Preprocessor.cpp revision cf29e0716bb3ecbbc15b74cd648367d6b075fdf0 
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 "clang/Lex/HeaderSearch.h"
30#include "clang/Lex/MacroInfo.h"
31#include "clang/Lex/Pragma.h"
32#include "clang/Lex/ScratchBuffer.h"
33#include "clang/Lex/LexDiagnostic.h"
34#include "clang/Basic/SourceManager.h"
35#include "clang/Basic/FileManager.h"
36#include "clang/Basic/TargetInfo.h"
37#include "llvm/ADT/APFloat.h"
38#include "llvm/ADT/SmallVector.h"
39#include "llvm/Support/MemoryBuffer.h"
40#include "llvm/Support/Streams.h"
41using namespace clang;
42
43//===----------------------------------------------------------------------===//
44
45PreprocessorFactory::~PreprocessorFactory() {}
46
47Preprocessor::Preprocessor(Diagnostic &diags, const LangOptions &opts,
48                           TargetInfo &target, SourceManager &SM,
49                           HeaderSearch &Headers,
50                           IdentifierInfoLookup* IILookup)
51  : Diags(diags), Features(opts), Target(target), FileMgr(Headers.getFileMgr()),
52    SourceMgr(SM), HeaderInfo(Headers), Identifiers(opts, IILookup),
53    CurPPLexer(0), CurDirLookup(0), Callbacks(0) {
54  ScratchBuf = new ScratchBuffer(SourceMgr);
55
56  // Clear stats.
57  NumDirectives = NumDefined = NumUndefined = NumPragma = 0;
58  NumIf = NumElse = NumEndif = 0;
59  NumEnteredSourceFiles = 0;
60  NumMacroExpanded = NumFnMacroExpanded = NumBuiltinMacroExpanded = 0;
61  NumFastMacroExpanded = NumTokenPaste = NumFastTokenPaste = 0;
62  MaxIncludeStackDepth = 0;
63  NumSkipped = 0;
64
65  // Default to discarding comments.
66  KeepComments = false;
67  KeepMacroComments = false;
68
69  // Macro expansion is enabled.
70  DisableMacroExpansion = false;
71  InMacroArgs = false;
72  NumCachedTokenLexers = 0;
73
74  CachedLexPos = 0;
75
76  // "Poison" __VA_ARGS__, which can only appear in the expansion of a macro.
77  // This gets unpoisoned where it is allowed.
78  (Ident__VA_ARGS__ = getIdentifierInfo("__VA_ARGS__"))->setIsPoisoned();
79
80  // Initialize the pragma handlers.
81  PragmaHandlers = new PragmaNamespace(0);
82  RegisterBuiltinPragmas();
83
84  // Initialize builtin macros like __LINE__ and friends.
85  RegisterBuiltinMacros();
86}
87
88Preprocessor::~Preprocessor() {
89  assert(BacktrackPositions.empty() && "EnableBacktrack/Backtrack imbalance!");
90
91  while (!IncludeMacroStack.empty()) {
92    delete IncludeMacroStack.back().TheLexer;
93    delete IncludeMacroStack.back().TheTokenLexer;
94    IncludeMacroStack.pop_back();
95  }
96
97  // Free any macro definitions.
98  for (llvm::DenseMap<IdentifierInfo*, MacroInfo*>::iterator I =
99       Macros.begin(), E = Macros.end(); I != E; ++I) {
100    // We don't need to free the MacroInfo objects directly.  These
101    // will be released when the BumpPtrAllocator 'BP' object gets
102    // destroyed. We still need to run the dstor, however, to free
103    // memory alocated by MacroInfo.
104    I->second->Destroy();
105    I->first->setHasMacroDefinition(false);
106  }
107
108  // Free any cached macro expanders.
109  for (unsigned i = 0, e = NumCachedTokenLexers; i != e; ++i)
110    delete TokenLexerCache[i];
111
112  // Release pragma information.
113  delete PragmaHandlers;
114
115  // Delete the scratch buffer info.
116  delete ScratchBuf;
117
118  delete Callbacks;
119}
120
121void Preprocessor::setPTHManager(PTHManager* pm) {
122  PTH.reset(pm);
123  FileMgr.setStatCache(PTH->createStatCache());
124}
125
126void Preprocessor::DumpToken(const Token &Tok, bool DumpFlags) const {
127  llvm::cerr << tok::getTokenName(Tok.getKind()) << " '"
128             << getSpelling(Tok) << "'";
129
130  if (!DumpFlags) return;
131
132  llvm::cerr << "\t";
133  if (Tok.isAtStartOfLine())
134    llvm::cerr << " [StartOfLine]";
135  if (Tok.hasLeadingSpace())
136    llvm::cerr << " [LeadingSpace]";
137  if (Tok.isExpandDisabled())
138    llvm::cerr << " [ExpandDisabled]";
139  if (Tok.needsCleaning()) {
140    const char *Start = SourceMgr.getCharacterData(Tok.getLocation());
141    llvm::cerr << " [UnClean='" << std::string(Start, Start+Tok.getLength())
142               << "']";
143  }
144
145  llvm::cerr << "\tLoc=<";
146  DumpLocation(Tok.getLocation());
147  llvm::cerr << ">";
148}
149
150void Preprocessor::DumpLocation(SourceLocation Loc) const {
151  Loc.dump(SourceMgr);
152}
153
154void Preprocessor::DumpMacro(const MacroInfo &MI) const {
155  llvm::cerr << "MACRO: ";
156  for (unsigned i = 0, e = MI.getNumTokens(); i != e; ++i) {
157    DumpToken(MI.getReplacementToken(i));
158    llvm::cerr << "  ";
159  }
160  llvm::cerr << "\n";
161}
162
163void Preprocessor::PrintStats() {
164  llvm::cerr << "\n*** Preprocessor Stats:\n";
165  llvm::cerr << NumDirectives << " directives found:\n";
166  llvm::cerr << "  " << NumDefined << " #define.\n";
167  llvm::cerr << "  " << NumUndefined << " #undef.\n";
168  llvm::cerr << "  #include/#include_next/#import:\n";
169  llvm::cerr << "    " << NumEnteredSourceFiles << " source files entered.\n";
170  llvm::cerr << "    " << MaxIncludeStackDepth << " max include stack depth\n";
171  llvm::cerr << "  " << NumIf << " #if/#ifndef/#ifdef.\n";
172  llvm::cerr << "  " << NumElse << " #else/#elif.\n";
173  llvm::cerr << "  " << NumEndif << " #endif.\n";
174  llvm::cerr << "  " << NumPragma << " #pragma.\n";
175  llvm::cerr << NumSkipped << " #if/#ifndef#ifdef regions skipped\n";
176
177  llvm::cerr << NumMacroExpanded << "/" << NumFnMacroExpanded << "/"
178             << NumBuiltinMacroExpanded << " obj/fn/builtin macros expanded, "
179             << NumFastMacroExpanded << " on the fast path.\n";
180  llvm::cerr << (NumFastTokenPaste+NumTokenPaste)
181             << " token paste (##) operations performed, "
182             << NumFastTokenPaste << " on the fast path.\n";
183}
184
185//===----------------------------------------------------------------------===//
186// Token Spelling
187//===----------------------------------------------------------------------===//
188
189
190/// getSpelling() - Return the 'spelling' of this token.  The spelling of a
191/// token are the characters used to represent the token in the source file
192/// after trigraph expansion and escaped-newline folding.  In particular, this
193/// wants to get the true, uncanonicalized, spelling of things like digraphs
194/// UCNs, etc.
195std::string Preprocessor::getSpelling(const Token &Tok) const {
196  assert((int)Tok.getLength() >= 0 && "Token character range is bogus!");
197
198  // If this token contains nothing interesting, return it directly.
199  const char* TokStart = SourceMgr.getCharacterData(Tok.getLocation());
200  if (!Tok.needsCleaning())
201    return std::string(TokStart, TokStart+Tok.getLength());
202
203  std::string Result;
204  Result.reserve(Tok.getLength());
205
206  // Otherwise, hard case, relex the characters into the string.
207  for (const char *Ptr = TokStart, *End = TokStart+Tok.getLength();
208       Ptr != End; ) {
209    unsigned CharSize;
210    Result.push_back(Lexer::getCharAndSizeNoWarn(Ptr, CharSize, Features));
211    Ptr += CharSize;
212  }
213  assert(Result.size() != unsigned(Tok.getLength()) &&
214         "NeedsCleaning flag set on something that didn't need cleaning!");
215  return Result;
216}
217
218/// getSpelling - This method is used to get the spelling of a token into a
219/// preallocated buffer, instead of as an std::string.  The caller is required
220/// to allocate enough space for the token, which is guaranteed to be at least
221/// Tok.getLength() bytes long.  The actual length of the token is returned.
222///
223/// Note that this method may do two possible things: it may either fill in
224/// the buffer specified with characters, or it may *change the input pointer*
225/// to point to a constant buffer with the data already in it (avoiding a
226/// copy).  The caller is not allowed to modify the returned buffer pointer
227/// if an internal buffer is returned.
228unsigned Preprocessor::getSpelling(const Token &Tok,
229                                   const char *&Buffer) const {
230  assert((int)Tok.getLength() >= 0 && "Token character range is bogus!");
231
232  // If this token is an identifier, just return the string from the identifier
233  // table, which is very quick.
234  if (const IdentifierInfo *II = Tok.getIdentifierInfo()) {
235    Buffer = II->getName();
236    return II->getLength();
237  }
238
239  // Otherwise, compute the start of the token in the input lexer buffer.
240  const char *TokStart = 0;
241
242  if (Tok.isLiteral())
243    TokStart = Tok.getLiteralData();
244
245  if (TokStart == 0)
246    TokStart = SourceMgr.getCharacterData(Tok.getLocation());
247
248  // If this token contains nothing interesting, return it directly.
249  if (!Tok.needsCleaning()) {
250    Buffer = TokStart;
251    return Tok.getLength();
252  }
253
254  // Otherwise, hard case, relex the characters into the string.
255  char *OutBuf = const_cast<char*>(Buffer);
256  for (const char *Ptr = TokStart, *End = TokStart+Tok.getLength();
257       Ptr != End; ) {
258    unsigned CharSize;
259    *OutBuf++ = Lexer::getCharAndSizeNoWarn(Ptr, CharSize, Features);
260    Ptr += CharSize;
261  }
262  assert(unsigned(OutBuf-Buffer) != Tok.getLength() &&
263         "NeedsCleaning flag set on something that didn't need cleaning!");
264
265  return OutBuf-Buffer;
266}
267
268
269/// CreateString - Plop the specified string into a scratch buffer and return a
270/// location for it.  If specified, the source location provides a source
271/// location for the token.
272void Preprocessor::CreateString(const char *Buf, unsigned Len, Token &Tok,
273                                SourceLocation InstantiationLoc) {
274  Tok.setLength(Len);
275
276  const char *DestPtr;
277  SourceLocation Loc = ScratchBuf->getToken(Buf, Len, DestPtr);
278
279  if (InstantiationLoc.isValid())
280    Loc = SourceMgr.createInstantiationLoc(Loc, InstantiationLoc,
281                                           InstantiationLoc, Len);
282  Tok.setLocation(Loc);
283
284  // If this is a literal token, set the pointer data.
285  if (Tok.isLiteral())
286    Tok.setLiteralData(DestPtr);
287}
288
289
290/// AdvanceToTokenCharacter - Given a location that specifies the start of a
291/// token, return a new location that specifies a character within the token.
292SourceLocation Preprocessor::AdvanceToTokenCharacter(SourceLocation TokStart,
293                                                     unsigned CharNo) {
294  // If they request the first char of the token, we're trivially done.
295  if (CharNo == 0) return TokStart;
296
297  // Figure out how many physical characters away the specified instantiation
298  // character is.  This needs to take into consideration newlines and
299  // trigraphs.
300  const char *TokPtr = SourceMgr.getCharacterData(TokStart);
301  unsigned PhysOffset = 0;
302
303  // The usual case is that tokens don't contain anything interesting.  Skip
304  // over the uninteresting characters.  If a token only consists of simple
305  // chars, this method is extremely fast.
306  while (CharNo && Lexer::isObviouslySimpleCharacter(*TokPtr))
307    ++TokPtr, --CharNo, ++PhysOffset;
308
309  // If we have a character that may be a trigraph or escaped newline, use a
310  // lexer to parse it correctly.
311  if (CharNo != 0) {
312    // Skip over characters the remaining characters.
313    for (; CharNo; --CharNo) {
314      unsigned Size;
315      Lexer::getCharAndSizeNoWarn(TokPtr, Size, Features);
316      TokPtr += Size;
317      PhysOffset += Size;
318    }
319  }
320
321  return TokStart.getFileLocWithOffset(PhysOffset);
322}
323
324
325//===----------------------------------------------------------------------===//
326// Preprocessor Initialization Methods
327//===----------------------------------------------------------------------===//
328
329// Append a #define line to Buf for Macro.  Macro should be of the form XXX,
330// in which case we emit "#define XXX 1" or "XXX=Y z W" in which case we emit
331// "#define XXX Y z W".  To get a #define with no value, use "XXX=".
332static void DefineBuiltinMacro(std::vector<char> &Buf, const char *Macro,
333                               const char *Command = "#define ") {
334  Buf.insert(Buf.end(), Command, Command+strlen(Command));
335  if (const char *Equal = strchr(Macro, '=')) {
336    // Turn the = into ' '.
337    Buf.insert(Buf.end(), Macro, Equal);
338    Buf.push_back(' ');
339    Buf.insert(Buf.end(), Equal+1, Equal+strlen(Equal));
340  } else {
341    // Push "macroname 1".
342    Buf.insert(Buf.end(), Macro, Macro+strlen(Macro));
343    Buf.push_back(' ');
344    Buf.push_back('1');
345  }
346  Buf.push_back('\n');
347}
348
349/// PickFP - This is used to pick a value based on the FP semantics of the
350/// specified FP model.
351template <typename T>
352static T PickFP(const llvm::fltSemantics *Sem, T IEEESingleVal,
353                T IEEEDoubleVal, T X87DoubleExtendedVal, T PPCDoubleDoubleVal) {
354  if (Sem == &llvm::APFloat::IEEEsingle)
355    return IEEESingleVal;
356  if (Sem == &llvm::APFloat::IEEEdouble)
357    return IEEEDoubleVal;
358  if (Sem == &llvm::APFloat::x87DoubleExtended)
359    return X87DoubleExtendedVal;
360  assert(Sem == &llvm::APFloat::PPCDoubleDouble);
361  return PPCDoubleDoubleVal;
362}
363
364static void DefineFloatMacros(std::vector<char> &Buf, const char *Prefix,
365                              const llvm::fltSemantics *Sem) {
366  const char *DenormMin, *Epsilon, *Max, *Min;
367  DenormMin = PickFP(Sem, "1.40129846e-45F", "4.9406564584124654e-324",
368                     "3.64519953188247460253e-4951L",
369                     "4.94065645841246544176568792868221e-324L");
370  int Digits = PickFP(Sem, 6, 15, 18, 31);
371  Epsilon = PickFP(Sem, "1.19209290e-7F", "2.2204460492503131e-16",
372                   "1.08420217248550443401e-19L",
373                   "4.94065645841246544176568792868221e-324L");
374  int HasInifinity = 1, HasQuietNaN = 1;
375  int MantissaDigits = PickFP(Sem, 24, 53, 64, 106);
376  int Min10Exp = PickFP(Sem, -37, -307, -4931, -291);
377  int Max10Exp = PickFP(Sem, 38, 308, 4932, 308);
378  int MinExp = PickFP(Sem, -125, -1021, -16381, -968);
379  int MaxExp = PickFP(Sem, 128, 1024, 16384, 1024);
380  Min = PickFP(Sem, "1.17549435e-38F", "2.2250738585072014e-308",
381               "3.36210314311209350626e-4932L",
382               "2.00416836000897277799610805135016e-292L");
383  Max = PickFP(Sem, "3.40282347e+38F", "1.7976931348623157e+308",
384               "1.18973149535723176502e+4932L",
385               "1.79769313486231580793728971405301e+308L");
386
387  char MacroBuf[60];
388  sprintf(MacroBuf, "__%s_DENORM_MIN__=%s", Prefix, DenormMin);
389  DefineBuiltinMacro(Buf, MacroBuf);
390  sprintf(MacroBuf, "__%s_DIG__=%d", Prefix, Digits);
391  DefineBuiltinMacro(Buf, MacroBuf);
392  sprintf(MacroBuf, "__%s_EPSILON__=%s", Prefix, Epsilon);
393  DefineBuiltinMacro(Buf, MacroBuf);
394  sprintf(MacroBuf, "__%s_HAS_INFINITY__=%d", Prefix, HasInifinity);
395  DefineBuiltinMacro(Buf, MacroBuf);
396  sprintf(MacroBuf, "__%s_HAS_QUIET_NAN__=%d", Prefix, HasQuietNaN);
397  DefineBuiltinMacro(Buf, MacroBuf);
398  sprintf(MacroBuf, "__%s_MANT_DIG__=%d", Prefix, MantissaDigits);
399  DefineBuiltinMacro(Buf, MacroBuf);
400  sprintf(MacroBuf, "__%s_MAX_10_EXP__=%d", Prefix, Max10Exp);
401  DefineBuiltinMacro(Buf, MacroBuf);
402  sprintf(MacroBuf, "__%s_MAX_EXP__=%d", Prefix, MaxExp);
403  DefineBuiltinMacro(Buf, MacroBuf);
404  sprintf(MacroBuf, "__%s_MAX__=%s", Prefix, Max);
405  DefineBuiltinMacro(Buf, MacroBuf);
406  sprintf(MacroBuf, "__%s_MIN_10_EXP__=(%d)", Prefix, Min10Exp);
407  DefineBuiltinMacro(Buf, MacroBuf);
408  sprintf(MacroBuf, "__%s_MIN_EXP__=(%d)", Prefix, MinExp);
409  DefineBuiltinMacro(Buf, MacroBuf);
410  sprintf(MacroBuf, "__%s_MIN__=%s", Prefix, Min);
411  DefineBuiltinMacro(Buf, MacroBuf);
412  sprintf(MacroBuf, "__%s_HAS_DENORM__=1", Prefix);
413  DefineBuiltinMacro(Buf, MacroBuf);
414}
415
416
417/// DefineTypeSize - Emit a macro to the predefines buffer that declares a macro
418/// named MacroName with the max value for a type with width 'TypeWidth' a
419/// signedness of 'isSigned' and with a value suffix of 'ValSuffix' (e.g. LL).
420static void DefineTypeSize(const char *MacroName, unsigned TypeWidth,
421                           const char *ValSuffix, bool isSigned,
422                           std::vector<char> &Buf) {
423  char MacroBuf[60];
424  uint64_t MaxVal;
425  if (isSigned)
426    MaxVal = (1LL << (TypeWidth - 1)) - 1;
427  else
428    MaxVal = ~0LL >> (64-TypeWidth);
429
430  sprintf(MacroBuf, "%s=%llu%s", MacroName, MaxVal, ValSuffix);
431  DefineBuiltinMacro(Buf, MacroBuf);
432}
433
434static void DefineType(const char *MacroName, TargetInfo::IntType Ty,
435                       std::vector<char> &Buf) {
436  char MacroBuf[60];
437  sprintf(MacroBuf, "%s=%s", MacroName, TargetInfo::getTypeName(Ty));
438  DefineBuiltinMacro(Buf, MacroBuf);
439}
440
441
442static void InitializePredefinedMacros(Preprocessor &PP,
443                                       std::vector<char> &Buf) {
444  char MacroBuf[60];
445  // Compiler version introspection macros.
446  DefineBuiltinMacro(Buf, "__llvm__=1");   // LLVM Backend
447  DefineBuiltinMacro(Buf, "__clang__=1");  // Clang Frontend
448
449  // Currently claim to be compatible with GCC 4.2.1-5621.
450  DefineBuiltinMacro(Buf, "__APPLE_CC__=5621");
451  DefineBuiltinMacro(Buf, "__GNUC_MINOR__=2");
452  DefineBuiltinMacro(Buf, "__GNUC_PATCHLEVEL__=1");
453  DefineBuiltinMacro(Buf, "__GNUC__=4");
454  DefineBuiltinMacro(Buf, "__GXX_ABI_VERSION=1002");
455  DefineBuiltinMacro(Buf, "__VERSION__=\"4.2.1 (Apple Computer, Inc. "
456                     "build 5621) (dot 3)\"");
457
458
459  // Initialize language-specific preprocessor defines.
460
461  // These should all be defined in the preprocessor according to the
462  // current language configuration.
463  if (!PP.getLangOptions().Microsoft)
464    DefineBuiltinMacro(Buf, "__STDC__=1");
465  if (PP.getLangOptions().AsmPreprocessor)
466    DefineBuiltinMacro(Buf, "__ASSEMBLER__=1");
467  if (PP.getLangOptions().C99 && !PP.getLangOptions().CPlusPlus)
468    DefineBuiltinMacro(Buf, "__STDC_VERSION__=199901L");
469  else if (0) // STDC94 ?
470    DefineBuiltinMacro(Buf, "__STDC_VERSION__=199409L");
471
472  DefineBuiltinMacro(Buf, "__STDC_HOSTED__=1");
473  if (PP.getLangOptions().ObjC1) {
474    DefineBuiltinMacro(Buf, "__OBJC__=1");
475    if (PP.getLangOptions().ObjCNonFragileABI)
476      DefineBuiltinMacro(Buf, "__OBJC2__=1");
477
478    if (PP.getLangOptions().getGCMode() == LangOptions::NonGC) {
479      DefineBuiltinMacro(Buf, "__weak=");
480      DefineBuiltinMacro(Buf, "__strong=");
481    } else {
482      DefineBuiltinMacro(Buf, "__weak=__attribute__((objc_gc(weak)))");
483      DefineBuiltinMacro(Buf, "__strong=__attribute__((objc_gc(strong)))");
484      DefineBuiltinMacro(Buf, "__OBJC_GC__=1");
485    }
486
487    if (PP.getLangOptions().NeXTRuntime)
488      DefineBuiltinMacro(Buf, "__NEXT_RUNTIME__=1");
489  }
490
491  // darwin_constant_cfstrings controls this. This is also dependent
492  // on other things like the runtime I believe.  This is set even for C code.
493  DefineBuiltinMacro(Buf, "__CONSTANT_CFSTRINGS__=1");
494
495  if (PP.getLangOptions().ObjC2)
496    DefineBuiltinMacro(Buf, "OBJC_NEW_PROPERTIES");
497
498  if (PP.getLangOptions().PascalStrings)
499    DefineBuiltinMacro(Buf, "__PASCAL_STRINGS__");
500
501  if (PP.getLangOptions().Blocks) {
502    DefineBuiltinMacro(Buf, "__block=__attribute__((__blocks__(byref)))");
503    DefineBuiltinMacro(Buf, "__BLOCKS__=1");
504  }
505
506  if (PP.getLangOptions().CPlusPlus) {
507    DefineBuiltinMacro(Buf, "__DEPRECATED=1");
508    DefineBuiltinMacro(Buf, "__EXCEPTIONS=1");
509    DefineBuiltinMacro(Buf, "__GNUG__=4");
510    DefineBuiltinMacro(Buf, "__GXX_WEAK__=1");
511    DefineBuiltinMacro(Buf, "__cplusplus=1");
512    DefineBuiltinMacro(Buf, "__private_extern__=extern");
513  }
514
515  // Filter out some microsoft extensions when trying to parse in ms-compat
516  // mode.
517  if (PP.getLangOptions().Microsoft) {
518    DefineBuiltinMacro(Buf, "_cdecl=__cdecl");
519    DefineBuiltinMacro(Buf, "__int8=__INT8_TYPE__");
520    DefineBuiltinMacro(Buf, "__int16=__INT16_TYPE__");
521    DefineBuiltinMacro(Buf, "__int32=__INT32_TYPE__");
522    DefineBuiltinMacro(Buf, "__int64=__INT64_TYPE__");
523  }
524
525  // Initialize target-specific preprocessor defines.
526  const TargetInfo &TI = PP.getTargetInfo();
527
528  // Define type sizing macros based on the target properties.
529  assert(TI.getCharWidth() == 8 && "Only support 8-bit char so far");
530  DefineBuiltinMacro(Buf, "__CHAR_BIT__=8");
531
532  unsigned IntMaxWidth;
533  const char *IntMaxSuffix;
534  if (TI.getIntMaxType() == TargetInfo::SignedLongLong) {
535    IntMaxWidth = TI.getLongLongWidth();
536    IntMaxSuffix = "LL";
537  } else if (TI.getIntMaxType() == TargetInfo::SignedLong) {
538    IntMaxWidth = TI.getLongWidth();
539    IntMaxSuffix = "L";
540  } else {
541    assert(TI.getIntMaxType() == TargetInfo::SignedInt);
542    IntMaxWidth = TI.getIntWidth();
543    IntMaxSuffix = "";
544  }
545
546  DefineTypeSize("__SCHAR_MAX__", TI.getCharWidth(), "", true, Buf);
547  DefineTypeSize("__SHRT_MAX__", TI.getShortWidth(), "", true, Buf);
548  DefineTypeSize("__INT_MAX__", TI.getIntWidth(), "", true, Buf);
549  DefineTypeSize("__LONG_MAX__", TI.getLongWidth(), "L", true, Buf);
550  DefineTypeSize("__LONG_LONG_MAX__", TI.getLongLongWidth(), "LL", true, Buf);
551  DefineTypeSize("__WCHAR_MAX__", TI.getWCharWidth(), "", true, Buf);
552  DefineTypeSize("__INTMAX_MAX__", IntMaxWidth, IntMaxSuffix, true, Buf);
553
554  DefineType("__INTMAX_TYPE__", TI.getIntMaxType(), Buf);
555  DefineType("__UINTMAX_TYPE__", TI.getUIntMaxType(), Buf);
556  DefineType("__PTRDIFF_TYPE__", TI.getPtrDiffType(0), Buf);
557  DefineType("__INTPTR_TYPE__", TI.getIntPtrType(), Buf);
558  DefineType("__SIZE_TYPE__", TI.getSizeType(), Buf);
559  DefineType("__WCHAR_TYPE__", TI.getWCharType(), Buf);
560  // FIXME: TargetInfo hookize __WINT_TYPE__.
561  DefineBuiltinMacro(Buf, "__WINT_TYPE__=int");
562
563  DefineFloatMacros(Buf, "FLT", &TI.getFloatFormat());
564  DefineFloatMacros(Buf, "DBL", &TI.getDoubleFormat());
565  DefineFloatMacros(Buf, "LDBL", &TI.getLongDoubleFormat());
566
567  // Define a __POINTER_WIDTH__ macro for stdint.h.
568  sprintf(MacroBuf, "__POINTER_WIDTH__=%d", (int)TI.getPointerWidth(0));
569  DefineBuiltinMacro(Buf, MacroBuf);
570
571  if (!TI.isCharSigned())
572    DefineBuiltinMacro(Buf, "__CHAR_UNSIGNED__");
573
574  // Define fixed-sized integer types for stdint.h
575  assert(TI.getCharWidth() == 8 && "unsupported target types");
576  assert(TI.getShortWidth() == 16 && "unsupported target types");
577  DefineBuiltinMacro(Buf, "__INT8_TYPE__=char");
578  DefineBuiltinMacro(Buf, "__INT16_TYPE__=short");
579
580  if (TI.getIntWidth() == 32)
581    DefineBuiltinMacro(Buf, "__INT32_TYPE__=int");
582  else {
583    assert(TI.getLongLongWidth() == 32 && "unsupported target types");
584    DefineBuiltinMacro(Buf, "__INT32_TYPE__=long long");
585  }
586
587  // 16-bit targets doesn't necessarily have a 64-bit type.
588  if (TI.getLongLongWidth() == 64)
589    DefineBuiltinMacro(Buf, "__INT64_TYPE__=long long");
590
591  // Add __builtin_va_list typedef.
592  {
593    const char *VAList = TI.getVAListDeclaration();
594    Buf.insert(Buf.end(), VAList, VAList+strlen(VAList));
595    Buf.push_back('\n');
596  }
597
598  if (const char *Prefix = TI.getUserLabelPrefix()) {
599    sprintf(MacroBuf, "__USER_LABEL_PREFIX__=%s", Prefix);
600    DefineBuiltinMacro(Buf, MacroBuf);
601  }
602
603  // Build configuration options.  FIXME: these should be controlled by
604  // command line options or something.
605  DefineBuiltinMacro(Buf, "__DYNAMIC__=1");
606  DefineBuiltinMacro(Buf, "__FINITE_MATH_ONLY__=0");
607  DefineBuiltinMacro(Buf, "__NO_INLINE__=1");
608  DefineBuiltinMacro(Buf, "__PIC__=1");
609
610  // Macros to control C99 numerics and <float.h>
611  DefineBuiltinMacro(Buf, "__FLT_EVAL_METHOD__=0");
612  DefineBuiltinMacro(Buf, "__FLT_RADIX__=2");
613  sprintf(MacroBuf, "__DECIMAL_DIG__=%d",
614          PickFP(&TI.getLongDoubleFormat(), -1/*FIXME*/, 17, 21, 33));
615  DefineBuiltinMacro(Buf, MacroBuf);
616
617  // Get other target #defines.
618  TI.getTargetDefines(Buf);
619
620  // FIXME: Should emit a #line directive here.
621}
622
623
624/// EnterMainSourceFile - Enter the specified FileID as the main source file,
625/// which implicitly adds the builtin defines etc.
626void Preprocessor::EnterMainSourceFile() {
627  // We do not allow the preprocessor to reenter the main file.  Doing so will
628  // cause FileID's to accumulate information from both runs (e.g. #line
629  // information) and predefined macros aren't guaranteed to be set properly.
630  assert(NumEnteredSourceFiles == 0 && "Cannot reenter the main file!");
631  FileID MainFileID = SourceMgr.getMainFileID();
632
633  // Enter the main file source buffer.
634  EnterSourceFile(MainFileID, 0);
635
636  // Tell the header info that the main file was entered.  If the file is later
637  // #imported, it won't be re-entered.
638  if (const FileEntry *FE = SourceMgr.getFileEntryForID(MainFileID))
639    HeaderInfo.IncrementIncludeCount(FE);
640
641  std::vector<char> PrologFile;
642  PrologFile.reserve(4080);
643
644  // Install things like __POWERPC__, __GNUC__, etc into the macro table.
645  InitializePredefinedMacros(*this, PrologFile);
646
647  // Add on the predefines from the driver.
648  PrologFile.insert(PrologFile.end(), Predefines.begin(), Predefines.end());
649
650  // Memory buffer must end with a null byte!
651  PrologFile.push_back(0);
652
653  // Now that we have emitted the predefined macros, #includes, etc into
654  // PrologFile, preprocess it to populate the initial preprocessor state.
655  llvm::MemoryBuffer *SB =
656    llvm::MemoryBuffer::getMemBufferCopy(&PrologFile.front(),&PrologFile.back(),
657                                         "<predefines>");
658  assert(SB && "Cannot fail to create predefined source buffer");
659  FileID FID = SourceMgr.createFileIDForMemBuffer(SB);
660  assert(!FID.isInvalid() && "Could not create FileID for predefines?");
661
662  // Start parsing the predefines.
663  EnterSourceFile(FID, 0);
664}
665
666
667//===----------------------------------------------------------------------===//
668// Lexer Event Handling.
669//===----------------------------------------------------------------------===//
670
671/// LookUpIdentifierInfo - Given a tok::identifier token, look up the
672/// identifier information for the token and install it into the token.
673IdentifierInfo *Preprocessor::LookUpIdentifierInfo(Token &Identifier,
674                                                   const char *BufPtr) {
675  assert(Identifier.is(tok::identifier) && "Not an identifier!");
676  assert(Identifier.getIdentifierInfo() == 0 && "Identinfo already exists!");
677
678  // Look up this token, see if it is a macro, or if it is a language keyword.
679  IdentifierInfo *II;
680  if (BufPtr && !Identifier.needsCleaning()) {
681    // No cleaning needed, just use the characters from the lexed buffer.
682    II = getIdentifierInfo(BufPtr, BufPtr+Identifier.getLength());
683  } else {
684    // Cleaning needed, alloca a buffer, clean into it, then use the buffer.
685    llvm::SmallVector<char, 64> IdentifierBuffer;
686    IdentifierBuffer.resize(Identifier.getLength());
687    const char *TmpBuf = &IdentifierBuffer[0];
688    unsigned Size = getSpelling(Identifier, TmpBuf);
689    II = getIdentifierInfo(TmpBuf, TmpBuf+Size);
690  }
691  Identifier.setIdentifierInfo(II);
692  return II;
693}
694
695
696/// HandleIdentifier - This callback is invoked when the lexer reads an
697/// identifier.  This callback looks up the identifier in the map and/or
698/// potentially macro expands it or turns it into a named token (like 'for').
699///
700/// Note that callers of this method are guarded by checking the
701/// IdentifierInfo's 'isHandleIdentifierCase' bit.  If this method changes, the
702/// IdentifierInfo methods that compute these properties will need to change to
703/// match.
704void Preprocessor::HandleIdentifier(Token &Identifier) {
705  assert(Identifier.getIdentifierInfo() &&
706         "Can't handle identifiers without identifier info!");
707
708  IdentifierInfo &II = *Identifier.getIdentifierInfo();
709
710  // If this identifier was poisoned, and if it was not produced from a macro
711  // expansion, emit an error.
712  if (II.isPoisoned() && CurPPLexer) {
713    if (&II != Ident__VA_ARGS__)   // We warn about __VA_ARGS__ with poisoning.
714      Diag(Identifier, diag::err_pp_used_poisoned_id);
715    else
716      Diag(Identifier, diag::ext_pp_bad_vaargs_use);
717  }
718
719  // If this is a macro to be expanded, do it.
720  if (MacroInfo *MI = getMacroInfo(&II)) {
721    if (!DisableMacroExpansion && !Identifier.isExpandDisabled()) {
722      if (MI->isEnabled()) {
723        if (!HandleMacroExpandedIdentifier(Identifier, MI))
724          return;
725      } else {
726        // C99 6.10.3.4p2 says that a disabled macro may never again be
727        // expanded, even if it's in a context where it could be expanded in the
728        // future.
729        Identifier.setFlag(Token::DisableExpand);
730      }
731    }
732  }
733
734  // C++ 2.11p2: If this is an alternative representation of a C++ operator,
735  // then we act as if it is the actual operator and not the textual
736  // representation of it.
737  if (II.isCPlusPlusOperatorKeyword())
738    Identifier.setIdentifierInfo(0);
739
740  // If this is an extension token, diagnose its use.
741  // We avoid diagnosing tokens that originate from macro definitions.
742  if (II.isExtensionToken() && Features.C99 && !DisableMacroExpansion)
743    Diag(Identifier, diag::ext_token_used);
744}
745