Preprocessor.cpp revision 6a170eb3ea6d6319277becabef68eb1a26bf8766 
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[MDNI] - 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/Basic/Diagnostic.h"
34#include "clang/Basic/SourceManager.h"
35#include "clang/Basic/TargetInfo.h"
36#include "llvm/ADT/APFloat.h"
37#include "llvm/ADT/SmallVector.h"
38#include "llvm/Support/MemoryBuffer.h"
39#include "llvm/Support/Streams.h"
40using namespace clang;
41
42//===----------------------------------------------------------------------===//
43
44PreprocessorFactory::~PreprocessorFactory() {}
45
46Preprocessor::Preprocessor(Diagnostic &diags, const LangOptions &opts,
47                           TargetInfo &target, SourceManager &SM,
48                           HeaderSearch &Headers,
49                           IdentifierInfoLookup* IILookup)
50  : Diags(diags), Features(opts), Target(target), FileMgr(Headers.getFileMgr()),
51    SourceMgr(SM), HeaderInfo(Headers), Identifiers(opts, IILookup),
52    CurPPLexer(0), CurDirLookup(0), Callbacks(0) {
53  ScratchBuf = new ScratchBuffer(SourceMgr);
54
55  // Clear stats.
56  NumDirectives = NumDefined = NumUndefined = NumPragma = 0;
57  NumIf = NumElse = NumEndif = 0;
58  NumEnteredSourceFiles = 0;
59  NumMacroExpanded = NumFnMacroExpanded = NumBuiltinMacroExpanded = 0;
60  NumFastMacroExpanded = NumTokenPaste = NumFastTokenPaste = 0;
61  MaxIncludeStackDepth = 0;
62  NumSkipped = 0;
63
64  // Default to discarding comments.
65  KeepComments = false;
66  KeepMacroComments = false;
67
68  // Macro expansion is enabled.
69  DisableMacroExpansion = false;
70  InMacroArgs = false;
71  NumCachedTokenLexers = 0;
72
73  CachedLexPos = 0;
74
75  // "Poison" __VA_ARGS__, which can only appear in the expansion of a macro.
76  // This gets unpoisoned where it is allowed.
77  (Ident__VA_ARGS__ = getIdentifierInfo("__VA_ARGS__"))->setIsPoisoned();
78
79  // Initialize the pragma handlers.
80  PragmaHandlers = new PragmaNamespace(0);
81  RegisterBuiltinPragmas();
82
83  // Initialize builtin macros like __LINE__ and friends.
84  RegisterBuiltinMacros();
85}
86
87Preprocessor::~Preprocessor() {
88  assert(BacktrackPositions.empty() && "EnableBacktrack/Backtrack imbalance!");
89
90  while (!IncludeMacroStack.empty()) {
91    delete IncludeMacroStack.back().TheLexer;
92    delete IncludeMacroStack.back().TheTokenLexer;
93    IncludeMacroStack.pop_back();
94  }
95
96  // Free any macro definitions.
97  for (llvm::DenseMap<IdentifierInfo*, MacroInfo*>::iterator I =
98       Macros.begin(), E = Macros.end(); I != E; ++I) {
99    // We don't need to free the MacroInfo objects directly.  These
100    // will be released when the BumpPtrAllocator 'BP' object gets
101    // destroyed. We still need to run the dstor, however, to free
102    // memory alocated by MacroInfo.
103    I->second->~MacroInfo();
104    I->first->setHasMacroDefinition(false);
105  }
106
107  // Free any cached macro expanders.
108  for (unsigned i = 0, e = NumCachedTokenLexers; i != e; ++i)
109    delete TokenLexerCache[i];
110
111  // Release pragma information.
112  delete PragmaHandlers;
113
114  // Delete the scratch buffer info.
115  delete ScratchBuf;
116
117  delete Callbacks;
118}
119
120void Preprocessor::DumpToken(const Token &Tok, bool DumpFlags) const {
121  llvm::cerr << tok::getTokenName(Tok.getKind()) << " '"
122             << getSpelling(Tok) << "'";
123
124  if (!DumpFlags) return;
125
126  llvm::cerr << "\t";
127  if (Tok.isAtStartOfLine())
128    llvm::cerr << " [StartOfLine]";
129  if (Tok.hasLeadingSpace())
130    llvm::cerr << " [LeadingSpace]";
131  if (Tok.isExpandDisabled())
132    llvm::cerr << " [ExpandDisabled]";
133  if (Tok.needsCleaning()) {
134    const char *Start = SourceMgr.getCharacterData(Tok.getLocation());
135    llvm::cerr << " [UnClean='" << std::string(Start, Start+Tok.getLength())
136               << "']";
137  }
138
139  llvm::cerr << "\tLoc=<";
140  DumpLocation(Tok.getLocation());
141  llvm::cerr << ">";
142}
143
144void Preprocessor::DumpLocation(SourceLocation Loc) const {
145  SourceLocation LogLoc = SourceMgr.getInstantiationLoc(Loc);
146  llvm::cerr << SourceMgr.getSourceName(LogLoc) << ':'
147             << SourceMgr.getLineNumber(LogLoc) << ':'
148             << SourceMgr.getColumnNumber(LogLoc);
149
150  SourceLocation SpellingLoc = SourceMgr.getSpellingLoc(Loc);
151  if (SpellingLoc != LogLoc) {
152    llvm::cerr << " <SpellingLoc=";
153    DumpLocation(SpellingLoc);
154    llvm::cerr << ">";
155  }
156}
157
158void Preprocessor::DumpMacro(const MacroInfo &MI) const {
159  llvm::cerr << "MACRO: ";
160  for (unsigned i = 0, e = MI.getNumTokens(); i != e; ++i) {
161    DumpToken(MI.getReplacementToken(i));
162    llvm::cerr << "  ";
163  }
164  llvm::cerr << "\n";
165}
166
167void Preprocessor::PrintStats() {
168  llvm::cerr << "\n*** Preprocessor Stats:\n";
169  llvm::cerr << NumDirectives << " directives found:\n";
170  llvm::cerr << "  " << NumDefined << " #define.\n";
171  llvm::cerr << "  " << NumUndefined << " #undef.\n";
172  llvm::cerr << "  #include/#include_next/#import:\n";
173  llvm::cerr << "    " << NumEnteredSourceFiles << " source files entered.\n";
174  llvm::cerr << "    " << MaxIncludeStackDepth << " max include stack depth\n";
175  llvm::cerr << "  " << NumIf << " #if/#ifndef/#ifdef.\n";
176  llvm::cerr << "  " << NumElse << " #else/#elif.\n";
177  llvm::cerr << "  " << NumEndif << " #endif.\n";
178  llvm::cerr << "  " << NumPragma << " #pragma.\n";
179  llvm::cerr << NumSkipped << " #if/#ifndef#ifdef regions skipped\n";
180
181  llvm::cerr << NumMacroExpanded << "/" << NumFnMacroExpanded << "/"
182             << NumBuiltinMacroExpanded << " obj/fn/builtin macros expanded, "
183             << NumFastMacroExpanded << " on the fast path.\n";
184  llvm::cerr << (NumFastTokenPaste+NumTokenPaste)
185             << " token paste (##) operations performed, "
186             << NumFastTokenPaste << " on the fast path.\n";
187}
188
189//===----------------------------------------------------------------------===//
190// Token Spelling
191//===----------------------------------------------------------------------===//
192
193
194/// getSpelling() - Return the 'spelling' of this token.  The spelling of a
195/// token are the characters used to represent the token in the source file
196/// after trigraph expansion and escaped-newline folding.  In particular, this
197/// wants to get the true, uncanonicalized, spelling of things like digraphs
198/// UCNs, etc.
199std::string Preprocessor::getSpelling(const Token &Tok) const {
200  assert((int)Tok.getLength() >= 0 && "Token character range is bogus!");
201  const char* TokStart;
202
203  if (PTH) {
204    if (unsigned Len = PTH->getSpelling(Tok.getLocation(), TokStart)) {
205      assert(!Tok.needsCleaning());
206      return std::string(TokStart, TokStart+Len);
207    }
208  }
209
210  // If this token contains nothing interesting, return it directly.
211  TokStart = SourceMgr.getCharacterData(Tok.getLocation());
212  if (!Tok.needsCleaning())
213    return std::string(TokStart, TokStart+Tok.getLength());
214
215  std::string Result;
216  Result.reserve(Tok.getLength());
217
218  // Otherwise, hard case, relex the characters into the string.
219  for (const char *Ptr = TokStart, *End = TokStart+Tok.getLength();
220       Ptr != End; ) {
221    unsigned CharSize;
222    Result.push_back(Lexer::getCharAndSizeNoWarn(Ptr, CharSize, Features));
223    Ptr += CharSize;
224  }
225  assert(Result.size() != unsigned(Tok.getLength()) &&
226         "NeedsCleaning flag set on something that didn't need cleaning!");
227  return Result;
228}
229
230/// getSpelling - This method is used to get the spelling of a token into a
231/// preallocated buffer, instead of as an std::string.  The caller is required
232/// to allocate enough space for the token, which is guaranteed to be at least
233/// Tok.getLength() bytes long.  The actual length of the token is returned.
234///
235/// Note that this method may do two possible things: it may either fill in
236/// the buffer specified with characters, or it may *change the input pointer*
237/// to point to a constant buffer with the data already in it (avoiding a
238/// copy).  The caller is not allowed to modify the returned buffer pointer
239/// if an internal buffer is returned.
240unsigned Preprocessor::getSpelling(const Token &Tok,
241                                   const char *&Buffer) const {
242  assert((int)Tok.getLength() >= 0 && "Token character range is bogus!");
243
244  // If this token is an identifier, just return the string from the identifier
245  // table, which is very quick.
246  if (const IdentifierInfo *II = Tok.getIdentifierInfo()) {
247    Buffer = II->getName();
248    return II->getLength();
249  }
250
251  // If using PTH, try and get the spelling from the PTH file.
252  if (PTH) {
253    unsigned Len;
254
255    if (CurPTHLexer) {
256      Len = CurPTHLexer.get()->getSpelling(Tok.getLocation(), Buffer);
257    } else {
258      Len = PTH->getSpelling(Tok.getLocation(), Buffer);
259    }
260
261    // Did we find a spelling?  If so return its length.  Otherwise fall
262    // back to the default behavior for getting the spelling by looking at
263    // at the source code.
264    if (Len)
265      return Len;
266  }
267
268  // Otherwise, compute the start of the token in the input lexer buffer.
269  const char *TokStart = SourceMgr.getCharacterData(Tok.getLocation());
270
271  // If this token contains nothing interesting, return it directly.
272  if (!Tok.needsCleaning()) {
273    Buffer = TokStart;
274    return Tok.getLength();
275  }
276  // Otherwise, hard case, relex the characters into the string.
277  char *OutBuf = const_cast<char*>(Buffer);
278  for (const char *Ptr = TokStart, *End = TokStart+Tok.getLength();
279       Ptr != End; ) {
280    unsigned CharSize;
281    *OutBuf++ = Lexer::getCharAndSizeNoWarn(Ptr, CharSize, Features);
282    Ptr += CharSize;
283  }
284  assert(unsigned(OutBuf-Buffer) != Tok.getLength() &&
285         "NeedsCleaning flag set on something that didn't need cleaning!");
286
287  return OutBuf-Buffer;
288}
289
290
291/// CreateString - Plop the specified string into a scratch buffer and return a
292/// location for it.  If specified, the source location provides a source
293/// location for the token.
294SourceLocation Preprocessor::
295CreateString(const char *Buf, unsigned Len, SourceLocation SLoc) {
296  if (SLoc.isValid())
297    return ScratchBuf->getToken(Buf, Len, SLoc);
298  return ScratchBuf->getToken(Buf, Len);
299}
300
301
302/// AdvanceToTokenCharacter - Given a location that specifies the start of a
303/// token, return a new location that specifies a character within the token.
304SourceLocation Preprocessor::AdvanceToTokenCharacter(SourceLocation TokStart,
305                                                     unsigned CharNo) {
306  // If they request the first char of the token, we're trivially done.  If this
307  // is a macro expansion, it doesn't make sense to point to a character within
308  // the instantiation point (the name).  We could point to the source
309  // character, but without also pointing to instantiation info, this is
310  // confusing.
311  if (CharNo == 0 || TokStart.isMacroID()) return TokStart;
312
313  // Figure out how many physical characters away the specified instantiation
314  // character is.  This needs to take into consideration newlines and
315  // trigraphs.
316  const char *TokPtr = SourceMgr.getCharacterData(TokStart);
317  unsigned PhysOffset = 0;
318
319  // The usual case is that tokens don't contain anything interesting.  Skip
320  // over the uninteresting characters.  If a token only consists of simple
321  // chars, this method is extremely fast.
322  while (CharNo && Lexer::isObviouslySimpleCharacter(*TokPtr))
323    ++TokPtr, --CharNo, ++PhysOffset;
324
325  // If we have a character that may be a trigraph or escaped newline, use a
326  // lexer to parse it correctly.
327  if (CharNo != 0) {
328    // Skip over characters the remaining characters.
329    for (; CharNo; --CharNo) {
330      unsigned Size;
331      Lexer::getCharAndSizeNoWarn(TokPtr, Size, Features);
332      TokPtr += Size;
333      PhysOffset += Size;
334    }
335  }
336
337  return TokStart.getFileLocWithOffset(PhysOffset);
338}
339
340
341//===----------------------------------------------------------------------===//
342// Preprocessor Initialization Methods
343//===----------------------------------------------------------------------===//
344
345// Append a #define line to Buf for Macro.  Macro should be of the form XXX,
346// in which case we emit "#define XXX 1" or "XXX=Y z W" in which case we emit
347// "#define XXX Y z W".  To get a #define with no value, use "XXX=".
348static void DefineBuiltinMacro(std::vector<char> &Buf, const char *Macro,
349                               const char *Command = "#define ") {
350  Buf.insert(Buf.end(), Command, Command+strlen(Command));
351  if (const char *Equal = strchr(Macro, '=')) {
352    // Turn the = into ' '.
353    Buf.insert(Buf.end(), Macro, Equal);
354    Buf.push_back(' ');
355    Buf.insert(Buf.end(), Equal+1, Equal+strlen(Equal));
356  } else {
357    // Push "macroname 1".
358    Buf.insert(Buf.end(), Macro, Macro+strlen(Macro));
359    Buf.push_back(' ');
360    Buf.push_back('1');
361  }
362  Buf.push_back('\n');
363}
364
365/// PickFP - This is used to pick a value based on the FP semantics of the
366/// specified FP model.
367template <typename T>
368static T PickFP(const llvm::fltSemantics *Sem, T IEEESingleVal,
369                T IEEEDoubleVal, T X87DoubleExtendedVal, T PPCDoubleDoubleVal) {
370  if (Sem == &llvm::APFloat::IEEEsingle)
371    return IEEESingleVal;
372  if (Sem == &llvm::APFloat::IEEEdouble)
373    return IEEEDoubleVal;
374  if (Sem == &llvm::APFloat::x87DoubleExtended)
375    return X87DoubleExtendedVal;
376  assert(Sem == &llvm::APFloat::PPCDoubleDouble);
377  return PPCDoubleDoubleVal;
378}
379
380static void DefineFloatMacros(std::vector<char> &Buf, const char *Prefix,
381                              const llvm::fltSemantics *Sem) {
382  const char *DenormMin, *Epsilon, *Max, *Min;
383  DenormMin = PickFP(Sem, "1.40129846e-45F", "4.9406564584124654e-324",
384                     "3.64519953188247460253e-4951L",
385                     "4.94065645841246544176568792868221e-324L");
386  int Digits = PickFP(Sem, 6, 15, 18, 31);
387  Epsilon = PickFP(Sem, "1.19209290e-7F", "2.2204460492503131e-16",
388                   "1.08420217248550443401e-19L",
389                   "4.94065645841246544176568792868221e-324L");
390  int HasInifinity = 1, HasQuietNaN = 1;
391  int MantissaDigits = PickFP(Sem, 24, 53, 64, 106);
392  int Min10Exp = PickFP(Sem, -37, -307, -4931, -291);
393  int Max10Exp = PickFP(Sem, 38, 308, 4932, 308);
394  int MinExp = PickFP(Sem, -125, -1021, -16381, -968);
395  int MaxExp = PickFP(Sem, 128, 1024, 16384, 1024);
396  Min = PickFP(Sem, "1.17549435e-38F", "2.2250738585072014e-308",
397               "3.36210314311209350626e-4932L",
398               "2.00416836000897277799610805135016e-292L");
399  Max = PickFP(Sem, "3.40282347e+38F", "1.7976931348623157e+308",
400               "1.18973149535723176502e+4932L",
401               "1.79769313486231580793728971405301e+308L");
402
403  char MacroBuf[60];
404  sprintf(MacroBuf, "__%s_DENORM_MIN__=%s", Prefix, DenormMin);
405  DefineBuiltinMacro(Buf, MacroBuf);
406  sprintf(MacroBuf, "__%s_DIG__=%d", Prefix, Digits);
407  DefineBuiltinMacro(Buf, MacroBuf);
408  sprintf(MacroBuf, "__%s_EPSILON__=%s", Prefix, Epsilon);
409  DefineBuiltinMacro(Buf, MacroBuf);
410  sprintf(MacroBuf, "__%s_HAS_INFINITY__=%d", Prefix, HasInifinity);
411  DefineBuiltinMacro(Buf, MacroBuf);
412  sprintf(MacroBuf, "__%s_HAS_QUIET_NAN__=%d", Prefix, HasQuietNaN);
413  DefineBuiltinMacro(Buf, MacroBuf);
414  sprintf(MacroBuf, "__%s_MANT_DIG__=%d", Prefix, MantissaDigits);
415  DefineBuiltinMacro(Buf, MacroBuf);
416  sprintf(MacroBuf, "__%s_MAX_10_EXP__=%d", Prefix, Max10Exp);
417  DefineBuiltinMacro(Buf, MacroBuf);
418  sprintf(MacroBuf, "__%s_MAX_EXP__=%d", Prefix, MaxExp);
419  DefineBuiltinMacro(Buf, MacroBuf);
420  sprintf(MacroBuf, "__%s_MAX__=%s", Prefix, Max);
421  DefineBuiltinMacro(Buf, MacroBuf);
422  sprintf(MacroBuf, "__%s_MIN_10_EXP__=(%d)", Prefix, Min10Exp);
423  DefineBuiltinMacro(Buf, MacroBuf);
424  sprintf(MacroBuf, "__%s_MIN_EXP__=(%d)", Prefix, MinExp);
425  DefineBuiltinMacro(Buf, MacroBuf);
426  sprintf(MacroBuf, "__%s_MIN__=%s", Prefix, Min);
427  DefineBuiltinMacro(Buf, MacroBuf);
428}
429
430
431static void InitializePredefinedMacros(Preprocessor &PP,
432                                       std::vector<char> &Buf) {
433  // Compiler version introspection macros.
434  DefineBuiltinMacro(Buf, "__llvm__=1");   // LLVM Backend
435  DefineBuiltinMacro(Buf, "__clang__=1");  // Clang Frontend
436
437  // Currently claim to be compatible with GCC 4.2.1-5621.
438  DefineBuiltinMacro(Buf, "__APPLE_CC__=5621");
439  DefineBuiltinMacro(Buf, "__GNUC_MINOR__=2");
440  DefineBuiltinMacro(Buf, "__GNUC_PATCHLEVEL__=1");
441  DefineBuiltinMacro(Buf, "__GNUC__=4");
442  DefineBuiltinMacro(Buf, "__GXX_ABI_VERSION=1002");
443  DefineBuiltinMacro(Buf, "__VERSION__=\"4.2.1 (Apple Computer, Inc. "
444                     "build 5621) (dot 3)\"");
445
446
447  // Initialize language-specific preprocessor defines.
448
449  // FIXME: Implement magic like cpp_init_builtins for things like __STDC__
450  // and __DATE__ etc.
451  // These should all be defined in the preprocessor according to the
452  // current language configuration.
453  if (!PP.getLangOptions().Microsoft)
454    DefineBuiltinMacro(Buf, "__STDC__=1");
455  if (PP.getLangOptions().AsmPreprocessor)
456    DefineBuiltinMacro(Buf, "__ASSEMBLER__=1");
457  if (PP.getLangOptions().C99 && !PP.getLangOptions().CPlusPlus)
458    DefineBuiltinMacro(Buf, "__STDC_VERSION__=199901L");
459  else if (0) // STDC94 ?
460    DefineBuiltinMacro(Buf, "__STDC_VERSION__=199409L");
461
462  DefineBuiltinMacro(Buf, "__STDC_HOSTED__=1");
463  if (PP.getLangOptions().ObjC1) {
464    DefineBuiltinMacro(Buf, "__OBJC__=1");
465
466    if (PP.getLangOptions().getGCMode() == LangOptions::NonGC) {
467      DefineBuiltinMacro(Buf, "__weak=");
468      DefineBuiltinMacro(Buf, "__strong=");
469    } else {
470      DefineBuiltinMacro(Buf, "__weak=__attribute__((objc_gc(weak)))");
471      DefineBuiltinMacro(Buf, "__strong=__attribute__((objc_gc(strong)))");
472      DefineBuiltinMacro(Buf, "__OBJC_GC__=1");
473    }
474
475    if (PP.getLangOptions().NeXTRuntime)
476      DefineBuiltinMacro(Buf, "__NEXT_RUNTIME__=1");
477  }
478
479  // darwin_constant_cfstrings controls this. This is also dependent
480  // on other things like the runtime I believe.  This is set even for C code.
481  DefineBuiltinMacro(Buf, "__CONSTANT_CFSTRINGS__=1");
482
483  if (PP.getLangOptions().ObjC2)
484    DefineBuiltinMacro(Buf, "OBJC_NEW_PROPERTIES");
485
486  if (PP.getLangOptions().PascalStrings)
487    DefineBuiltinMacro(Buf, "__PASCAL_STRINGS__");
488
489  if (PP.getLangOptions().Blocks) {
490    DefineBuiltinMacro(Buf, "__block=__attribute__((__blocks__(byref)))");
491    DefineBuiltinMacro(Buf, "__BLOCKS__=1");
492  }
493
494  if (PP.getLangOptions().CPlusPlus) {
495    DefineBuiltinMacro(Buf, "__DEPRECATED=1");
496    DefineBuiltinMacro(Buf, "__EXCEPTIONS=1");
497    DefineBuiltinMacro(Buf, "__GNUG__=4");
498    DefineBuiltinMacro(Buf, "__GXX_WEAK__=1");
499    DefineBuiltinMacro(Buf, "__cplusplus=1");
500    DefineBuiltinMacro(Buf, "__private_extern__=extern");
501  }
502
503  // Filter out some microsoft extensions when trying to parse in ms-compat
504  // mode.
505  if (PP.getLangOptions().Microsoft) {
506    DefineBuiltinMacro(Buf, "_cdecl=__cdecl");
507    DefineBuiltinMacro(Buf, "__int8=char");
508    DefineBuiltinMacro(Buf, "__int16=short");
509    DefineBuiltinMacro(Buf, "__int32=int");
510    DefineBuiltinMacro(Buf, "__int64=long long");
511  }
512
513
514  // Initialize target-specific preprocessor defines.
515  const TargetInfo &TI = PP.getTargetInfo();
516
517  // Define type sizing macros based on the target properties.
518  assert(TI.getCharWidth() == 8 && "Only support 8-bit char so far");
519  DefineBuiltinMacro(Buf, "__CHAR_BIT__=8");
520  DefineBuiltinMacro(Buf, "__SCHAR_MAX__=127");
521
522  assert(TI.getWCharWidth() == 32 && "Only support 32-bit wchar so far");
523  DefineBuiltinMacro(Buf, "__WCHAR_MAX__=2147483647");
524  DefineBuiltinMacro(Buf, "__WCHAR_TYPE__=int");
525  DefineBuiltinMacro(Buf, "__WINT_TYPE__=int");
526
527  assert(TI.getShortWidth() == 16 && "Only support 16-bit short so far");
528  DefineBuiltinMacro(Buf, "__SHRT_MAX__=32767");
529
530  if (TI.getIntWidth() == 32)
531    DefineBuiltinMacro(Buf, "__INT_MAX__=2147483647");
532  else if (TI.getIntWidth() == 16)
533    DefineBuiltinMacro(Buf, "__INT_MAX__=32767");
534  else
535    assert(0 && "Unknown integer size");
536
537  if (TI.getLongLongWidth() == 64)
538    DefineBuiltinMacro(Buf, "__LONG_LONG_MAX__=9223372036854775807LL");
539  else if (TI.getLongLongWidth() == 32)
540    DefineBuiltinMacro(Buf, "__LONG_LONG_MAX__=2147483647L");
541
542  if (TI.getLongWidth() == 32)
543    DefineBuiltinMacro(Buf, "__LONG_MAX__=2147483647L");
544  else if (TI.getLongWidth() == 64)
545    DefineBuiltinMacro(Buf, "__LONG_MAX__=9223372036854775807L");
546  else if (TI.getLongWidth() == 16)
547    DefineBuiltinMacro(Buf, "__LONG_MAX__=32767L");
548  else
549    assert(0 && "Unknown long size");
550  char MacroBuf[60];
551  sprintf(MacroBuf, "__INTMAX_MAX__=%lld",
552          (TI.getIntMaxType() == TargetInfo::UnsignedLongLong?
553           (1LL << (TI.getLongLongWidth() - 1)) :
554           ((1LL << (TI.getLongLongWidth() - 2)) - 1)));
555  DefineBuiltinMacro(Buf, MacroBuf);
556
557  if (TI.getIntMaxType() == TargetInfo::UnsignedLongLong)
558    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=unsigned long long int");
559  else if (TI.getIntMaxType() == TargetInfo::SignedLongLong)
560    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=long long int");
561  else if (TI.getIntMaxType() == TargetInfo::UnsignedLong)
562    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=unsigned long int");
563  else if (TI.getIntMaxType() == TargetInfo::SignedLong)
564    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=long int");
565  else if (TI.getIntMaxType() == TargetInfo::UnsignedInt)
566    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=unsigned int");
567  else
568    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=int");
569
570  if (TI.getUIntMaxType() == TargetInfo::UnsignedLongLong)
571    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=unsigned long long int");
572  else if (TI.getUIntMaxType() == TargetInfo::SignedLongLong)
573    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=long long int");
574  else if (TI.getUIntMaxType() == TargetInfo::UnsignedLong)
575    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=unsigned long int");
576  else if (TI.getUIntMaxType() == TargetInfo::SignedLong)
577    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=long int");
578  else if (TI.getUIntMaxType() == TargetInfo::UnsignedInt)
579    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=unsigned int");
580  else
581    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=int");
582
583  if (TI.getPtrDiffType(0) == TargetInfo::UnsignedLongLong)
584    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=unsigned long long int");
585  else if (TI.getPtrDiffType(0) == TargetInfo::SignedLongLong)
586    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=long long int");
587  else if (TI.getPtrDiffType(0) == TargetInfo::UnsignedLong)
588    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=unsigned long int");
589  else if (TI.getPtrDiffType(0) == TargetInfo::SignedLong)
590    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=long int");
591  else if (TI.getPtrDiffType(0) == TargetInfo::UnsignedInt)
592    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=unsigned int");
593  else
594    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=int");
595
596  if (TI.getSizeType() == TargetInfo::UnsignedLongLong)
597    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=unsigned long long int");
598  else if (TI.getSizeType() == TargetInfo::SignedLongLong)
599    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=long long int");
600  else if (TI.getSizeType() == TargetInfo::UnsignedLong)
601    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=unsigned long int");
602  else if (TI.getSizeType() == TargetInfo::SignedLong)
603    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=long int");
604  else if (TI.getSizeType() == TargetInfo::UnsignedInt)
605    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=unsigned int");
606  else if (TI.getSizeType() == TargetInfo::SignedInt)
607    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=int");
608  else
609    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=unsigned short");
610
611  DefineFloatMacros(Buf, "FLT", &TI.getFloatFormat());
612  DefineFloatMacros(Buf, "DBL", &TI.getDoubleFormat());
613  DefineFloatMacros(Buf, "LDBL", &TI.getLongDoubleFormat());
614
615
616  // Add __builtin_va_list typedef.
617  {
618    const char *VAList = TI.getVAListDeclaration();
619    Buf.insert(Buf.end(), VAList, VAList+strlen(VAList));
620    Buf.push_back('\n');
621  }
622
623  if (const char *Prefix = TI.getUserLabelPrefix()) {
624    sprintf(MacroBuf, "__USER_LABEL_PREFIX__=%s", Prefix);
625    DefineBuiltinMacro(Buf, MacroBuf);
626  }
627
628  // Build configuration options.  FIXME: these should be controlled by
629  // command line options or something.
630  DefineBuiltinMacro(Buf, "__DYNAMIC__=1");
631  DefineBuiltinMacro(Buf, "__FINITE_MATH_ONLY__=0");
632  DefineBuiltinMacro(Buf, "__NO_INLINE__=1");
633  DefineBuiltinMacro(Buf, "__PIC__=1");
634
635  // Macros to control C99 numerics and <float.h>
636  DefineBuiltinMacro(Buf, "__FLT_EVAL_METHOD__=0");
637  DefineBuiltinMacro(Buf, "__FLT_RADIX__=2");
638  sprintf(MacroBuf, "__DECIMAL_DIG__=%d",
639          PickFP(&TI.getLongDoubleFormat(), -1/*FIXME*/, 17, 21, 33));
640  DefineBuiltinMacro(Buf, MacroBuf);
641
642  // Get other target #defines.
643  TI.getTargetDefines(Buf);
644
645  // FIXME: Should emit a #line directive here.
646}
647
648
649/// EnterMainSourceFile - Enter the specified FileID as the main source file,
650/// which implicitly adds the builtin defines etc.
651void Preprocessor::EnterMainSourceFile() {
652
653  FileID MainFileID = SourceMgr.getMainFileID();
654
655  // Enter the main file source buffer.
656  EnterSourceFile(MainFileID, 0);
657
658  // Tell the header info that the main file was entered.  If the file is later
659  // #imported, it won't be re-entered.
660  if (const FileEntry *FE = SourceMgr.getFileEntryForID(MainFileID))
661    HeaderInfo.IncrementIncludeCount(FE);
662
663  std::vector<char> PrologFile;
664  PrologFile.reserve(4080);
665
666  // Install things like __POWERPC__, __GNUC__, etc into the macro table.
667  InitializePredefinedMacros(*this, PrologFile);
668
669  // Add on the predefines from the driver.
670  PrologFile.insert(PrologFile.end(), Predefines.begin(), Predefines.end());
671
672  // Memory buffer must end with a null byte!
673  PrologFile.push_back(0);
674
675  // Now that we have emitted the predefined macros, #includes, etc into
676  // PrologFile, preprocess it to populate the initial preprocessor state.
677  llvm::MemoryBuffer *SB =
678    llvm::MemoryBuffer::getMemBufferCopy(&PrologFile.front(),&PrologFile.back(),
679                                         "<predefines>");
680  assert(SB && "Cannot fail to create predefined source buffer");
681  FileID FID = SourceMgr.createFileIDForMemBuffer(SB);
682  assert(!FID.isInvalid() && "Could not create FileID for predefines?");
683
684  // Start parsing the predefines.
685  EnterSourceFile(FID, 0);
686}
687
688
689//===----------------------------------------------------------------------===//
690// Lexer Event Handling.
691//===----------------------------------------------------------------------===//
692
693/// LookUpIdentifierInfo - Given a tok::identifier token, look up the
694/// identifier information for the token and install it into the token.
695IdentifierInfo *Preprocessor::LookUpIdentifierInfo(Token &Identifier,
696                                                   const char *BufPtr) {
697  assert(Identifier.is(tok::identifier) && "Not an identifier!");
698  assert(Identifier.getIdentifierInfo() == 0 && "Identinfo already exists!");
699
700  // Look up this token, see if it is a macro, or if it is a language keyword.
701  IdentifierInfo *II;
702  if (BufPtr && !Identifier.needsCleaning()) {
703    // No cleaning needed, just use the characters from the lexed buffer.
704    II = getIdentifierInfo(BufPtr, BufPtr+Identifier.getLength());
705  } else {
706    // Cleaning needed, alloca a buffer, clean into it, then use the buffer.
707    llvm::SmallVector<char, 64> IdentifierBuffer;
708    IdentifierBuffer.resize(Identifier.getLength());
709    const char *TmpBuf = &IdentifierBuffer[0];
710    unsigned Size = getSpelling(Identifier, TmpBuf);
711    II = getIdentifierInfo(TmpBuf, TmpBuf+Size);
712  }
713  Identifier.setIdentifierInfo(II);
714  return II;
715}
716
717
718/// HandleIdentifier - This callback is invoked when the lexer reads an
719/// identifier.  This callback looks up the identifier in the map and/or
720/// potentially macro expands it or turns it into a named token (like 'for').
721///
722/// Note that callers of this method are guarded by checking the
723/// IdentifierInfo's 'isHandleIdentifierCase' bit.  If this method changes, the
724/// IdentifierInfo methods that compute these properties will need to change to
725/// match.
726void Preprocessor::HandleIdentifier(Token &Identifier) {
727  assert(Identifier.getIdentifierInfo() &&
728         "Can't handle identifiers without identifier info!");
729
730  IdentifierInfo &II = *Identifier.getIdentifierInfo();
731
732  // If this identifier was poisoned, and if it was not produced from a macro
733  // expansion, emit an error.
734  if (II.isPoisoned() && CurPPLexer) {
735    if (&II != Ident__VA_ARGS__)   // We warn about __VA_ARGS__ with poisoning.
736      Diag(Identifier, diag::err_pp_used_poisoned_id);
737    else
738      Diag(Identifier, diag::ext_pp_bad_vaargs_use);
739  }
740
741  // If this is a macro to be expanded, do it.
742  if (MacroInfo *MI = getMacroInfo(&II)) {
743    if (!DisableMacroExpansion && !Identifier.isExpandDisabled()) {
744      if (MI->isEnabled()) {
745        if (!HandleMacroExpandedIdentifier(Identifier, MI))
746          return;
747      } else {
748        // C99 6.10.3.4p2 says that a disabled macro may never again be
749        // expanded, even if it's in a context where it could be expanded in the
750        // future.
751        Identifier.setFlag(Token::DisableExpand);
752      }
753    }
754  }
755
756  // C++ 2.11p2: If this is an alternative representation of a C++ operator,
757  // then we act as if it is the actual operator and not the textual
758  // representation of it.
759  if (II.isCPlusPlusOperatorKeyword())
760    Identifier.setIdentifierInfo(0);
761
762  // Change the kind of this identifier to the appropriate token kind, e.g.
763  // turning "for" into a keyword.
764  Identifier.setKind(II.getTokenID());
765
766  // If this is an extension token, diagnose its use.
767  // We avoid diagnosing tokens that originate from macro definitions.
768  if (II.isExtensionToken() && Features.C99 && !DisableMacroExpansion)
769    Diag(Identifier, diag::ext_token_used);
770}
771