Preprocessor.cpp revision f02f6f0ee3e05b958bcf67f00f4503671d67eccd 
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  : Diags(diags), Features(opts), Target(target), FileMgr(Headers.getFileMgr()),
50    SourceMgr(SM), HeaderInfo(Headers), Identifiers(opts),
51    CurPPLexer(0), CurDirLookup(0), Callbacks(0) {
52  ScratchBuf = new ScratchBuffer(SourceMgr);
53
54  // Clear stats.
55  NumDirectives = NumDefined = NumUndefined = NumPragma = 0;
56  NumIf = NumElse = NumEndif = 0;
57  NumEnteredSourceFiles = 0;
58  NumMacroExpanded = NumFnMacroExpanded = NumBuiltinMacroExpanded = 0;
59  NumFastMacroExpanded = NumTokenPaste = NumFastTokenPaste = 0;
60  MaxIncludeStackDepth = 0;
61  NumSkipped = 0;
62
63  // Default to discarding comments.
64  KeepComments = false;
65  KeepMacroComments = false;
66
67  // Macro expansion is enabled.
68  DisableMacroExpansion = false;
69  InMacroArgs = false;
70  NumCachedTokenLexers = 0;
71
72  CachedLexPos = 0;
73
74  // "Poison" __VA_ARGS__, which can only appear in the expansion of a macro.
75  // This gets unpoisoned where it is allowed.
76  (Ident__VA_ARGS__ = getIdentifierInfo("__VA_ARGS__"))->setIsPoisoned();
77
78  // Initialize the pragma handlers.
79  PragmaHandlers = new PragmaNamespace(0);
80  RegisterBuiltinPragmas();
81
82  // Initialize builtin macros like __LINE__ and friends.
83  RegisterBuiltinMacros();
84}
85
86Preprocessor::~Preprocessor() {
87  assert(BacktrackPositions.empty() && "EnableBacktrack/Backtrack imbalance!");
88
89  while (!IncludeMacroStack.empty()) {
90    delete IncludeMacroStack.back().TheLexer;
91    delete IncludeMacroStack.back().TheTokenLexer;
92    IncludeMacroStack.pop_back();
93  }
94
95  // Free any macro definitions.
96  for (llvm::DenseMap<IdentifierInfo*, MacroInfo*>::iterator I =
97       Macros.begin(), E = Macros.end(); I != E; ++I) {
98    // We don't need to free the MacroInfo objects directly.  These
99    // will be released when the BumpPtrAllocator 'BP' object gets
100    // destroyed.
101    I->first->setHasMacroDefinition(false);
102  }
103
104  // Free any cached macro expanders.
105  for (unsigned i = 0, e = NumCachedTokenLexers; i != e; ++i)
106    delete TokenLexerCache[i];
107
108  // Release pragma information.
109  delete PragmaHandlers;
110
111  // Delete the scratch buffer info.
112  delete ScratchBuf;
113
114  delete Callbacks;
115}
116
117void Preprocessor::DumpToken(const Token &Tok, bool DumpFlags) const {
118  llvm::cerr << tok::getTokenName(Tok.getKind()) << " '"
119             << getSpelling(Tok) << "'";
120
121  if (!DumpFlags) return;
122
123  llvm::cerr << "\t";
124  if (Tok.isAtStartOfLine())
125    llvm::cerr << " [StartOfLine]";
126  if (Tok.hasLeadingSpace())
127    llvm::cerr << " [LeadingSpace]";
128  if (Tok.isExpandDisabled())
129    llvm::cerr << " [ExpandDisabled]";
130  if (Tok.needsCleaning()) {
131    const char *Start = SourceMgr.getCharacterData(Tok.getLocation());
132    llvm::cerr << " [UnClean='" << std::string(Start, Start+Tok.getLength())
133               << "']";
134  }
135
136  llvm::cerr << "\tLoc=<";
137  DumpLocation(Tok.getLocation());
138  llvm::cerr << ">";
139}
140
141void Preprocessor::DumpLocation(SourceLocation Loc) const {
142  SourceLocation LogLoc = SourceMgr.getLogicalLoc(Loc);
143  llvm::cerr << SourceMgr.getSourceName(LogLoc) << ':'
144             << SourceMgr.getLineNumber(LogLoc) << ':'
145             << SourceMgr.getColumnNumber(LogLoc);
146
147  SourceLocation PhysLoc = SourceMgr.getPhysicalLoc(Loc);
148  if (PhysLoc != LogLoc) {
149    llvm::cerr << " <PhysLoc=";
150    DumpLocation(PhysLoc);
151    llvm::cerr << ">";
152  }
153}
154
155void Preprocessor::DumpMacro(const MacroInfo &MI) const {
156  llvm::cerr << "MACRO: ";
157  for (unsigned i = 0, e = MI.getNumTokens(); i != e; ++i) {
158    DumpToken(MI.getReplacementToken(i));
159    llvm::cerr << "  ";
160  }
161  llvm::cerr << "\n";
162}
163
164void Preprocessor::PrintStats() {
165  llvm::cerr << "\n*** Preprocessor Stats:\n";
166  llvm::cerr << NumDirectives << " directives found:\n";
167  llvm::cerr << "  " << NumDefined << " #define.\n";
168  llvm::cerr << "  " << NumUndefined << " #undef.\n";
169  llvm::cerr << "  #include/#include_next/#import:\n";
170  llvm::cerr << "    " << NumEnteredSourceFiles << " source files entered.\n";
171  llvm::cerr << "    " << MaxIncludeStackDepth << " max include stack depth\n";
172  llvm::cerr << "  " << NumIf << " #if/#ifndef/#ifdef.\n";
173  llvm::cerr << "  " << NumElse << " #else/#elif.\n";
174  llvm::cerr << "  " << NumEndif << " #endif.\n";
175  llvm::cerr << "  " << NumPragma << " #pragma.\n";
176  llvm::cerr << NumSkipped << " #if/#ifndef#ifdef regions skipped\n";
177
178  llvm::cerr << NumMacroExpanded << "/" << NumFnMacroExpanded << "/"
179             << NumBuiltinMacroExpanded << " obj/fn/builtin macros expanded, "
180             << NumFastMacroExpanded << " on the fast path.\n";
181  llvm::cerr << (NumFastTokenPaste+NumTokenPaste)
182             << " token paste (##) operations performed, "
183             << NumFastTokenPaste << " on the fast path.\n";
184}
185
186//===----------------------------------------------------------------------===//
187// Token Spelling
188//===----------------------------------------------------------------------===//
189
190
191/// getSpelling() - Return the 'spelling' of this token.  The spelling of a
192/// token are the characters used to represent the token in the source file
193/// after trigraph expansion and escaped-newline folding.  In particular, this
194/// wants to get the true, uncanonicalized, spelling of things like digraphs
195/// UCNs, etc.
196std::string Preprocessor::getSpelling(const Token &Tok) const {
197  assert((int)Tok.getLength() >= 0 && "Token character range is bogus!");
198  const char* TokStart;
199
200  if (PTH) {
201    SourceLocation sloc = SourceMgr.getPhysicalLoc(Tok.getLocation());
202    unsigned fid = sloc.getFileID();
203    unsigned fpos = SourceMgr.getFullFilePos(sloc);
204    if (unsigned len = PTH->getSpelling(fid, fpos, 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      // We perform the const_cast<> here because we will only have a PTHLexer
257      // when grabbing a stream of tokens from the PTH file (and thus the
258      // Preprocessor state is allowed to change).  The PTHLexer can assume we are
259      // getting token spellings in the order of tokens, and thus can update
260      // its internal state so that it can quickly fetch spellings from the PTH
261      // file.
262      len =
263        const_cast<PTHLexer*>(CurPTHLexer.get())->getSpelling(Tok.getLocation(),
264                                                              Buffer);
265    }
266    else {
267      SourceLocation sloc = SourceMgr.getPhysicalLoc(Tok.getLocation());
268      unsigned fid = sloc.getFileID();
269      unsigned fpos = SourceMgr.getFullFilePos(sloc);
270      len = PTH->getSpelling(fid, fpos, Buffer);
271    }
272
273    // Did we find a spelling?  If so return its length.  Otherwise fall
274    // back to the default behavior for getting the spelling by looking at
275    // at the source code.
276    if (len)
277      return len;
278  }
279
280  // Otherwise, compute the start of the token in the input lexer buffer.
281  const char *TokStart = SourceMgr.getCharacterData(Tok.getLocation());
282
283  // If this token contains nothing interesting, return it directly.
284  if (!Tok.needsCleaning()) {
285    Buffer = TokStart;
286    return Tok.getLength();
287  }
288  // Otherwise, hard case, relex the characters into the string.
289  char *OutBuf = const_cast<char*>(Buffer);
290  for (const char *Ptr = TokStart, *End = TokStart+Tok.getLength();
291       Ptr != End; ) {
292    unsigned CharSize;
293    *OutBuf++ = Lexer::getCharAndSizeNoWarn(Ptr, CharSize, Features);
294    Ptr += CharSize;
295  }
296  assert(unsigned(OutBuf-Buffer) != Tok.getLength() &&
297         "NeedsCleaning flag set on something that didn't need cleaning!");
298
299  return OutBuf-Buffer;
300}
301
302
303/// CreateString - Plop the specified string into a scratch buffer and return a
304/// location for it.  If specified, the source location provides a source
305/// location for the token.
306SourceLocation Preprocessor::
307CreateString(const char *Buf, unsigned Len, SourceLocation SLoc) {
308  if (SLoc.isValid())
309    return ScratchBuf->getToken(Buf, Len, SLoc);
310  return ScratchBuf->getToken(Buf, Len);
311}
312
313
314/// AdvanceToTokenCharacter - Given a location that specifies the start of a
315/// token, return a new location that specifies a character within the token.
316SourceLocation Preprocessor::AdvanceToTokenCharacter(SourceLocation TokStart,
317                                                     unsigned CharNo) {
318  // If they request the first char of the token, we're trivially done.  If this
319  // is a macro expansion, it doesn't make sense to point to a character within
320  // the instantiation point (the name).  We could point to the source
321  // character, but without also pointing to instantiation info, this is
322  // confusing.
323  if (CharNo == 0 || TokStart.isMacroID()) return TokStart;
324
325  // Figure out how many physical characters away the specified logical
326  // character is.  This needs to take into consideration newlines and
327  // trigraphs.
328  const char *TokPtr = SourceMgr.getCharacterData(TokStart);
329  unsigned PhysOffset = 0;
330
331  // The usual case is that tokens don't contain anything interesting.  Skip
332  // over the uninteresting characters.  If a token only consists of simple
333  // chars, this method is extremely fast.
334  while (CharNo && Lexer::isObviouslySimpleCharacter(*TokPtr))
335    ++TokPtr, --CharNo, ++PhysOffset;
336
337  // If we have a character that may be a trigraph or escaped newline, create a
338  // lexer to parse it correctly.
339  if (CharNo != 0) {
340    // Create a lexer starting at this token position.
341    Lexer TheLexer(TokStart, *this, TokPtr);
342    Token Tok;
343    // Skip over characters the remaining characters.
344    const char *TokStartPtr = TokPtr;
345    for (; CharNo; --CharNo)
346      TheLexer.getAndAdvanceChar(TokPtr, Tok);
347
348    PhysOffset += TokPtr-TokStartPtr;
349  }
350
351  return TokStart.getFileLocWithOffset(PhysOffset);
352}
353
354
355//===----------------------------------------------------------------------===//
356// Preprocessor Initialization Methods
357//===----------------------------------------------------------------------===//
358
359// Append a #define line to Buf for Macro.  Macro should be of the form XXX,
360// in which case we emit "#define XXX 1" or "XXX=Y z W" in which case we emit
361// "#define XXX Y z W".  To get a #define with no value, use "XXX=".
362static void DefineBuiltinMacro(std::vector<char> &Buf, const char *Macro,
363                               const char *Command = "#define ") {
364  Buf.insert(Buf.end(), Command, Command+strlen(Command));
365  if (const char *Equal = strchr(Macro, '=')) {
366    // Turn the = into ' '.
367    Buf.insert(Buf.end(), Macro, Equal);
368    Buf.push_back(' ');
369    Buf.insert(Buf.end(), Equal+1, Equal+strlen(Equal));
370  } else {
371    // Push "macroname 1".
372    Buf.insert(Buf.end(), Macro, Macro+strlen(Macro));
373    Buf.push_back(' ');
374    Buf.push_back('1');
375  }
376  Buf.push_back('\n');
377}
378
379/// PickFP - This is used to pick a value based on the FP semantics of the
380/// specified FP model.
381template <typename T>
382static T PickFP(const llvm::fltSemantics *Sem, T IEEESingleVal,
383                T IEEEDoubleVal, T X87DoubleExtendedVal, T PPCDoubleDoubleVal) {
384  if (Sem == &llvm::APFloat::IEEEsingle)
385    return IEEESingleVal;
386  if (Sem == &llvm::APFloat::IEEEdouble)
387    return IEEEDoubleVal;
388  if (Sem == &llvm::APFloat::x87DoubleExtended)
389    return X87DoubleExtendedVal;
390  assert(Sem == &llvm::APFloat::PPCDoubleDouble);
391  return PPCDoubleDoubleVal;
392}
393
394static void DefineFloatMacros(std::vector<char> &Buf, const char *Prefix,
395                              const llvm::fltSemantics *Sem) {
396  const char *DenormMin, *Epsilon, *Max, *Min;
397  DenormMin = PickFP(Sem, "1.40129846e-45F", "4.9406564584124654e-324",
398                     "3.64519953188247460253e-4951L",
399                     "4.94065645841246544176568792868221e-324L");
400  int Digits = PickFP(Sem, 6, 15, 18, 31);
401  Epsilon = PickFP(Sem, "1.19209290e-7F", "2.2204460492503131e-16",
402                   "1.08420217248550443401e-19L",
403                   "4.94065645841246544176568792868221e-324L");
404  int HasInifinity = 1, HasQuietNaN = 1;
405  int MantissaDigits = PickFP(Sem, 24, 53, 64, 106);
406  int Min10Exp = PickFP(Sem, -37, -307, -4931, -291);
407  int Max10Exp = PickFP(Sem, 38, 308, 4932, 308);
408  int MinExp = PickFP(Sem, -125, -1021, -16381, -968);
409  int MaxExp = PickFP(Sem, 128, 1024, 16384, 1024);
410  Min = PickFP(Sem, "1.17549435e-38F", "2.2250738585072014e-308",
411               "3.36210314311209350626e-4932L",
412               "2.00416836000897277799610805135016e-292L");
413  Max = PickFP(Sem, "3.40282347e+38F", "1.7976931348623157e+308",
414               "1.18973149535723176502e+4932L",
415               "1.79769313486231580793728971405301e+308L");
416
417  char MacroBuf[60];
418  sprintf(MacroBuf, "__%s_DENORM_MIN__=%s", Prefix, DenormMin);
419  DefineBuiltinMacro(Buf, MacroBuf);
420  sprintf(MacroBuf, "__%s_DIG__=%d", Prefix, Digits);
421  DefineBuiltinMacro(Buf, MacroBuf);
422  sprintf(MacroBuf, "__%s_EPSILON__=%s", Prefix, Epsilon);
423  DefineBuiltinMacro(Buf, MacroBuf);
424  sprintf(MacroBuf, "__%s_HAS_INFINITY__=%d", Prefix, HasInifinity);
425  DefineBuiltinMacro(Buf, MacroBuf);
426  sprintf(MacroBuf, "__%s_HAS_QUIET_NAN__=%d", Prefix, HasQuietNaN);
427  DefineBuiltinMacro(Buf, MacroBuf);
428  sprintf(MacroBuf, "__%s_MANT_DIG__=%d", Prefix, MantissaDigits);
429  DefineBuiltinMacro(Buf, MacroBuf);
430  sprintf(MacroBuf, "__%s_MAX_10_EXP__=%d", Prefix, Max10Exp);
431  DefineBuiltinMacro(Buf, MacroBuf);
432  sprintf(MacroBuf, "__%s_MAX_EXP__=%d", Prefix, MaxExp);
433  DefineBuiltinMacro(Buf, MacroBuf);
434  sprintf(MacroBuf, "__%s_MAX__=%s", Prefix, Max);
435  DefineBuiltinMacro(Buf, MacroBuf);
436  sprintf(MacroBuf, "__%s_MIN_10_EXP__=(%d)", Prefix, Min10Exp);
437  DefineBuiltinMacro(Buf, MacroBuf);
438  sprintf(MacroBuf, "__%s_MIN_EXP__=(%d)", Prefix, MinExp);
439  DefineBuiltinMacro(Buf, MacroBuf);
440  sprintf(MacroBuf, "__%s_MIN__=%s", Prefix, Min);
441  DefineBuiltinMacro(Buf, MacroBuf);
442}
443
444
445static void InitializePredefinedMacros(Preprocessor &PP,
446                                       std::vector<char> &Buf) {
447  // Compiler version introspection macros.
448  DefineBuiltinMacro(Buf, "__llvm__=1");   // LLVM Backend
449  DefineBuiltinMacro(Buf, "__clang__=1");  // Clang Frontend
450
451  // Currently claim to be compatible with GCC 4.2.1-5621.
452  DefineBuiltinMacro(Buf, "__APPLE_CC__=5621");
453  DefineBuiltinMacro(Buf, "__GNUC_MINOR__=2");
454  DefineBuiltinMacro(Buf, "__GNUC_PATCHLEVEL__=1");
455  DefineBuiltinMacro(Buf, "__GNUC__=4");
456  DefineBuiltinMacro(Buf, "__GXX_ABI_VERSION=1002");
457  DefineBuiltinMacro(Buf, "__VERSION__=\"4.2.1 (Apple Computer, Inc. "
458                     "build 5621) (dot 3)\"");
459
460
461  // Initialize language-specific preprocessor defines.
462
463  // FIXME: Implement magic like cpp_init_builtins for things like __STDC__
464  // and __DATE__ etc.
465  // These should all be defined in the preprocessor according to the
466  // current language configuration.
467  if (!PP.getLangOptions().Microsoft)
468    DefineBuiltinMacro(Buf, "__STDC__=1");
469  if (PP.getLangOptions().AsmPreprocessor)
470    DefineBuiltinMacro(Buf, "__ASSEMBLER__=1");
471  if (PP.getLangOptions().C99 && !PP.getLangOptions().CPlusPlus)
472    DefineBuiltinMacro(Buf, "__STDC_VERSION__=199901L");
473  else if (0) // STDC94 ?
474    DefineBuiltinMacro(Buf, "__STDC_VERSION__=199409L");
475
476  DefineBuiltinMacro(Buf, "__STDC_HOSTED__=1");
477  if (PP.getLangOptions().ObjC1) {
478    DefineBuiltinMacro(Buf, "__OBJC__=1");
479
480    if (PP.getLangOptions().getGCMode() == LangOptions::NonGC) {
481      DefineBuiltinMacro(Buf, "__weak=");
482      DefineBuiltinMacro(Buf, "__strong=");
483    } else {
484      DefineBuiltinMacro(Buf, "__weak=__attribute__((objc_gc(weak)))");
485      DefineBuiltinMacro(Buf, "__strong=__attribute__((objc_gc(strong)))");
486      DefineBuiltinMacro(Buf, "__OBJC_GC__=1");
487    }
488
489    if (PP.getLangOptions().NeXTRuntime)
490      DefineBuiltinMacro(Buf, "__NEXT_RUNTIME__=1");
491  }
492
493  // darwin_constant_cfstrings controls this. This is also dependent
494  // on other things like the runtime I believe.  This is set even for C code.
495  DefineBuiltinMacro(Buf, "__CONSTANT_CFSTRINGS__=1");
496
497  if (PP.getLangOptions().ObjC2)
498    DefineBuiltinMacro(Buf, "OBJC_NEW_PROPERTIES");
499
500  if (PP.getLangOptions().PascalStrings)
501    DefineBuiltinMacro(Buf, "__PASCAL_STRINGS__");
502
503  if (PP.getLangOptions().Blocks) {
504    DefineBuiltinMacro(Buf, "__block=__attribute__((__blocks__(byref)))");
505    DefineBuiltinMacro(Buf, "__BLOCKS__=1");
506  }
507
508  if (PP.getLangOptions().CPlusPlus) {
509    DefineBuiltinMacro(Buf, "__DEPRECATED=1");
510    DefineBuiltinMacro(Buf, "__EXCEPTIONS=1");
511    DefineBuiltinMacro(Buf, "__GNUG__=4");
512    DefineBuiltinMacro(Buf, "__GXX_WEAK__=1");
513    DefineBuiltinMacro(Buf, "__cplusplus=1");
514    DefineBuiltinMacro(Buf, "__private_extern__=extern");
515  }
516
517  // Filter out some microsoft extensions when trying to parse in ms-compat
518  // mode.
519  if (PP.getLangOptions().Microsoft) {
520    DefineBuiltinMacro(Buf, "_cdecl=__cdecl");
521    DefineBuiltinMacro(Buf, "__int8=char");
522    DefineBuiltinMacro(Buf, "__int16=short");
523    DefineBuiltinMacro(Buf, "__int32=int");
524    DefineBuiltinMacro(Buf, "__int64=long long");
525  }
526
527
528  // Initialize target-specific preprocessor defines.
529  const TargetInfo &TI = PP.getTargetInfo();
530
531  // Define type sizing macros based on the target properties.
532  assert(TI.getCharWidth() == 8 && "Only support 8-bit char so far");
533  DefineBuiltinMacro(Buf, "__CHAR_BIT__=8");
534  DefineBuiltinMacro(Buf, "__SCHAR_MAX__=127");
535
536  assert(TI.getWCharWidth() == 32 && "Only support 32-bit wchar so far");
537  DefineBuiltinMacro(Buf, "__WCHAR_MAX__=2147483647");
538  DefineBuiltinMacro(Buf, "__WCHAR_TYPE__=int");
539  DefineBuiltinMacro(Buf, "__WINT_TYPE__=int");
540
541  assert(TI.getShortWidth() == 16 && "Only support 16-bit short so far");
542  DefineBuiltinMacro(Buf, "__SHRT_MAX__=32767");
543
544  if (TI.getIntWidth() == 32)
545    DefineBuiltinMacro(Buf, "__INT_MAX__=2147483647");
546  else if (TI.getIntWidth() == 16)
547    DefineBuiltinMacro(Buf, "__INT_MAX__=32767");
548  else
549    assert(0 && "Unknown integer size");
550
551  if (TI.getLongLongWidth() == 64)
552    DefineBuiltinMacro(Buf, "__LONG_LONG_MAX__=9223372036854775807LL");
553  else if (TI.getLongLongWidth() == 32)
554    DefineBuiltinMacro(Buf, "__LONG_LONG_MAX__=2147483647L");
555
556  if (TI.getLongWidth() == 32)
557    DefineBuiltinMacro(Buf, "__LONG_MAX__=2147483647L");
558  else if (TI.getLongWidth() == 64)
559    DefineBuiltinMacro(Buf, "__LONG_MAX__=9223372036854775807L");
560  else if (TI.getLongWidth() == 16)
561    DefineBuiltinMacro(Buf, "__LONG_MAX__=32767L");
562  else
563    assert(0 && "Unknown long size");
564  char MacroBuf[60];
565  sprintf(MacroBuf, "__INTMAX_MAX__=%lld",
566          (TI.getIntMaxType() == TargetInfo::UnsignedLongLong?
567           (1LL << (TI.getLongLongWidth() - 1)) :
568           ((1LL << (TI.getLongLongWidth() - 2)) - 1)));
569  DefineBuiltinMacro(Buf, MacroBuf);
570
571  if (TI.getIntMaxType() == TargetInfo::UnsignedLongLong)
572    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=unsigned long long int");
573  else if (TI.getIntMaxType() == TargetInfo::SignedLongLong)
574    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=long long int");
575  else if (TI.getIntMaxType() == TargetInfo::UnsignedLong)
576    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=unsigned long int");
577  else if (TI.getIntMaxType() == TargetInfo::SignedLong)
578    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=long int");
579  else if (TI.getIntMaxType() == TargetInfo::UnsignedInt)
580    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=unsigned int");
581  else
582    DefineBuiltinMacro(Buf, "__INTMAX_TYPE__=int");
583
584  if (TI.getUIntMaxType() == TargetInfo::UnsignedLongLong)
585    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=unsigned long long int");
586  else if (TI.getUIntMaxType() == TargetInfo::SignedLongLong)
587    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=long long int");
588  else if (TI.getUIntMaxType() == TargetInfo::UnsignedLong)
589    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=unsigned long int");
590  else if (TI.getUIntMaxType() == TargetInfo::SignedLong)
591    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=long int");
592  else if (TI.getUIntMaxType() == TargetInfo::UnsignedInt)
593    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=unsigned int");
594  else
595    DefineBuiltinMacro(Buf, "__UINTMAX_TYPE__=int");
596
597  if (TI.getPtrDiffType(0) == TargetInfo::UnsignedLongLong)
598    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=unsigned long long int");
599  else if (TI.getPtrDiffType(0) == TargetInfo::SignedLongLong)
600    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=long long int");
601  else if (TI.getPtrDiffType(0) == TargetInfo::UnsignedLong)
602    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=unsigned long int");
603  else if (TI.getPtrDiffType(0) == TargetInfo::SignedLong)
604    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=long int");
605  else if (TI.getPtrDiffType(0) == TargetInfo::UnsignedInt)
606    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=unsigned int");
607  else
608    DefineBuiltinMacro(Buf, "__PTRDIFF_TYPE__=int");
609
610  if (TI.getSizeType() == TargetInfo::UnsignedLongLong)
611    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=unsigned long long int");
612  else if (TI.getSizeType() == TargetInfo::SignedLongLong)
613    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=long long int");
614  else if (TI.getSizeType() == TargetInfo::UnsignedLong)
615    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=unsigned long int");
616  else if (TI.getSizeType() == TargetInfo::SignedLong)
617    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=long int");
618  else if (TI.getSizeType() == TargetInfo::UnsignedInt)
619    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=unsigned int");
620  else if (TI.getSizeType() == TargetInfo::SignedInt)
621    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=int");
622  else
623    DefineBuiltinMacro(Buf, "__SIZE_TYPE__=unsigned short");
624
625  DefineFloatMacros(Buf, "FLT", &TI.getFloatFormat());
626  DefineFloatMacros(Buf, "DBL", &TI.getDoubleFormat());
627  DefineFloatMacros(Buf, "LDBL", &TI.getLongDoubleFormat());
628
629
630  // Add __builtin_va_list typedef.
631  {
632    const char *VAList = TI.getVAListDeclaration();
633    Buf.insert(Buf.end(), VAList, VAList+strlen(VAList));
634    Buf.push_back('\n');
635  }
636
637  if (const char *Prefix = TI.getUserLabelPrefix()) {
638    sprintf(MacroBuf, "__USER_LABEL_PREFIX__=%s", Prefix);
639    DefineBuiltinMacro(Buf, MacroBuf);
640  }
641
642  // Build configuration options.  FIXME: these should be controlled by
643  // command line options or something.
644  DefineBuiltinMacro(Buf, "__DYNAMIC__=1");
645  DefineBuiltinMacro(Buf, "__FINITE_MATH_ONLY__=0");
646  DefineBuiltinMacro(Buf, "__NO_INLINE__=1");
647  DefineBuiltinMacro(Buf, "__PIC__=1");
648
649  // Macros to control C99 numerics and <float.h>
650  DefineBuiltinMacro(Buf, "__FLT_EVAL_METHOD__=0");
651  DefineBuiltinMacro(Buf, "__FLT_RADIX__=2");
652  sprintf(MacroBuf, "__DECIMAL_DIG__=%d",
653          PickFP(&TI.getLongDoubleFormat(), -1/*FIXME*/, 17, 21, 33));
654  DefineBuiltinMacro(Buf, MacroBuf);
655
656  // Get other target #defines.
657  TI.getTargetDefines(Buf);
658
659  // FIXME: Should emit a #line directive here.
660}
661
662
663/// EnterMainSourceFile - Enter the specified FileID as the main source file,
664/// which implicitly adds the builtin defines etc.
665void Preprocessor::EnterMainSourceFile() {
666
667  unsigned MainFileID = SourceMgr.getMainFileID();
668
669  // Enter the main file source buffer.
670  EnterSourceFile(MainFileID, 0);
671
672  // Tell the header info that the main file was entered.  If the file is later
673  // #imported, it won't be re-entered.
674  if (const FileEntry *FE =
675        SourceMgr.getFileEntryForLoc(SourceLocation::getFileLoc(MainFileID, 0)))
676    HeaderInfo.IncrementIncludeCount(FE);
677
678  std::vector<char> PrologFile;
679  PrologFile.reserve(4080);
680
681  // Install things like __POWERPC__, __GNUC__, etc into the macro table.
682  InitializePredefinedMacros(*this, PrologFile);
683
684  // Add on the predefines from the driver.
685  PrologFile.insert(PrologFile.end(), Predefines.begin(), Predefines.end());
686
687  // Memory buffer must end with a null byte!
688  PrologFile.push_back(0);
689
690  // Now that we have emitted the predefined macros, #includes, etc into
691  // PrologFile, preprocess it to populate the initial preprocessor state.
692  llvm::MemoryBuffer *SB =
693    llvm::MemoryBuffer::getMemBufferCopy(&PrologFile.front(),&PrologFile.back(),
694                                         "<predefines>");
695  assert(SB && "Cannot fail to create predefined source buffer");
696  unsigned FileID = SourceMgr.createFileIDForMemBuffer(SB);
697  assert(FileID && "Could not create FileID for predefines?");
698
699  // Start parsing the predefines.
700  EnterSourceFile(FileID, 0);
701}
702
703
704//===----------------------------------------------------------------------===//
705// Lexer Event Handling.
706//===----------------------------------------------------------------------===//
707
708/// LookUpIdentifierInfo - Given a tok::identifier token, look up the
709/// identifier information for the token and install it into the token.
710IdentifierInfo *Preprocessor::LookUpIdentifierInfo(Token &Identifier,
711                                                   const char *BufPtr) {
712  assert(Identifier.is(tok::identifier) && "Not an identifier!");
713  assert(Identifier.getIdentifierInfo() == 0 && "Identinfo already exists!");
714
715  // Look up this token, see if it is a macro, or if it is a language keyword.
716  IdentifierInfo *II;
717  if (BufPtr && !Identifier.needsCleaning()) {
718    // No cleaning needed, just use the characters from the lexed buffer.
719    II = getIdentifierInfo(BufPtr, BufPtr+Identifier.getLength());
720  } else {
721    // Cleaning needed, alloca a buffer, clean into it, then use the buffer.
722    llvm::SmallVector<char, 64> IdentifierBuffer;
723    IdentifierBuffer.resize(Identifier.getLength());
724    const char *TmpBuf = &IdentifierBuffer[0];
725    unsigned Size = getSpelling(Identifier, TmpBuf);
726    II = getIdentifierInfo(TmpBuf, TmpBuf+Size);
727  }
728  Identifier.setIdentifierInfo(II);
729  return II;
730}
731
732
733/// HandleIdentifier - This callback is invoked when the lexer reads an
734/// identifier.  This callback looks up the identifier in the map and/or
735/// potentially macro expands it or turns it into a named token (like 'for').
736void Preprocessor::HandleIdentifier(Token &Identifier) {
737  assert(Identifier.getIdentifierInfo() &&
738         "Can't handle identifiers without identifier info!");
739
740  IdentifierInfo &II = *Identifier.getIdentifierInfo();
741
742  // If this identifier was poisoned, and if it was not produced from a macro
743  // expansion, emit an error.
744  if (II.isPoisoned() && CurPPLexer) {
745    if (&II != Ident__VA_ARGS__)   // We warn about __VA_ARGS__ with poisoning.
746      Diag(Identifier, diag::err_pp_used_poisoned_id);
747    else
748      Diag(Identifier, diag::ext_pp_bad_vaargs_use);
749  }
750
751  // If this is a macro to be expanded, do it.
752  if (MacroInfo *MI = getMacroInfo(&II)) {
753    if (!DisableMacroExpansion && !Identifier.isExpandDisabled()) {
754      if (MI->isEnabled()) {
755        if (!HandleMacroExpandedIdentifier(Identifier, MI))
756          return;
757      } else {
758        // C99 6.10.3.4p2 says that a disabled macro may never again be
759        // expanded, even if it's in a context where it could be expanded in the
760        // future.
761        Identifier.setFlag(Token::DisableExpand);
762      }
763    }
764  }
765
766  // C++ 2.11p2: If this is an alternative representation of a C++ operator,
767  // then we act as if it is the actual operator and not the textual
768  // representation of it.
769  if (II.isCPlusPlusOperatorKeyword())
770    Identifier.setIdentifierInfo(0);
771
772  // Change the kind of this identifier to the appropriate token kind, e.g.
773  // turning "for" into a keyword.
774  Identifier.setKind(II.getTokenID());
775
776  // If this is an extension token, diagnose its use.
777  // We avoid diagnosing tokens that originate from macro definitions.
778  if (II.isExtensionToken() && Features.C99 && !DisableMacroExpansion)
779    Diag(Identifier, diag::ext_token_used);
780}
781