpts)
Given a token range, produce a corresponding CharSourceRange that is not a token range.
Definition: Lexer.h:380
unsigned Size
Size of the preamble in bytes.
Definition: Lexer.h:61
Defines the clang::TokenKind enum and support functions.
Defines the clang::SourceLocation class and associated facilities.
Not within a conflict marker.
Definition: Lexer.h:45
void SetCommentRetentionState(bool Mode)
SetCommentRetentionMode - Change the comment retention mode of the lexer to the specified mode...
Definition: Lexer.h:228
void setLocation(SourceLocation L)
Definition: Token.h:132
A normal or diff3 conflict marker, initiated by at least 7 "<"s, separated by at least 7 "="s or "|"s...
Definition: Lexer.h:49
A trivial tuple used to represent a source range.
Defines the PreprocessorLexer interface.
void SetKeepWhitespaceMode(bool Val)
SetKeepWhitespaceMode - This method lets clients enable or disable whitespace retention mode...
Definition: Lexer.h:213
SourceLocation getBegin() const
This class handles loading and caching of source files into memory.
Engages in a tight little dance with the lexer to efficiently preprocess tokens.
Definition: Preprocessor.h:127