lib/Lex/Preprocessor.cpp

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