1 //===--- Preprocessor.h - C Language Family Preprocessor --------*- C++ -*-===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
10 // This file defines the Preprocessor interface.
12 //===----------------------------------------------------------------------===//
14 #ifndef LLVM_CLANG_LEX_PREPROCESSOR_H
15 #define LLVM_CLANG_LEX_PREPROCESSOR_H
17 #include "clang/Lex/Lexer.h"
18 #include "clang/Lex/PTHLexer.h"
19 #include "clang/Lex/PPCallbacks.h"
20 #include "clang/Lex/TokenLexer.h"
21 #include "clang/Lex/PTHManager.h"
22 #include "clang/Basic/Diagnostic.h"
23 #include "clang/Basic/IdentifierTable.h"
24 #include "clang/Basic/SourceLocation.h"
25 #include "llvm/ADT/DenseMap.h"
26 #include "llvm/ADT/OwningPtr.h"
27 #include "llvm/Support/Allocator.h"
35 class PragmaNamespace;
40 class DirectoryLookup;
42 /// Preprocessor - This object engages in a tight little dance with the lexer to
43 /// efficiently preprocess tokens. Lexers know only about tokens within a
44 /// single source file, and don't know anything about preprocessor-level issues
45 /// like the #include stack, token expansion, etc.
49 const LangOptions &Features;
52 SourceManager &SourceMgr;
53 ScratchBuffer *ScratchBuf;
54 HeaderSearch &HeaderInfo;
56 /// PTH - An optional PTHManager object used for getting tokens from
57 /// a token cache rather than lexing the original source file.
58 llvm::OwningPtr<PTHManager> PTH;
60 /// BP - A BumpPtrAllocator object used to quickly allocate and release
61 /// objects internal to the Preprocessor.
62 llvm::BumpPtrAllocator BP;
64 /// Identifiers for builtin macros and other builtins.
65 IdentifierInfo *Ident__LINE__, *Ident__FILE__; // __LINE__, __FILE__
66 IdentifierInfo *Ident__DATE__, *Ident__TIME__; // __DATE__, __TIME__
67 IdentifierInfo *Ident__INCLUDE_LEVEL__; // __INCLUDE_LEVEL__
68 IdentifierInfo *Ident__BASE_FILE__; // __BASE_FILE__
69 IdentifierInfo *Ident__TIMESTAMP__; // __TIMESTAMP__
70 IdentifierInfo *Ident__COUNTER__; // __COUNTER__
71 IdentifierInfo *Ident_Pragma, *Ident__VA_ARGS__; // _Pragma, __VA_ARGS__
73 SourceLocation DATELoc, TIMELoc;
74 unsigned CounterValue; // Next __COUNTER__ value.
77 /// MaxIncludeStackDepth - Maximum depth of #includes.
78 MaxAllowedIncludeStackDepth = 200
81 // State that is set before the preprocessor begins.
82 bool KeepComments : 1;
83 bool KeepMacroComments : 1;
85 // State that changes while the preprocessor runs:
86 bool DisableMacroExpansion : 1; // True if macro expansion is disabled.
87 bool InMacroArgs : 1; // True if parsing fn macro invocation args.
89 /// Identifiers - This is mapping/lookup information for all identifiers in
90 /// the program, including program keywords.
91 IdentifierTable Identifiers;
93 /// Selectors - This table contains all the selectors in the program. Unlike
94 /// IdentifierTable above, this table *isn't* populated by the preprocessor.
95 /// It is declared/instantiated here because it's role/lifetime is
96 /// conceptually similar the IdentifierTable. In addition, the current control
97 /// flow (in clang::ParseAST()), make it convenient to put here.
98 /// FIXME: Make sure the lifetime of Identifiers/Selectors *isn't* tied to
99 /// the lifetime fo the preprocessor.
100 SelectorTable Selectors;
102 /// PragmaHandlers - This tracks all of the pragmas that the client registered
103 /// with this preprocessor.
104 PragmaNamespace *PragmaHandlers;
106 /// CurLexer - This is the current top of the stack that we're lexing from if
107 /// not expanding a macro and we are lexing directly from source code.
108 /// Only one of CurLexer, CurPTHLexer, or CurTokenLexer will be non-null.
109 llvm::OwningPtr<Lexer> CurLexer;
111 /// CurPTHLexer - This is the current top of stack that we're lexing from if
112 /// not expanding from a macro and we are lexing from a PTH cache.
113 /// Only one of CurLexer, CurPTHLexer, or CurTokenLexer will be non-null.
114 llvm::OwningPtr<PTHLexer> CurPTHLexer;
116 /// CurPPLexer - This is the current top of the stack what we're lexing from
117 /// if not expanding a macro. This is an alias for either CurLexer or
119 PreprocessorLexer* CurPPLexer;
121 /// CurLookup - The DirectoryLookup structure used to find the current
122 /// FileEntry, if CurLexer is non-null and if applicable. This allows us to
123 /// implement #include_next and find directory-specific properties.
124 const DirectoryLookup *CurDirLookup;
126 /// CurTokenLexer - This is the current macro we are expanding, if we are
127 /// expanding a macro. One of CurLexer and CurTokenLexer must be null.
128 llvm::OwningPtr<TokenLexer> CurTokenLexer;
130 /// IncludeMacroStack - This keeps track of the stack of files currently
131 /// #included, and macros currently being expanded from, not counting
132 /// CurLexer/CurTokenLexer.
133 struct IncludeStackInfo {
135 PTHLexer *ThePTHLexer;
136 PreprocessorLexer *ThePPLexer;
137 TokenLexer *TheTokenLexer;
138 const DirectoryLookup *TheDirLookup;
140 IncludeStackInfo(Lexer *L, PTHLexer* P, PreprocessorLexer* PPL,
141 TokenLexer* TL, const DirectoryLookup *D)
142 : TheLexer(L), ThePTHLexer(P), ThePPLexer(PPL), TheTokenLexer(TL),
145 std::vector<IncludeStackInfo> IncludeMacroStack;
147 /// Callbacks - These are actions invoked when some preprocessor activity is
148 /// encountered (e.g. a file is #included, etc).
149 PPCallbacks *Callbacks;
151 /// Macros - For each IdentifierInfo with 'HasMacro' set, we keep a mapping
152 /// to the actual definition of the macro.
153 llvm::DenseMap<IdentifierInfo*, MacroInfo*> Macros;
155 /// MICache - A "freelist" of MacroInfo objects that can be reused for quick
157 std::vector<MacroInfo*> MICache;
159 // Various statistics we track for performance analysis.
160 unsigned NumDirectives, NumIncluded, NumDefined, NumUndefined, NumPragma;
161 unsigned NumIf, NumElse, NumEndif;
162 unsigned NumEnteredSourceFiles, MaxIncludeStackDepth;
163 unsigned NumMacroExpanded, NumFnMacroExpanded, NumBuiltinMacroExpanded;
164 unsigned NumFastMacroExpanded, NumTokenPaste, NumFastTokenPaste;
167 /// Predefines - This string is the predefined macros that preprocessor
168 /// should use from the command line etc.
169 std::string Predefines;
171 /// TokenLexerCache - Cache macro expanders to reduce malloc traffic.
172 enum { TokenLexerCacheSize = 8 };
173 unsigned NumCachedTokenLexers;
174 TokenLexer *TokenLexerCache[TokenLexerCacheSize];
176 private: // Cached tokens state.
177 typedef std::vector<Token> CachedTokensTy;
179 /// CachedTokens - Cached tokens are stored here when we do backtracking or
180 /// lookahead. They are "lexed" by the CachingLex() method.
181 CachedTokensTy CachedTokens;
183 /// CachedLexPos - The position of the cached token that CachingLex() should
184 /// "lex" next. If it points beyond the CachedTokens vector, it means that
185 /// a normal Lex() should be invoked.
186 CachedTokensTy::size_type CachedLexPos;
188 /// BacktrackPositions - Stack of backtrack positions, allowing nested
189 /// backtracks. The EnableBacktrackAtThisPos() method pushes a position to
190 /// indicate where CachedLexPos should be set when the BackTrack() method is
191 /// invoked (at which point the last position is popped).
192 std::vector<CachedTokensTy::size_type> BacktrackPositions;
195 Preprocessor(Diagnostic &diags, const LangOptions &opts, TargetInfo &target,
196 SourceManager &SM, HeaderSearch &Headers,
197 IdentifierInfoLookup* IILookup = 0);
201 Diagnostic &getDiagnostics() const { return *Diags; }
202 void setDiagnostics(Diagnostic &D) { Diags = &D; }
205 const LangOptions &getLangOptions() const { return Features; }
206 TargetInfo &getTargetInfo() const { return Target; }
207 FileManager &getFileManager() const { return FileMgr; }
208 SourceManager &getSourceManager() const { return SourceMgr; }
209 HeaderSearch &getHeaderSearchInfo() const { return HeaderInfo; }
211 IdentifierTable &getIdentifierTable() { return Identifiers; }
212 SelectorTable &getSelectorTable() { return Selectors; }
213 llvm::BumpPtrAllocator &getPreprocessorAllocator() { return BP; }
215 void setPTHManager(PTHManager* pm);
217 PTHManager *getPTHManager() { return PTH.get(); }
219 /// SetCommentRetentionState - Control whether or not the preprocessor retains
220 /// comments in output.
221 void SetCommentRetentionState(bool KeepComments, bool KeepMacroComments) {
222 this->KeepComments = KeepComments | KeepMacroComments;
223 this->KeepMacroComments = KeepMacroComments;
226 bool getCommentRetentionState() const { return KeepComments; }
228 /// isCurrentLexer - Return true if we are lexing directly from the specified
230 bool isCurrentLexer(const PreprocessorLexer *L) const {
231 return CurPPLexer == L;
234 /// getCurrentLexer - Return the current file lexer being lexed from. Note
235 /// that this ignores any potentially active macro expansions and _Pragma
236 /// expansions going on at the time.
237 PreprocessorLexer *getCurrentFileLexer() const;
239 /// getPPCallbacks/setPPCallbacks - Accessors for preprocessor callbacks.
240 /// Note that this class takes ownership of any PPCallbacks object given to
242 PPCallbacks *getPPCallbacks() const { return Callbacks; }
243 void setPPCallbacks(PPCallbacks *C) {
245 C = new PPChainedCallbacks(C, Callbacks);
249 /// getMacroInfo - Given an identifier, return the MacroInfo it is #defined to
250 /// or null if it isn't #define'd.
251 MacroInfo *getMacroInfo(IdentifierInfo *II) const {
252 return II->hasMacroDefinition() ? Macros.find(II)->second : 0;
255 /// setMacroInfo - Specify a macro for this identifier.
257 void setMacroInfo(IdentifierInfo *II, MacroInfo *MI);
259 /// macro_iterator/macro_begin/macro_end - This allows you to walk the current
260 /// state of the macro table. This visits every currently-defined macro.
261 typedef llvm::DenseMap<IdentifierInfo*,
262 MacroInfo*>::const_iterator macro_iterator;
263 macro_iterator macro_begin() const { return Macros.begin(); }
264 macro_iterator macro_end() const { return Macros.end(); }
268 const std::string &getPredefines() const { return Predefines; }
269 /// setPredefines - Set the predefines for this Preprocessor. These
270 /// predefines are automatically injected when parsing the main file.
271 void setPredefines(const char *P) { Predefines = P; }
272 void setPredefines(const std::string &P) { Predefines = P; }
274 /// getIdentifierInfo - Return information about the specified preprocessor
275 /// identifier token. The version of this method that takes two character
276 /// pointers is preferred unless the identifier is already available as a
277 /// string (this avoids allocation and copying of memory to construct an
279 IdentifierInfo *getIdentifierInfo(const char *NameStart,
280 const char *NameEnd) {
281 return &Identifiers.get(NameStart, NameEnd);
283 IdentifierInfo *getIdentifierInfo(const char *NameStr) {
284 return getIdentifierInfo(NameStr, NameStr+strlen(NameStr));
287 /// AddPragmaHandler - Add the specified pragma handler to the preprocessor.
288 /// If 'Namespace' is non-null, then it is a token required to exist on the
289 /// pragma line before the pragma string starts, e.g. "STDC" or "GCC".
290 void AddPragmaHandler(const char *Namespace, PragmaHandler *Handler);
292 /// RemovePragmaHandler - Remove the specific pragma handler from
293 /// the preprocessor. If \arg Namespace is non-null, then it should
294 /// be the namespace that \arg Handler was added to. It is an error
295 /// to remove a handler that has not been registered.
296 void RemovePragmaHandler(const char *Namespace, PragmaHandler *Handler);
298 /// EnterMainSourceFile - Enter the specified FileID as the main source file,
299 /// which implicitly adds the builtin defines etc.
300 void EnterMainSourceFile();
302 /// EnterSourceFile - Add a source file to the top of the include stack and
303 /// start lexing tokens from it instead of the current buffer. If isMainFile
304 /// is true, this is the main file for the translation unit.
305 void EnterSourceFile(FileID CurFileID, const DirectoryLookup *Dir);
307 /// EnterMacro - Add a Macro to the top of the include stack and start lexing
308 /// tokens from it instead of the current buffer. Args specifies the
309 /// tokens input to a function-like macro.
311 /// ILEnd specifies the location of the ')' for a function-like macro or the
312 /// identifier for an object-like macro.
313 void EnterMacro(Token &Identifier, SourceLocation ILEnd, MacroArgs *Args);
315 /// EnterTokenStream - Add a "macro" context to the top of the include stack,
316 /// which will cause the lexer to start returning the specified tokens.
318 /// If DisableMacroExpansion is true, tokens lexed from the token stream will
319 /// not be subject to further macro expansion. Otherwise, these tokens will
320 /// be re-macro-expanded when/if expansion is enabled.
322 /// If OwnsTokens is false, this method assumes that the specified stream of
323 /// tokens has a permanent owner somewhere, so they do not need to be copied.
324 /// If it is true, it assumes the array of tokens is allocated with new[] and
327 void EnterTokenStream(const Token *Toks, unsigned NumToks,
328 bool DisableMacroExpansion, bool OwnsTokens);
330 /// RemoveTopOfLexerStack - Pop the current lexer/macro exp off the top of the
331 /// lexer stack. This should only be used in situations where the current
332 /// state of the top-of-stack lexer is known.
333 void RemoveTopOfLexerStack();
335 /// EnableBacktrackAtThisPos - From the point that this method is called, and
336 /// until CommitBacktrackedTokens() or Backtrack() is called, the Preprocessor
337 /// keeps track of the lexed tokens so that a subsequent Backtrack() call will
338 /// make the Preprocessor re-lex the same tokens.
340 /// Nested backtracks are allowed, meaning that EnableBacktrackAtThisPos can
341 /// be called multiple times and CommitBacktrackedTokens/Backtrack calls will
342 /// be combined with the EnableBacktrackAtThisPos calls in reverse order.
344 /// NOTE: *DO NOT* forget to call either CommitBacktrackedTokens or Backtrack
345 /// at some point after EnableBacktrackAtThisPos. If you don't, caching of
346 /// tokens will continue indefinitely.
348 void EnableBacktrackAtThisPos();
350 /// CommitBacktrackedTokens - Disable the last EnableBacktrackAtThisPos call.
351 void CommitBacktrackedTokens();
353 /// Backtrack - Make Preprocessor re-lex the tokens that were lexed since
354 /// EnableBacktrackAtThisPos() was previously called.
357 /// isBacktrackEnabled - True if EnableBacktrackAtThisPos() was called and
358 /// caching of tokens is on.
359 bool isBacktrackEnabled() const { return !BacktrackPositions.empty(); }
361 /// Lex - To lex a token from the preprocessor, just pull a token from the
362 /// current lexer or macro object.
363 void Lex(Token &Result) {
365 CurLexer->Lex(Result);
366 else if (CurPTHLexer)
367 CurPTHLexer->Lex(Result);
368 else if (CurTokenLexer)
369 CurTokenLexer->Lex(Result);
374 /// LexNonComment - Lex a token. If it's a comment, keep lexing until we get
375 /// something not a comment. This is useful in -E -C mode where comments
376 /// would foul up preprocessor directive handling.
377 void LexNonComment(Token &Result) {
380 while (Result.getKind() == tok::comment);
383 /// LexUnexpandedToken - This is just like Lex, but this disables macro
384 /// expansion of identifier tokens.
385 void LexUnexpandedToken(Token &Result) {
386 // Disable macro expansion.
387 bool OldVal = DisableMacroExpansion;
388 DisableMacroExpansion = true;
393 DisableMacroExpansion = OldVal;
396 /// LookAhead - This peeks ahead N tokens and returns that token without
397 /// consuming any tokens. LookAhead(0) returns the next token that would be
398 /// returned by Lex(), LookAhead(1) returns the token after it, etc. This
399 /// returns normal tokens after phase 5. As such, it is equivalent to using
400 /// 'Lex', not 'LexUnexpandedToken'.
401 const Token &LookAhead(unsigned N) {
402 if (CachedLexPos + N < CachedTokens.size())
403 return CachedTokens[CachedLexPos+N];
405 return PeekAhead(N+1);
408 /// RevertCachedTokens - When backtracking is enabled and tokens are cached,
409 /// this allows to revert a specific number of tokens.
410 /// Note that the number of tokens being reverted should be up to the last
411 /// backtrack position, not more.
412 void RevertCachedTokens(unsigned N) {
413 assert(isBacktrackEnabled() &&
414 "Should only be called when tokens are cached for backtracking");
415 assert(signed(CachedLexPos) - signed(N) >= signed(BacktrackPositions.back())
416 && "Should revert tokens up to the last backtrack position, not more");
417 assert(signed(CachedLexPos) - signed(N) >= 0 &&
418 "Corrupted backtrack positions ?");
422 /// EnterToken - Enters a token in the token stream to be lexed next. If
423 /// BackTrack() is called afterwards, the token will remain at the insertion
425 void EnterToken(const Token &Tok) {
426 EnterCachingLexMode();
427 CachedTokens.insert(CachedTokens.begin()+CachedLexPos, Tok);
430 /// AnnotateCachedTokens - We notify the Preprocessor that if it is caching
431 /// tokens (because backtrack is enabled) it should replace the most recent
432 /// cached tokens with the given annotation token. This function has no effect
433 /// if backtracking is not enabled.
435 /// Note that the use of this function is just for optimization; so that the
436 /// cached tokens doesn't get re-parsed and re-resolved after a backtrack is
438 void AnnotateCachedTokens(const Token &Tok) {
439 assert(Tok.isAnnotation() && "Expected annotation token");
440 if (CachedLexPos != 0 && isBacktrackEnabled())
441 AnnotatePreviousCachedTokens(Tok);
444 /// \brief Replace the last token with an annotation token.
446 /// Like AnnotateCachedTokens(), this routine replaces an
447 /// already-parsed (and resolved) token with an annotation
448 /// token. However, this routine only replaces the last token with
449 /// the annotation token; it does not affect any other cached
450 /// tokens. This function has no effect if backtracking is not
452 void ReplaceLastTokenWithAnnotation(const Token &Tok) {
453 assert(Tok.isAnnotation() && "Expected annotation token");
454 if (CachedLexPos != 0 && isBacktrackEnabled())
455 CachedTokens[CachedLexPos-1] = Tok;
458 /// Diag - Forwarding function for diagnostics. This emits a diagnostic at
459 /// the specified Token's location, translating the token's start
460 /// position in the current buffer into a SourcePosition object for rendering.
461 DiagnosticBuilder Diag(SourceLocation Loc, unsigned DiagID) {
462 return Diags->Report(FullSourceLoc(Loc, getSourceManager()), DiagID);
465 DiagnosticBuilder Diag(const Token &Tok, unsigned DiagID) {
466 return Diags->Report(FullSourceLoc(Tok.getLocation(), getSourceManager()),
470 /// getSpelling() - Return the 'spelling' of the Tok token. The spelling of a
471 /// token is the characters used to represent the token in the source file
472 /// after trigraph expansion and escaped-newline folding. In particular, this
473 /// wants to get the true, uncanonicalized, spelling of things like digraphs
475 std::string getSpelling(const Token &Tok) const;
477 /// getSpelling - This method is used to get the spelling of a token into a
478 /// preallocated buffer, instead of as an std::string. The caller is required
479 /// to allocate enough space for the token, which is guaranteed to be at least
480 /// Tok.getLength() bytes long. The length of the actual result is returned.
482 /// Note that this method may do two possible things: it may either fill in
483 /// the buffer specified with characters, or it may *change the input pointer*
484 /// to point to a constant buffer with the data already in it (avoiding a
485 /// copy). The caller is not allowed to modify the returned buffer pointer
486 /// if an internal buffer is returned.
487 unsigned getSpelling(const Token &Tok, const char *&Buffer) const;
489 /// getSpellingOfSingleCharacterNumericConstant - Tok is a numeric constant
490 /// with length 1, return the character.
491 char getSpellingOfSingleCharacterNumericConstant(const Token &Tok) const {
492 assert(Tok.is(tok::numeric_constant) &&
493 Tok.getLength() == 1 && "Called on unsupported token");
494 assert(!Tok.needsCleaning() && "Token can't need cleaning with length 1");
496 // If the token is carrying a literal data pointer, just use it.
497 if (const char *D = Tok.getLiteralData())
500 // Otherwise, fall back on getCharacterData, which is slower, but always
502 return *SourceMgr.getCharacterData(Tok.getLocation());
505 /// CreateString - Plop the specified string into a scratch buffer and set the
506 /// specified token's location and length to it. If specified, the source
507 /// location provides a location of the instantiation point of the token.
508 void CreateString(const char *Buf, unsigned Len,
509 Token &Tok, SourceLocation SourceLoc = SourceLocation());
511 /// \brief Computes the source location just past the end of the
512 /// token at this source location.
514 /// This routine can be used to produce a source location that
515 /// points just past the end of the token referenced by \p Loc, and
516 /// is generally used when a diagnostic needs to point just after a
517 /// token where it expected something different that it received. If
518 /// the returned source location would not be meaningful (e.g., if
519 /// it points into a macro), this routine returns an invalid
521 SourceLocation getLocForEndOfToken(SourceLocation Loc);
523 /// DumpToken - Print the token to stderr, used for debugging.
525 void DumpToken(const Token &Tok, bool DumpFlags = false) const;
526 void DumpLocation(SourceLocation Loc) const;
527 void DumpMacro(const MacroInfo &MI) const;
529 /// AdvanceToTokenCharacter - Given a location that specifies the start of a
530 /// token, return a new location that specifies a character within the token.
531 SourceLocation AdvanceToTokenCharacter(SourceLocation TokStart,unsigned Char);
533 /// IncrementPasteCounter - Increment the counters for the number of token
534 /// paste operations performed. If fast was specified, this is a 'fast paste'
537 void IncrementPasteCounter(bool isFast) {
546 /// HandleMicrosoftCommentPaste - When the macro expander pastes together a
547 /// comment (/##/) in microsoft mode, this method handles updating the current
548 /// state, returning the token on the next source line.
549 void HandleMicrosoftCommentPaste(Token &Tok);
551 //===--------------------------------------------------------------------===//
552 // Preprocessor callback methods. These are invoked by a lexer as various
553 // directives and events are found.
555 /// LookUpIdentifierInfo - Given a tok::identifier token, look up the
556 /// identifier information for the token and install it into the token.
557 IdentifierInfo *LookUpIdentifierInfo(Token &Identifier,
558 const char *BufPtr = 0);
560 /// HandleIdentifier - This callback is invoked when the lexer reads an
561 /// identifier and has filled in the tokens IdentifierInfo member. This
562 /// callback potentially macro expands it or turns it into a named token (like
564 void HandleIdentifier(Token &Identifier);
567 /// HandleEndOfFile - This callback is invoked when the lexer hits the end of
568 /// the current file. This either returns the EOF token and returns true, or
569 /// pops a level off the include stack and returns false, at which point the
570 /// client should call lex again.
571 bool HandleEndOfFile(Token &Result, bool isEndOfMacro = false);
573 /// HandleEndOfTokenLexer - This callback is invoked when the current
574 /// TokenLexer hits the end of its token stream.
575 bool HandleEndOfTokenLexer(Token &Result);
577 /// HandleDirective - This callback is invoked when the lexer sees a # token
578 /// at the start of a line. This consumes the directive, modifies the
579 /// lexer/preprocessor state, and advances the lexer(s) so that the next token
580 /// read is the correct one.
581 void HandleDirective(Token &Result);
583 /// CheckEndOfDirective - Ensure that the next token is a tok::eom token. If
584 /// not, emit a diagnostic and consume up until the eom. If EnableMacros is
585 /// true, then we consider macros that expand to zero tokens as being ok.
586 void CheckEndOfDirective(const char *Directive, bool EnableMacros = false);
588 /// DiscardUntilEndOfDirective - Read and discard all tokens remaining on the
589 /// current line until the tok::eom token is found.
590 void DiscardUntilEndOfDirective();
592 /// SawDateOrTime - This returns true if the preprocessor has seen a use of
593 /// __DATE__ or __TIME__ in the file so far.
594 bool SawDateOrTime() const {
595 return DATELoc != SourceLocation() || TIMELoc != SourceLocation();
597 unsigned getCounterValue() const { return CounterValue; }
598 void setCounterValue(unsigned V) { CounterValue = V; }
600 /// AllocateMacroInfo - Allocate a new MacroInfo object with the provide
602 MacroInfo* AllocateMacroInfo(SourceLocation L);
606 void PushIncludeMacroStack() {
607 IncludeMacroStack.push_back(IncludeStackInfo(CurLexer.take(),
610 CurTokenLexer.take(),
615 void PopIncludeMacroStack() {
616 CurLexer.reset(IncludeMacroStack.back().TheLexer);
617 CurPTHLexer.reset(IncludeMacroStack.back().ThePTHLexer);
618 CurPPLexer = IncludeMacroStack.back().ThePPLexer;
619 CurTokenLexer.reset(IncludeMacroStack.back().TheTokenLexer);
620 CurDirLookup = IncludeMacroStack.back().TheDirLookup;
621 IncludeMacroStack.pop_back();
624 /// ReleaseMacroInfo - Release the specified MacroInfo. This memory will
625 /// be reused for allocating new MacroInfo objects.
626 void ReleaseMacroInfo(MacroInfo* MI);
628 /// isInPrimaryFile - Return true if we're in the top-level file, not in a
630 bool isInPrimaryFile() const;
632 /// ReadMacroName - Lex and validate a macro name, which occurs after a
633 /// #define or #undef. This emits a diagnostic, sets the token kind to eom,
634 /// and discards the rest of the macro line if the macro name is invalid.
635 void ReadMacroName(Token &MacroNameTok, char isDefineUndef = 0);
637 /// ReadMacroDefinitionArgList - The ( starting an argument list of a macro
638 /// definition has just been read. Lex the rest of the arguments and the
639 /// closing ), updating MI with what we learn. Return true if an error occurs
640 /// parsing the arg list.
641 bool ReadMacroDefinitionArgList(MacroInfo *MI);
643 /// SkipExcludedConditionalBlock - We just read a #if or related directive and
644 /// decided that the subsequent tokens are in the #if'd out portion of the
645 /// file. Lex the rest of the file, until we see an #endif. If
646 /// FoundNonSkipPortion is true, then we have already emitted code for part of
647 /// this #if directive, so #else/#elif blocks should never be entered. If
648 /// FoundElse is false, then #else directives are ok, if not, then we have
649 /// already seen one so a #else directive is a duplicate. When this returns,
650 /// the caller can lex the first valid token.
651 void SkipExcludedConditionalBlock(SourceLocation IfTokenLoc,
652 bool FoundNonSkipPortion, bool FoundElse);
654 /// PTHSkipExcludedConditionalBlock - A fast PTH version of
655 /// SkipExcludedConditionalBlock.
656 void PTHSkipExcludedConditionalBlock();
658 /// EvaluateDirectiveExpression - Evaluate an integer constant expression that
659 /// may occur after a #if or #elif directive and return it as a bool. If the
660 /// expression is equivalent to "!defined(X)" return X in IfNDefMacro.
661 bool EvaluateDirectiveExpression(IdentifierInfo *&IfNDefMacro);
663 /// RegisterBuiltinPragmas - Install the standard preprocessor pragmas:
664 /// #pragma GCC poison/system_header/dependency and #pragma once.
665 void RegisterBuiltinPragmas();
667 /// RegisterBuiltinMacros - Register builtin macros, such as __LINE__ with the
668 /// identifier table.
669 void RegisterBuiltinMacros();
670 IdentifierInfo *RegisterBuiltinMacro(const char *Name);
672 /// HandleMacroExpandedIdentifier - If an identifier token is read that is to
673 /// be expanded as a macro, handle it and return the next token as 'Tok'. If
674 /// the macro should not be expanded return true, otherwise return false.
675 bool HandleMacroExpandedIdentifier(Token &Tok, MacroInfo *MI);
677 /// isNextPPTokenLParen - Determine whether the next preprocessor token to be
678 /// lexed is a '('. If so, consume the token and return true, if not, this
679 /// method should have no observable side-effect on the lexed tokens.
680 bool isNextPPTokenLParen();
682 /// ReadFunctionLikeMacroArgs - After reading "MACRO(", this method is
683 /// invoked to read all of the formal arguments specified for the macro
684 /// invocation. This returns null on error.
685 MacroArgs *ReadFunctionLikeMacroArgs(Token &MacroName, MacroInfo *MI,
686 SourceLocation &InstantiationEnd);
688 /// ExpandBuiltinMacro - If an identifier token is read that is to be expanded
689 /// as a builtin macro, handle it and return the next token as 'Tok'.
690 void ExpandBuiltinMacro(Token &Tok);
692 /// Handle_Pragma - Read a _Pragma directive, slice it up, process it, then
693 /// return the first token after the directive. The _Pragma token has just
694 /// been read into 'Tok'.
695 void Handle_Pragma(Token &Tok);
697 /// EnterSourceFileWithLexer - Add a lexer to the top of the include stack and
698 /// start lexing tokens from it instead of the current buffer.
699 void EnterSourceFileWithLexer(Lexer *TheLexer, const DirectoryLookup *Dir);
701 /// EnterSourceFileWithPTH - Add a lexer to the top of the include stack and
702 /// start getting tokens from it using the PTH cache.
703 void EnterSourceFileWithPTH(PTHLexer *PL, const DirectoryLookup *Dir);
705 /// GetIncludeFilenameSpelling - Turn the specified lexer token into a fully
706 /// checked and spelled filename, e.g. as an operand of #include. This returns
707 /// true if the input filename was in <>'s or false if it were in ""'s. The
708 /// caller is expected to provide a buffer that is large enough to hold the
709 /// spelling of the filename, but is also expected to handle the case when
710 /// this method decides to use a different buffer.
711 bool GetIncludeFilenameSpelling(SourceLocation Loc,
712 const char *&BufStart, const char *&BufEnd);
714 /// LookupFile - Given a "foo" or <foo> reference, look up the indicated file,
715 /// return null on failure. isAngled indicates whether the file reference is
716 /// for system #include's or not (i.e. using <> instead of "").
717 const FileEntry *LookupFile(const char *FilenameStart,const char *FilenameEnd,
718 bool isAngled, const DirectoryLookup *FromDir,
719 const DirectoryLookup *&CurDir);
723 /// IsFileLexer - Returns true if we are lexing from a file and not a
724 /// pragma or a macro.
725 static bool IsFileLexer(const Lexer* L, const PreprocessorLexer* P) {
726 return L ? !L->isPragmaLexer() : P != 0;
729 static bool IsFileLexer(const IncludeStackInfo& I) {
730 return IsFileLexer(I.TheLexer, I.ThePPLexer);
733 bool IsFileLexer() const {
734 return IsFileLexer(CurLexer.get(), CurPPLexer);
737 //===--------------------------------------------------------------------===//
739 void CachingLex(Token &Result);
740 bool InCachingLexMode() const { return CurPPLexer == 0 && CurTokenLexer == 0;}
741 void EnterCachingLexMode();
742 void ExitCachingLexMode() {
743 if (InCachingLexMode())
744 RemoveTopOfLexerStack();
746 const Token &PeekAhead(unsigned N);
747 void AnnotatePreviousCachedTokens(const Token &Tok);
749 //===--------------------------------------------------------------------===//
750 /// Handle*Directive - implement the various preprocessor directives. These
751 /// should side-effect the current preprocessor object so that the next call
752 /// to Lex() will return the appropriate token next.
753 void HandleLineDirective(Token &Tok);
754 void HandleDigitDirective(Token &Tok);
755 void HandleUserDiagnosticDirective(Token &Tok, bool isWarning);
756 void HandleIdentSCCSDirective(Token &Tok);
759 void HandleIncludeDirective(Token &Tok,
760 const DirectoryLookup *LookupFrom = 0,
761 bool isImport = false);
762 void HandleIncludeNextDirective(Token &Tok);
763 void HandleIncludeMacrosDirective(Token &Tok);
764 void HandleImportDirective(Token &Tok);
767 void HandleDefineDirective(Token &Tok);
768 void HandleUndefDirective(Token &Tok);
769 // HandleAssertDirective(Token &Tok);
770 // HandleUnassertDirective(Token &Tok);
772 // Conditional Inclusion.
773 void HandleIfdefDirective(Token &Tok, bool isIfndef,
774 bool ReadAnyTokensBeforeDirective);
775 void HandleIfDirective(Token &Tok, bool ReadAnyTokensBeforeDirective);
776 void HandleEndifDirective(Token &Tok);
777 void HandleElseDirective(Token &Tok);
778 void HandleElifDirective(Token &Tok);
781 void HandlePragmaDirective();
783 void HandlePragmaOnce(Token &OnceTok);
784 void HandlePragmaMark();
785 void HandlePragmaPoison(Token &PoisonTok);
786 void HandlePragmaSystemHeader(Token &SysHeaderTok);
787 void HandlePragmaDependency(Token &DependencyTok);
788 void HandlePragmaComment(Token &CommentTok);
791 /// PreprocessorFactory - A generic factory interface for lazily creating
792 /// Preprocessor objects on-demand when they are needed.
793 class PreprocessorFactory {
795 virtual ~PreprocessorFactory();
796 virtual Preprocessor* CreatePreprocessor() = 0;
799 } // end namespace clang