GlobalModuleIndex.cpp revision 87f9d81d0ab806dcf6ca50a0c068dcb2ba7f51b3 
1//===--- GlobalModuleIndex.cpp - Global Module Index ------------*- C++ -*-===//
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 GlobalModuleIndex class.
11//
12//===----------------------------------------------------------------------===//
13
14#include "ASTReaderInternals.h"
15#include "clang/Basic/FileManager.h"
16#include "clang/Basic/OnDiskHashTable.h"
17#include "clang/Serialization/ASTBitCodes.h"
18#include "clang/Serialization/GlobalModuleIndex.h"
19#include "clang/Serialization/Module.h"
20#include "llvm/ADT/DenseMap.h"
21#include "llvm/ADT/MapVector.h"
22#include "llvm/ADT/SmallString.h"
23#include "llvm/ADT/StringExtras.h"
24#include "llvm/Bitcode/BitstreamReader.h"
25#include "llvm/Bitcode/BitstreamWriter.h"
26#include "llvm/Support/FileSystem.h"
27#include "llvm/Support/LockFileManager.h"
28#include "llvm/Support/MemoryBuffer.h"
29#include "llvm/Support/PathV2.h"
30#include <cstdio>
31using namespace clang;
32using namespace serialization;
33
34//----------------------------------------------------------------------------//
35// Shared constants
36//----------------------------------------------------------------------------//
37namespace {
38  enum {
39    /// \brief The block containing the index.
40    GLOBAL_INDEX_BLOCK_ID = llvm::bitc::FIRST_APPLICATION_BLOCKID
41  };
42
43  /// \brief Describes the record types in the index.
44  enum IndexRecordTypes {
45    /// \brief Contains version information and potentially other metadata,
46    /// used to determine if we can read this global index file.
47    INDEX_METADATA,
48    /// \brief Describes a module, including its file name and dependencies.
49    MODULE,
50    /// \brief The index for identifiers.
51    IDENTIFIER_INDEX
52  };
53}
54
55/// \brief The name of the global index file.
56static const char * const IndexFileName = "modules.idx";
57
58/// \brief The global index file version.
59static const unsigned CurrentVersion = 1;
60
61//----------------------------------------------------------------------------//
62// Global module index reader.
63//----------------------------------------------------------------------------//
64
65namespace {
66
67/// \brief Trait used to read the identifier index from the on-disk hash
68/// table.
69class IdentifierIndexReaderTrait {
70public:
71  typedef StringRef external_key_type;
72  typedef StringRef internal_key_type;
73  typedef SmallVector<unsigned, 2> data_type;
74
75  static bool EqualKey(const internal_key_type& a, const internal_key_type& b) {
76    return a == b;
77  }
78
79  static unsigned ComputeHash(const internal_key_type& a) {
80    return llvm::HashString(a);
81  }
82
83  static std::pair<unsigned, unsigned>
84  ReadKeyDataLength(const unsigned char*& d) {
85    using namespace clang::io;
86    unsigned KeyLen = ReadUnalignedLE16(d);
87    unsigned DataLen = ReadUnalignedLE16(d);
88    return std::make_pair(KeyLen, DataLen);
89  }
90
91  static const internal_key_type&
92  GetInternalKey(const external_key_type& x) { return x; }
93
94  static const external_key_type&
95  GetExternalKey(const internal_key_type& x) { return x; }
96
97  static internal_key_type ReadKey(const unsigned char* d, unsigned n) {
98    return StringRef((const char *)d, n);
99  }
100
101  static data_type ReadData(const internal_key_type& k,
102                            const unsigned char* d,
103                            unsigned DataLen) {
104    using namespace clang::io;
105
106    data_type Result;
107    while (DataLen > 0) {
108      unsigned ID = ReadUnalignedLE32(d);
109      Result.push_back(ID);
110      DataLen -= 4;
111    }
112
113    return Result;
114  }
115};
116
117typedef OnDiskChainedHashTable<IdentifierIndexReaderTrait> IdentifierIndexTable;
118
119}
120
121GlobalModuleIndex::GlobalModuleIndex(llvm::MemoryBuffer *Buffer,
122                                     llvm::BitstreamCursor Cursor)
123  : Buffer(Buffer), IdentifierIndex(),
124    NumIdentifierLookups(), NumIdentifierLookupHits()
125{
126  // Read the global index.
127  bool InGlobalIndexBlock = false;
128  bool Done = false;
129  while (!Done) {
130    llvm::BitstreamEntry Entry = Cursor.advance();
131
132    switch (Entry.Kind) {
133    case llvm::BitstreamEntry::Error:
134      return;
135
136    case llvm::BitstreamEntry::EndBlock:
137      if (InGlobalIndexBlock) {
138        InGlobalIndexBlock = false;
139        Done = true;
140        continue;
141      }
142      return;
143
144
145    case llvm::BitstreamEntry::Record:
146      // Entries in the global index block are handled below.
147      if (InGlobalIndexBlock)
148        break;
149
150      return;
151
152    case llvm::BitstreamEntry::SubBlock:
153      if (!InGlobalIndexBlock && Entry.ID == GLOBAL_INDEX_BLOCK_ID) {
154        if (Cursor.EnterSubBlock(GLOBAL_INDEX_BLOCK_ID))
155          return;
156
157        InGlobalIndexBlock = true;
158      } else if (Cursor.SkipBlock()) {
159        return;
160      }
161      continue;
162    }
163
164    SmallVector<uint64_t, 64> Record;
165    StringRef Blob;
166    switch ((IndexRecordTypes)Cursor.readRecord(Entry.ID, Record, &Blob)) {
167    case INDEX_METADATA:
168      // Make sure that the version matches.
169      if (Record.size() < 1 || Record[0] != CurrentVersion)
170        return;
171      break;
172
173    case MODULE: {
174      unsigned Idx = 0;
175      unsigned ID = Record[Idx++];
176
177      // Make room for this module's information.
178      if (ID == Modules.size())
179        Modules.push_back(ModuleInfo());
180      else
181        Modules.resize(ID + 1);
182
183      // Size/modification time for this module file at the time the
184      // global index was built.
185      Modules[ID].Size = Record[Idx++];
186      Modules[ID].ModTime = Record[Idx++];
187
188      // File name.
189      unsigned NameLen = Record[Idx++];
190      Modules[ID].FileName.assign(Record.begin() + Idx,
191                                  Record.begin() + Idx + NameLen);
192      Idx += NameLen;
193
194      // Dependencies
195      unsigned NumDeps = Record[Idx++];
196      Modules[ID].Dependencies.insert(Modules[ID].Dependencies.end(),
197                                      Record.begin() + Idx,
198                                      Record.begin() + Idx + NumDeps);
199      Idx += NumDeps;
200
201      // Make sure we're at the end of the record.
202      assert(Idx == Record.size() && "More module info?");
203
204      // Record this module as an unresolved module.
205      UnresolvedModules[llvm::sys::path::stem(Modules[ID].FileName)] = ID;
206      break;
207    }
208
209    case IDENTIFIER_INDEX:
210      // Wire up the identifier index.
211      if (Record[0]) {
212        IdentifierIndex = IdentifierIndexTable::Create(
213                            (const unsigned char *)Blob.data() + Record[0],
214                            (const unsigned char *)Blob.data(),
215                            IdentifierIndexReaderTrait());
216      }
217      break;
218    }
219  }
220}
221
222GlobalModuleIndex::~GlobalModuleIndex() { }
223
224std::pair<GlobalModuleIndex *, GlobalModuleIndex::ErrorCode>
225GlobalModuleIndex::readIndex(StringRef Path) {
226  // Load the index file, if it's there.
227  llvm::SmallString<128> IndexPath;
228  IndexPath += Path;
229  llvm::sys::path::append(IndexPath, IndexFileName);
230
231  llvm::OwningPtr<llvm::MemoryBuffer> Buffer;
232  if (llvm::MemoryBuffer::getFile(IndexPath, Buffer) != llvm::errc::success)
233    return std::make_pair((GlobalModuleIndex *)0, EC_NotFound);
234
235  /// \brief The bitstream reader from which we'll read the AST file.
236  llvm::BitstreamReader Reader((const unsigned char *)Buffer->getBufferStart(),
237                               (const unsigned char *)Buffer->getBufferEnd());
238
239  /// \brief The main bitstream cursor for the main block.
240  llvm::BitstreamCursor Cursor(Reader);
241
242  // Sniff for the signature.
243  if (Cursor.Read(8) != 'B' ||
244      Cursor.Read(8) != 'C' ||
245      Cursor.Read(8) != 'G' ||
246      Cursor.Read(8) != 'I') {
247    return std::make_pair((GlobalModuleIndex *)0, EC_IOError);
248  }
249
250  return std::make_pair(new GlobalModuleIndex(Buffer.take(), Cursor), EC_None);
251}
252
253void
254GlobalModuleIndex::getKnownModules(SmallVectorImpl<ModuleFile *> &ModuleFiles) {
255  ModuleFiles.clear();
256  for (unsigned I = 0, N = Modules.size(); I != N; ++I) {
257    if (ModuleFile *MF = Modules[I].File)
258      ModuleFiles.push_back(MF);
259  }
260}
261
262void GlobalModuleIndex::getModuleDependencies(
263       ModuleFile *File,
264       SmallVectorImpl<ModuleFile *> &Dependencies) {
265  // Look for information about this module file.
266  llvm::DenseMap<ModuleFile *, unsigned>::iterator Known
267    = ModulesByFile.find(File);
268  if (Known == ModulesByFile.end())
269    return;
270
271  // Record dependencies.
272  Dependencies.clear();
273  ArrayRef<unsigned> StoredDependencies = Modules[Known->second].Dependencies;
274  for (unsigned I = 0, N = StoredDependencies.size(); I != N; ++I) {
275    if (ModuleFile *MF = Modules[I].File)
276      Dependencies.push_back(MF);
277  }
278}
279
280bool GlobalModuleIndex::lookupIdentifier(StringRef Name, HitSet &Hits) {
281  Hits.clear();
282
283  // If there's no identifier index, there is nothing we can do.
284  if (!IdentifierIndex)
285    return false;
286
287  // Look into the identifier index.
288  ++NumIdentifierLookups;
289  IdentifierIndexTable &Table
290    = *static_cast<IdentifierIndexTable *>(IdentifierIndex);
291  IdentifierIndexTable::iterator Known = Table.find(Name);
292  if (Known == Table.end()) {
293    return true;
294  }
295
296  SmallVector<unsigned, 2> ModuleIDs = *Known;
297  for (unsigned I = 0, N = ModuleIDs.size(); I != N; ++I) {
298    if (ModuleFile *MF = Modules[ModuleIDs[I]].File)
299      Hits.insert(MF);
300  }
301
302  ++NumIdentifierLookupHits;
303  return true;
304}
305
306bool GlobalModuleIndex::loadedModuleFile(ModuleFile *File) {
307  // Look for the module in the global module index based on the module name.
308  StringRef Name = llvm::sys::path::stem(File->FileName);
309  llvm::StringMap<unsigned>::iterator Known = UnresolvedModules.find(Name);
310  if (Known == UnresolvedModules.end()) {
311    return true;
312  }
313
314  // Rectify this module with the global module index.
315  ModuleInfo &Info = Modules[Known->second];
316
317  //  If the size and modification time match what we expected, record this
318  // module file.
319  bool Failed = true;
320  if (File->File->getSize() == Info.Size &&
321      File->File->getModificationTime() == Info.ModTime) {
322    Info.File = File;
323    ModulesByFile[File] = Known->second;
324
325    Failed = false;
326  }
327
328  // One way or another, we have resolved this module file.
329  UnresolvedModules.erase(Known);
330  return Failed;
331}
332
333void GlobalModuleIndex::printStats() {
334  std::fprintf(stderr, "*** Global Module Index Statistics:\n");
335  if (NumIdentifierLookups) {
336    fprintf(stderr, "  %u / %u identifier lookups succeeded (%f%%)\n",
337            NumIdentifierLookupHits, NumIdentifierLookups,
338            (double)NumIdentifierLookupHits*100.0/NumIdentifierLookups);
339  }
340  std::fprintf(stderr, "\n");
341}
342
343//----------------------------------------------------------------------------//
344// Global module index writer.
345//----------------------------------------------------------------------------//
346
347namespace {
348  /// \brief Provides information about a specific module file.
349  struct ModuleFileInfo {
350    /// \brief The numberic ID for this module file.
351    unsigned ID;
352
353    /// \brief The set of modules on which this module depends. Each entry is
354    /// a module ID.
355    SmallVector<unsigned, 4> Dependencies;
356  };
357
358  /// \brief Builder that generates the global module index file.
359  class GlobalModuleIndexBuilder {
360    FileManager &FileMgr;
361
362    /// \brief Mapping from files to module file information.
363    typedef llvm::MapVector<const FileEntry *, ModuleFileInfo> ModuleFilesMap;
364
365    /// \brief Information about each of the known module files.
366    ModuleFilesMap ModuleFiles;
367
368    /// \brief Mapping from identifiers to the list of module file IDs that
369    /// consider this identifier to be interesting.
370    typedef llvm::StringMap<SmallVector<unsigned, 2> > InterestingIdentifierMap;
371
372    /// \brief A mapping from all interesting identifiers to the set of module
373    /// files in which those identifiers are considered interesting.
374    InterestingIdentifierMap InterestingIdentifiers;
375
376    /// \brief Write the block-info block for the global module index file.
377    void emitBlockInfoBlock(llvm::BitstreamWriter &Stream);
378
379    /// \brief Retrieve the module file information for the given file.
380    ModuleFileInfo &getModuleFileInfo(const FileEntry *File) {
381      llvm::MapVector<const FileEntry *, ModuleFileInfo>::iterator Known
382        = ModuleFiles.find(File);
383      if (Known != ModuleFiles.end())
384        return Known->second;
385
386      unsigned NewID = ModuleFiles.size();
387      ModuleFileInfo &Info = ModuleFiles[File];
388      Info.ID = NewID;
389      return Info;
390    }
391
392  public:
393    explicit GlobalModuleIndexBuilder(FileManager &FileMgr) : FileMgr(FileMgr){}
394
395    /// \brief Load the contents of the given module file into the builder.
396    ///
397    /// \returns true if an error occurred, false otherwise.
398    bool loadModuleFile(const FileEntry *File);
399
400    /// \brief Write the index to the given bitstream.
401    void writeIndex(llvm::BitstreamWriter &Stream);
402  };
403}
404
405static void emitBlockID(unsigned ID, const char *Name,
406                        llvm::BitstreamWriter &Stream,
407                        SmallVectorImpl<uint64_t> &Record) {
408  Record.clear();
409  Record.push_back(ID);
410  Stream.EmitRecord(llvm::bitc::BLOCKINFO_CODE_SETBID, Record);
411
412  // Emit the block name if present.
413  if (Name == 0 || Name[0] == 0) return;
414  Record.clear();
415  while (*Name)
416    Record.push_back(*Name++);
417  Stream.EmitRecord(llvm::bitc::BLOCKINFO_CODE_BLOCKNAME, Record);
418}
419
420static void emitRecordID(unsigned ID, const char *Name,
421                         llvm::BitstreamWriter &Stream,
422                         SmallVectorImpl<uint64_t> &Record) {
423  Record.clear();
424  Record.push_back(ID);
425  while (*Name)
426    Record.push_back(*Name++);
427  Stream.EmitRecord(llvm::bitc::BLOCKINFO_CODE_SETRECORDNAME, Record);
428}
429
430void
431GlobalModuleIndexBuilder::emitBlockInfoBlock(llvm::BitstreamWriter &Stream) {
432  SmallVector<uint64_t, 64> Record;
433  Stream.EnterSubblock(llvm::bitc::BLOCKINFO_BLOCK_ID, 3);
434
435#define BLOCK(X) emitBlockID(X ## _ID, #X, Stream, Record)
436#define RECORD(X) emitRecordID(X, #X, Stream, Record)
437  BLOCK(GLOBAL_INDEX_BLOCK);
438  RECORD(INDEX_METADATA);
439  RECORD(MODULE);
440  RECORD(IDENTIFIER_INDEX);
441#undef RECORD
442#undef BLOCK
443
444  Stream.ExitBlock();
445}
446
447namespace {
448  class InterestingASTIdentifierLookupTrait
449    : public serialization::reader::ASTIdentifierLookupTraitBase {
450
451  public:
452    /// \brief The identifier and whether it is "interesting".
453    typedef std::pair<StringRef, bool> data_type;
454
455    data_type ReadData(const internal_key_type& k,
456                       const unsigned char* d,
457                       unsigned DataLen) {
458      // The first bit indicates whether this identifier is interesting.
459      // That's all we care about.
460      using namespace clang::io;
461      unsigned RawID = ReadUnalignedLE32(d);
462      bool IsInteresting = RawID & 0x01;
463      return std::make_pair(k, IsInteresting);
464    }
465  };
466}
467
468bool GlobalModuleIndexBuilder::loadModuleFile(const FileEntry *File) {
469  // Open the module file.
470  OwningPtr<llvm::MemoryBuffer> Buffer;
471  std::string ErrorStr;
472  Buffer.reset(FileMgr.getBufferForFile(File, &ErrorStr, /*isVolatile=*/true));
473  if (!Buffer) {
474    return true;
475  }
476
477  // Initialize the input stream
478  llvm::BitstreamReader InStreamFile;
479  llvm::BitstreamCursor InStream;
480  InStreamFile.init((const unsigned char *)Buffer->getBufferStart(),
481                  (const unsigned char *)Buffer->getBufferEnd());
482  InStream.init(InStreamFile);
483
484  // Sniff for the signature.
485  if (InStream.Read(8) != 'C' ||
486      InStream.Read(8) != 'P' ||
487      InStream.Read(8) != 'C' ||
488      InStream.Read(8) != 'H') {
489    return true;
490  }
491
492  // Record this module file and assign it a unique ID (if it doesn't have
493  // one already).
494  unsigned ID = getModuleFileInfo(File).ID;
495
496  // Search for the blocks and records we care about.
497  enum { Other, ControlBlock, ASTBlock } State = Other;
498  bool Done = false;
499  while (!Done) {
500    llvm::BitstreamEntry Entry = InStream.advance();
501    switch (Entry.Kind) {
502    case llvm::BitstreamEntry::Error:
503      Done = true;
504      continue;
505
506    case llvm::BitstreamEntry::Record:
507      // In the 'other' state, just skip the record. We don't care.
508      if (State == Other) {
509        InStream.skipRecord(Entry.ID);
510        continue;
511      }
512
513      // Handle potentially-interesting records below.
514      break;
515
516    case llvm::BitstreamEntry::SubBlock:
517      if (Entry.ID == CONTROL_BLOCK_ID) {
518        if (InStream.EnterSubBlock(CONTROL_BLOCK_ID))
519          return true;
520
521        // Found the control block.
522        State = ControlBlock;
523        continue;
524      }
525
526      if (Entry.ID == AST_BLOCK_ID) {
527        if (InStream.EnterSubBlock(AST_BLOCK_ID))
528          return true;
529
530        // Found the AST block.
531        State = ASTBlock;
532        continue;
533      }
534
535      if (InStream.SkipBlock())
536        return true;
537
538      continue;
539
540    case llvm::BitstreamEntry::EndBlock:
541      State = Other;
542      continue;
543    }
544
545    // Read the given record.
546    SmallVector<uint64_t, 64> Record;
547    StringRef Blob;
548    unsigned Code = InStream.readRecord(Entry.ID, Record, &Blob);
549
550    // Handle module dependencies.
551    if (State == ControlBlock && Code == IMPORTS) {
552      // Load each of the imported PCH files.
553      unsigned Idx = 0, N = Record.size();
554      while (Idx < N) {
555        // Read information about the AST file.
556
557        // Skip the imported kind
558        ++Idx;
559
560        // Skip the import location
561        ++Idx;
562
563        // Load stored size/modification time.
564        off_t StoredSize = (off_t)Record[Idx++];
565        time_t StoredModTime = (time_t)Record[Idx++];
566
567        // Retrieve the imported file name.
568        unsigned Length = Record[Idx++];
569        SmallString<128> ImportedFile(Record.begin() + Idx,
570                                      Record.begin() + Idx + Length);
571        Idx += Length;
572
573        // Find the imported module file.
574        const FileEntry *DependsOnFile
575          = FileMgr.getFile(ImportedFile, /*openFile=*/false,
576                            /*cacheFailure=*/false);
577        if (!DependsOnFile ||
578            (StoredSize != DependsOnFile->getSize()) ||
579            (StoredModTime != DependsOnFile->getModificationTime()))
580          return true;
581
582        // Record the dependency.
583        unsigned DependsOnID = getModuleFileInfo(DependsOnFile).ID;
584        getModuleFileInfo(File).Dependencies.push_back(DependsOnID);
585      }
586
587      continue;
588    }
589
590    // Handle the identifier table
591    if (State == ASTBlock && Code == IDENTIFIER_TABLE && Record[0] > 0) {
592      typedef OnDiskChainedHashTable<InterestingASTIdentifierLookupTrait>
593        InterestingIdentifierTable;
594      llvm::OwningPtr<InterestingIdentifierTable>
595        Table(InterestingIdentifierTable::Create(
596                (const unsigned char *)Blob.data() + Record[0],
597                (const unsigned char *)Blob.data()));
598      for (InterestingIdentifierTable::data_iterator D = Table->data_begin(),
599                                                     DEnd = Table->data_end();
600           D != DEnd; ++D) {
601        std::pair<StringRef, bool> Ident = *D;
602        if (Ident.second)
603          InterestingIdentifiers[Ident.first].push_back(ID);
604        else
605          (void)InterestingIdentifiers[Ident.first];
606      }
607    }
608
609    // We don't care about this record.
610  }
611
612  return false;
613}
614
615namespace {
616
617/// \brief Trait used to generate the identifier index as an on-disk hash
618/// table.
619class IdentifierIndexWriterTrait {
620public:
621  typedef StringRef key_type;
622  typedef StringRef key_type_ref;
623  typedef SmallVector<unsigned, 2> data_type;
624  typedef const SmallVector<unsigned, 2> &data_type_ref;
625
626  static unsigned ComputeHash(key_type_ref Key) {
627    return llvm::HashString(Key);
628  }
629
630  std::pair<unsigned,unsigned>
631  EmitKeyDataLength(raw_ostream& Out, key_type_ref Key, data_type_ref Data) {
632    unsigned KeyLen = Key.size();
633    unsigned DataLen = Data.size() * 4;
634    clang::io::Emit16(Out, KeyLen);
635    clang::io::Emit16(Out, DataLen);
636    return std::make_pair(KeyLen, DataLen);
637  }
638
639  void EmitKey(raw_ostream& Out, key_type_ref Key, unsigned KeyLen) {
640    Out.write(Key.data(), KeyLen);
641  }
642
643  void EmitData(raw_ostream& Out, key_type_ref Key, data_type_ref Data,
644                unsigned DataLen) {
645    for (unsigned I = 0, N = Data.size(); I != N; ++I)
646      clang::io::Emit32(Out, Data[I]);
647  }
648};
649
650}
651
652void GlobalModuleIndexBuilder::writeIndex(llvm::BitstreamWriter &Stream) {
653  using namespace llvm;
654
655  // Emit the file header.
656  Stream.Emit((unsigned)'B', 8);
657  Stream.Emit((unsigned)'C', 8);
658  Stream.Emit((unsigned)'G', 8);
659  Stream.Emit((unsigned)'I', 8);
660
661  // Write the block-info block, which describes the records in this bitcode
662  // file.
663  emitBlockInfoBlock(Stream);
664
665  Stream.EnterSubblock(GLOBAL_INDEX_BLOCK_ID, 3);
666
667  // Write the metadata.
668  SmallVector<uint64_t, 2> Record;
669  Record.push_back(CurrentVersion);
670  Stream.EmitRecord(INDEX_METADATA, Record);
671
672  // Write the set of known module files.
673  for (ModuleFilesMap::iterator M = ModuleFiles.begin(),
674                                MEnd = ModuleFiles.end();
675       M != MEnd; ++M) {
676    Record.clear();
677    Record.push_back(M->second.ID);
678    Record.push_back(M->first->getSize());
679    Record.push_back(M->first->getModificationTime());
680
681    // File name
682    StringRef Name(M->first->getName());
683    Record.push_back(Name.size());
684    Record.append(Name.begin(), Name.end());
685
686    // Dependencies
687    Record.push_back(M->second.Dependencies.size());
688    Record.append(M->second.Dependencies.begin(), M->second.Dependencies.end());
689    Stream.EmitRecord(MODULE, Record);
690  }
691
692  // Write the identifier -> module file mapping.
693  {
694    OnDiskChainedHashTableGenerator<IdentifierIndexWriterTrait> Generator;
695    IdentifierIndexWriterTrait Trait;
696
697    // Populate the hash table.
698    for (InterestingIdentifierMap::iterator I = InterestingIdentifiers.begin(),
699                                            IEnd = InterestingIdentifiers.end();
700         I != IEnd; ++I) {
701      Generator.insert(I->first(), I->second, Trait);
702    }
703
704    // Create the on-disk hash table in a buffer.
705    SmallString<4096> IdentifierTable;
706    uint32_t BucketOffset;
707    {
708      llvm::raw_svector_ostream Out(IdentifierTable);
709      // Make sure that no bucket is at offset 0
710      clang::io::Emit32(Out, 0);
711      BucketOffset = Generator.Emit(Out, Trait);
712    }
713
714    // Create a blob abbreviation
715    BitCodeAbbrev *Abbrev = new BitCodeAbbrev();
716    Abbrev->Add(BitCodeAbbrevOp(IDENTIFIER_INDEX));
717    Abbrev->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 32));
718    Abbrev->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Blob));
719    unsigned IDTableAbbrev = Stream.EmitAbbrev(Abbrev);
720
721    // Write the identifier table
722    Record.clear();
723    Record.push_back(IDENTIFIER_INDEX);
724    Record.push_back(BucketOffset);
725    Stream.EmitRecordWithBlob(IDTableAbbrev, Record, IdentifierTable.str());
726  }
727
728  Stream.ExitBlock();
729}
730
731GlobalModuleIndex::ErrorCode
732GlobalModuleIndex::writeIndex(FileManager &FileMgr, StringRef Path) {
733  llvm::SmallString<128> IndexPath;
734  IndexPath += Path;
735  llvm::sys::path::append(IndexPath, IndexFileName);
736
737  // Coordinate building the global index file with other processes that might
738  // try to do the same.
739  llvm::LockFileManager Locked(IndexPath);
740  switch (Locked) {
741  case llvm::LockFileManager::LFS_Error:
742    return EC_IOError;
743
744  case llvm::LockFileManager::LFS_Owned:
745    // We're responsible for building the index ourselves. Do so below.
746    break;
747
748  case llvm::LockFileManager::LFS_Shared:
749    // Someone else is responsible for building the index. We don't care
750    // when they finish, so we're done.
751    return EC_Building;
752  }
753
754  // The module index builder.
755  GlobalModuleIndexBuilder Builder(FileMgr);
756
757  // Load each of the module files.
758  llvm::error_code EC;
759  for (llvm::sys::fs::directory_iterator D(Path, EC), DEnd;
760       D != DEnd && !EC;
761       D.increment(EC)) {
762    // If this isn't a module file, we don't care.
763    if (llvm::sys::path::extension(D->path()) != ".pcm") {
764      // ... unless it's a .pcm.lock file, which indicates that someone is
765      // in the process of rebuilding a module. They'll rebuild the index
766      // at the end of that translation unit, so we don't have to.
767      if (llvm::sys::path::extension(D->path()) == ".pcm.lock")
768        return EC_Building;
769
770      continue;
771    }
772
773    // If we can't find the module file, skip it.
774    const FileEntry *ModuleFile = FileMgr.getFile(D->path());
775    if (!ModuleFile)
776      continue;
777
778    // Load this module file.
779    if (Builder.loadModuleFile(ModuleFile))
780      return EC_IOError;
781  }
782
783  // The output buffer, into which the global index will be written.
784  SmallVector<char, 16> OutputBuffer;
785  {
786    llvm::BitstreamWriter OutputStream(OutputBuffer);
787    Builder.writeIndex(OutputStream);
788  }
789
790  // Write the global index file to a temporary file.
791  llvm::SmallString<128> IndexTmpPath;
792  int TmpFD;
793  if (llvm::sys::fs::unique_file(IndexPath + "-%%%%%%%%", TmpFD, IndexTmpPath))
794    return EC_IOError;
795
796  // Open the temporary global index file for output.
797  llvm::raw_fd_ostream Out(TmpFD, true);
798  if (Out.has_error())
799    return EC_IOError;
800
801  // Write the index.
802  Out.write(OutputBuffer.data(), OutputBuffer.size());
803  Out.close();
804  if (Out.has_error())
805    return EC_IOError;
806
807  // Remove the old index file. It isn't relevant any more.
808  bool OldIndexExisted;
809  llvm::sys::fs::remove(IndexPath.str(), OldIndexExisted);
810
811  // Rename the newly-written index file to the proper name.
812  if (llvm::sys::fs::rename(IndexTmpPath.str(), IndexPath.str())) {
813    // Rename failed; just remove the
814    llvm::sys::fs::remove(IndexTmpPath.str(), OldIndexExisted);
815    return EC_IOError;
816  }
817
818  // We're done.
819  return EC_None;
820}
821
822namespace {
823  class GlobalIndexIdentifierIterator : public IdentifierIterator {
824    /// \brief The current position within the identifier lookup table.
825    IdentifierIndexTable::key_iterator Current;
826
827    /// \brief The end position within the identifier lookup table.
828    IdentifierIndexTable::key_iterator End;
829
830  public:
831    explicit GlobalIndexIdentifierIterator(IdentifierIndexTable &Idx) {
832      Current = Idx.key_begin();
833      End = Idx.key_end();
834    }
835
836    virtual StringRef Next() {
837      if (Current == End)
838        return StringRef();
839
840      StringRef Result = *Current;
841      ++Current;
842      return Result;
843    }
844  };
845}
846
847IdentifierIterator *GlobalModuleIndex::createIdentifierIterator() const {
848  IdentifierIndexTable &Table =
849    *static_cast<IdentifierIndexTable *>(IdentifierIndex);
850  return new GlobalIndexIdentifierIterator(Table);
851}
852