Preprocessor.cpp revision 6be16fe900bdd1e5f677d23ae34fffead5bcfc77 
1//===--- Preprocess.cpp - C Language Family Preprocessor Implementation ---===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10//  This file implements the Preprocessor interface.
11//
12//===----------------------------------------------------------------------===//
13//
14// Options to support:
15//   -H       - Print the name of each header file used.
16//   -d[DNI] - Dump various things.
17//   -fworking-directory - #line's with preprocessor's working dir.
18//   -fpreprocessed
19//   -dependency-file,-M,-MM,-MF,-MG,-MP,-MT,-MQ,-MD,-MMD
20//   -W*
21//   -w
22//
23// Messages to emit:
24//   "Multiple include guards may be useful for:\n"
25//
26//===----------------------------------------------------------------------===//
27
28#include "clang/Lex/Preprocessor.h"
29#include "MacroArgs.h"
30#include "clang/Lex/ExternalPreprocessorSource.h"
31#include "clang/Lex/HeaderSearch.h"
32#include "clang/Lex/MacroInfo.h"
33#include "clang/Lex/Pragma.h"
34#include "clang/Lex/PreprocessingRecord.h"
35#include "clang/Lex/ScratchBuffer.h"
36#include "clang/Lex/LexDiagnostic.h"
37#include "clang/Lex/CodeCompletionHandler.h"
38#include "clang/Lex/ModuleLoader.h"
39#include "clang/Basic/SourceManager.h"
40#include "clang/Basic/FileManager.h"
41#include "clang/Basic/TargetInfo.h"
42#include "llvm/ADT/APFloat.h"
43#include "llvm/ADT/SmallVector.h"
44#include "llvm/Support/MemoryBuffer.h"
45#include "llvm/Support/raw_ostream.h"
46#include "llvm/Support/Capacity.h"
47using namespace clang;
48
49//===----------------------------------------------------------------------===//
50ExternalPreprocessorSource::~ExternalPreprocessorSource() { }
51
52Preprocessor::Preprocessor(Diagnostic &diags, const LangOptions &opts,
53                           const TargetInfo &target, SourceManager &SM,
54                           HeaderSearch &Headers, ModuleLoader &TheModuleLoader,
55                           IdentifierInfoLookup* IILookup,
56                           bool OwnsHeaders)
57  : Diags(&diags), Features(opts), Target(target),FileMgr(Headers.getFileMgr()),
58    SourceMgr(SM), HeaderInfo(Headers), TheModuleLoader(TheModuleLoader),
59    ExternalSource(0),
60    Identifiers(opts, IILookup), BuiltinInfo(Target), CodeComplete(0),
61    CodeCompletionFile(0), SkipMainFilePreamble(0, true), CurPPLexer(0),
62    CurDirLookup(0), Callbacks(0), MacroArgCache(0), Record(0), MIChainHead(0),
63    MICache(0) {
64  ScratchBuf = new ScratchBuffer(SourceMgr);
65  CounterValue = 0; // __COUNTER__ starts at 0.
66  OwnsHeaderSearch = OwnsHeaders;
67
68  // Clear stats.
69  NumDirectives = NumDefined = NumUndefined = NumPragma = 0;
70  NumIf = NumElse = NumEndif = 0;
71  NumEnteredSourceFiles = 0;
72  NumMacroExpanded = NumFnMacroExpanded = NumBuiltinMacroExpanded = 0;
73  NumFastMacroExpanded = NumTokenPaste = NumFastTokenPaste = 0;
74  MaxIncludeStackDepth = 0;
75  NumSkipped = 0;
76
77  // Default to discarding comments.
78  KeepComments = false;
79  KeepMacroComments = false;
80
81  // Macro expansion is enabled.
82  DisableMacroExpansion = false;
83  InMacroArgs = false;
84  NumCachedTokenLexers = 0;
85
86  CachedLexPos = 0;
87
88  // We haven't read anything from the external source.
89  ReadMacrosFromExternalSource = false;
90
91  LexDepth = 0;
92
93  // "Poison" __VA_ARGS__, which can only appear in the expansion of a macro.
94  // This gets unpoisoned where it is allowed.
95  (Ident__VA_ARGS__ = getIdentifierInfo("__VA_ARGS__"))->setIsPoisoned();
96  SetPoisonReason(Ident__VA_ARGS__,diag::ext_pp_bad_vaargs_use);
97
98  // Initialize the pragma handlers.
99  PragmaHandlers = new PragmaNamespace(StringRef());
100  RegisterBuiltinPragmas();
101
102  // Initialize builtin macros like __LINE__ and friends.
103  RegisterBuiltinMacros();
104
105  if(Features.Borland) {
106    Ident__exception_info        = getIdentifierInfo("_exception_info");
107    Ident___exception_info       = getIdentifierInfo("__exception_info");
108    Ident_GetExceptionInfo       = getIdentifierInfo("GetExceptionInformation");
109    Ident__exception_code        = getIdentifierInfo("_exception_code");
110    Ident___exception_code       = getIdentifierInfo("__exception_code");
111    Ident_GetExceptionCode       = getIdentifierInfo("GetExceptionCode");
112    Ident__abnormal_termination  = getIdentifierInfo("_abnormal_termination");
113    Ident___abnormal_termination = getIdentifierInfo("__abnormal_termination");
114    Ident_AbnormalTermination    = getIdentifierInfo("AbnormalTermination");
115  } else {
116    Ident__exception_info = Ident__exception_code = Ident__abnormal_termination = 0;
117    Ident___exception_info = Ident___exception_code = Ident___abnormal_termination = 0;
118    Ident_GetExceptionInfo = Ident_GetExceptionCode = Ident_AbnormalTermination = 0;
119  }
120
121}
122
123Preprocessor::~Preprocessor() {
124  assert(BacktrackPositions.empty() && "EnableBacktrack/Backtrack imbalance!");
125  assert(MacroExpandingLexersStack.empty() && MacroExpandedTokens.empty() &&
126         "Preprocessor::HandleEndOfTokenLexer should have cleared those");
127
128  while (!IncludeMacroStack.empty()) {
129    delete IncludeMacroStack.back().TheLexer;
130    delete IncludeMacroStack.back().TheTokenLexer;
131    IncludeMacroStack.pop_back();
132  }
133
134  // Free any macro definitions.
135  for (MacroInfoChain *I = MIChainHead ; I ; I = I->Next)
136    I->MI.Destroy();
137
138  // Free any cached macro expanders.
139  for (unsigned i = 0, e = NumCachedTokenLexers; i != e; ++i)
140    delete TokenLexerCache[i];
141
142  // Free any cached MacroArgs.
143  for (MacroArgs *ArgList = MacroArgCache; ArgList; )
144    ArgList = ArgList->deallocate();
145
146  // Release pragma information.
147  delete PragmaHandlers;
148
149  // Delete the scratch buffer info.
150  delete ScratchBuf;
151
152  // Delete the header search info, if we own it.
153  if (OwnsHeaderSearch)
154    delete &HeaderInfo;
155
156  delete Callbacks;
157}
158
159void Preprocessor::setPTHManager(PTHManager* pm) {
160  PTH.reset(pm);
161  FileMgr.addStatCache(PTH->createStatCache());
162}
163
164void Preprocessor::DumpToken(const Token &Tok, bool DumpFlags) const {
165  llvm::errs() << tok::getTokenName(Tok.getKind()) << " '"
166               << getSpelling(Tok) << "'";
167
168  if (!DumpFlags) return;
169
170  llvm::errs() << "\t";
171  if (Tok.isAtStartOfLine())
172    llvm::errs() << " [StartOfLine]";
173  if (Tok.hasLeadingSpace())
174    llvm::errs() << " [LeadingSpace]";
175  if (Tok.isExpandDisabled())
176    llvm::errs() << " [ExpandDisabled]";
177  if (Tok.needsCleaning()) {
178    const char *Start = SourceMgr.getCharacterData(Tok.getLocation());
179    llvm::errs() << " [UnClean='" << StringRef(Start, Tok.getLength())
180                 << "']";
181  }
182
183  llvm::errs() << "\tLoc=<";
184  DumpLocation(Tok.getLocation());
185  llvm::errs() << ">";
186}
187
188void Preprocessor::DumpLocation(SourceLocation Loc) const {
189  Loc.dump(SourceMgr);
190}
191
192void Preprocessor::DumpMacro(const MacroInfo &MI) const {
193  llvm::errs() << "MACRO: ";
194  for (unsigned i = 0, e = MI.getNumTokens(); i != e; ++i) {
195    DumpToken(MI.getReplacementToken(i));
196    llvm::errs() << "  ";
197  }
198  llvm::errs() << "\n";
199}
200
201void Preprocessor::PrintStats() {
202  llvm::errs() << "\n*** Preprocessor Stats:\n";
203  llvm::errs() << NumDirectives << " directives found:\n";
204  llvm::errs() << "  " << NumDefined << " #define.\n";
205  llvm::errs() << "  " << NumUndefined << " #undef.\n";
206  llvm::errs() << "  #include/#include_next/#import:\n";
207  llvm::errs() << "    " << NumEnteredSourceFiles << " source files entered.\n";
208  llvm::errs() << "    " << MaxIncludeStackDepth << " max include stack depth\n";
209  llvm::errs() << "  " << NumIf << " #if/#ifndef/#ifdef.\n";
210  llvm::errs() << "  " << NumElse << " #else/#elif.\n";
211  llvm::errs() << "  " << NumEndif << " #endif.\n";
212  llvm::errs() << "  " << NumPragma << " #pragma.\n";
213  llvm::errs() << NumSkipped << " #if/#ifndef#ifdef regions skipped\n";
214
215  llvm::errs() << NumMacroExpanded << "/" << NumFnMacroExpanded << "/"
216             << NumBuiltinMacroExpanded << " obj/fn/builtin macros expanded, "
217             << NumFastMacroExpanded << " on the fast path.\n";
218  llvm::errs() << (NumFastTokenPaste+NumTokenPaste)
219             << " token paste (##) operations performed, "
220             << NumFastTokenPaste << " on the fast path.\n";
221}
222
223Preprocessor::macro_iterator
224Preprocessor::macro_begin(bool IncludeExternalMacros) const {
225  if (IncludeExternalMacros && ExternalSource &&
226      !ReadMacrosFromExternalSource) {
227    ReadMacrosFromExternalSource = true;
228    ExternalSource->ReadDefinedMacros();
229  }
230
231  return Macros.begin();
232}
233
234size_t Preprocessor::getTotalMemory() const {
235  return BP.getTotalMemory()
236    + llvm::capacity_in_bytes(MacroExpandedTokens)
237    + Predefines.capacity() /* Predefines buffer. */
238    + llvm::capacity_in_bytes(Macros)
239    + llvm::capacity_in_bytes(PragmaPushMacroInfo)
240    + llvm::capacity_in_bytes(PoisonReasons)
241    + llvm::capacity_in_bytes(CommentHandlers);
242}
243
244Preprocessor::macro_iterator
245Preprocessor::macro_end(bool IncludeExternalMacros) const {
246  if (IncludeExternalMacros && ExternalSource &&
247      !ReadMacrosFromExternalSource) {
248    ReadMacrosFromExternalSource = true;
249    ExternalSource->ReadDefinedMacros();
250  }
251
252  return Macros.end();
253}
254
255bool Preprocessor::SetCodeCompletionPoint(const FileEntry *File,
256                                          unsigned TruncateAtLine,
257                                          unsigned TruncateAtColumn) {
258  using llvm::MemoryBuffer;
259
260  CodeCompletionFile = File;
261
262  // Okay to clear out the code-completion point by passing NULL.
263  if (!CodeCompletionFile)
264    return false;
265
266  // Load the actual file's contents.
267  bool Invalid = false;
268  const MemoryBuffer *Buffer = SourceMgr.getMemoryBufferForFile(File, &Invalid);
269  if (Invalid)
270    return true;
271
272  // Find the byte position of the truncation point.
273  const char *Position = Buffer->getBufferStart();
274  for (unsigned Line = 1; Line < TruncateAtLine; ++Line) {
275    for (; *Position; ++Position) {
276      if (*Position != '\r' && *Position != '\n')
277        continue;
278
279      // Eat \r\n or \n\r as a single line.
280      if ((Position[1] == '\r' || Position[1] == '\n') &&
281          Position[0] != Position[1])
282        ++Position;
283      ++Position;
284      break;
285    }
286  }
287
288  Position += TruncateAtColumn - 1;
289
290  // Truncate the buffer.
291  if (Position < Buffer->getBufferEnd()) {
292    StringRef Data(Buffer->getBufferStart(),
293                         Position-Buffer->getBufferStart());
294    MemoryBuffer *TruncatedBuffer
295      = MemoryBuffer::getMemBufferCopy(Data, Buffer->getBufferIdentifier());
296    SourceMgr.overrideFileContents(File, TruncatedBuffer);
297  }
298
299  return false;
300}
301
302bool Preprocessor::isCodeCompletionFile(SourceLocation FileLoc) const {
303  return CodeCompletionFile && FileLoc.isFileID() &&
304    SourceMgr.getFileEntryForID(SourceMgr.getFileID(FileLoc))
305      == CodeCompletionFile;
306}
307
308void Preprocessor::CodeCompleteNaturalLanguage() {
309  SetCodeCompletionPoint(0, 0, 0);
310  getDiagnostics().setSuppressAllDiagnostics(true);
311  if (CodeComplete)
312    CodeComplete->CodeCompleteNaturalLanguage();
313}
314
315/// getSpelling - This method is used to get the spelling of a token into a
316/// SmallVector. Note that the returned StringRef may not point to the
317/// supplied buffer if a copy can be avoided.
318StringRef Preprocessor::getSpelling(const Token &Tok,
319                                          SmallVectorImpl<char> &Buffer,
320                                          bool *Invalid) const {
321  // NOTE: this has to be checked *before* testing for an IdentifierInfo.
322  if (Tok.isNot(tok::raw_identifier)) {
323    // Try the fast path.
324    if (const IdentifierInfo *II = Tok.getIdentifierInfo())
325      return II->getName();
326  }
327
328  // Resize the buffer if we need to copy into it.
329  if (Tok.needsCleaning())
330    Buffer.resize(Tok.getLength());
331
332  const char *Ptr = Buffer.data();
333  unsigned Len = getSpelling(Tok, Ptr, Invalid);
334  return StringRef(Ptr, Len);
335}
336
337/// CreateString - Plop the specified string into a scratch buffer and return a
338/// location for it.  If specified, the source location provides a source
339/// location for the token.
340void Preprocessor::CreateString(const char *Buf, unsigned Len, Token &Tok,
341                                SourceLocation ExpansionLoc) {
342  Tok.setLength(Len);
343
344  const char *DestPtr;
345  SourceLocation Loc = ScratchBuf->getToken(Buf, Len, DestPtr);
346
347  if (ExpansionLoc.isValid())
348    Loc = SourceMgr.createExpansionLoc(Loc, ExpansionLoc, ExpansionLoc, Len);
349  Tok.setLocation(Loc);
350
351  // If this is a raw identifier or a literal token, set the pointer data.
352  if (Tok.is(tok::raw_identifier))
353    Tok.setRawIdentifierData(DestPtr);
354  else if (Tok.isLiteral())
355    Tok.setLiteralData(DestPtr);
356}
357
358
359
360//===----------------------------------------------------------------------===//
361// Preprocessor Initialization Methods
362//===----------------------------------------------------------------------===//
363
364
365/// EnterMainSourceFile - Enter the specified FileID as the main source file,
366/// which implicitly adds the builtin defines etc.
367void Preprocessor::EnterMainSourceFile() {
368  // We do not allow the preprocessor to reenter the main file.  Doing so will
369  // cause FileID's to accumulate information from both runs (e.g. #line
370  // information) and predefined macros aren't guaranteed to be set properly.
371  assert(NumEnteredSourceFiles == 0 && "Cannot reenter the main file!");
372  FileID MainFileID = SourceMgr.getMainFileID();
373
374  // Enter the main file source buffer.
375  EnterSourceFile(MainFileID, 0, SourceLocation());
376
377  // If we've been asked to skip bytes in the main file (e.g., as part of a
378  // precompiled preamble), do so now.
379  if (SkipMainFilePreamble.first > 0)
380    CurLexer->SkipBytes(SkipMainFilePreamble.first,
381                        SkipMainFilePreamble.second);
382
383  // Tell the header info that the main file was entered.  If the file is later
384  // #imported, it won't be re-entered.
385  if (const FileEntry *FE = SourceMgr.getFileEntryForID(MainFileID))
386    HeaderInfo.IncrementIncludeCount(FE);
387
388  // Preprocess Predefines to populate the initial preprocessor state.
389  llvm::MemoryBuffer *SB =
390    llvm::MemoryBuffer::getMemBufferCopy(Predefines, "<built-in>");
391  assert(SB && "Cannot create predefined source buffer");
392  FileID FID = SourceMgr.createFileIDForMemBuffer(SB);
393  assert(!FID.isInvalid() && "Could not create FileID for predefines?");
394
395  // Start parsing the predefines.
396  EnterSourceFile(FID, 0, SourceLocation());
397}
398
399void Preprocessor::EndSourceFile() {
400  // Notify the client that we reached the end of the source file.
401  if (Callbacks)
402    Callbacks->EndOfMainFile();
403}
404
405//===----------------------------------------------------------------------===//
406// Lexer Event Handling.
407//===----------------------------------------------------------------------===//
408
409/// LookUpIdentifierInfo - Given a tok::raw_identifier token, look up the
410/// identifier information for the token and install it into the token,
411/// updating the token kind accordingly.
412IdentifierInfo *Preprocessor::LookUpIdentifierInfo(Token &Identifier) const {
413  assert(Identifier.getRawIdentifierData() != 0 && "No raw identifier data!");
414
415  // Look up this token, see if it is a macro, or if it is a language keyword.
416  IdentifierInfo *II;
417  if (!Identifier.needsCleaning()) {
418    // No cleaning needed, just use the characters from the lexed buffer.
419    II = getIdentifierInfo(StringRef(Identifier.getRawIdentifierData(),
420                                           Identifier.getLength()));
421  } else {
422    // Cleaning needed, alloca a buffer, clean into it, then use the buffer.
423    llvm::SmallString<64> IdentifierBuffer;
424    StringRef CleanedStr = getSpelling(Identifier, IdentifierBuffer);
425    II = getIdentifierInfo(CleanedStr);
426  }
427
428  // Update the token info (identifier info and appropriate token kind).
429  Identifier.setIdentifierInfo(II);
430  Identifier.setKind(II->getTokenID());
431
432  return II;
433}
434
435void Preprocessor::SetPoisonReason(IdentifierInfo *II, unsigned DiagID) {
436  PoisonReasons[II] = DiagID;
437}
438
439void Preprocessor::PoisonSEHIdentifiers(bool Poison) {
440  assert(Ident__exception_code && Ident__exception_info);
441  assert(Ident___exception_code && Ident___exception_info);
442  Ident__exception_code->setIsPoisoned(Poison);
443  Ident___exception_code->setIsPoisoned(Poison);
444  Ident_GetExceptionCode->setIsPoisoned(Poison);
445  Ident__exception_info->setIsPoisoned(Poison);
446  Ident___exception_info->setIsPoisoned(Poison);
447  Ident_GetExceptionInfo->setIsPoisoned(Poison);
448  Ident__abnormal_termination->setIsPoisoned(Poison);
449  Ident___abnormal_termination->setIsPoisoned(Poison);
450  Ident_AbnormalTermination->setIsPoisoned(Poison);
451}
452
453void Preprocessor::HandlePoisonedIdentifier(Token & Identifier) {
454  assert(Identifier.getIdentifierInfo() &&
455         "Can't handle identifiers without identifier info!");
456  llvm::DenseMap<IdentifierInfo*,unsigned>::const_iterator it =
457    PoisonReasons.find(Identifier.getIdentifierInfo());
458  if(it == PoisonReasons.end())
459    Diag(Identifier, diag::err_pp_used_poisoned_id);
460  else
461    Diag(Identifier,it->second) << Identifier.getIdentifierInfo();
462}
463
464/// HandleIdentifier - This callback is invoked when the lexer reads an
465/// identifier.  This callback looks up the identifier in the map and/or
466/// potentially macro expands it or turns it into a named token (like 'for').
467///
468/// Note that callers of this method are guarded by checking the
469/// IdentifierInfo's 'isHandleIdentifierCase' bit.  If this method changes, the
470/// IdentifierInfo methods that compute these properties will need to change to
471/// match.
472void Preprocessor::HandleIdentifier(Token &Identifier) {
473  assert(Identifier.getIdentifierInfo() &&
474         "Can't handle identifiers without identifier info!");
475
476  IdentifierInfo &II = *Identifier.getIdentifierInfo();
477
478  // If this identifier was poisoned, and if it was not produced from a macro
479  // expansion, emit an error.
480  if (II.isPoisoned() && CurPPLexer) {
481    HandlePoisonedIdentifier(Identifier);
482  }
483
484  // If this is a macro to be expanded, do it.
485  if (MacroInfo *MI = getMacroInfo(&II)) {
486    if (!DisableMacroExpansion && !Identifier.isExpandDisabled()) {
487      if (MI->isEnabled()) {
488        if (!HandleMacroExpandedIdentifier(Identifier, MI))
489          return;
490      } else {
491        // C99 6.10.3.4p2 says that a disabled macro may never again be
492        // expanded, even if it's in a context where it could be expanded in the
493        // future.
494        Identifier.setFlag(Token::DisableExpand);
495      }
496    }
497  }
498
499  // C++ 2.11p2: If this is an alternative representation of a C++ operator,
500  // then we act as if it is the actual operator and not the textual
501  // representation of it.
502  if (II.isCPlusPlusOperatorKeyword())
503    Identifier.setIdentifierInfo(0);
504
505  // If this is an extension token, diagnose its use.
506  // We avoid diagnosing tokens that originate from macro definitions.
507  // FIXME: This warning is disabled in cases where it shouldn't be,
508  // like "#define TY typeof", "TY(1) x".
509  if (II.isExtensionToken() && !DisableMacroExpansion)
510    Diag(Identifier, diag::ext_token_used);
511}
512
513void Preprocessor::HandleModuleImport(Token &Import) {
514  // The token sequence
515  //
516  //   __import__ identifier
517  //
518  // indicates a module import directive. We load the module and then
519  // leave the token sequence for the parser.
520  Token ModuleNameTok = LookAhead(0);
521  if (ModuleNameTok.getKind() != tok::identifier)
522    return;
523
524  (void)TheModuleLoader.loadModule(Import.getLocation(),
525                                   *ModuleNameTok.getIdentifierInfo(),
526                                   ModuleNameTok.getLocation());
527
528  // FIXME: Transmogrify __import__ into some kind of AST-only __import__ that
529  // is not recognized by the preprocessor but is recognized by the parser.
530  // It would also be useful to stash the ModuleKey somewhere, so we don't try
531  // to load the module twice.
532}
533
534void Preprocessor::AddCommentHandler(CommentHandler *Handler) {
535  assert(Handler && "NULL comment handler");
536  assert(std::find(CommentHandlers.begin(), CommentHandlers.end(), Handler) ==
537         CommentHandlers.end() && "Comment handler already registered");
538  CommentHandlers.push_back(Handler);
539}
540
541void Preprocessor::RemoveCommentHandler(CommentHandler *Handler) {
542  std::vector<CommentHandler *>::iterator Pos
543  = std::find(CommentHandlers.begin(), CommentHandlers.end(), Handler);
544  assert(Pos != CommentHandlers.end() && "Comment handler not registered");
545  CommentHandlers.erase(Pos);
546}
547
548bool Preprocessor::HandleComment(Token &result, SourceRange Comment) {
549  bool AnyPendingTokens = false;
550  for (std::vector<CommentHandler *>::iterator H = CommentHandlers.begin(),
551       HEnd = CommentHandlers.end();
552       H != HEnd; ++H) {
553    if ((*H)->HandleComment(*this, Comment))
554      AnyPendingTokens = true;
555  }
556  if (!AnyPendingTokens || getCommentRetentionState())
557    return false;
558  Lex(result);
559  return true;
560}
561
562ModuleLoader::~ModuleLoader() { }
563
564CommentHandler::~CommentHandler() { }
565
566CodeCompletionHandler::~CodeCompletionHandler() { }
567
568void Preprocessor::createPreprocessingRecord(
569                                      bool IncludeNestedMacroExpansions) {
570  if (Record)
571    return;
572
573  Record = new PreprocessingRecord(IncludeNestedMacroExpansions);
574  addPPCallbacks(Record);
575}
576