GlobalModuleIndex.cpp revision 263508 
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/Path.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.c_str(), Buffer) !=
233      llvm::errc::success)
234    return std::make_pair((GlobalModuleIndex *)0, EC_NotFound);
235
236  /// \brief The bitstream reader from which we'll read the AST file.
237  llvm::BitstreamReader Reader((const unsigned char *)Buffer->getBufferStart(),
238                               (const unsigned char *)Buffer->getBufferEnd());
239
240  /// \brief The main bitstream cursor for the main block.
241  llvm::BitstreamCursor Cursor(Reader);
242
243  // Sniff for the signature.
244  if (Cursor.Read(8) != 'B' ||
245      Cursor.Read(8) != 'C' ||
246      Cursor.Read(8) != 'G' ||
247      Cursor.Read(8) != 'I') {
248    return std::make_pair((GlobalModuleIndex *)0, EC_IOError);
249  }
250
251  return std::make_pair(new GlobalModuleIndex(Buffer.take(), Cursor), EC_None);
252}
253
254void
255GlobalModuleIndex::getKnownModules(SmallVectorImpl<ModuleFile *> &ModuleFiles) {
256  ModuleFiles.clear();
257  for (unsigned I = 0, N = Modules.size(); I != N; ++I) {
258    if (ModuleFile *MF = Modules[I].File)
259      ModuleFiles.push_back(MF);
260  }
261}
262
263void GlobalModuleIndex::getModuleDependencies(
264       ModuleFile *File,
265       SmallVectorImpl<ModuleFile *> &Dependencies) {
266  // Look for information about this module file.
267  llvm::DenseMap<ModuleFile *, unsigned>::iterator Known
268    = ModulesByFile.find(File);
269  if (Known == ModulesByFile.end())
270    return;
271
272  // Record dependencies.
273  Dependencies.clear();
274  ArrayRef<unsigned> StoredDependencies = Modules[Known->second].Dependencies;
275  for (unsigned I = 0, N = StoredDependencies.size(); I != N; ++I) {
276    if (ModuleFile *MF = Modules[I].File)
277      Dependencies.push_back(MF);
278  }
279}
280
281bool GlobalModuleIndex::lookupIdentifier(StringRef Name, HitSet &Hits) {
282  Hits.clear();
283
284  // If there's no identifier index, there is nothing we can do.
285  if (!IdentifierIndex)
286    return false;
287
288  // Look into the identifier index.
289  ++NumIdentifierLookups;
290  IdentifierIndexTable &Table
291    = *static_cast<IdentifierIndexTable *>(IdentifierIndex);
292  IdentifierIndexTable::iterator Known = Table.find(Name);
293  if (Known == Table.end()) {
294    return true;
295  }
296
297  SmallVector<unsigned, 2> ModuleIDs = *Known;
298  for (unsigned I = 0, N = ModuleIDs.size(); I != N; ++I) {
299    if (ModuleFile *MF = Modules[ModuleIDs[I]].File)
300      Hits.insert(MF);
301  }
302
303  ++NumIdentifierLookupHits;
304  return true;
305}
306
307bool GlobalModuleIndex::loadedModuleFile(ModuleFile *File) {
308  // Look for the module in the global module index based on the module name.
309  StringRef Name = llvm::sys::path::stem(File->FileName);
310  llvm::StringMap<unsigned>::iterator Known = UnresolvedModules.find(Name);
311  if (Known == UnresolvedModules.end()) {
312    return true;
313  }
314
315  // Rectify this module with the global module index.
316  ModuleInfo &Info = Modules[Known->second];
317
318  //  If the size and modification time match what we expected, record this
319  // module file.
320  bool Failed = true;
321  if (File->File->getSize() == Info.Size &&
322      File->File->getModificationTime() == Info.ModTime) {
323    Info.File = File;
324    ModulesByFile[File] = Known->second;
325
326    Failed = false;
327  }
328
329  // One way or another, we have resolved this module file.
330  UnresolvedModules.erase(Known);
331  return Failed;
332}
333
334void GlobalModuleIndex::printStats() {
335  std::fprintf(stderr, "*** Global Module Index Statistics:\n");
336  if (NumIdentifierLookups) {
337    fprintf(stderr, "  %u / %u identifier lookups succeeded (%f%%)\n",
338            NumIdentifierLookupHits, NumIdentifierLookups,
339            (double)NumIdentifierLookupHits*100.0/NumIdentifierLookups);
340  }
341  std::fprintf(stderr, "\n");
342}
343
344//----------------------------------------------------------------------------//
345// Global module index writer.
346//----------------------------------------------------------------------------//
347
348namespace {
349  /// \brief Provides information about a specific module file.
350  struct ModuleFileInfo {
351    /// \brief The numberic ID for this module file.
352    unsigned ID;
353
354    /// \brief The set of modules on which this module depends. Each entry is
355    /// a module ID.
356    SmallVector<unsigned, 4> Dependencies;
357  };
358
359  /// \brief Builder that generates the global module index file.
360  class GlobalModuleIndexBuilder {
361    FileManager &FileMgr;
362
363    /// \brief Mapping from files to module file information.
364    typedef llvm::MapVector<const FileEntry *, ModuleFileInfo> ModuleFilesMap;
365
366    /// \brief Information about each of the known module files.
367    ModuleFilesMap ModuleFiles;
368
369    /// \brief Mapping from identifiers to the list of module file IDs that
370    /// consider this identifier to be interesting.
371    typedef llvm::StringMap<SmallVector<unsigned, 2> > InterestingIdentifierMap;
372
373    /// \brief A mapping from all interesting identifiers to the set of module
374    /// files in which those identifiers are considered interesting.
375    InterestingIdentifierMap InterestingIdentifiers;
376
377    /// \brief Write the block-info block for the global module index file.
378    void emitBlockInfoBlock(llvm::BitstreamWriter &Stream);
379
380    /// \brief Retrieve the module file information for the given file.
381    ModuleFileInfo &getModuleFileInfo(const FileEntry *File) {
382      llvm::MapVector<const FileEntry *, ModuleFileInfo>::iterator Known
383        = ModuleFiles.find(File);
384      if (Known != ModuleFiles.end())
385        return Known->second;
386
387      unsigned NewID = ModuleFiles.size();
388      ModuleFileInfo &Info = ModuleFiles[File];
389      Info.ID = NewID;
390      return Info;
391    }
392
393  public:
394    explicit GlobalModuleIndexBuilder(FileManager &FileMgr) : FileMgr(FileMgr){}
395
396    /// \brief Load the contents of the given module file into the builder.
397    ///
398    /// \returns true if an error occurred, false otherwise.
399    bool loadModuleFile(const FileEntry *File);
400
401    /// \brief Write the index to the given bitstream.
402    void writeIndex(llvm::BitstreamWriter &Stream);
403  };
404}
405
406static void emitBlockID(unsigned ID, const char *Name,
407                        llvm::BitstreamWriter &Stream,
408                        SmallVectorImpl<uint64_t> &Record) {
409  Record.clear();
410  Record.push_back(ID);
411  Stream.EmitRecord(llvm::bitc::BLOCKINFO_CODE_SETBID, Record);
412
413  // Emit the block name if present.
414  if (Name == 0 || Name[0] == 0) return;
415  Record.clear();
416  while (*Name)
417    Record.push_back(*Name++);
418  Stream.EmitRecord(llvm::bitc::BLOCKINFO_CODE_BLOCKNAME, Record);
419}
420
421static void emitRecordID(unsigned ID, const char *Name,
422                         llvm::BitstreamWriter &Stream,
423                         SmallVectorImpl<uint64_t> &Record) {
424  Record.clear();
425  Record.push_back(ID);
426  while (*Name)
427    Record.push_back(*Name++);
428  Stream.EmitRecord(llvm::bitc::BLOCKINFO_CODE_SETRECORDNAME, Record);
429}
430
431void
432GlobalModuleIndexBuilder::emitBlockInfoBlock(llvm::BitstreamWriter &Stream) {
433  SmallVector<uint64_t, 64> Record;
434  Stream.EnterSubblock(llvm::bitc::BLOCKINFO_BLOCK_ID, 3);
435
436#define BLOCK(X) emitBlockID(X ## _ID, #X, Stream, Record)
437#define RECORD(X) emitRecordID(X, #X, Stream, Record)
438  BLOCK(GLOBAL_INDEX_BLOCK);
439  RECORD(INDEX_METADATA);
440  RECORD(MODULE);
441  RECORD(IDENTIFIER_INDEX);
442#undef RECORD
443#undef BLOCK
444
445  Stream.ExitBlock();
446}
447
448namespace {
449  class InterestingASTIdentifierLookupTrait
450    : public serialization::reader::ASTIdentifierLookupTraitBase {
451
452  public:
453    /// \brief The identifier and whether it is "interesting".
454    typedef std::pair<StringRef, bool> data_type;
455
456    data_type ReadData(const internal_key_type& k,
457                       const unsigned char* d,
458                       unsigned DataLen) {
459      // The first bit indicates whether this identifier is interesting.
460      // That's all we care about.
461      using namespace clang::io;
462      unsigned RawID = ReadUnalignedLE32(d);
463      bool IsInteresting = RawID & 0x01;
464      return std::make_pair(k, IsInteresting);
465    }
466  };
467}
468
469bool GlobalModuleIndexBuilder::loadModuleFile(const FileEntry *File) {
470  // Open the module file.
471  OwningPtr<llvm::MemoryBuffer> Buffer;
472  std::string ErrorStr;
473  Buffer.reset(FileMgr.getBufferForFile(File, &ErrorStr, /*isVolatile=*/true));
474  if (!Buffer) {
475    return true;
476  }
477
478  // Initialize the input stream
479  llvm::BitstreamReader InStreamFile;
480  llvm::BitstreamCursor InStream;
481  InStreamFile.init((const unsigned char *)Buffer->getBufferStart(),
482                  (const unsigned char *)Buffer->getBufferEnd());
483  InStream.init(InStreamFile);
484
485  // Sniff for the signature.
486  if (InStream.Read(8) != 'C' ||
487      InStream.Read(8) != 'P' ||
488      InStream.Read(8) != 'C' ||
489      InStream.Read(8) != 'H') {
490    return true;
491  }
492
493  // Record this module file and assign it a unique ID (if it doesn't have
494  // one already).
495  unsigned ID = getModuleFileInfo(File).ID;
496
497  // Search for the blocks and records we care about.
498  enum { Other, ControlBlock, ASTBlock } State = Other;
499  bool Done = false;
500  while (!Done) {
501    llvm::BitstreamEntry Entry = InStream.advance();
502    switch (Entry.Kind) {
503    case llvm::BitstreamEntry::Error:
504      Done = true;
505      continue;
506
507    case llvm::BitstreamEntry::Record:
508      // In the 'other' state, just skip the record. We don't care.
509      if (State == Other) {
510        InStream.skipRecord(Entry.ID);
511        continue;
512      }
513
514      // Handle potentially-interesting records below.
515      break;
516
517    case llvm::BitstreamEntry::SubBlock:
518      if (Entry.ID == CONTROL_BLOCK_ID) {
519        if (InStream.EnterSubBlock(CONTROL_BLOCK_ID))
520          return true;
521
522        // Found the control block.
523        State = ControlBlock;
524        continue;
525      }
526
527      if (Entry.ID == AST_BLOCK_ID) {
528        if (InStream.EnterSubBlock(AST_BLOCK_ID))
529          return true;
530
531        // Found the AST block.
532        State = ASTBlock;
533        continue;
534      }
535
536      if (InStream.SkipBlock())
537        return true;
538
539      continue;
540
541    case llvm::BitstreamEntry::EndBlock:
542      State = Other;
543      continue;
544    }
545
546    // Read the given record.
547    SmallVector<uint64_t, 64> Record;
548    StringRef Blob;
549    unsigned Code = InStream.readRecord(Entry.ID, Record, &Blob);
550
551    // Handle module dependencies.
552    if (State == ControlBlock && Code == IMPORTS) {
553      // Load each of the imported PCH files.
554      unsigned Idx = 0, N = Record.size();
555      while (Idx < N) {
556        // Read information about the AST file.
557
558        // Skip the imported kind
559        ++Idx;
560
561        // Skip the import location
562        ++Idx;
563
564        // Load stored size/modification time.
565        off_t StoredSize = (off_t)Record[Idx++];
566        time_t StoredModTime = (time_t)Record[Idx++];
567
568        // Retrieve the imported file name.
569        unsigned Length = Record[Idx++];
570        SmallString<128> ImportedFile(Record.begin() + Idx,
571                                      Record.begin() + Idx + Length);
572        Idx += Length;
573
574        // Find the imported module file.
575        const FileEntry *DependsOnFile
576          = FileMgr.getFile(ImportedFile, /*openFile=*/false,
577                            /*cacheFailure=*/false);
578        if (!DependsOnFile ||
579            (StoredSize != DependsOnFile->getSize()) ||
580            (StoredModTime != DependsOnFile->getModificationTime()))
581          return true;
582
583        // Record the dependency.
584        unsigned DependsOnID = getModuleFileInfo(DependsOnFile).ID;
585        getModuleFileInfo(File).Dependencies.push_back(DependsOnID);
586      }
587
588      continue;
589    }
590
591    // Handle the identifier table
592    if (State == ASTBlock && Code == IDENTIFIER_TABLE && Record[0] > 0) {
593      typedef OnDiskChainedHashTable<InterestingASTIdentifierLookupTrait>
594        InterestingIdentifierTable;
595      llvm::OwningPtr<InterestingIdentifierTable>
596        Table(InterestingIdentifierTable::Create(
597                (const unsigned char *)Blob.data() + Record[0],
598                (const unsigned char *)Blob.data()));
599      for (InterestingIdentifierTable::data_iterator D = Table->data_begin(),
600                                                     DEnd = Table->data_end();
601           D != DEnd; ++D) {
602        std::pair<StringRef, bool> Ident = *D;
603        if (Ident.second)
604          InterestingIdentifiers[Ident.first].push_back(ID);
605        else
606          (void)InterestingIdentifiers[Ident.first];
607      }
608    }
609
610    // We don't care about this record.
611  }
612
613  return false;
614}
615
616namespace {
617
618/// \brief Trait used to generate the identifier index as an on-disk hash
619/// table.
620class IdentifierIndexWriterTrait {
621public:
622  typedef StringRef key_type;
623  typedef StringRef key_type_ref;
624  typedef SmallVector<unsigned, 2> data_type;
625  typedef const SmallVector<unsigned, 2> &data_type_ref;
626
627  static unsigned ComputeHash(key_type_ref Key) {
628    return llvm::HashString(Key);
629  }
630
631  std::pair<unsigned,unsigned>
632  EmitKeyDataLength(raw_ostream& Out, key_type_ref Key, data_type_ref Data) {
633    unsigned KeyLen = Key.size();
634    unsigned DataLen = Data.size() * 4;
635    clang::io::Emit16(Out, KeyLen);
636    clang::io::Emit16(Out, DataLen);
637    return std::make_pair(KeyLen, DataLen);
638  }
639
640  void EmitKey(raw_ostream& Out, key_type_ref Key, unsigned KeyLen) {
641    Out.write(Key.data(), KeyLen);
642  }
643
644  void EmitData(raw_ostream& Out, key_type_ref Key, data_type_ref Data,
645                unsigned DataLen) {
646    for (unsigned I = 0, N = Data.size(); I != N; ++I)
647      clang::io::Emit32(Out, Data[I]);
648  }
649};
650
651}
652
653void GlobalModuleIndexBuilder::writeIndex(llvm::BitstreamWriter &Stream) {
654  using namespace llvm;
655
656  // Emit the file header.
657  Stream.Emit((unsigned)'B', 8);
658  Stream.Emit((unsigned)'C', 8);
659  Stream.Emit((unsigned)'G', 8);
660  Stream.Emit((unsigned)'I', 8);
661
662  // Write the block-info block, which describes the records in this bitcode
663  // file.
664  emitBlockInfoBlock(Stream);
665
666  Stream.EnterSubblock(GLOBAL_INDEX_BLOCK_ID, 3);
667
668  // Write the metadata.
669  SmallVector<uint64_t, 2> Record;
670  Record.push_back(CurrentVersion);
671  Stream.EmitRecord(INDEX_METADATA, Record);
672
673  // Write the set of known module files.
674  for (ModuleFilesMap::iterator M = ModuleFiles.begin(),
675                                MEnd = ModuleFiles.end();
676       M != MEnd; ++M) {
677    Record.clear();
678    Record.push_back(M->second.ID);
679    Record.push_back(M->first->getSize());
680    Record.push_back(M->first->getModificationTime());
681
682    // File name
683    StringRef Name(M->first->getName());
684    Record.push_back(Name.size());
685    Record.append(Name.begin(), Name.end());
686
687    // Dependencies
688    Record.push_back(M->second.Dependencies.size());
689    Record.append(M->second.Dependencies.begin(), M->second.Dependencies.end());
690    Stream.EmitRecord(MODULE, Record);
691  }
692
693  // Write the identifier -> module file mapping.
694  {
695    OnDiskChainedHashTableGenerator<IdentifierIndexWriterTrait> Generator;
696    IdentifierIndexWriterTrait Trait;
697
698    // Populate the hash table.
699    for (InterestingIdentifierMap::iterator I = InterestingIdentifiers.begin(),
700                                            IEnd = InterestingIdentifiers.end();
701         I != IEnd; ++I) {
702      Generator.insert(I->first(), I->second, Trait);
703    }
704
705    // Create the on-disk hash table in a buffer.
706    SmallString<4096> IdentifierTable;
707    uint32_t BucketOffset;
708    {
709      llvm::raw_svector_ostream Out(IdentifierTable);
710      // Make sure that no bucket is at offset 0
711      clang::io::Emit32(Out, 0);
712      BucketOffset = Generator.Emit(Out, Trait);
713    }
714
715    // Create a blob abbreviation
716    BitCodeAbbrev *Abbrev = new BitCodeAbbrev();
717    Abbrev->Add(BitCodeAbbrevOp(IDENTIFIER_INDEX));
718    Abbrev->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 32));
719    Abbrev->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Blob));
720    unsigned IDTableAbbrev = Stream.EmitAbbrev(Abbrev);
721
722    // Write the identifier table
723    Record.clear();
724    Record.push_back(IDENTIFIER_INDEX);
725    Record.push_back(BucketOffset);
726    Stream.EmitRecordWithBlob(IDTableAbbrev, Record, IdentifierTable.str());
727  }
728
729  Stream.ExitBlock();
730}
731
732GlobalModuleIndex::ErrorCode
733GlobalModuleIndex::writeIndex(FileManager &FileMgr, StringRef Path) {
734  llvm::SmallString<128> IndexPath;
735  IndexPath += Path;
736  llvm::sys::path::append(IndexPath, IndexFileName);
737
738  // Coordinate building the global index file with other processes that might
739  // try to do the same.
740  llvm::LockFileManager Locked(IndexPath);
741  switch (Locked) {
742  case llvm::LockFileManager::LFS_Error:
743    return EC_IOError;
744
745  case llvm::LockFileManager::LFS_Owned:
746    // We're responsible for building the index ourselves. Do so below.
747    break;
748
749  case llvm::LockFileManager::LFS_Shared:
750    // Someone else is responsible for building the index. We don't care
751    // when they finish, so we're done.
752    return EC_Building;
753  }
754
755  // The module index builder.
756  GlobalModuleIndexBuilder Builder(FileMgr);
757
758  // Load each of the module files.
759  llvm::error_code EC;
760  for (llvm::sys::fs::directory_iterator D(Path, EC), DEnd;
761       D != DEnd && !EC;
762       D.increment(EC)) {
763    // If this isn't a module file, we don't care.
764    if (llvm::sys::path::extension(D->path()) != ".pcm") {
765      // ... unless it's a .pcm.lock file, which indicates that someone is
766      // in the process of rebuilding a module. They'll rebuild the index
767      // at the end of that translation unit, so we don't have to.
768      if (llvm::sys::path::extension(D->path()) == ".pcm.lock")
769        return EC_Building;
770
771      continue;
772    }
773
774    // If we can't find the module file, skip it.
775    const FileEntry *ModuleFile = FileMgr.getFile(D->path());
776    if (!ModuleFile)
777      continue;
778
779    // Load this module file.
780    if (Builder.loadModuleFile(ModuleFile))
781      return EC_IOError;
782  }
783
784  // The output buffer, into which the global index will be written.
785  SmallVector<char, 16> OutputBuffer;
786  {
787    llvm::BitstreamWriter OutputStream(OutputBuffer);
788    Builder.writeIndex(OutputStream);
789  }
790
791  // Write the global index file to a temporary file.
792  llvm::SmallString<128> IndexTmpPath;
793  int TmpFD;
794  if (llvm::sys::fs::createUniqueFile(IndexPath + "-%%%%%%%%", TmpFD,
795                                      IndexTmpPath))
796    return EC_IOError;
797
798  // Open the temporary global index file for output.
799  llvm::raw_fd_ostream Out(TmpFD, true);
800  if (Out.has_error())
801    return EC_IOError;
802
803  // Write the index.
804  Out.write(OutputBuffer.data(), OutputBuffer.size());
805  Out.close();
806  if (Out.has_error())
807    return EC_IOError;
808
809  // Remove the old index file. It isn't relevant any more.
810  bool OldIndexExisted;
811  llvm::sys::fs::remove(IndexPath.str(), OldIndexExisted);
812
813  // Rename the newly-written index file to the proper name.
814  if (llvm::sys::fs::rename(IndexTmpPath.str(), IndexPath.str())) {
815    // Rename failed; just remove the
816    llvm::sys::fs::remove(IndexTmpPath.str(), OldIndexExisted);
817    return EC_IOError;
818  }
819
820  // We're done.
821  return EC_None;
822}
823
824namespace {
825  class GlobalIndexIdentifierIterator : public IdentifierIterator {
826    /// \brief The current position within the identifier lookup table.
827    IdentifierIndexTable::key_iterator Current;
828
829    /// \brief The end position within the identifier lookup table.
830    IdentifierIndexTable::key_iterator End;
831
832  public:
833    explicit GlobalIndexIdentifierIterator(IdentifierIndexTable &Idx) {
834      Current = Idx.key_begin();
835      End = Idx.key_end();
836    }
837
838    virtual StringRef Next() {
839      if (Current == End)
840        return StringRef();
841
842      StringRef Result = *Current;
843      ++Current;
844      return Result;
845    }
846  };
847}
848
849IdentifierIterator *GlobalModuleIndex::createIdentifierIterator() const {
850  IdentifierIndexTable &Table =
851    *static_cast<IdentifierIndexTable *>(IdentifierIndex);
852  return new GlobalIndexIdentifierIterator(Table);
853}
854