clang  10.0.0svn
Lexer.h
Go to the documentation of this file.
1 //===- Lexer.h - C Language Family Lexer ------------------------*- C++ -*-===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // This file defines the Lexer interface.
10 //
11 //===----------------------------------------------------------------------===//
12 
13 #ifndef LLVM_CLANG_LEX_LEXER_H
14 #define LLVM_CLANG_LEX_LEXER_H
15 
18 #include "clang/Basic/TokenKinds.h"
20 #include "clang/Lex/Token.h"
21 #include "llvm/ADT/Optional.h"
22 #include "llvm/ADT/SmallVector.h"
23 #include "llvm/ADT/StringRef.h"
24 #include <cassert>
25 #include <cstdint>
26 #include <string>
27 
28 namespace llvm {
29 
30 class MemoryBuffer;
31 
32 } // namespace llvm
33 
34 namespace clang {
35 
36 class DiagnosticBuilder;
37 class Preprocessor;
38 class SourceManager;
39 
40 /// ConflictMarkerKind - Kinds of conflict marker which the lexer might be
41 /// recovering from.
43  /// Not within a conflict marker.
45 
46  /// A normal or diff3 conflict marker, initiated by at least 7 "<"s,
47  /// separated by at least 7 "="s or "|"s, and terminated by at least 7 ">"s.
49 
50  /// A Perforce-style conflict marker, initiated by 4 ">"s,
51  /// separated by 4 "="s, and terminated by 4 "<"s.
53 };
54 
55 /// Describes the bounds (start, size) of the preamble and a flag required by
56 /// PreprocessorOptions::PrecompiledPreambleBytes.
57 /// The preamble includes the BOM, if any.
59  /// Size of the preamble in bytes.
60  unsigned Size;
61 
62  /// Whether the preamble ends at the start of a new line.
63  ///
64  /// Used to inform the lexer as to whether it's starting at the beginning of
65  /// a line after skipping the preamble.
67 
68  PreambleBounds(unsigned Size, bool PreambleEndsAtStartOfLine)
69  : Size(Size), PreambleEndsAtStartOfLine(PreambleEndsAtStartOfLine) {}
70 };
71 
72 /// Lexer - This provides a simple interface that turns a text buffer into a
73 /// stream of tokens. This provides no support for file reading or buffering,
74 /// or buffering/seeking of tokens, only forward lexing is supported. It relies
75 /// on the specified Preprocessor object to handle preprocessor directives, etc.
76 class Lexer : public PreprocessorLexer {
77  friend class Preprocessor;
78 
79  void anchor() override;
80 
81  //===--------------------------------------------------------------------===//
82  // Constant configuration values for this lexer.
83 
84  // Start of the buffer.
85  const char *BufferStart;
86 
87  // End of the buffer.
88  const char *BufferEnd;
89 
90  // Location for start of file.
91  SourceLocation FileLoc;
92 
93  // LangOpts enabled by this language (cache).
94  LangOptions LangOpts;
95 
96  // True if lexer for _Pragma handling.
97  bool Is_PragmaLexer;
98 
99  //===--------------------------------------------------------------------===//
100  // Context-specific lexing flags set by the preprocessor.
101  //
102 
103  /// ExtendedTokenMode - The lexer can optionally keep comments and whitespace
104  /// and return them as tokens. This is used for -C and -CC modes, and
105  /// whitespace preservation can be useful for some clients that want to lex
106  /// the file in raw mode and get every character from the file.
107  ///
108  /// When this is set to 2 it returns comments and whitespace. When set to 1
109  /// it returns comments, when it is set to 0 it returns normal tokens only.
110  unsigned char ExtendedTokenMode;
111 
112  //===--------------------------------------------------------------------===//
113  // Context that changes as the file is lexed.
114  // NOTE: any state that mutates when in raw mode must have save/restore code
115  // in Lexer::isNextPPTokenLParen.
116 
117  // BufferPtr - Current pointer into the buffer. This is the next character
118  // to be lexed.
119  const char *BufferPtr;
120 
121  // IsAtStartOfLine - True if the next lexed token should get the "start of
122  // line" flag set on it.
123  bool IsAtStartOfLine;
124 
125  bool IsAtPhysicalStartOfLine;
126 
127  bool HasLeadingSpace;
128 
129  bool HasLeadingEmptyMacro;
130 
131  // CurrentConflictMarkerState - The kind of conflict marker we are handling.
132  ConflictMarkerKind CurrentConflictMarkerState;
133 
134  void InitLexer(const char *BufStart, const char *BufPtr, const char *BufEnd);
135 
136 public:
137  /// Lexer constructor - Create a new lexer object for the specified buffer
138  /// with the specified preprocessor managing the lexing process. This lexer
139  /// assumes that the associated file buffer and Preprocessor objects will
140  /// outlive it, so it doesn't take ownership of either of them.
141  Lexer(FileID FID, const llvm::MemoryBuffer *InputFile, Preprocessor &PP);
142 
143  /// Lexer constructor - Create a new raw lexer object. This object is only
144  /// suitable for calls to 'LexFromRawLexer'. This lexer assumes that the
145  /// text range will outlive it, so it doesn't take ownership of it.
146  Lexer(SourceLocation FileLoc, const LangOptions &LangOpts,
147  const char *BufStart, const char *BufPtr, const char *BufEnd);
148 
149  /// Lexer constructor - Create a new raw lexer object. This object is only
150  /// suitable for calls to 'LexFromRawLexer'. This lexer assumes that the
151  /// text range will outlive it, so it doesn't take ownership of it.
152  Lexer(FileID FID, const llvm::MemoryBuffer *FromFile,
153  const SourceManager &SM, const LangOptions &LangOpts);
154 
155  Lexer(const Lexer &) = delete;
156  Lexer &operator=(const Lexer &) = delete;
157 
158  /// Create_PragmaLexer: Lexer constructor - Create a new lexer object for
159  /// _Pragma expansion. This has a variety of magic semantics that this method
160  /// sets up. It returns a new'd Lexer that must be delete'd when done.
161  static Lexer *Create_PragmaLexer(SourceLocation SpellingLoc,
162  SourceLocation ExpansionLocStart,
163  SourceLocation ExpansionLocEnd,
164  unsigned TokLen, Preprocessor &PP);
165 
166  /// getLangOpts - Return the language features currently enabled.
167  /// NOTE: this lexer modifies features as a file is parsed!
168  const LangOptions &getLangOpts() const { return LangOpts; }
169 
170  /// getFileLoc - Return the File Location for the file we are lexing out of.
171  /// The physical location encodes the location where the characters come from,
172  /// the virtual location encodes where we should *claim* the characters came
173  /// from. Currently this is only used by _Pragma handling.
174  SourceLocation getFileLoc() const { return FileLoc; }
175 
176 private:
177  /// Lex - Return the next token in the file. If this is the end of file, it
178  /// return the tok::eof token. This implicitly involves the preprocessor.
179  bool Lex(Token &Result);
180 
181 public:
182  /// isPragmaLexer - Returns true if this Lexer is being used to lex a pragma.
183  bool isPragmaLexer() const { return Is_PragmaLexer; }
184 
185 private:
186  /// IndirectLex - An indirect call to 'Lex' that can be invoked via
187  /// the PreprocessorLexer interface.
188  void IndirectLex(Token &Result) override { Lex(Result); }
189 
190 public:
191  /// LexFromRawLexer - Lex a token from a designated raw lexer (one with no
192  /// associated preprocessor object. Return true if the 'next character to
193  /// read' pointer points at the end of the lexer buffer, false otherwise.
194  bool LexFromRawLexer(Token &Result) {
195  assert(LexingRawMode && "Not already in raw mode!");
196  Lex(Result);
197  // Note that lexing to the end of the buffer doesn't implicitly delete the
198  // lexer when in raw mode.
199  return BufferPtr == BufferEnd;
200  }
201 
202  /// isKeepWhitespaceMode - Return true if the lexer should return tokens for
203  /// every character in the file, including whitespace and comments. This
204  /// should only be used in raw mode, as the preprocessor is not prepared to
205  /// deal with the excess tokens.
206  bool isKeepWhitespaceMode() const {
207  return ExtendedTokenMode > 1;
208  }
209 
210  /// SetKeepWhitespaceMode - This method lets clients enable or disable
211  /// whitespace retention mode.
212  void SetKeepWhitespaceMode(bool Val) {
213  assert((!Val || LexingRawMode || LangOpts.TraditionalCPP) &&
214  "Can only retain whitespace in raw mode or -traditional-cpp");
215  ExtendedTokenMode = Val ? 2 : 0;
216  }
217 
218  /// inKeepCommentMode - Return true if the lexer should return comments as
219  /// tokens.
220  bool inKeepCommentMode() const {
221  return ExtendedTokenMode > 0;
222  }
223 
224  /// SetCommentRetentionMode - Change the comment retention mode of the lexer
225  /// to the specified mode. This is really only useful when lexing in raw
226  /// mode, because otherwise the lexer needs to manage this.
227  void SetCommentRetentionState(bool Mode) {
228  assert(!isKeepWhitespaceMode() &&
229  "Can't play with comment retention state when retaining whitespace");
230  ExtendedTokenMode = Mode ? 1 : 0;
231  }
232 
233  /// Sets the extended token mode back to its initial value, according to the
234  /// language options and preprocessor. This controls whether the lexer
235  /// produces comment and whitespace tokens.
236  ///
237  /// This requires the lexer to have an associated preprocessor. A standalone
238  /// lexer has nothing to reset to.
239  void resetExtendedTokenMode();
240 
241  /// Gets source code buffer.
242  StringRef getBuffer() const {
243  return StringRef(BufferStart, BufferEnd - BufferStart);
244  }
245 
246  /// ReadToEndOfLine - Read the rest of the current preprocessor line as an
247  /// uninterpreted string. This switches the lexer out of directive mode.
248  void ReadToEndOfLine(SmallVectorImpl<char> *Result = nullptr);
249 
250 
251  /// Diag - Forwarding function for diagnostics. This translate a source
252  /// position in the current buffer into a SourceLocation object for rendering.
253  DiagnosticBuilder Diag(const char *Loc, unsigned DiagID) const;
254 
255  /// getSourceLocation - Return a source location identifier for the specified
256  /// offset in the current file.
257  SourceLocation getSourceLocation(const char *Loc, unsigned TokLen = 1) const;
258 
259  /// getSourceLocation - Return a source location for the next character in
260  /// the current file.
262  return getSourceLocation(BufferPtr);
263  }
264 
265  /// Return the current location in the buffer.
266  const char *getBufferLocation() const { return BufferPtr; }
267 
268  /// Returns the current lexing offset.
270  assert(BufferPtr >= BufferStart && "Invalid buffer state");
271  return BufferPtr - BufferStart;
272  }
273 
274  /// Skip over \p NumBytes bytes.
275  ///
276  /// If the skip is successful, the next token will be lexed from the new
277  /// offset. The lexer also assumes that we skipped to the start of the line.
278  ///
279  /// \returns true if the skip failed (new offset would have been past the
280  /// end of the buffer), false otherwise.
281  bool skipOver(unsigned NumBytes);
282 
283  /// Stringify - Convert the specified string into a C string by i) escaping
284  /// '\\' and " characters and ii) replacing newline character(s) with "\\n".
285  /// If Charify is true, this escapes the ' character instead of ".
286  static std::string Stringify(StringRef Str, bool Charify = false);
287 
288  /// Stringify - Convert the specified string into a C string by i) escaping
289  /// '\\' and " characters and ii) replacing newline character(s) with "\\n".
290  static void Stringify(SmallVectorImpl<char> &Str);
291 
292  /// getSpelling - This method is used to get the spelling of a token into a
293  /// preallocated buffer, instead of as an std::string. The caller is required
294  /// to allocate enough space for the token, which is guaranteed to be at least
295  /// Tok.getLength() bytes long. The length of the actual result is returned.
296  ///
297  /// Note that this method may do two possible things: it may either fill in
298  /// the buffer specified with characters, or it may *change the input pointer*
299  /// to point to a constant buffer with the data already in it (avoiding a
300  /// copy). The caller is not allowed to modify the returned buffer pointer
301  /// if an internal buffer is returned.
302  static unsigned getSpelling(const Token &Tok, const char *&Buffer,
303  const SourceManager &SourceMgr,
304  const LangOptions &LangOpts,
305  bool *Invalid = nullptr);
306 
307  /// getSpelling() - Return the 'spelling' of the Tok token. The spelling of a
308  /// token is the characters used to represent the token in the source file
309  /// after trigraph expansion and escaped-newline folding. In particular, this
310  /// wants to get the true, uncanonicalized, spelling of things like digraphs
311  /// UCNs, etc.
312  static std::string getSpelling(const Token &Tok,
313  const SourceManager &SourceMgr,
314  const LangOptions &LangOpts,
315  bool *Invalid = nullptr);
316 
317  /// getSpelling - This method is used to get the spelling of the
318  /// token at the given source location. If, as is usually true, it
319  /// is not necessary to copy any data, then the returned string may
320  /// not point into the provided buffer.
321  ///
322  /// This method lexes at the expansion depth of the given
323  /// location and does not jump to the expansion or spelling
324  /// location.
325  static StringRef getSpelling(SourceLocation loc,
326  SmallVectorImpl<char> &buffer,
327  const SourceManager &SM,
328  const LangOptions &options,
329  bool *invalid = nullptr);
330 
331  /// MeasureTokenLength - Relex the token at the specified location and return
332  /// its length in bytes in the input file. If the token needs cleaning (e.g.
333  /// includes a trigraph or an escaped newline) then this count includes bytes
334  /// that are part of that.
335  static unsigned MeasureTokenLength(SourceLocation Loc,
336  const SourceManager &SM,
337  const LangOptions &LangOpts);
338 
339  /// Relex the token at the specified location.
340  /// \returns true if there was a failure, false on success.
341  static bool getRawToken(SourceLocation Loc, Token &Result,
342  const SourceManager &SM,
343  const LangOptions &LangOpts,
344  bool IgnoreWhiteSpace = false);
345 
346  /// Given a location any where in a source buffer, find the location
347  /// that corresponds to the beginning of the token in which the original
348  /// source location lands.
349  static SourceLocation GetBeginningOfToken(SourceLocation Loc,
350  const SourceManager &SM,
351  const LangOptions &LangOpts);
352 
353  /// Get the physical length (including trigraphs and escaped newlines) of the
354  /// first \p Characters characters of the token starting at TokStart.
355  static unsigned getTokenPrefixLength(SourceLocation TokStart,
356  unsigned CharNo,
357  const SourceManager &SM,
358  const LangOptions &LangOpts);
359 
360  /// AdvanceToTokenCharacter - If the current SourceLocation specifies a
361  /// location at the start of a token, return a new location that specifies a
362  /// character within the token. This handles trigraphs and escaped newlines.
364  unsigned Characters,
365  const SourceManager &SM,
366  const LangOptions &LangOpts) {
367  return TokStart.getLocWithOffset(
368  getTokenPrefixLength(TokStart, Characters, SM, LangOpts));
369  }
370 
371  /// Computes the source location just past the end of the
372  /// token at this source location.
373  ///
374  /// This routine can be used to produce a source location that
375  /// points just past the end of the token referenced by \p Loc, and
376  /// is generally used when a diagnostic needs to point just after a
377  /// token where it expected something different that it received. If
378  /// the returned source location would not be meaningful (e.g., if
379  /// it points into a macro), this routine returns an invalid
380  /// source location.
381  ///
382  /// \param Offset an offset from the end of the token, where the source
383  /// location should refer to. The default offset (0) produces a source
384  /// location pointing just past the end of the token; an offset of 1 produces
385  /// a source location pointing to the last character in the token, etc.
386  static SourceLocation getLocForEndOfToken(SourceLocation Loc, unsigned Offset,
387  const SourceManager &SM,
388  const LangOptions &LangOpts);
389 
390  /// Given a token range, produce a corresponding CharSourceRange that
391  /// is not a token range. This allows the source range to be used by
392  /// components that don't have access to the lexer and thus can't find the
393  /// end of the range for themselves.
395  const SourceManager &SM,
396  const LangOptions &LangOpts) {
397  SourceLocation End = getLocForEndOfToken(Range.getEnd(), 0, SM, LangOpts);
398  return End.isInvalid() ? CharSourceRange()
399  : CharSourceRange::getCharRange(
400  Range.getBegin(), End);
401  }
403  const SourceManager &SM,
404  const LangOptions &LangOpts) {
405  return Range.isTokenRange()
406  ? getAsCharRange(Range.getAsRange(), SM, LangOpts)
407  : Range;
408  }
409 
410  /// Returns true if the given MacroID location points at the first
411  /// token of the macro expansion.
412  ///
413  /// \param MacroBegin If non-null and function returns true, it is set to
414  /// begin location of the macro.
415  static bool isAtStartOfMacroExpansion(SourceLocation loc,
416  const SourceManager &SM,
417  const LangOptions &LangOpts,
418  SourceLocation *MacroBegin = nullptr);
419 
420  /// Returns true if the given MacroID location points at the last
421  /// token of the macro expansion.
422  ///
423  /// \param MacroEnd If non-null and function returns true, it is set to
424  /// end location of the macro.
425  static bool isAtEndOfMacroExpansion(SourceLocation loc,
426  const SourceManager &SM,
427  const LangOptions &LangOpts,
428  SourceLocation *MacroEnd = nullptr);
429 
430  /// Accepts a range and returns a character range with file locations.
431  ///
432  /// Returns a null range if a part of the range resides inside a macro
433  /// expansion or the range does not reside on the same FileID.
434  ///
435  /// This function is trying to deal with macros and return a range based on
436  /// file locations. The cases where it can successfully handle macros are:
437  ///
438  /// -begin or end range lies at the start or end of a macro expansion, in
439  /// which case the location will be set to the expansion point, e.g:
440  /// \#define M 1 2
441  /// a M
442  /// If you have a range [a, 2] (where 2 came from the macro), the function
443  /// will return a range for "a M"
444  /// if you have range [a, 1], the function will fail because the range
445  /// overlaps with only a part of the macro
446  ///
447  /// -The macro is a function macro and the range can be mapped to the macro
448  /// arguments, e.g:
449  /// \#define M 1 2
450  /// \#define FM(x) x
451  /// FM(a b M)
452  /// if you have range [b, 2], the function will return the file range "b M"
453  /// inside the macro arguments.
454  /// if you have range [a, 2], the function will return the file range
455  /// "FM(a b M)" since the range includes all of the macro expansion.
456  static CharSourceRange makeFileCharRange(CharSourceRange Range,
457  const SourceManager &SM,
458  const LangOptions &LangOpts);
459 
460  /// Returns a string for the source that the range encompasses.
461  static StringRef getSourceText(CharSourceRange Range,
462  const SourceManager &SM,
463  const LangOptions &LangOpts,
464  bool *Invalid = nullptr);
465 
466  /// Retrieve the name of the immediate macro expansion.
467  ///
468  /// This routine starts from a source location, and finds the name of the macro
469  /// responsible for its immediate expansion. It looks through any intervening
470  /// macro argument expansions to compute this. It returns a StringRef which
471  /// refers to the SourceManager-owned buffer of the source where that macro
472  /// name is spelled. Thus, the result shouldn't out-live that SourceManager.
473  static StringRef getImmediateMacroName(SourceLocation Loc,
474  const SourceManager &SM,
475  const LangOptions &LangOpts);
476 
477  /// Retrieve the name of the immediate macro expansion.
478  ///
479  /// This routine starts from a source location, and finds the name of the
480  /// macro responsible for its immediate expansion. It looks through any
481  /// intervening macro argument expansions to compute this. It returns a
482  /// StringRef which refers to the SourceManager-owned buffer of the source
483  /// where that macro name is spelled. Thus, the result shouldn't out-live
484  /// that SourceManager.
485  ///
486  /// This differs from Lexer::getImmediateMacroName in that any macro argument
487  /// location will result in the topmost function macro that accepted it.
488  /// e.g.
489  /// \code
490  /// MAC1( MAC2(foo) )
491  /// \endcode
492  /// for location of 'foo' token, this function will return "MAC1" while
493  /// Lexer::getImmediateMacroName will return "MAC2".
494  static StringRef getImmediateMacroNameForDiagnostics(
495  SourceLocation Loc, const SourceManager &SM, const LangOptions &LangOpts);
496 
497  /// Compute the preamble of the given file.
498  ///
499  /// The preamble of a file contains the initial comments, include directives,
500  /// and other preprocessor directives that occur before the code in this
501  /// particular file actually begins. The preamble of the main source file is
502  /// a potential prefix header.
503  ///
504  /// \param Buffer The memory buffer containing the file's contents.
505  ///
506  /// \param MaxLines If non-zero, restrict the length of the preamble
507  /// to fewer than this number of lines.
508  ///
509  /// \returns The offset into the file where the preamble ends and the rest
510  /// of the file begins along with a boolean value indicating whether
511  /// the preamble ends at the beginning of a new line.
512  static PreambleBounds ComputePreamble(StringRef Buffer,
513  const LangOptions &LangOpts,
514  unsigned MaxLines = 0);
515 
516  /// Finds the token that comes right after the given location.
517  ///
518  /// Returns the next token, or none if the location is inside a macro.
519  static Optional<Token> findNextToken(SourceLocation Loc,
520  const SourceManager &SM,
521  const LangOptions &LangOpts);
522 
523  /// Checks that the given token is the first token that occurs after
524  /// the given location (this excludes comments and whitespace). Returns the
525  /// location immediately after the specified token. If the token is not found
526  /// or the location is inside a macro, the returned source location will be
527  /// invalid.
528  static SourceLocation findLocationAfterToken(SourceLocation loc,
529  tok::TokenKind TKind,
530  const SourceManager &SM,
531  const LangOptions &LangOpts,
532  bool SkipTrailingWhitespaceAndNewLine);
533 
534  /// Returns true if the given character could appear in an identifier.
535  static bool isIdentifierBodyChar(char c, const LangOptions &LangOpts);
536 
537  /// Checks whether new line pointed by Str is preceded by escape
538  /// sequence.
539  static bool isNewLineEscaped(const char *BufferStart, const char *Str);
540 
541  /// getCharAndSizeNoWarn - Like the getCharAndSize method, but does not ever
542  /// emit a warning.
543  static inline char getCharAndSizeNoWarn(const char *Ptr, unsigned &Size,
544  const LangOptions &LangOpts) {
545  // If this is not a trigraph and not a UCN or escaped newline, return
546  // quickly.
547  if (isObviouslySimpleCharacter(Ptr[0])) {
548  Size = 1;
549  return *Ptr;
550  }
551 
552  Size = 0;
553  return getCharAndSizeSlowNoWarn(Ptr, Size, LangOpts);
554  }
555 
556  /// Returns the leading whitespace for line that corresponds to the given
557  /// location \p Loc.
558  static StringRef getIndentationForLine(SourceLocation Loc,
559  const SourceManager &SM);
560 
561 private:
562  //===--------------------------------------------------------------------===//
563  // Internal implementation interfaces.
564 
565  /// LexTokenInternal - Internal interface to lex a preprocessing token. Called
566  /// by Lex.
567  ///
568  bool LexTokenInternal(Token &Result, bool TokAtPhysicalStartOfLine);
569 
570  bool CheckUnicodeWhitespace(Token &Result, uint32_t C, const char *CurPtr);
571 
572  /// Given that a token begins with the Unicode character \p C, figure out
573  /// what kind of token it is and dispatch to the appropriate lexing helper
574  /// function.
575  bool LexUnicode(Token &Result, uint32_t C, const char *CurPtr);
576 
577  /// FormTokenWithChars - When we lex a token, we have identified a span
578  /// starting at BufferPtr, going to TokEnd that forms the token. This method
579  /// takes that range and assigns it to the token as its location and size. In
580  /// addition, since tokens cannot overlap, this also updates BufferPtr to be
581  /// TokEnd.
582  void FormTokenWithChars(Token &Result, const char *TokEnd,
584  unsigned TokLen = TokEnd-BufferPtr;
585  Result.setLength(TokLen);
586  Result.setLocation(getSourceLocation(BufferPtr, TokLen));
587  Result.setKind(Kind);
588  BufferPtr = TokEnd;
589  }
590 
591  /// isNextPPTokenLParen - Return 1 if the next unexpanded token will return a
592  /// tok::l_paren token, 0 if it is something else and 2 if there are no more
593  /// tokens in the buffer controlled by this lexer.
594  unsigned isNextPPTokenLParen();
595 
596  //===--------------------------------------------------------------------===//
597  // Lexer character reading interfaces.
598 
599  // This lexer is built on two interfaces for reading characters, both of which
600  // automatically provide phase 1/2 translation. getAndAdvanceChar is used
601  // when we know that we will be reading a character from the input buffer and
602  // that this character will be part of the result token. This occurs in (f.e.)
603  // string processing, because we know we need to read until we find the
604  // closing '"' character.
605  //
606  // The second interface is the combination of getCharAndSize with
607  // ConsumeChar. getCharAndSize reads a phase 1/2 translated character,
608  // returning it and its size. If the lexer decides that this character is
609  // part of the current token, it calls ConsumeChar on it. This two stage
610  // approach allows us to emit diagnostics for characters (e.g. warnings about
611  // trigraphs), knowing that they only are emitted if the character is
612  // consumed.
613 
614  /// isObviouslySimpleCharacter - Return true if the specified character is
615  /// obviously the same in translation phase 1 and translation phase 3. This
616  /// can return false for characters that end up being the same, but it will
617  /// never return true for something that needs to be mapped.
618  static bool isObviouslySimpleCharacter(char C) {
619  return C != '?' && C != '\\';
620  }
621 
622  /// getAndAdvanceChar - Read a single 'character' from the specified buffer,
623  /// advance over it, and return it. This is tricky in several cases. Here we
624  /// just handle the trivial case and fall-back to the non-inlined
625  /// getCharAndSizeSlow method to handle the hard case.
626  inline char getAndAdvanceChar(const char *&Ptr, Token &Tok) {
627  // If this is not a trigraph and not a UCN or escaped newline, return
628  // quickly.
629  if (isObviouslySimpleCharacter(Ptr[0])) return *Ptr++;
630 
631  unsigned Size = 0;
632  char C = getCharAndSizeSlow(Ptr, Size, &Tok);
633  Ptr += Size;
634  return C;
635  }
636 
637  /// ConsumeChar - When a character (identified by getCharAndSize) is consumed
638  /// and added to a given token, check to see if there are diagnostics that
639  /// need to be emitted or flags that need to be set on the token. If so, do
640  /// it.
641  const char *ConsumeChar(const char *Ptr, unsigned Size, Token &Tok) {
642  // Normal case, we consumed exactly one token. Just return it.
643  if (Size == 1)
644  return Ptr+Size;
645 
646  // Otherwise, re-lex the character with a current token, allowing
647  // diagnostics to be emitted and flags to be set.
648  Size = 0;
649  getCharAndSizeSlow(Ptr, Size, &Tok);
650  return Ptr+Size;
651  }
652 
653  /// getCharAndSize - Peek a single 'character' from the specified buffer,
654  /// get its size, and return it. This is tricky in several cases. Here we
655  /// just handle the trivial case and fall-back to the non-inlined
656  /// getCharAndSizeSlow method to handle the hard case.
657  inline char getCharAndSize(const char *Ptr, unsigned &Size) {
658  // If this is not a trigraph and not a UCN or escaped newline, return
659  // quickly.
660  if (isObviouslySimpleCharacter(Ptr[0])) {
661  Size = 1;
662  return *Ptr;
663  }
664 
665  Size = 0;
666  return getCharAndSizeSlow(Ptr, Size);
667  }
668 
669  /// getCharAndSizeSlow - Handle the slow/uncommon case of the getCharAndSize
670  /// method.
671  char getCharAndSizeSlow(const char *Ptr, unsigned &Size,
672  Token *Tok = nullptr);
673 
674  /// getEscapedNewLineSize - Return the size of the specified escaped newline,
675  /// or 0 if it is not an escaped newline. P[-1] is known to be a "\" on entry
676  /// to this function.
677  static unsigned getEscapedNewLineSize(const char *P);
678 
679  /// SkipEscapedNewLines - If P points to an escaped newline (or a series of
680  /// them), skip over them and return the first non-escaped-newline found,
681  /// otherwise return P.
682  static const char *SkipEscapedNewLines(const char *P);
683 
684  /// getCharAndSizeSlowNoWarn - Same as getCharAndSizeSlow, but never emits a
685  /// diagnostic.
686  static char getCharAndSizeSlowNoWarn(const char *Ptr, unsigned &Size,
687  const LangOptions &LangOpts);
688 
689  //===--------------------------------------------------------------------===//
690  // Other lexer functions.
691 
692  void SetByteOffset(unsigned Offset, bool StartOfLine);
693 
694  void PropagateLineStartLeadingSpaceInfo(Token &Result);
695 
696  const char *LexUDSuffix(Token &Result, const char *CurPtr,
697  bool IsStringLiteral);
698 
699  // Helper functions to lex the remainder of a token of the specific type.
700  bool LexIdentifier (Token &Result, const char *CurPtr);
701  bool LexNumericConstant (Token &Result, const char *CurPtr);
702  bool LexStringLiteral (Token &Result, const char *CurPtr,
703  tok::TokenKind Kind);
704  bool LexRawStringLiteral (Token &Result, const char *CurPtr,
705  tok::TokenKind Kind);
706  bool LexAngledStringLiteral(Token &Result, const char *CurPtr);
707  bool LexCharConstant (Token &Result, const char *CurPtr,
708  tok::TokenKind Kind);
709  bool LexEndOfFile (Token &Result, const char *CurPtr);
710  bool SkipWhitespace (Token &Result, const char *CurPtr,
711  bool &TokAtPhysicalStartOfLine);
712  bool SkipLineComment (Token &Result, const char *CurPtr,
713  bool &TokAtPhysicalStartOfLine);
714  bool SkipBlockComment (Token &Result, const char *CurPtr,
715  bool &TokAtPhysicalStartOfLine);
716  bool SaveLineComment (Token &Result, const char *CurPtr);
717 
718  bool IsStartOfConflictMarker(const char *CurPtr);
719  bool HandleEndOfConflictMarker(const char *CurPtr);
720 
721  bool lexEditorPlaceholder(Token &Result, const char *CurPtr);
722 
723  bool isCodeCompletionPoint(const char *CurPtr) const;
724  void cutOffLexing() { BufferPtr = BufferEnd; }
725 
726  bool isHexaLiteral(const char *Start, const LangOptions &LangOpts);
727 
728  void codeCompleteIncludedFile(const char *PathStart,
729  const char *CompletionPoint, bool IsAngled);
730 
731  /// Read a universal character name.
732  ///
733  /// \param StartPtr The position in the source buffer after the initial '\'.
734  /// If the UCN is syntactically well-formed (but not
735  /// necessarily valid), this parameter will be updated to
736  /// point to the character after the UCN.
737  /// \param SlashLoc The position in the source buffer of the '\'.
738  /// \param Result The token being formed. Pass \c nullptr to suppress
739  /// diagnostics and handle token formation in the caller.
740  ///
741  /// \return The Unicode codepoint specified by the UCN, or 0 if the UCN is
742  /// invalid.
743  uint32_t tryReadUCN(const char *&StartPtr, const char *SlashLoc, Token *Result);
744 
745  /// Try to consume a UCN as part of an identifier at the current
746  /// location.
747  /// \param CurPtr Initially points to the range of characters in the source
748  /// buffer containing the '\'. Updated to point past the end of
749  /// the UCN on success.
750  /// \param Size The number of characters occupied by the '\' (including
751  /// trigraphs and escaped newlines).
752  /// \param Result The token being produced. Marked as containing a UCN on
753  /// success.
754  /// \return \c true if a UCN was lexed and it produced an acceptable
755  /// identifier character, \c false otherwise.
756  bool tryConsumeIdentifierUCN(const char *&CurPtr, unsigned Size,
757  Token &Result);
758 
759  /// Try to consume an identifier character encoded in UTF-8.
760  /// \param CurPtr Points to the start of the (potential) UTF-8 code unit
761  /// sequence. On success, updated to point past the end of it.
762  /// \return \c true if a UTF-8 sequence mapping to an acceptable identifier
763  /// character was lexed, \c false otherwise.
764  bool tryConsumeIdentifierUTF8Char(const char *&CurPtr);
765 };
766 
767 } // namespace clang
768 
769 #endif // LLVM_CLANG_LEX_LEXER_H
Describes the bounds (start, size) of the preamble and a flag required by PreprocessorOptions::Precom...
Definition: Lexer.h:58
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:76
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:266
Specialize PointerLikeTypeTraits to allow LazyGenerationalUpdatePtr to be placed into a PointerUnion...
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:194
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:543
ConflictMarkerKind
ConflictMarkerKind - Kinds of conflict marker which the lexer might be recovering from...
Definition: Lexer.h:42
static CharSourceRange getAsCharRange(CharSourceRange Range, const SourceManager &SM, const LangOptions &LangOpts)
Definition: Lexer.h:402
Token - This structure provides full information about a lexed token.
Definition: Token.h:34
void setKind(tok::TokenKind K)
Definition: Token.h:93
SourceLocation getFileLoc() const
getFileLoc - Return the File Location for the file we are lexing out of.
Definition: Lexer.h:174
Keeps track of the various options that can be enabled, which controls the dialect of C or C++ that i...
Definition: LangOptions.h:49
bool isPragmaLexer() const
isPragmaLexer - Returns true if this Lexer is being used to lex a pragma.
Definition: Lexer.h:183
A Perforce-style conflict marker, initiated by 4 ">"s, separated by 4 "="s, and terminated by 4 "<"s...
Definition: Lexer.h:52
SourceLocation getSourceLocation() override
getSourceLocation - Return a source location for the next character in the current file...
Definition: Lexer.h:261
const FormatToken & Tok
StringRef getBuffer() const
Gets source code buffer.
Definition: Lexer.h:242
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:363
bool PreambleEndsAtStartOfLine
Whether the preamble ends at the start of a new line.
Definition: Lexer.h:66
A little helper class used to produce diagnostics.
Definition: Diagnostic.h:1043
const LangOptions & getLangOpts() const
getLangOpts - Return the language features currently enabled.
Definition: Lexer.h:168
unsigned Offset
Definition: Format.cpp:1758
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:206
SourceLocation getEnd() const
const SourceManager & SM
Definition: Format.cpp:1616
Kind
Encodes a location in the source.
void setLength(unsigned Len)
Definition: Token.h:135
PreambleBounds(unsigned Size, bool PreambleEndsAtStartOfLine)
Definition: Lexer.h:68
TokenKind
Provides a simple uniform namespace for tokens from all C languages.
Definition: TokenKinds.h:24
bool inKeepCommentMode() const
inKeepCommentMode - Return true if the lexer should return comments as tokens.
Definition: Lexer.h:220
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:394
unsigned getCurrentBufferOffset()
Returns the current lexing offset.
Definition: Lexer.h:269
unsigned Size
Size of the preamble in bytes.
Definition: Lexer.h:60
Defines the clang::TokenKind enum and support functions.
Defines the clang::SourceLocation class and associated facilities.
Not within a conflict marker.
Definition: Lexer.h:44
void SetCommentRetentionState(bool Mode)
SetCommentRetentionMode - Change the comment retention mode of the lexer to the specified mode...
Definition: Lexer.h:227
void setLocation(SourceLocation L)
Definition: Token.h:134
A normal or diff3 conflict marker, initiated by at least 7 "<"s, separated by at least 7 "="s or "|"s...
Definition: Lexer.h:48
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:212
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:125