Coverage Report

Created: 2022-01-18 06:27

/Users/buildslave/jenkins/workspace/coverage/llvm-project/clang/include/clang/Lex/Lexer.h
Line
Count
Source (jump to first uncovered line)
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
16
#include "clang/Basic/LangOptions.h"
17
#include "clang/Basic/SourceLocation.h"
18
#include "clang/Basic/TokenKinds.h"
19
#include "clang/Lex/PreprocessorLexer.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 MemoryBufferRef;
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.
42
enum ConflictMarkerKind {
43
  /// Not within a conflict marker.
44
  CMK_None,
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.
48
  CMK_Normal,
49
50
  /// A Perforce-style conflict marker, initiated by 4 ">"s,
51
  /// separated by 4 "="s, and terminated by 4 "<"s.
52
  CMK_Perforce
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.
58
struct PreambleBounds {
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.
66
  bool PreambleEndsAtStartOfLine;
67
68
  PreambleBounds(unsigned Size, bool PreambleEndsAtStartOfLine)
69
1.38k
      : 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
  /// True if this is the first time we're lexing the input file.
132
  bool IsFirstTimeLexingFile;
133
134
  // NewLinePtr - A pointer to new line character '\n' being lexed. For '\r\n',
135
  // it also points to '\n.'
136
  const char *NewLinePtr;
137
138
  // CurrentConflictMarkerState - The kind of conflict marker we are handling.
139
  ConflictMarkerKind CurrentConflictMarkerState;
140
141
  void InitLexer(const char *BufStart, const char *BufPtr, const char *BufEnd);
142
143
public:
144
  /// Lexer constructor - Create a new lexer object for the specified buffer
145
  /// with the specified preprocessor managing the lexing process.  This lexer
146
  /// assumes that the associated file buffer and Preprocessor objects will
147
  /// outlive it, so it doesn't take ownership of either of them.
148
  Lexer(FileID FID, const llvm::MemoryBufferRef &InputFile, Preprocessor &PP,
149
        bool IsFirstIncludeOfFile = true);
150
151
  /// Lexer constructor - Create a new raw lexer object.  This object is only
152
  /// suitable for calls to 'LexFromRawLexer'.  This lexer assumes that the
153
  /// text range will outlive it, so it doesn't take ownership of it.
154
  Lexer(SourceLocation FileLoc, const LangOptions &LangOpts,
155
        const char *BufStart, const char *BufPtr, const char *BufEnd,
156
        bool IsFirstIncludeOfFile = true);
157
158
  /// Lexer constructor - Create a new raw lexer object.  This object is only
159
  /// suitable for calls to 'LexFromRawLexer'.  This lexer assumes that the
160
  /// text range will outlive it, so it doesn't take ownership of it.
161
  Lexer(FileID FID, const llvm::MemoryBufferRef &FromFile,
162
        const SourceManager &SM, const LangOptions &LangOpts,
163
        bool IsFirstIncludeOfFile = true);
164
165
  Lexer(const Lexer &) = delete;
166
  Lexer &operator=(const Lexer &) = delete;
167
168
  /// Create_PragmaLexer: Lexer constructor - Create a new lexer object for
169
  /// _Pragma expansion.  This has a variety of magic semantics that this method
170
  /// sets up.  It returns a new'd Lexer that must be delete'd when done.
171
  static Lexer *Create_PragmaLexer(SourceLocation SpellingLoc,
172
                                   SourceLocation ExpansionLocStart,
173
                                   SourceLocation ExpansionLocEnd,
174
                                   unsigned TokLen, Preprocessor &PP);
175
176
  /// getLangOpts - Return the language features currently enabled.
177
  /// NOTE: this lexer modifies features as a file is parsed!
178
20.7M
  const LangOptions &getLangOpts() const { return LangOpts; }
179
180
  /// getFileLoc - Return the File Location for the file we are lexing out of.
181
  /// The physical location encodes the location where the characters come from,
182
  /// the virtual location encodes where we should *claim* the characters came
183
  /// from.  Currently this is only used by _Pragma handling.
184
1.52M
  SourceLocation getFileLoc() const { return FileLoc; }
185
186
private:
187
  /// Lex - Return the next token in the file.  If this is the end of file, it
188
  /// return the tok::eof token.  This implicitly involves the preprocessor.
189
  bool Lex(Token &Result);
190
191
public:
192
  /// isPragmaLexer - Returns true if this Lexer is being used to lex a pragma.
193
3.16M
  bool isPragmaLexer() const { return Is_PragmaLexer; }
194
195
private:
196
  /// IndirectLex - An indirect call to 'Lex' that can be invoked via
197
  ///  the PreprocessorLexer interface.
198
0
  void IndirectLex(Token &Result) override { Lex(Result); }
199
200
public:
201
  /// LexFromRawLexer - Lex a token from a designated raw lexer (one with no
202
  /// associated preprocessor object.  Return true if the 'next character to
203
  /// read' pointer points at the end of the lexer buffer, false otherwise.
204
65.0M
  bool LexFromRawLexer(Token &Result) {
205
65.0M
    assert(LexingRawMode && "Not already in raw mode!");
206
0
    Lex(Result);
207
    // Note that lexing to the end of the buffer doesn't implicitly delete the
208
    // lexer when in raw mode.
209
65.0M
    return BufferPtr == BufferEnd;
210
65.0M
  }
211
212
  /// isKeepWhitespaceMode - Return true if the lexer should return tokens for
213
  /// every character in the file, including whitespace and comments.  This
214
  /// should only be used in raw mode, as the preprocessor is not prepared to
215
  /// deal with the excess tokens.
216
829M
  bool isKeepWhitespaceMode() const {
217
829M
    return ExtendedTokenMode > 1;
218
829M
  }
219
220
  /// SetKeepWhitespaceMode - This method lets clients enable or disable
221
  /// whitespace retention mode.
222
78.0M
  void SetKeepWhitespaceMode(bool Val) {
223
78.0M
    assert((!Val || LexingRawMode || LangOpts.TraditionalCPP) &&
224
78.0M
           "Can only retain whitespace in raw mode or -traditional-cpp");
225
78.0M
    ExtendedTokenMode = Val ? 
264.8k
:
078.0M
;
226
78.0M
  }
227
228
  /// inKeepCommentMode - Return true if the lexer should return comments as
229
  /// tokens.
230
119M
  bool inKeepCommentMode() const {
231
119M
    return ExtendedTokenMode > 0;
232
119M
  }
233
234
  /// SetCommentRetentionMode - Change the comment retention mode of the lexer
235
  /// to the specified mode.  This is really only useful when lexing in raw
236
  /// mode, because otherwise the lexer needs to manage this.
237
177M
  void SetCommentRetentionState(bool Mode) {
238
177M
    assert(!isKeepWhitespaceMode() &&
239
177M
           "Can't play with comment retention state when retaining whitespace");
240
177M
    ExtendedTokenMode = Mode ? 
150.2M
:
0127M
;
241
177M
  }
242
243
  /// Sets the extended token mode back to its initial value, according to the
244
  /// language options and preprocessor. This controls whether the lexer
245
  /// produces comment and whitespace tokens.
246
  ///
247
  /// This requires the lexer to have an associated preprocessor. A standalone
248
  /// lexer has nothing to reset to.
249
  void resetExtendedTokenMode();
250
251
  /// Gets source code buffer.
252
5.21k
  StringRef getBuffer() const {
253
5.21k
    return StringRef(BufferStart, BufferEnd - BufferStart);
254
5.21k
  }
255
256
  /// ReadToEndOfLine - Read the rest of the current preprocessor line as an
257
  /// uninterpreted string.  This switches the lexer out of directive mode.
258
  void ReadToEndOfLine(SmallVectorImpl<char> *Result = nullptr);
259
260
261
  /// Diag - Forwarding function for diagnostics.  This translate a source
262
  /// position in the current buffer into a SourceLocation object for rendering.
263
  DiagnosticBuilder Diag(const char *Loc, unsigned DiagID) const;
264
265
  /// getSourceLocation - Return a source location identifier for the specified
266
  /// offset in the current file.
267
  SourceLocation getSourceLocation(const char *Loc, unsigned TokLen = 1) const;
268
269
  /// getSourceLocation - Return a source location for the next character in
270
  /// the current file.
271
1.67M
  SourceLocation getSourceLocation() override {
272
1.67M
    return getSourceLocation(BufferPtr);
273
1.67M
  }
274
275
  /// Return the current location in the buffer.
276
48.5k
  const char *getBufferLocation() const { return BufferPtr; }
277
278
  /// Returns the current lexing offset.
279
39.2k
  unsigned getCurrentBufferOffset() {
280
39.2k
    assert(BufferPtr >= BufferStart && "Invalid buffer state");
281
0
    return BufferPtr - BufferStart;
282
39.2k
  }
283
284
  /// Skip over \p NumBytes bytes.
285
  ///
286
  /// If the skip is successful, the next token will be lexed from the new
287
  /// offset. The lexer also assumes that we skipped to the start of the line.
288
  ///
289
  /// \returns true if the skip failed (new offset would have been past the
290
  /// end of the buffer), false otherwise.
291
  bool skipOver(unsigned NumBytes);
292
293
  /// Stringify - Convert the specified string into a C string by i) escaping
294
  /// '\\' and " characters and ii) replacing newline character(s) with "\\n".
295
  /// If Charify is true, this escapes the ' character instead of ".
296
  static std::string Stringify(StringRef Str, bool Charify = false);
297
298
  /// Stringify - Convert the specified string into a C string by i) escaping
299
  /// '\\' and " characters and ii) replacing newline character(s) with "\\n".
300
  static void Stringify(SmallVectorImpl<char> &Str);
301
302
  /// getSpelling - This method is used to get the spelling of a token into a
303
  /// preallocated buffer, instead of as an std::string.  The caller is required
304
  /// to allocate enough space for the token, which is guaranteed to be at least
305
  /// Tok.getLength() bytes long.  The length of the actual result is returned.
306
  ///
307
  /// Note that this method may do two possible things: it may either fill in
308
  /// the buffer specified with characters, or it may *change the input pointer*
309
  /// to point to a constant buffer with the data already in it (avoiding a
310
  /// copy).  The caller is not allowed to modify the returned buffer pointer
311
  /// if an internal buffer is returned.
312
  static unsigned getSpelling(const Token &Tok, const char *&Buffer,
313
                              const SourceManager &SourceMgr,
314
                              const LangOptions &LangOpts,
315
                              bool *Invalid = nullptr);
316
317
  /// getSpelling() - Return the 'spelling' of the Tok token.  The spelling of a
318
  /// token is the characters used to represent the token in the source file
319
  /// after trigraph expansion and escaped-newline folding.  In particular, this
320
  /// wants to get the true, uncanonicalized, spelling of things like digraphs
321
  /// UCNs, etc.
322
  static std::string getSpelling(const Token &Tok,
323
                                 const SourceManager &SourceMgr,
324
                                 const LangOptions &LangOpts,
325
                                 bool *Invalid = nullptr);
326
327
  /// getSpelling - This method is used to get the spelling of the
328
  /// token at the given source location.  If, as is usually true, it
329
  /// is not necessary to copy any data, then the returned string may
330
  /// not point into the provided buffer.
331
  ///
332
  /// This method lexes at the expansion depth of the given
333
  /// location and does not jump to the expansion or spelling
334
  /// location.
335
  static StringRef getSpelling(SourceLocation loc,
336
                               SmallVectorImpl<char> &buffer,
337
                               const SourceManager &SM,
338
                               const LangOptions &options,
339
                               bool *invalid = nullptr);
340
341
  /// MeasureTokenLength - Relex the token at the specified location and return
342
  /// its length in bytes in the input file.  If the token needs cleaning (e.g.
343
  /// includes a trigraph or an escaped newline) then this count includes bytes
344
  /// that are part of that.
345
  static unsigned MeasureTokenLength(SourceLocation Loc,
346
                                     const SourceManager &SM,
347
                                     const LangOptions &LangOpts);
348
349
  /// Relex the token at the specified location.
350
  /// \returns true if there was a failure, false on success.
351
  static bool getRawToken(SourceLocation Loc, Token &Result,
352
                          const SourceManager &SM,
353
                          const LangOptions &LangOpts,
354
                          bool IgnoreWhiteSpace = false);
355
356
  /// Given a location any where in a source buffer, find the location
357
  /// that corresponds to the beginning of the token in which the original
358
  /// source location lands.
359
  static SourceLocation GetBeginningOfToken(SourceLocation Loc,
360
                                            const SourceManager &SM,
361
                                            const LangOptions &LangOpts);
362
363
  /// Get the physical length (including trigraphs and escaped newlines) of the
364
  /// first \p Characters characters of the token starting at TokStart.
365
  static unsigned getTokenPrefixLength(SourceLocation TokStart,
366
                                       unsigned CharNo,
367
                                       const SourceManager &SM,
368
                                       const LangOptions &LangOpts);
369
370
  /// AdvanceToTokenCharacter - If the current SourceLocation specifies a
371
  /// location at the start of a token, return a new location that specifies a
372
  /// character within the token.  This handles trigraphs and escaped newlines.
373
  static SourceLocation AdvanceToTokenCharacter(SourceLocation TokStart,
374
                                                unsigned Characters,
375
                                                const SourceManager &SM,
376
51.2k
                                                const LangOptions &LangOpts) {
377
51.2k
    return TokStart.getLocWithOffset(
378
51.2k
        getTokenPrefixLength(TokStart, Characters, SM, LangOpts));
379
51.2k
  }
380
381
  /// Computes the source location just past the end of the
382
  /// token at this source location.
383
  ///
384
  /// This routine can be used to produce a source location that
385
  /// points just past the end of the token referenced by \p Loc, and
386
  /// is generally used when a diagnostic needs to point just after a
387
  /// token where it expected something different that it received. If
388
  /// the returned source location would not be meaningful (e.g., if
389
  /// it points into a macro), this routine returns an invalid
390
  /// source location.
391
  ///
392
  /// \param Offset an offset from the end of the token, where the source
393
  /// location should refer to. The default offset (0) produces a source
394
  /// location pointing just past the end of the token; an offset of 1 produces
395
  /// a source location pointing to the last character in the token, etc.
396
  static SourceLocation getLocForEndOfToken(SourceLocation Loc, unsigned Offset,
397
                                            const SourceManager &SM,
398
                                            const LangOptions &LangOpts);
399
400
  /// Given a token range, produce a corresponding CharSourceRange that
401
  /// is not a token range. This allows the source range to be used by
402
  /// components that don't have access to the lexer and thus can't find the
403
  /// end of the range for themselves.
404
  static CharSourceRange getAsCharRange(SourceRange Range,
405
                                        const SourceManager &SM,
406
6.50k
                                        const LangOptions &LangOpts) {
407
6.50k
    SourceLocation End = getLocForEndOfToken(Range.getEnd(), 0, SM, LangOpts);
408
6.50k
    return End.isInvalid() ? 
CharSourceRange()1
409
6.50k
                           : CharSourceRange::getCharRange(
410
6.50k
                                 Range.getBegin(), End);
411
6.50k
  }
412
  static CharSourceRange getAsCharRange(CharSourceRange Range,
413
                                        const SourceManager &SM,
414
2.06k
                                        const LangOptions &LangOpts) {
415
2.06k
    return Range.isTokenRange()
416
2.06k
               ? 
getAsCharRange(Range.getAsRange(), SM, LangOpts)2.02k
417
2.06k
               : 
Range41
;
418
2.06k
  }
419
420
  /// Returns true if the given MacroID location points at the first
421
  /// token of the macro expansion.
422
  ///
423
  /// \param MacroBegin If non-null and function returns true, it is set to
424
  /// begin location of the macro.
425
  static bool isAtStartOfMacroExpansion(SourceLocation loc,
426
                                        const SourceManager &SM,
427
                                        const LangOptions &LangOpts,
428
                                        SourceLocation *MacroBegin = nullptr);
429
430
  /// Returns true if the given MacroID location points at the last
431
  /// token of the macro expansion.
432
  ///
433
  /// \param MacroEnd If non-null and function returns true, it is set to
434
  /// end location of the macro.
435
  static bool isAtEndOfMacroExpansion(SourceLocation loc,
436
                                      const SourceManager &SM,
437
                                      const LangOptions &LangOpts,
438
                                      SourceLocation *MacroEnd = nullptr);
439
440
  /// Accepts a range and returns a character range with file locations.
441
  ///
442
  /// Returns a null range if a part of the range resides inside a macro
443
  /// expansion or the range does not reside on the same FileID.
444
  ///
445
  /// This function is trying to deal with macros and return a range based on
446
  /// file locations. The cases where it can successfully handle macros are:
447
  ///
448
  /// -begin or end range lies at the start or end of a macro expansion, in
449
  ///  which case the location will be set to the expansion point, e.g:
450
  ///    \#define M 1 2
451
  ///    a M
452
  /// If you have a range [a, 2] (where 2 came from the macro), the function
453
  /// will return a range for "a M"
454
  /// if you have range [a, 1], the function will fail because the range
455
  /// overlaps with only a part of the macro
456
  ///
457
  /// -The macro is a function macro and the range can be mapped to the macro
458
  ///  arguments, e.g:
459
  ///    \#define M 1 2
460
  ///    \#define FM(x) x
461
  ///    FM(a b M)
462
  /// if you have range [b, 2], the function will return the file range "b M"
463
  /// inside the macro arguments.
464
  /// if you have range [a, 2], the function will return the file range
465
  /// "FM(a b M)" since the range includes all of the macro expansion.
466
  static CharSourceRange makeFileCharRange(CharSourceRange Range,
467
                                           const SourceManager &SM,
468
                                           const LangOptions &LangOpts);
469
470
  /// Returns a string for the source that the range encompasses.
471
  static StringRef getSourceText(CharSourceRange Range,
472
                                 const SourceManager &SM,
473
                                 const LangOptions &LangOpts,
474
                                 bool *Invalid = nullptr);
475
476
  /// Retrieve the name of the immediate macro expansion.
477
  ///
478
  /// This routine starts from a source location, and finds the name of the macro
479
  /// responsible for its immediate expansion. It looks through any intervening
480
  /// macro argument expansions to compute this. It returns a StringRef which
481
  /// refers to the SourceManager-owned buffer of the source where that macro
482
  /// name is spelled. Thus, the result shouldn't out-live that SourceManager.
483
  static StringRef getImmediateMacroName(SourceLocation Loc,
484
                                         const SourceManager &SM,
485
                                         const LangOptions &LangOpts);
486
487
  /// Retrieve the name of the immediate macro expansion.
488
  ///
489
  /// This routine starts from a source location, and finds the name of the
490
  /// macro responsible for its immediate expansion. It looks through any
491
  /// intervening macro argument expansions to compute this. It returns a
492
  /// StringRef which refers to the SourceManager-owned buffer of the source
493
  /// where that macro name is spelled. Thus, the result shouldn't out-live
494
  /// that SourceManager.
495
  ///
496
  /// This differs from Lexer::getImmediateMacroName in that any macro argument
497
  /// location will result in the topmost function macro that accepted it.
498
  /// e.g.
499
  /// \code
500
  ///   MAC1( MAC2(foo) )
501
  /// \endcode
502
  /// for location of 'foo' token, this function will return "MAC1" while
503
  /// Lexer::getImmediateMacroName will return "MAC2".
504
  static StringRef getImmediateMacroNameForDiagnostics(
505
      SourceLocation Loc, const SourceManager &SM, const LangOptions &LangOpts);
506
507
  /// Compute the preamble of the given file.
508
  ///
509
  /// The preamble of a file contains the initial comments, include directives,
510
  /// and other preprocessor directives that occur before the code in this
511
  /// particular file actually begins. The preamble of the main source file is
512
  /// a potential prefix header.
513
  ///
514
  /// \param Buffer The memory buffer containing the file's contents.
515
  ///
516
  /// \param MaxLines If non-zero, restrict the length of the preamble
517
  /// to fewer than this number of lines.
518
  ///
519
  /// \returns The offset into the file where the preamble ends and the rest
520
  /// of the file begins along with a boolean value indicating whether
521
  /// the preamble ends at the beginning of a new line.
522
  static PreambleBounds ComputePreamble(StringRef Buffer,
523
                                        const LangOptions &LangOpts,
524
                                        unsigned MaxLines = 0);
525
526
  /// Finds the token that comes right after the given location.
527
  ///
528
  /// Returns the next token, or none if the location is inside a macro.
529
  static Optional<Token> findNextToken(SourceLocation Loc,
530
                                       const SourceManager &SM,
531
                                       const LangOptions &LangOpts);
532
533
  /// Checks that the given token is the first token that occurs after
534
  /// the given location (this excludes comments and whitespace). Returns the
535
  /// location immediately after the specified token. If the token is not found
536
  /// or the location is inside a macro, the returned source location will be
537
  /// invalid.
538
  static SourceLocation findLocationAfterToken(SourceLocation loc,
539
                                         tok::TokenKind TKind,
540
                                         const SourceManager &SM,
541
                                         const LangOptions &LangOpts,
542
                                         bool SkipTrailingWhitespaceAndNewLine);
543
544
  /// Returns true if the given character could appear in an identifier.
545
  static bool isAsciiIdentifierContinueChar(char c,
546
                                            const LangOptions &LangOpts);
547
548
  /// Checks whether new line pointed by Str is preceded by escape
549
  /// sequence.
550
  static bool isNewLineEscaped(const char *BufferStart, const char *Str);
551
552
  /// getCharAndSizeNoWarn - Like the getCharAndSize method, but does not ever
553
  /// emit a warning.
554
  static inline char getCharAndSizeNoWarn(const char *Ptr, unsigned &Size,
555
261k
                                          const LangOptions &LangOpts) {
556
    // If this is not a trigraph and not a UCN or escaped newline, return
557
    // quickly.
558
261k
    if (isObviouslySimpleCharacter(Ptr[0])) {
559
251k
      Size = 1;
560
251k
      return *Ptr;
561
251k
    }
562
563
10.5k
    Size = 0;
564
10.5k
    return getCharAndSizeSlowNoWarn(Ptr, Size, LangOpts);
565
261k
  }
566
567
  /// Returns the leading whitespace for line that corresponds to the given
568
  /// location \p Loc.
569
  static StringRef getIndentationForLine(SourceLocation Loc,
570
                                         const SourceManager &SM);
571
572
  /// Check if this is the first time we're lexing the input file.
573
14
  bool isFirstTimeLexingFile() const { return IsFirstTimeLexingFile; }
574
575
private:
576
  //===--------------------------------------------------------------------===//
577
  // Internal implementation interfaces.
578
579
  /// LexTokenInternal - Internal interface to lex a preprocessing token. Called
580
  /// by Lex.
581
  ///
582
  bool LexTokenInternal(Token &Result, bool TokAtPhysicalStartOfLine);
583
584
  bool CheckUnicodeWhitespace(Token &Result, uint32_t C, const char *CurPtr);
585
586
  bool LexUnicodeIdentifierStart(Token &Result, uint32_t C, const char *CurPtr);
587
588
  /// FormTokenWithChars - When we lex a token, we have identified a span
589
  /// starting at BufferPtr, going to TokEnd that forms the token.  This method
590
  /// takes that range and assigns it to the token as its location and size.  In
591
  /// addition, since tokens cannot overlap, this also updates BufferPtr to be
592
  /// TokEnd.
593
  void FormTokenWithChars(Token &Result, const char *TokEnd,
594
1.67G
                          tok::TokenKind Kind) {
595
1.67G
    unsigned TokLen = TokEnd-BufferPtr;
596
1.67G
    Result.setLength(TokLen);
597
1.67G
    Result.setLocation(getSourceLocation(BufferPtr, TokLen));
598
1.67G
    Result.setKind(Kind);
599
1.67G
    BufferPtr = TokEnd;
600
1.67G
  }
601
602
  /// isNextPPTokenLParen - Return 1 if the next unexpanded token will return a
603
  /// tok::l_paren token, 0 if it is something else and 2 if there are no more
604
  /// tokens in the buffer controlled by this lexer.
605
  unsigned isNextPPTokenLParen();
606
607
  //===--------------------------------------------------------------------===//
608
  // Lexer character reading interfaces.
609
610
  // This lexer is built on two interfaces for reading characters, both of which
611
  // automatically provide phase 1/2 translation.  getAndAdvanceChar is used
612
  // when we know that we will be reading a character from the input buffer and
613
  // that this character will be part of the result token. This occurs in (f.e.)
614
  // string processing, because we know we need to read until we find the
615
  // closing '"' character.
616
  //
617
  // The second interface is the combination of getCharAndSize with
618
  // ConsumeChar.  getCharAndSize reads a phase 1/2 translated character,
619
  // returning it and its size.  If the lexer decides that this character is
620
  // part of the current token, it calls ConsumeChar on it.  This two stage
621
  // approach allows us to emit diagnostics for characters (e.g. warnings about
622
  // trigraphs), knowing that they only are emitted if the character is
623
  // consumed.
624
625
  /// isObviouslySimpleCharacter - Return true if the specified character is
626
  /// obviously the same in translation phase 1 and translation phase 3.  This
627
  /// can return false for characters that end up being the same, but it will
628
  /// never return true for something that needs to be mapped.
629
3.22G
  static bool isObviouslySimpleCharacter(char C) {
630
3.22G
    return C != '?' && 
C != '\\'3.22G
;
631
3.22G
  }
632
633
  /// getAndAdvanceChar - Read a single 'character' from the specified buffer,
634
  /// advance over it, and return it.  This is tricky in several cases.  Here we
635
  /// just handle the trivial case and fall-back to the non-inlined
636
  /// getCharAndSizeSlow method to handle the hard case.
637
2.04G
  inline char getAndAdvanceChar(const char *&Ptr, Token &Tok) {
638
    // If this is not a trigraph and not a UCN or escaped newline, return
639
    // quickly.
640
2.04G
    if (isObviouslySimpleCharacter(Ptr[0])) 
return *Ptr++2.03G
;
641
642
6.96M
    unsigned Size = 0;
643
6.96M
    char C = getCharAndSizeSlow(Ptr, Size, &Tok);
644
6.96M
    Ptr += Size;
645
6.96M
    return C;
646
2.04G
  }
647
648
  /// ConsumeChar - When a character (identified by getCharAndSize) is consumed
649
  /// and added to a given token, check to see if there are diagnostics that
650
  /// need to be emitted or flags that need to be set on the token.  If so, do
651
  /// it.
652
194M
  const char *ConsumeChar(const char *Ptr, unsigned Size, Token &Tok) {
653
    // Normal case, we consumed exactly one token.  Just return it.
654
194M
    if (Size == 1)
655
194M
      return Ptr+Size;
656
657
    // Otherwise, re-lex the character with a current token, allowing
658
    // diagnostics to be emitted and flags to be set.
659
100
    Size = 0;
660
100
    getCharAndSizeSlow(Ptr, Size, &Tok);
661
100
    return Ptr+Size;
662
194M
  }
663
664
  /// getCharAndSize - Peek a single 'character' from the specified buffer,
665
  /// get its size, and return it.  This is tricky in several cases.  Here we
666
  /// just handle the trivial case and fall-back to the non-inlined
667
  /// getCharAndSizeSlow method to handle the hard case.
668
1.18G
  inline char getCharAndSize(const char *Ptr, unsigned &Size) {
669
    // If this is not a trigraph and not a UCN or escaped newline, return
670
    // quickly.
671
1.18G
    if (isObviouslySimpleCharacter(Ptr[0])) {
672
1.18G
      Size = 1;
673
1.18G
      return *Ptr;
674
1.18G
    }
675
676
7.08k
    Size = 0;
677
7.08k
    return getCharAndSizeSlow(Ptr, Size);
678
1.18G
  }
679
680
  /// getCharAndSizeSlow - Handle the slow/uncommon case of the getCharAndSize
681
  /// method.
682
  char getCharAndSizeSlow(const char *Ptr, unsigned &Size,
683
                          Token *Tok = nullptr);
684
685
  /// getEscapedNewLineSize - Return the size of the specified escaped newline,
686
  /// or 0 if it is not an escaped newline. P[-1] is known to be a "\" on entry
687
  /// to this function.
688
  static unsigned getEscapedNewLineSize(const char *P);
689
690
  /// SkipEscapedNewLines - If P points to an escaped newline (or a series of
691
  /// them), skip over them and return the first non-escaped-newline found,
692
  /// otherwise return P.
693
  static const char *SkipEscapedNewLines(const char *P);
694
695
  /// getCharAndSizeSlowNoWarn - Same as getCharAndSizeSlow, but never emits a
696
  /// diagnostic.
697
  static char getCharAndSizeSlowNoWarn(const char *Ptr, unsigned &Size,
698
                                       const LangOptions &LangOpts);
699
700
  //===--------------------------------------------------------------------===//
701
  // Other lexer functions.
702
703
  void SetByteOffset(unsigned Offset, bool StartOfLine);
704
705
  void PropagateLineStartLeadingSpaceInfo(Token &Result);
706
707
  const char *LexUDSuffix(Token &Result, const char *CurPtr,
708
                          bool IsStringLiteral);
709
710
  // Helper functions to lex the remainder of a token of the specific type.
711
712
  // This function handles both ASCII and Unicode identifiers after
713
  // the first codepoint of the identifyier has been parsed.
714
  bool LexIdentifierContinue(Token &Result, const char *CurPtr);
715
716
  bool LexNumericConstant    (Token &Result, const char *CurPtr);
717
  bool LexStringLiteral      (Token &Result, const char *CurPtr,
718
                              tok::TokenKind Kind);
719
  bool LexRawStringLiteral   (Token &Result, const char *CurPtr,
720
                              tok::TokenKind Kind);
721
  bool LexAngledStringLiteral(Token &Result, const char *CurPtr);
722
  bool LexCharConstant       (Token &Result, const char *CurPtr,
723
                              tok::TokenKind Kind);
724
  bool LexEndOfFile          (Token &Result, const char *CurPtr);
725
  bool SkipWhitespace        (Token &Result, const char *CurPtr,
726
                              bool &TokAtPhysicalStartOfLine);
727
  bool SkipLineComment       (Token &Result, const char *CurPtr,
728
                              bool &TokAtPhysicalStartOfLine);
729
  bool SkipBlockComment      (Token &Result, const char *CurPtr,
730
                              bool &TokAtPhysicalStartOfLine);
731
  bool SaveLineComment       (Token &Result, const char *CurPtr);
732
733
  bool IsStartOfConflictMarker(const char *CurPtr);
734
  bool HandleEndOfConflictMarker(const char *CurPtr);
735
736
  bool lexEditorPlaceholder(Token &Result, const char *CurPtr);
737
738
  bool isCodeCompletionPoint(const char *CurPtr) const;
739
66
  void cutOffLexing() { BufferPtr = BufferEnd; }
740
741
  bool isHexaLiteral(const char *Start, const LangOptions &LangOpts);
742
743
  void codeCompleteIncludedFile(const char *PathStart,
744
                                const char *CompletionPoint, bool IsAngled);
745
746
  /// Read a universal character name.
747
  ///
748
  /// \param StartPtr The position in the source buffer after the initial '\'.
749
  ///                 If the UCN is syntactically well-formed (but not 
750
  ///                 necessarily valid), this parameter will be updated to
751
  ///                 point to the character after the UCN.
752
  /// \param SlashLoc The position in the source buffer of the '\'.
753
  /// \param Result   The token being formed. Pass \c nullptr to suppress
754
  ///                 diagnostics and handle token formation in the caller.
755
  ///
756
  /// \return The Unicode codepoint specified by the UCN, or 0 if the UCN is
757
  ///         invalid.
758
  uint32_t tryReadUCN(const char *&StartPtr, const char *SlashLoc, Token *Result);
759
760
  /// Try to consume a UCN as part of an identifier at the current
761
  /// location.
762
  /// \param CurPtr Initially points to the range of characters in the source
763
  ///               buffer containing the '\'. Updated to point past the end of
764
  ///               the UCN on success.
765
  /// \param Size The number of characters occupied by the '\' (including
766
  ///             trigraphs and escaped newlines).
767
  /// \param Result The token being produced. Marked as containing a UCN on
768
  ///               success.
769
  /// \return \c true if a UCN was lexed and it produced an acceptable
770
  ///         identifier character, \c false otherwise.
771
  bool tryConsumeIdentifierUCN(const char *&CurPtr, unsigned Size,
772
                               Token &Result);
773
774
  /// Try to consume an identifier character encoded in UTF-8.
775
  /// \param CurPtr Points to the start of the (potential) UTF-8 code unit
776
  ///        sequence. On success, updated to point past the end of it.
777
  /// \return \c true if a UTF-8 sequence mapping to an acceptable identifier
778
  ///         character was lexed, \c false otherwise.
779
  bool tryConsumeIdentifierUTF8Char(const char *&CurPtr);
780
};
781
782
} // namespace clang
783
784
#endif // LLVM_CLANG_LEX_LEXER_H