aboutsummaryrefslogtreecommitdiff
path: root/lib/Lex/PTHLexer.cpp
blob: 3b949d0ab40a15c3db36f856a4d9f241f3a02d01 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
//===--- PTHLexer.cpp - Lex from a token stream ---------------------------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file implements the PTHLexer interface.
//
//===----------------------------------------------------------------------===//

#include "clang/Basic/TokenKinds.h"
#include "clang/Basic/FileManager.h"
#include "clang/Basic/IdentifierTable.h"
#include "clang/Basic/OnDiskHashTable.h"
#include "clang/Lex/LexDiagnostic.h"
#include "clang/Lex/PTHLexer.h"
#include "clang/Lex/Preprocessor.h"
#include "clang/Lex/PTHManager.h"
#include "clang/Lex/Token.h"
#include "clang/Lex/Preprocessor.h"
#include "llvm/ADT/OwningPtr.h"
#include "llvm/ADT/StringExtras.h"
#include "llvm/ADT/StringMap.h"
#include "llvm/Support/MemoryBuffer.h"
#include <sys/stat.h>
using namespace clang;
using namespace clang::io;

#define DISK_TOKEN_SIZE (1+1+2+4+4)

//===----------------------------------------------------------------------===//
// PTHLexer methods.
//===----------------------------------------------------------------------===//

PTHLexer::PTHLexer(Preprocessor &PP, FileID FID, const unsigned char *D,
                   const unsigned char *ppcond, PTHManager &PM)
  : PreprocessorLexer(&PP, FID), TokBuf(D), CurPtr(D), LastHashTokPtr(0),
    PPCond(ppcond), CurPPCondPtr(ppcond), PTHMgr(PM) {

  FileStartLoc = PP.getSourceManager().getLocForStartOfFile(FID);
}

void PTHLexer::Lex(Token& Tok) {
LexNextToken:

  //===--------------------------------------==//
  // Read the raw token data.
  //===--------------------------------------==//

  // Shadow CurPtr into an automatic variable.
  const unsigned char *CurPtrShadow = CurPtr;

  // Read in the data for the token.
  unsigned Word0 = ReadLE32(CurPtrShadow);
  uint32_t IdentifierID = ReadLE32(CurPtrShadow);
  uint32_t FileOffset = ReadLE32(CurPtrShadow);

  tok::TokenKind TKind = (tok::TokenKind) (Word0 & 0xFF);
  Token::TokenFlags TFlags = (Token::TokenFlags) ((Word0 >> 8) & 0xFF);
  uint32_t Len = Word0 >> 16;

  CurPtr = CurPtrShadow;

  //===--------------------------------------==//
  // Construct the token itself.
  //===--------------------------------------==//

  Tok.startToken();
  Tok.setKind(TKind);
  Tok.setFlag(TFlags);
  assert(!LexingRawMode);
  Tok.setLocation(FileStartLoc.getFileLocWithOffset(FileOffset));
  Tok.setLength(Len);

  // Handle identifiers.
  if (Tok.isLiteral()) {
    Tok.setLiteralData((const char*) (PTHMgr.SpellingBase + IdentifierID));
  }
  else if (IdentifierID) {
    MIOpt.ReadToken();
    IdentifierInfo *II = PTHMgr.GetIdentifierInfo(IdentifierID-1);

    Tok.setIdentifierInfo(II);

    // Change the kind of this identifier to the appropriate token kind, e.g.
    // turning "for" into a keyword.
    Tok.setKind(II->getTokenID());

    if (II->isHandleIdentifierCase())
      PP->HandleIdentifier(Tok);
    return;
  }

  //===--------------------------------------==//
  // Process the token.
  //===--------------------------------------==//
  if (TKind == tok::eof) {
    // Save the end-of-file token.
    EofToken = Tok;

    Preprocessor *PPCache = PP;

    assert(!ParsingPreprocessorDirective);
    assert(!LexingRawMode);

    // FIXME: Issue diagnostics similar to Lexer.
    if (PP->HandleEndOfFile(Tok, false))
      return;

    assert(PPCache && "Raw buffer::LexEndOfFile should return a token");
    return PPCache->Lex(Tok);
  }

  if (TKind == tok::hash && Tok.isAtStartOfLine()) {
    LastHashTokPtr = CurPtr - DISK_TOKEN_SIZE;
    assert(!LexingRawMode);
    PP->HandleDirective(Tok);

    if (PP->isCurrentLexer(this))
      goto LexNextToken;

    return PP->Lex(Tok);
  }

  if (TKind == tok::eom) {
    assert(ParsingPreprocessorDirective);
    ParsingPreprocessorDirective = false;
    return;
  }

  MIOpt.ReadToken();
}

// FIXME: We can just grab the last token instead of storing a copy
// into EofToken.
void PTHLexer::getEOF(Token& Tok) {
  assert(EofToken.is(tok::eof));
  Tok = EofToken;
}

void PTHLexer::DiscardToEndOfLine() {
  assert(ParsingPreprocessorDirective && ParsingFilename == false &&
         "Must be in a preprocessing directive!");

  // We assume that if the preprocessor wishes to discard to the end of
  // the line that it also means to end the current preprocessor directive.
  ParsingPreprocessorDirective = false;

  // Skip tokens by only peeking at their token kind and the flags.
  // We don't need to actually reconstruct full tokens from the token buffer.
  // This saves some copies and it also reduces IdentifierInfo* lookup.
  const unsigned char* p = CurPtr;
  while (1) {
    // Read the token kind.  Are we at the end of the file?
    tok::TokenKind x = (tok::TokenKind) (uint8_t) *p;
    if (x == tok::eof) break;

    // Read the token flags.  Are we at the start of the next line?
    Token::TokenFlags y = (Token::TokenFlags) (uint8_t) p[1];
    if (y & Token::StartOfLine) break;

    // Skip to the next token.
    p += DISK_TOKEN_SIZE;
  }

  CurPtr = p;
}

/// SkipBlock - Used by Preprocessor to skip the current conditional block.
bool PTHLexer::SkipBlock() {
  assert(CurPPCondPtr && "No cached PP conditional information.");
  assert(LastHashTokPtr && "No known '#' token.");

  const unsigned char* HashEntryI = 0;
  uint32_t Offset;
  uint32_t TableIdx;

  do {
    // Read the token offset from the side-table.
    Offset = ReadLE32(CurPPCondPtr);

    // Read the target table index from the side-table.
    TableIdx = ReadLE32(CurPPCondPtr);

    // Compute the actual memory address of the '#' token data for this entry.
    HashEntryI = TokBuf + Offset;

    // Optmization: "Sibling jumping".  #if...#else...#endif blocks can
    //  contain nested blocks.  In the side-table we can jump over these
    //  nested blocks instead of doing a linear search if the next "sibling"
    //  entry is not at a location greater than LastHashTokPtr.
    if (HashEntryI < LastHashTokPtr && TableIdx) {
      // In the side-table we are still at an entry for a '#' token that
      // is earlier than the last one we saw.  Check if the location we would
      // stride gets us closer.
      const unsigned char* NextPPCondPtr =
        PPCond + TableIdx*(sizeof(uint32_t)*2);
      assert(NextPPCondPtr >= CurPPCondPtr);
      // Read where we should jump to.
      uint32_t TmpOffset = ReadLE32(NextPPCondPtr);
      const unsigned char* HashEntryJ = TokBuf + TmpOffset;

      if (HashEntryJ <= LastHashTokPtr) {
        // Jump directly to the next entry in the side table.
        HashEntryI = HashEntryJ;
        Offset = TmpOffset;
        TableIdx = ReadLE32(NextPPCondPtr);
        CurPPCondPtr = NextPPCondPtr;
      }
    }
  }
  while (HashEntryI < LastHashTokPtr);
  assert(HashEntryI == LastHashTokPtr && "No PP-cond entry found for '#'");
  assert(TableIdx && "No jumping from #endifs.");

  // Update our side-table iterator.
  const unsigned char* NextPPCondPtr = PPCond + TableIdx*(sizeof(uint32_t)*2);
  assert(NextPPCondPtr >= CurPPCondPtr);
  CurPPCondPtr = NextPPCondPtr;

  // Read where we should jump to.
  HashEntryI = TokBuf + ReadLE32(NextPPCondPtr);
  uint32_t NextIdx = ReadLE32(NextPPCondPtr);

  // By construction NextIdx will be zero if this is a #endif.  This is useful
  // to know to obviate lexing another token.
  bool isEndif = NextIdx == 0;

  // This case can occur when we see something like this:
  //
  //  #if ...
  //   /* a comment or nothing */
  //  #elif
  //
  // If we are skipping the first #if block it will be the case that CurPtr
  // already points 'elif'.  Just return.

  if (CurPtr > HashEntryI) {
    assert(CurPtr == HashEntryI + DISK_TOKEN_SIZE);
    // Did we reach a #endif?  If so, go ahead and consume that token as well.
    if (isEndif)
      CurPtr += DISK_TOKEN_SIZE*2;
    else
      LastHashTokPtr = HashEntryI;

    return isEndif;
  }

  // Otherwise, we need to advance.  Update CurPtr to point to the '#' token.
  CurPtr = HashEntryI;

  // Update the location of the last observed '#'.  This is useful if we
  // are skipping multiple blocks.
  LastHashTokPtr = CurPtr;

  // Skip the '#' token.
  assert(((tok::TokenKind)*CurPtr) == tok::hash);
  CurPtr += DISK_TOKEN_SIZE;

  // Did we reach a #endif?  If so, go ahead and consume that token as well.
  if (isEndif) { CurPtr += DISK_TOKEN_SIZE*2; }

  return isEndif;
}

SourceLocation PTHLexer::getSourceLocation() {
  // getSourceLocation is not on the hot path.  It is used to get the location
  // of the next token when transitioning back to this lexer when done
  // handling a #included file.  Just read the necessary data from the token
  // data buffer to construct the SourceLocation object.
  // NOTE: This is a virtual function; hence it is defined out-of-line.
  const unsigned char *OffsetPtr = CurPtr + (DISK_TOKEN_SIZE - 4);
  uint32_t Offset = ReadLE32(OffsetPtr);
  return FileStartLoc.getFileLocWithOffset(Offset);
}

//===----------------------------------------------------------------------===//
// PTH file lookup: map from strings to file data.
//===----------------------------------------------------------------------===//

/// PTHFileLookup - This internal data structure is used by the PTHManager
///  to map from FileEntry objects managed by FileManager to offsets within
///  the PTH file.
namespace {
class PTHFileData {
  const uint32_t TokenOff;
  const uint32_t PPCondOff;
public:
  PTHFileData(uint32_t tokenOff, uint32_t ppCondOff)
    : TokenOff(tokenOff), PPCondOff(ppCondOff) {}

  uint32_t getTokenOffset() const { return TokenOff; }
  uint32_t getPPCondOffset() const { return PPCondOff; }
};


class PTHFileLookupCommonTrait {
public:
  typedef std::pair<unsigned char, const char*> internal_key_type;

  static unsigned ComputeHash(internal_key_type x) {
    return llvm::HashString(x.second);
  }

  static std::pair<unsigned, unsigned>
  ReadKeyDataLength(const unsigned char*& d) {
    unsigned keyLen = (unsigned) ReadUnalignedLE16(d);
    unsigned dataLen = (unsigned) *(d++);
    return std::make_pair(keyLen, dataLen);
  }

  static internal_key_type ReadKey(const unsigned char* d, unsigned) {
    unsigned char k = *(d++); // Read the entry kind.
    return std::make_pair(k, (const char*) d);
  }
};

class PTHFileLookupTrait : public PTHFileLookupCommonTrait {
public:
  typedef const FileEntry* external_key_type;
  typedef PTHFileData      data_type;

  static internal_key_type GetInternalKey(const FileEntry* FE) {
    return std::make_pair((unsigned char) 0x1, FE->getName());
  }

  static bool EqualKey(internal_key_type a, internal_key_type b) {
    return a.first == b.first && strcmp(a.second, b.second) == 0;
  }

  static PTHFileData ReadData(const internal_key_type& k,
                              const unsigned char* d, unsigned) {
    assert(k.first == 0x1 && "Only file lookups can match!");
    uint32_t x = ::ReadUnalignedLE32(d);
    uint32_t y = ::ReadUnalignedLE32(d);
    return PTHFileData(x, y);
  }
};

class PTHStringLookupTrait {
public:
  typedef uint32_t
          data_type;

  typedef const std::pair<const char*, unsigned>
          external_key_type;

  typedef external_key_type internal_key_type;

  static bool EqualKey(const internal_key_type& a,
                       const internal_key_type& b) {
    return (a.second == b.second) ? memcmp(a.first, b.first, a.second) == 0
                                  : false;
  }

  static unsigned ComputeHash(const internal_key_type& a) {
    return llvm::HashString(llvm::StringRef(a.first, a.second));
  }

  // This hopefully will just get inlined and removed by the optimizer.
  static const internal_key_type&
  GetInternalKey(const external_key_type& x) { return x; }

  static std::pair<unsigned, unsigned>
  ReadKeyDataLength(const unsigned char*& d) {
    return std::make_pair((unsigned) ReadUnalignedLE16(d), sizeof(uint32_t));
  }

  static std::pair<const char*, unsigned>
  ReadKey(const unsigned char* d, unsigned n) {
      assert(n >= 2 && d[n-1] == '\0');
      return std::make_pair((const char*) d, n-1);
    }

  static uint32_t ReadData(const internal_key_type& k, const unsigned char* d,
                           unsigned) {
    return ::ReadUnalignedLE32(d);
  }
};

} // end anonymous namespace

typedef OnDiskChainedHashTable<PTHFileLookupTrait>   PTHFileLookup;
typedef OnDiskChainedHashTable<PTHStringLookupTrait> PTHStringIdLookup;

//===----------------------------------------------------------------------===//
// PTHManager methods.
//===----------------------------------------------------------------------===//

PTHManager::PTHManager(const llvm::MemoryBuffer* buf, void* fileLookup,
                       const unsigned char* idDataTable,
                       IdentifierInfo** perIDCache,
                       void* stringIdLookup, unsigned numIds,
                       const unsigned char* spellingBase,
                       const char* originalSourceFile)
: Buf(buf), PerIDCache(perIDCache), FileLookup(fileLookup),
  IdDataTable(idDataTable), StringIdLookup(stringIdLookup),
  NumIds(numIds), PP(0), SpellingBase(spellingBase),
  OriginalSourceFile(originalSourceFile) {}

PTHManager::~PTHManager() {
  delete Buf;
  delete (PTHFileLookup*) FileLookup;
  delete (PTHStringIdLookup*) StringIdLookup;
  free(PerIDCache);
}

static void InvalidPTH(Diagnostic &Diags, const char *Msg) {
  Diags.Report(Diags.getCustomDiagID(Diagnostic::Error, Msg));
}

PTHManager* PTHManager::Create(const std::string& file, Diagnostic &Diags) {
  // Memory map the PTH file.
  llvm::OwningPtr<llvm::MemoryBuffer>
  File(llvm::MemoryBuffer::getFile(file.c_str()));

  if (!File) {
    Diags.Report(diag::err_invalid_pth_file) << file;
    return 0;
  }

  // Get the buffer ranges and check if there are at least three 32-bit
  // words at the end of the file.
  const unsigned char* BufBeg = (unsigned char*)File->getBufferStart();
  const unsigned char* BufEnd = (unsigned char*)File->getBufferEnd();

  // Check the prologue of the file.
  if ((BufEnd - BufBeg) < (signed) (sizeof("cfe-pth") + 3 + 4) ||
      memcmp(BufBeg, "cfe-pth", sizeof("cfe-pth") - 1) != 0) {
    Diags.Report(diag::err_invalid_pth_file) << file;
    return 0;
  }

  // Read the PTH version.
  const unsigned char *p = BufBeg + (sizeof("cfe-pth") - 1);
  unsigned Version = ReadLE32(p);

  if (Version < PTHManager::Version) {
    InvalidPTH(Diags,
        Version < PTHManager::Version
        ? "PTH file uses an older PTH format that is no longer supported"
        : "PTH file uses a newer PTH format that cannot be read");
    return 0;
  }

  // Compute the address of the index table at the end of the PTH file.
  const unsigned char *PrologueOffset = p;

  if (PrologueOffset >= BufEnd) {
    Diags.Report(diag::err_invalid_pth_file) << file;
    return 0;
  }

  // Construct the file lookup table.  This will be used for mapping from
  // FileEntry*'s to cached tokens.
  const unsigned char* FileTableOffset = PrologueOffset + sizeof(uint32_t)*2;
  const unsigned char* FileTable = BufBeg + ReadLE32(FileTableOffset);

  if (!(FileTable > BufBeg && FileTable < BufEnd)) {
    Diags.Report(diag::err_invalid_pth_file) << file;
    return 0; // FIXME: Proper error diagnostic?
  }

  llvm::OwningPtr<PTHFileLookup> FL(PTHFileLookup::Create(FileTable, BufBeg));

  // Warn if the PTH file is empty.  We still want to create a PTHManager
  // as the PTH could be used with -include-pth.
  if (FL->isEmpty())
    InvalidPTH(Diags, "PTH file contains no cached source data");

  // Get the location of the table mapping from persistent ids to the
  // data needed to reconstruct identifiers.
  const unsigned char* IDTableOffset = PrologueOffset + sizeof(uint32_t)*0;
  const unsigned char* IData = BufBeg + ReadLE32(IDTableOffset);

  if (!(IData >= BufBeg && IData < BufEnd)) {
    Diags.Report(diag::err_invalid_pth_file) << file;
    return 0;
  }

  // Get the location of the hashtable mapping between strings and
  // persistent IDs.
  const unsigned char* StringIdTableOffset = PrologueOffset + sizeof(uint32_t)*1;
  const unsigned char* StringIdTable = BufBeg + ReadLE32(StringIdTableOffset);
  if (!(StringIdTable >= BufBeg && StringIdTable < BufEnd)) {
    Diags.Report(diag::err_invalid_pth_file) << file;
    return 0;
  }

  llvm::OwningPtr<PTHStringIdLookup> SL(PTHStringIdLookup::Create(StringIdTable,
                                                                  BufBeg));

  // Get the location of the spelling cache.
  const unsigned char* spellingBaseOffset = PrologueOffset + sizeof(uint32_t)*3;
  const unsigned char* spellingBase = BufBeg + ReadLE32(spellingBaseOffset);
  if (!(spellingBase >= BufBeg && spellingBase < BufEnd)) {
    Diags.Report(diag::err_invalid_pth_file) << file;
    return 0;
  }

  // Get the number of IdentifierInfos and pre-allocate the identifier cache.
  uint32_t NumIds = ReadLE32(IData);

  // Pre-allocate the peristent ID -> IdentifierInfo* cache.  We use calloc()
  // so that we in the best case only zero out memory once when the OS returns
  // us new pages.
  IdentifierInfo** PerIDCache = 0;

  if (NumIds) {
    PerIDCache = (IdentifierInfo**)calloc(NumIds, sizeof(*PerIDCache));
    if (!PerIDCache) {
      InvalidPTH(Diags, "Could not allocate memory for processing PTH file");
      return 0;
    }
  }

  // Compute the address of the original source file.
  const unsigned char* originalSourceBase = PrologueOffset + sizeof(uint32_t)*4;
  unsigned len = ReadUnalignedLE16(originalSourceBase);
  if (!len) originalSourceBase = 0;

  // Create the new PTHManager.
  return new PTHManager(File.take(), FL.take(), IData, PerIDCache,
                        SL.take(), NumIds, spellingBase,
                        (const char*) originalSourceBase);
}

IdentifierInfo* PTHManager::LazilyCreateIdentifierInfo(unsigned PersistentID) {
  // Look in the PTH file for the string data for the IdentifierInfo object.
  const unsigned char* TableEntry = IdDataTable + sizeof(uint32_t)*PersistentID;
  const unsigned char* IDData =
    (const unsigned char*)Buf->getBufferStart() + ReadLE32(TableEntry);
  assert(IDData < (const unsigned char*)Buf->getBufferEnd());

  // Allocate the object.
  std::pair<IdentifierInfo,const unsigned char*> *Mem =
    Alloc.Allocate<std::pair<IdentifierInfo,const unsigned char*> >();

  Mem->second = IDData;
  assert(IDData[0] != '\0');
  IdentifierInfo *II = new ((void*) Mem) IdentifierInfo();

  // Store the new IdentifierInfo in the cache.
  PerIDCache[PersistentID] = II;
  assert(II->getNameStart() && II->getNameStart()[0] != '\0');
  return II;
}

IdentifierInfo* PTHManager::get(llvm::StringRef Name) {
  PTHStringIdLookup& SL = *((PTHStringIdLookup*)StringIdLookup);
  // Double check our assumption that the last character isn't '\0'.
  assert(Name.empty() || Name.data()[Name.size()-1] != '\0');
  PTHStringIdLookup::iterator I = SL.find(std::make_pair(Name.data(),
                                                         Name.size()));
  if (I == SL.end()) // No identifier found?
    return 0;

  // Match found.  Return the identifier!
  assert(*I > 0);
  return GetIdentifierInfo(*I-1);
}

PTHLexer *PTHManager::CreateLexer(FileID FID) {
  const FileEntry *FE = PP->getSourceManager().getFileEntryForID(FID);
  if (!FE)
    return 0;

  // Lookup the FileEntry object in our file lookup data structure.  It will
  // return a variant that indicates whether or not there is an offset within
  // the PTH file that contains cached tokens.
  PTHFileLookup& PFL = *((PTHFileLookup*)FileLookup);
  PTHFileLookup::iterator I = PFL.find(FE);

  if (I == PFL.end()) // No tokens available?
    return 0;

  const PTHFileData& FileData = *I;

  const unsigned char *BufStart = (const unsigned char *)Buf->getBufferStart();
  // Compute the offset of the token data within the buffer.
  const unsigned char* data = BufStart + FileData.getTokenOffset();

  // Get the location of pp-conditional table.
  const unsigned char* ppcond = BufStart + FileData.getPPCondOffset();
  uint32_t Len = ReadLE32(ppcond);
  if (Len == 0) ppcond = 0;

  assert(PP && "No preprocessor set yet!");
  return new PTHLexer(*PP, FID, data, ppcond, *this);
}

//===----------------------------------------------------------------------===//
// 'stat' caching.
//===----------------------------------------------------------------------===//

namespace {
class PTHStatData {
public:
  const bool hasStat;
  const ino_t ino;
  const dev_t dev;
  const mode_t mode;
  const time_t mtime;
  const off_t size;

  PTHStatData(ino_t i, dev_t d, mode_t mo, time_t m, off_t s)
  : hasStat(true), ino(i), dev(d), mode(mo), mtime(m), size(s) {}

  PTHStatData()
    : hasStat(false), ino(0), dev(0), mode(0), mtime(0), size(0) {}
};

class PTHStatLookupTrait : public PTHFileLookupCommonTrait {
public:
  typedef const char* external_key_type;  // const char*
  typedef PTHStatData data_type;

  static internal_key_type GetInternalKey(const char *path) {
    // The key 'kind' doesn't matter here because it is ignored in EqualKey.
    return std::make_pair((unsigned char) 0x0, path);
  }

  static bool EqualKey(internal_key_type a, internal_key_type b) {
    // When doing 'stat' lookups we don't care about the kind of 'a' and 'b',
    // just the paths.
    return strcmp(a.second, b.second) == 0;
  }

  static data_type ReadData(const internal_key_type& k, const unsigned char* d,
                            unsigned) {

    if (k.first /* File or Directory */) {
      if (k.first == 0x1 /* File */) d += 4 * 2; // Skip the first 2 words.
      ino_t ino = (ino_t) ReadUnalignedLE32(d);
      dev_t dev = (dev_t) ReadUnalignedLE32(d);
      mode_t mode = (mode_t) ReadUnalignedLE16(d);
      time_t mtime = (time_t) ReadUnalignedLE64(d);
      return data_type(ino, dev, mode, mtime, (off_t) ReadUnalignedLE64(d));
    }

    // Negative stat.  Don't read anything.
    return data_type();
  }
};

class PTHStatCache : public StatSysCallCache {
  typedef OnDiskChainedHashTable<PTHStatLookupTrait> CacheTy;
  CacheTy Cache;

public:
  PTHStatCache(PTHFileLookup &FL) :
    Cache(FL.getNumBuckets(), FL.getNumEntries(), FL.getBuckets(),
          FL.getBase()) {}

  ~PTHStatCache() {}

  int stat(const char *path, struct stat *buf) {
    // Do the lookup for the file's data in the PTH file.
    CacheTy::iterator I = Cache.find(path);

    // If we don't get a hit in the PTH file just forward to 'stat'.
    if (I == Cache.end())
      return StatSysCallCache::stat(path, buf);

    const PTHStatData& Data = *I;

    if (!Data.hasStat)
      return 1;

    buf->st_ino = Data.ino;
    buf->st_dev = Data.dev;
    buf->st_mtime = Data.mtime;
    buf->st_mode = Data.mode;
    buf->st_size = Data.size;
    return 0;
  }
};
} // end anonymous namespace

StatSysCallCache *PTHManager::createStatCache() {
  return new PTHStatCache(*((PTHFileLookup*) FileLookup));
}