ASTReader.h revision bef1a7b9c175d37e4a727e6ce68bd05232fa6970
1//===--- ASTReader.h - AST File Reader --------------------------*- 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 defines the ASTReader class, which reads AST files. 11// 12//===----------------------------------------------------------------------===// 13 14#ifndef LLVM_CLANG_FRONTEND_AST_READER_H 15#define LLVM_CLANG_FRONTEND_AST_READER_H 16 17#include "clang/Serialization/ASTBitCodes.h" 18#include "clang/Sema/ExternalSemaSource.h" 19#include "clang/AST/DeclarationName.h" 20#include "clang/AST/DeclObjC.h" 21#include "clang/AST/TemplateBase.h" 22#include "clang/Lex/ExternalPreprocessorSource.h" 23#include "clang/Lex/PreprocessingRecord.h" 24#include "clang/Basic/Diagnostic.h" 25#include "clang/Basic/IdentifierTable.h" 26#include "clang/Basic/SourceManager.h" 27#include "llvm/ADT/APFloat.h" 28#include "llvm/ADT/APInt.h" 29#include "llvm/ADT/APSInt.h" 30#include "llvm/ADT/OwningPtr.h" 31#include "llvm/ADT/SmallVector.h" 32#include "llvm/ADT/StringRef.h" 33#include "llvm/Bitcode/BitstreamReader.h" 34#include "llvm/System/DataTypes.h" 35#include <deque> 36#include <map> 37#include <string> 38#include <utility> 39#include <vector> 40 41namespace llvm { 42 class MemoryBuffer; 43} 44 45namespace clang { 46 47class AddrLabelExpr; 48class ASTConsumer; 49class ASTContext; 50class ASTIdentifierIterator; 51class Attr; 52class Decl; 53class DeclContext; 54class NestedNameSpecifier; 55class CXXBaseSpecifier; 56class CXXBaseOrMemberInitializer; 57class GotoStmt; 58class LabelStmt; 59class MacroDefinition; 60class NamedDecl; 61class Preprocessor; 62class Sema; 63class SwitchCase; 64class ASTDeserializationListener; 65class ASTReader; 66class ASTDeclReader; 67class ASTStmtReader; 68class ASTIdentifierLookupTrait; 69class TypeLocReader; 70struct HeaderFileInfo; 71 72struct PCHPredefinesBlock { 73 /// \brief The file ID for this predefines buffer in a PCH file. 74 FileID BufferID; 75 76 /// \brief This predefines buffer in a PCH file. 77 llvm::StringRef Data; 78}; 79typedef llvm::SmallVector<PCHPredefinesBlock, 2> PCHPredefinesBlocks; 80 81/// \brief Abstract interface for callback invocations by the ASTReader. 82/// 83/// While reading an AST file, the ASTReader will call the methods of the 84/// listener to pass on specific information. Some of the listener methods can 85/// return true to indicate to the ASTReader that the information (and 86/// consequently the AST file) is invalid. 87class ASTReaderListener { 88public: 89 virtual ~ASTReaderListener(); 90 91 /// \brief Receives the language options. 92 /// 93 /// \returns true to indicate the options are invalid or false otherwise. 94 virtual bool ReadLanguageOptions(const LangOptions &LangOpts) { 95 return false; 96 } 97 98 /// \brief Receives the target triple. 99 /// 100 /// \returns true to indicate the target triple is invalid or false otherwise. 101 virtual bool ReadTargetTriple(llvm::StringRef Triple) { 102 return false; 103 } 104 105 /// \brief Receives the contents of the predefines buffer. 106 /// 107 /// \param Buffers Information about the predefines buffers. 108 /// 109 /// \param OriginalFileName The original file name for the AST file, which 110 /// will appear as an entry in the predefines buffer. 111 /// 112 /// \param SuggestedPredefines If necessary, additional definitions are added 113 /// here. 114 /// 115 /// \returns true to indicate the predefines are invalid or false otherwise. 116 virtual bool ReadPredefinesBuffer(const PCHPredefinesBlocks &Buffers, 117 llvm::StringRef OriginalFileName, 118 std::string &SuggestedPredefines) { 119 return false; 120 } 121 122 /// \brief Receives a HeaderFileInfo entry. 123 virtual void ReadHeaderFileInfo(const HeaderFileInfo &HFI, unsigned ID) {} 124 125 /// \brief Receives __COUNTER__ value. 126 virtual void ReadCounter(unsigned Value) {} 127}; 128 129/// \brief ASTReaderListener implementation to validate the information of 130/// the PCH file against an initialized Preprocessor. 131class PCHValidator : public ASTReaderListener { 132 Preprocessor &PP; 133 ASTReader &Reader; 134 135 unsigned NumHeaderInfos; 136 137public: 138 PCHValidator(Preprocessor &PP, ASTReader &Reader) 139 : PP(PP), Reader(Reader), NumHeaderInfos(0) {} 140 141 virtual bool ReadLanguageOptions(const LangOptions &LangOpts); 142 virtual bool ReadTargetTriple(llvm::StringRef Triple); 143 virtual bool ReadPredefinesBuffer(const PCHPredefinesBlocks &Buffers, 144 llvm::StringRef OriginalFileName, 145 std::string &SuggestedPredefines); 146 virtual void ReadHeaderFileInfo(const HeaderFileInfo &HFI, unsigned ID); 147 virtual void ReadCounter(unsigned Value); 148 149private: 150 void Error(const char *Msg); 151}; 152 153/// \brief Reads an AST files chain containing the contents of a translation 154/// unit. 155/// 156/// The ASTReader class reads bitstreams (produced by the ASTWriter 157/// class) containing the serialized representation of a given 158/// abstract syntax tree and its supporting data structures. An 159/// instance of the ASTReader can be attached to an ASTContext object, 160/// which will provide access to the contents of the AST files. 161/// 162/// The AST reader provides lazy de-serialization of declarations, as 163/// required when traversing the AST. Only those AST nodes that are 164/// actually required will be de-serialized. 165class ASTReader 166 : public ExternalPreprocessorSource, 167 public ExternalPreprocessingRecordSource, 168 public ExternalSemaSource, 169 public IdentifierInfoLookup, 170 public ExternalIdentifierLookup, 171 public ExternalSLocEntrySource { 172public: 173 enum ASTReadResult { Success, Failure, IgnorePCH }; 174 /// \brief Types of AST files. 175 enum ASTFileType { 176 Module, ///< File is a module proper. 177 PCH, ///< File is a PCH file treated as such. 178 Preamble, ///< File is a PCH file treated as the preamble. 179 MainFile ///< File is a PCH file treated as the actual main file. 180 }; 181 friend class PCHValidator; 182 friend class ASTDeclReader; 183 friend class ASTStmtReader; 184 friend class ASTIdentifierIterator; 185 friend class ASTIdentifierLookupTrait; 186 friend class TypeLocReader; 187private: 188 /// \brief The receiver of some callbacks invoked by ASTReader. 189 llvm::OwningPtr<ASTReaderListener> Listener; 190 191 /// \brief The receiver of deserialization events. 192 ASTDeserializationListener *DeserializationListener; 193 194 SourceManager &SourceMgr; 195 FileManager &FileMgr; 196 Diagnostic &Diags; 197 198 /// \brief The semantic analysis object that will be processing the 199 /// AST files and the translation unit that uses it. 200 Sema *SemaObj; 201 202 /// \brief The preprocessor that will be loading the source file. 203 Preprocessor *PP; 204 205 /// \brief The AST context into which we'll read the AST files. 206 ASTContext *Context; 207 208 /// \brief The AST consumer. 209 ASTConsumer *Consumer; 210 211 /// \brief Information that is needed for every module. 212 struct PerFileData { 213 PerFileData(ASTFileType Ty); 214 ~PerFileData(); 215 216 // === General information === 217 218 /// \brief The type of this AST file. 219 ASTFileType Type; 220 221 /// \brief The file name of the AST file. 222 std::string FileName; 223 224 /// \brief The memory buffer that stores the data associated with 225 /// this AST file. 226 llvm::OwningPtr<llvm::MemoryBuffer> Buffer; 227 228 /// \brief The size of this file, in bits. 229 uint64_t SizeInBits; 230 231 /// \brief The bitstream reader from which we'll read the AST file. 232 llvm::BitstreamReader StreamFile; 233 234 /// \brief The main bitstream cursor for the main block. 235 llvm::BitstreamCursor Stream; 236 237 // === Source Locations === 238 239 /// \brief Cursor used to read source location entries. 240 llvm::BitstreamCursor SLocEntryCursor; 241 242 /// \brief The number of source location entries in this AST file. 243 unsigned LocalNumSLocEntries; 244 245 /// \brief Offsets for all of the source location entries in the 246 /// AST file. 247 const uint32_t *SLocOffsets; 248 249 /// \brief The entire size of this module's source location offset range. 250 unsigned LocalSLocSize; 251 252 // === Identifiers === 253 254 /// \brief The number of identifiers in this AST file. 255 unsigned LocalNumIdentifiers; 256 257 /// \brief Offsets into the identifier table data. 258 /// 259 /// This array is indexed by the identifier ID (-1), and provides 260 /// the offset into IdentifierTableData where the string data is 261 /// stored. 262 const uint32_t *IdentifierOffsets; 263 264 /// \brief Actual data for the on-disk hash table. 265 /// 266 /// This pointer points into a memory buffer, where the on-disk hash 267 /// table for identifiers actually lives. 268 const char *IdentifierTableData; 269 270 /// \brief A pointer to an on-disk hash table of opaque type 271 /// IdentifierHashTable. 272 void *IdentifierLookupTable; 273 274 // === Macros === 275 276 /// \brief The cursor to the start of the preprocessor block, which stores 277 /// all of the macro definitions. 278 llvm::BitstreamCursor MacroCursor; 279 280 /// \brief The offset of the start of the set of defined macros. 281 uint64_t MacroStartOffset; 282 283 /// \brief The number of macro definitions in this file. 284 unsigned LocalNumMacroDefinitions; 285 286 /// \brief Offsets of all of the macro definitions in the preprocessing 287 /// record in the AST file. 288 const uint32_t *MacroDefinitionOffsets; 289 290 // === Selectors === 291 292 /// \brief The number of selectors new to this file. 293 /// 294 /// This is the number of entries in SelectorOffsets. 295 unsigned LocalNumSelectors; 296 297 /// \brief Offsets into the selector lookup table's data array 298 /// where each selector resides. 299 const uint32_t *SelectorOffsets; 300 301 /// \brief A pointer to the character data that comprises the selector table 302 /// 303 /// The SelectorOffsets table refers into this memory. 304 const unsigned char *SelectorLookupTableData; 305 306 /// \brief A pointer to an on-disk hash table of opaque type 307 /// ASTSelectorLookupTable. 308 /// 309 /// This hash table provides the IDs of all selectors, and the associated 310 /// instance and factory methods. 311 void *SelectorLookupTable; 312 313 /// \brief Method selectors used in a @selector expression. Used for 314 /// implementation of -Wselector. 315 llvm::SmallVector<uint64_t, 64> ReferencedSelectorsData; 316 317 // === Declarations === 318 319 /// DeclsCursor - This is a cursor to the start of the DECLS_BLOCK block. It 320 /// has read all the abbreviations at the start of the block and is ready to 321 /// jump around with these in context. 322 llvm::BitstreamCursor DeclsCursor; 323 324 /// \brief The number of declarations in this AST file. 325 unsigned LocalNumDecls; 326 327 /// \brief Offset of each declaration within the bitstream, indexed 328 /// by the declaration ID (-1). 329 const uint32_t *DeclOffsets; 330 331 /// \brief A snapshot of the pending instantiations in the chain. 332 /// 333 /// This record tracks the instantiations that Sema has to perform at the 334 /// end of the TU. It consists of a pair of values for every pending 335 /// instantiation where the first value is the ID of the decl and the second 336 /// is the instantiation location. 337 llvm::SmallVector<uint64_t, 64> PendingInstantiations; 338 339 // === Types === 340 341 /// \brief The number of types in this AST file. 342 unsigned LocalNumTypes; 343 344 /// \brief Offset of each type within the bitstream, indexed by the 345 /// type ID, or the representation of a Type*. 346 const uint32_t *TypeOffsets; 347 348 // === Miscellaneous === 349 350 /// \brief The AST stat cache installed for this file, if any. 351 /// 352 /// The dynamic type of this stat cache is always ASTStatCache 353 void *StatCache; 354 355 /// \brief The number of preallocated preprocessing entities in the 356 /// preprocessing record. 357 unsigned NumPreallocatedPreprocessingEntities; 358 359 /// \brief The next module in source order. 360 PerFileData *NextInSource; 361 362 /// \brief All the modules that loaded this one. Can contain NULL for 363 /// directly loaded modules. 364 llvm::SmallVector<PerFileData *, 1> Loaders; 365 }; 366 367 /// \brief All loaded modules, indexed by name. 368 llvm::StringMap<PerFileData*> Modules; 369 370 /// \brief The first module in source order. 371 PerFileData *FirstInSource; 372 373 /// \brief The chain of AST files. The first entry is the one named by the 374 /// user, the last one is the one that doesn't depend on anything further. 375 /// That is, the entry I was created with -include-pch I+1. 376 llvm::SmallVector<PerFileData*, 2> Chain; 377 378 /// \brief SLocEntries that we're going to preload. 379 llvm::SmallVector<uint64_t, 64> PreloadSLocEntries; 380 381 /// \brief Types that have already been loaded from the chain. 382 /// 383 /// When the pointer at index I is non-NULL, the type with 384 /// ID = (I + 1) << FastQual::Width has already been loaded 385 std::vector<QualType> TypesLoaded; 386 387 /// \brief Map that provides the ID numbers of each type within the 388 /// output stream, plus those deserialized from a chained PCH. 389 /// 390 /// The ID numbers of types are consecutive (in order of discovery) 391 /// and start at 1. 0 is reserved for NULL. When types are actually 392 /// stored in the stream, the ID number is shifted by 2 bits to 393 /// allow for the const/volatile qualifiers. 394 /// 395 /// Keys in the map never have const/volatile qualifiers. 396 serialization::TypeIdxMap TypeIdxs; 397 398 /// \brief Declarations that have already been loaded from the chain. 399 /// 400 /// When the pointer at index I is non-NULL, the declaration with ID 401 /// = I + 1 has already been loaded. 402 std::vector<Decl *> DeclsLoaded; 403 404 typedef std::pair<PerFileData *, uint64_t> FileOffset; 405 typedef llvm::SmallVector<FileOffset, 2> FileOffsetsTy; 406 typedef llvm::DenseMap<serialization::DeclID, FileOffsetsTy> 407 DeclUpdateOffsetsMap; 408 /// \brief Declarations that have modifications residing in a later file 409 /// in the chain. 410 DeclUpdateOffsetsMap DeclUpdateOffsets; 411 412 typedef llvm::DenseMap<serialization::DeclID, 413 std::pair<PerFileData *, uint64_t> > 414 DeclReplacementMap; 415 /// \brief Declarations that have been replaced in a later file in the chain. 416 DeclReplacementMap ReplacedDecls; 417 418 /// \brief Information about the contents of a DeclContext. 419 struct DeclContextInfo { 420 void *NameLookupTableData; // a ASTDeclContextNameLookupTable. 421 const serialization::KindDeclIDPair *LexicalDecls; 422 unsigned NumLexicalDecls; 423 }; 424 // In a full chain, there could be multiple updates to every decl context, 425 // so this is a vector. However, typically a chain is only two elements long, 426 // with only one file containing updates, so there will be only one update 427 // per decl context. 428 typedef llvm::SmallVector<DeclContextInfo, 1> DeclContextInfos; 429 typedef llvm::DenseMap<const DeclContext *, DeclContextInfos> 430 DeclContextOffsetsMap; 431 // Updates for visible decls can occur for other contexts than just the 432 // TU, and when we read those update records, the actual context will not 433 // be available yet (unless it's the TU), so have this pending map using the 434 // ID as a key. It will be realized when the context is actually loaded. 435 typedef llvm::SmallVector<void *, 1> DeclContextVisibleUpdates; 436 typedef llvm::DenseMap<serialization::DeclID, DeclContextVisibleUpdates> 437 DeclContextVisibleUpdatesPending; 438 439 /// \brief Offsets of the lexical and visible declarations for each 440 /// DeclContext. 441 DeclContextOffsetsMap DeclContextOffsets; 442 443 /// \brief Updates to the visible declarations of declaration contexts that 444 /// haven't been loaded yet. 445 DeclContextVisibleUpdatesPending PendingVisibleUpdates; 446 447 typedef llvm::SmallVector<CXXRecordDecl *, 4> ForwardRefs; 448 typedef llvm::DenseMap<const CXXRecordDecl *, ForwardRefs> 449 PendingForwardRefsMap; 450 /// \brief Forward references that have a definition but the definition decl 451 /// is still initializing. When the definition gets read it will update 452 /// the DefinitionData pointer of all pending references. 453 PendingForwardRefsMap PendingForwardRefs; 454 455 typedef llvm::DenseMap<serialization::DeclID, serialization::DeclID> 456 FirstLatestDeclIDMap; 457 /// \brief Map of first declarations from a chained PCH that point to the 458 /// most recent declarations in another AST file. 459 FirstLatestDeclIDMap FirstLatestDeclIDs; 460 461 /// \brief Read the records that describe the contents of declcontexts. 462 bool ReadDeclContextStorage(llvm::BitstreamCursor &Cursor, 463 const std::pair<uint64_t, uint64_t> &Offsets, 464 DeclContextInfo &Info); 465 466 /// \brief A vector containing identifiers that have already been 467 /// loaded. 468 /// 469 /// If the pointer at index I is non-NULL, then it refers to the 470 /// IdentifierInfo for the identifier with ID=I+1 that has already 471 /// been loaded. 472 std::vector<IdentifierInfo *> IdentifiersLoaded; 473 474 /// \brief A vector containing selectors that have already been loaded. 475 /// 476 /// This vector is indexed by the Selector ID (-1). NULL selector 477 /// entries indicate that the particular selector ID has not yet 478 /// been loaded. 479 llvm::SmallVector<Selector, 16> SelectorsLoaded; 480 481 /// \brief The macro definitions we have already loaded. 482 llvm::SmallVector<MacroDefinition *, 16> MacroDefinitionsLoaded; 483 484 /// \name CodeGen-relevant special data 485 /// \brief Fields containing data that is relevant to CodeGen. 486 //@{ 487 488 /// \brief The IDs of all declarations that fulfill the criteria of 489 /// "interesting" decls. 490 /// 491 /// This contains the data loaded from all EXTERNAL_DEFINITIONS blocks in the 492 /// chain. The referenced declarations are deserialized and passed to the 493 /// consumer eagerly. 494 llvm::SmallVector<uint64_t, 16> ExternalDefinitions; 495 496 /// \brief The IDs of all tentative definitions stored in the the chain. 497 /// 498 /// Sema keeps track of all tentative definitions in a TU because it has to 499 /// complete them and pass them on to CodeGen. Thus, tentative definitions in 500 /// the PCH chain must be eagerly deserialized. 501 llvm::SmallVector<uint64_t, 16> TentativeDefinitions; 502 503 /// \brief The IDs of all CXXRecordDecls stored in the chain whose VTables are 504 /// used. 505 /// 506 /// CodeGen has to emit VTables for these records, so they have to be eagerly 507 /// deserialized. 508 llvm::SmallVector<uint64_t, 64> VTableUses; 509 510 //@} 511 512 /// \name Diagnostic-relevant special data 513 /// \brief Fields containing data that is used for generating diagnostics 514 //@{ 515 516 /// \brief A snapshot of Sema's unused file-scoped variable tracking, for 517 /// generating warnings. 518 llvm::SmallVector<uint64_t, 16> UnusedFileScopedDecls; 519 520 /// \brief A snapshot of Sema's weak undeclared identifier tracking, for 521 /// generating warnings. 522 llvm::SmallVector<uint64_t, 64> WeakUndeclaredIdentifiers; 523 524 /// \brief The IDs of type aliases for ext_vectors that exist in the chain. 525 /// 526 /// Used by Sema for finding sugared names for ext_vectors in diagnostics. 527 llvm::SmallVector<uint64_t, 4> ExtVectorDecls; 528 529 //@} 530 531 /// \name Sema-relevant special data 532 /// \brief Fields containing data that is used for semantic analysis 533 //@{ 534 535 /// \brief The IDs of all locally scoped external decls in the chain. 536 /// 537 /// Sema tracks these to validate that the types are consistent across all 538 /// local external declarations. 539 llvm::SmallVector<uint64_t, 16> LocallyScopedExternalDecls; 540 541 /// \brief The IDs of all dynamic class declarations in the chain. 542 /// 543 /// Sema tracks these because it checks for the key functions being defined 544 /// at the end of the TU, in which case it directs CodeGen to emit the VTable. 545 llvm::SmallVector<uint64_t, 16> DynamicClasses; 546 547 /// \brief The IDs of the declarations Sema stores directly. 548 /// 549 /// Sema tracks a few important decls, such as namespace std, directly. 550 llvm::SmallVector<uint64_t, 4> SemaDeclRefs; 551 552 /// \brief The IDs of the types ASTContext stores directly. 553 /// 554 /// The AST context tracks a few important types, such as va_list, directly. 555 llvm::SmallVector<uint64_t, 16> SpecialTypes; 556 557 //@} 558 559 /// \brief The original file name that was used to build the primary AST file, 560 /// which may have been modified for relocatable-pch support. 561 std::string OriginalFileName; 562 563 /// \brief The actual original file name that was used to build the primary 564 /// AST file. 565 std::string ActualOriginalFileName; 566 567 /// \brief Whether this precompiled header is a relocatable PCH file. 568 bool RelocatablePCH; 569 570 /// \brief The system include root to be used when loading the 571 /// precompiled header. 572 const char *isysroot; 573 574 /// \brief Whether to disable the normal validation performed on precompiled 575 /// headers when they are loaded. 576 bool DisableValidation; 577 578 /// \brief Mapping from switch-case IDs in the chain to switch-case statements 579 /// 580 /// Statements usually don't have IDs, but switch cases need them, so that the 581 /// switch statement can refer to them. 582 std::map<unsigned, SwitchCase *> SwitchCaseStmts; 583 584 /// \brief Mapping from label statement IDs in the chain to label statements. 585 /// 586 /// Statements usually don't have IDs, but labeled statements need them, so 587 /// that goto statements and address-of-label expressions can refer to them. 588 std::map<unsigned, LabelStmt *> LabelStmts; 589 590 /// \brief Mapping from label IDs to the set of "goto" statements 591 /// that point to that label before the label itself has been 592 /// de-serialized. 593 std::multimap<unsigned, GotoStmt *> UnresolvedGotoStmts; 594 595 /// \brief Mapping from label IDs to the set of address label 596 /// expressions that point to that label before the label itself has 597 /// been de-serialized. 598 std::multimap<unsigned, AddrLabelExpr *> UnresolvedAddrLabelExprs; 599 600 /// \brief The number of stat() calls that hit/missed the stat 601 /// cache. 602 unsigned NumStatHits, NumStatMisses; 603 604 /// \brief The number of source location entries de-serialized from 605 /// the PCH file. 606 unsigned NumSLocEntriesRead; 607 608 /// \brief The number of source location entries in the chain. 609 unsigned TotalNumSLocEntries; 610 611 /// \brief The next offset for a SLocEntry after everything in this reader. 612 unsigned NextSLocOffset; 613 614 /// \brief The number of statements (and expressions) de-serialized 615 /// from the chain. 616 unsigned NumStatementsRead; 617 618 /// \brief The total number of statements (and expressions) stored 619 /// in the chain. 620 unsigned TotalNumStatements; 621 622 /// \brief The number of macros de-serialized from the chain. 623 unsigned NumMacrosRead; 624 625 /// \brief The total number of macros stored in the chain. 626 unsigned TotalNumMacros; 627 628 /// \brief The number of selectors that have been read. 629 unsigned NumSelectorsRead; 630 631 /// \brief The number of method pool entries that have been read. 632 unsigned NumMethodPoolEntriesRead; 633 634 /// \brief The number of times we have looked up a selector in the method 635 /// pool and not found anything interesting. 636 unsigned NumMethodPoolMisses; 637 638 /// \brief The total number of method pool entries in the selector table. 639 unsigned TotalNumMethodPoolEntries; 640 641 /// Number of lexical decl contexts read/total. 642 unsigned NumLexicalDeclContextsRead, TotalLexicalDeclContexts; 643 644 /// Number of visible decl contexts read/total. 645 unsigned NumVisibleDeclContextsRead, TotalVisibleDeclContexts; 646 647 /// \brief Number of Decl/types that are currently deserializing. 648 unsigned NumCurrentElementsDeserializing; 649 650 /// \brief An IdentifierInfo that has been loaded but whose top-level 651 /// declarations of the same name have not (yet) been loaded. 652 struct PendingIdentifierInfo { 653 IdentifierInfo *II; 654 llvm::SmallVector<uint32_t, 4> DeclIDs; 655 }; 656 657 /// \brief The set of identifiers that were read while the AST reader was 658 /// (recursively) loading declarations. 659 /// 660 /// The declarations on the identifier chain for these identifiers will be 661 /// loaded once the recursive loading has completed. 662 std::deque<PendingIdentifierInfo> PendingIdentifierInfos; 663 664 /// \brief Contains declarations and definitions that will be 665 /// "interesting" to the ASTConsumer, when we get that AST consumer. 666 /// 667 /// "Interesting" declarations are those that have data that may 668 /// need to be emitted, such as inline function definitions or 669 /// Objective-C protocols. 670 std::deque<Decl *> InterestingDecls; 671 672 /// \brief When reading a Stmt tree, Stmt operands are placed in this stack. 673 llvm::SmallVector<Stmt *, 16> StmtStack; 674 675 /// \brief What kind of records we are reading. 676 enum ReadingKind { 677 Read_Decl, Read_Type, Read_Stmt 678 }; 679 680 /// \brief What kind of records we are reading. 681 ReadingKind ReadingKind; 682 683 /// \brief RAII object to change the reading kind. 684 class ReadingKindTracker { 685 ASTReader &Reader; 686 enum ReadingKind PrevKind; 687 688 ReadingKindTracker(const ReadingKindTracker&); // do not implement 689 ReadingKindTracker &operator=(const ReadingKindTracker&);// do not implement 690 691 public: 692 ReadingKindTracker(enum ReadingKind newKind, ASTReader &reader) 693 : Reader(reader), PrevKind(Reader.ReadingKind) { 694 Reader.ReadingKind = newKind; 695 } 696 697 ~ReadingKindTracker() { Reader.ReadingKind = PrevKind; } 698 }; 699 700 /// \brief All predefines buffers in the chain, to be treated as if 701 /// concatenated. 702 PCHPredefinesBlocks PCHPredefinesBuffers; 703 704 /// \brief Suggested contents of the predefines buffer, after this 705 /// PCH file has been processed. 706 /// 707 /// In most cases, this string will be empty, because the predefines 708 /// buffer computed to build the PCH file will be identical to the 709 /// predefines buffer computed from the command line. However, when 710 /// there are differences that the PCH reader can work around, this 711 /// predefines buffer may contain additional definitions. 712 std::string SuggestedPredefines; 713 714 /// \brief Reads a statement from the specified cursor. 715 Stmt *ReadStmtFromStream(PerFileData &F); 716 717 void MaybeAddSystemRootToFilename(std::string &Filename); 718 719 ASTReadResult ReadASTCore(llvm::StringRef FileName, ASTFileType Type); 720 ASTReadResult ReadASTBlock(PerFileData &F); 721 bool CheckPredefinesBuffers(); 722 bool ParseLineTable(PerFileData &F, llvm::SmallVectorImpl<uint64_t> &Record); 723 ASTReadResult ReadSourceManagerBlock(PerFileData &F); 724 ASTReadResult ReadSLocEntryRecord(unsigned ID); 725 PerFileData *SLocCursorForID(unsigned ID); 726 SourceLocation getImportLocation(PerFileData *F); 727 bool ParseLanguageOptions(const llvm::SmallVectorImpl<uint64_t> &Record); 728 729 struct RecordLocation { 730 RecordLocation(PerFileData *M, uint64_t O) 731 : F(M), Offset(O) {} 732 PerFileData *F; 733 uint64_t Offset; 734 }; 735 736 QualType ReadTypeRecord(unsigned Index); 737 RecordLocation TypeCursorForIndex(unsigned Index); 738 void LoadedDecl(unsigned Index, Decl *D); 739 Decl *ReadDeclRecord(unsigned Index, serialization::DeclID ID); 740 RecordLocation DeclCursorForIndex(unsigned Index, serialization::DeclID ID); 741 742 void PassInterestingDeclsToConsumer(); 743 744 /// \brief Produce an error diagnostic and return true. 745 /// 746 /// This routine should only be used for fatal errors that have to 747 /// do with non-routine failures (e.g., corrupted AST file). 748 void Error(const char *Msg); 749 750 ASTReader(const ASTReader&); // do not implement 751 ASTReader &operator=(const ASTReader &); // do not implement 752public: 753 typedef llvm::SmallVector<uint64_t, 64> RecordData; 754 755 /// \brief Load the AST file and validate its contents against the given 756 /// Preprocessor. 757 /// 758 /// \param PP the preprocessor associated with the context in which this 759 /// precompiled header will be loaded. 760 /// 761 /// \param Context the AST context that this precompiled header will be 762 /// loaded into. 763 /// 764 /// \param isysroot If non-NULL, the system include path specified by the 765 /// user. This is only used with relocatable PCH files. If non-NULL, 766 /// a relocatable PCH file will use the default path "/". 767 /// 768 /// \param DisableValidation If true, the AST reader will suppress most 769 /// of its regular consistency checking, allowing the use of precompiled 770 /// headers that cannot be determined to be compatible. 771 ASTReader(Preprocessor &PP, ASTContext *Context, const char *isysroot = 0, 772 bool DisableValidation = false); 773 774 /// \brief Load the AST file without using any pre-initialized Preprocessor. 775 /// 776 /// The necessary information to initialize a Preprocessor later can be 777 /// obtained by setting a ASTReaderListener. 778 /// 779 /// \param SourceMgr the source manager into which the AST file will be loaded 780 /// 781 /// \param FileMgr the file manager into which the AST file will be loaded. 782 /// 783 /// \param Diags the diagnostics system to use for reporting errors and 784 /// warnings relevant to loading the AST file. 785 /// 786 /// \param isysroot If non-NULL, the system include path specified by the 787 /// user. This is only used with relocatable PCH files. If non-NULL, 788 /// a relocatable PCH file will use the default path "/". 789 /// 790 /// \param DisableValidation If true, the AST reader will suppress most 791 /// of its regular consistency checking, allowing the use of precompiled 792 /// headers that cannot be determined to be compatible. 793 ASTReader(SourceManager &SourceMgr, FileManager &FileMgr, 794 Diagnostic &Diags, const char *isysroot = 0, 795 bool DisableValidation = false); 796 ~ASTReader(); 797 798 /// \brief Load the precompiled header designated by the given file 799 /// name. 800 ASTReadResult ReadAST(const std::string &FileName, ASTFileType Type); 801 802 /// \brief Set the AST callbacks listener. 803 void setListener(ASTReaderListener *listener) { 804 Listener.reset(listener); 805 } 806 807 /// \brief Set the AST deserialization listener. 808 void setDeserializationListener(ASTDeserializationListener *Listener); 809 810 /// \brief Set the Preprocessor to use. 811 void setPreprocessor(Preprocessor &pp); 812 813 /// \brief Sets and initializes the given Context. 814 void InitializeContext(ASTContext &Context); 815 816 /// \brief Retrieve the name of the named (primary) AST file 817 const std::string &getFileName() const { return Chain[0]->FileName; } 818 819 /// \brief Retrieve the name of the original source file name 820 const std::string &getOriginalSourceFile() { return OriginalFileName; } 821 822 /// \brief Retrieve the name of the original source file name directly from 823 /// the AST file, without actually loading the AST file. 824 static std::string getOriginalSourceFile(const std::string &ASTFileName, 825 Diagnostic &Diags); 826 827 /// \brief Returns the suggested contents of the predefines buffer, 828 /// which contains a (typically-empty) subset of the predefines 829 /// build prior to including the precompiled header. 830 const std::string &getSuggestedPredefines() { return SuggestedPredefines; } 831 832 /// \brief Read preprocessed entities into the 833 virtual void ReadPreprocessedEntities(); 834 835 /// \brief Returns the number of source locations found in the chain. 836 unsigned getTotalNumSLocs() const { 837 return TotalNumSLocEntries; 838 } 839 840 /// \brief Returns the next SLocEntry offset after the chain. 841 unsigned getNextSLocOffset() const { 842 return NextSLocOffset; 843 } 844 845 /// \brief Returns the number of identifiers found in the chain. 846 unsigned getTotalNumIdentifiers() const { 847 return static_cast<unsigned>(IdentifiersLoaded.size()); 848 } 849 850 /// \brief Returns the number of types found in the chain. 851 unsigned getTotalNumTypes() const { 852 return static_cast<unsigned>(TypesLoaded.size()); 853 } 854 855 /// \brief Returns the number of declarations found in the chain. 856 unsigned getTotalNumDecls() const { 857 return static_cast<unsigned>(DeclsLoaded.size()); 858 } 859 860 /// \brief Returns the number of selectors found in the chain. 861 unsigned getTotalNumSelectors() const { 862 return static_cast<unsigned>(SelectorsLoaded.size()); 863 } 864 865 /// \brief Returns the number of macro definitions found in the chain. 866 unsigned getTotalNumMacroDefinitions() const { 867 return static_cast<unsigned>(MacroDefinitionsLoaded.size()); 868 } 869 870 /// \brief Reads a TemplateArgumentLocInfo appropriate for the 871 /// given TemplateArgument kind. 872 TemplateArgumentLocInfo 873 GetTemplateArgumentLocInfo(PerFileData &F, TemplateArgument::ArgKind Kind, 874 const RecordData &Record, unsigned &Idx); 875 876 /// \brief Reads a TemplateArgumentLoc. 877 TemplateArgumentLoc 878 ReadTemplateArgumentLoc(PerFileData &F, 879 const RecordData &Record, unsigned &Idx); 880 881 /// \brief Reads a declarator info from the given record. 882 TypeSourceInfo *GetTypeSourceInfo(PerFileData &F, 883 const RecordData &Record, unsigned &Idx); 884 885 /// \brief Resolve and return the translation unit declaration. 886 TranslationUnitDecl *GetTranslationUnitDecl(); 887 888 /// \brief Resolve a type ID into a type, potentially building a new 889 /// type. 890 QualType GetType(serialization::TypeID ID); 891 892 /// \brief Returns the type ID associated with the given type. 893 /// If the type didn't come from the AST file the ID that is returned is 894 /// marked as "doesn't exist in AST". 895 serialization::TypeID GetTypeID(QualType T) const; 896 897 /// \brief Returns the type index associated with the given type. 898 /// If the type didn't come from the AST file the index that is returned is 899 /// marked as "doesn't exist in AST". 900 serialization::TypeIdx GetTypeIdx(QualType T) const; 901 902 /// \brief Resolve a declaration ID into a declaration, potentially 903 /// building a new declaration. 904 Decl *GetDecl(serialization::DeclID ID); 905 virtual Decl *GetExternalDecl(uint32_t ID); 906 907 /// \brief Resolve the offset of a statement into a statement. 908 /// 909 /// This operation will read a new statement from the external 910 /// source each time it is called, and is meant to be used via a 911 /// LazyOffsetPtr (which is used by Decls for the body of functions, etc). 912 virtual Stmt *GetExternalDeclStmt(uint64_t Offset); 913 914 /// ReadBlockAbbrevs - Enter a subblock of the specified BlockID with the 915 /// specified cursor. Read the abbreviations that are at the top of the block 916 /// and then leave the cursor pointing into the block. 917 bool ReadBlockAbbrevs(llvm::BitstreamCursor &Cursor, unsigned BlockID); 918 919 /// \brief Finds all the visible declarations with a given name. 920 /// The current implementation of this method just loads the entire 921 /// lookup table as unmaterialized references. 922 virtual DeclContext::lookup_result 923 FindExternalVisibleDeclsByName(const DeclContext *DC, 924 DeclarationName Name); 925 926 virtual void MaterializeVisibleDecls(const DeclContext *DC); 927 928 /// \brief Read all of the declarations lexically stored in a 929 /// declaration context. 930 /// 931 /// \param DC The declaration context whose declarations will be 932 /// read. 933 /// 934 /// \param Decls Vector that will contain the declarations loaded 935 /// from the external source. The caller is responsible for merging 936 /// these declarations with any declarations already stored in the 937 /// declaration context. 938 /// 939 /// \returns true if there was an error while reading the 940 /// declarations for this declaration context. 941 virtual bool FindExternalLexicalDecls(const DeclContext *DC, 942 bool (*isKindWeWant)(Decl::Kind), 943 llvm::SmallVectorImpl<Decl*> &Decls); 944 945 /// \brief Notify ASTReader that we started deserialization of 946 /// a decl or type so until FinishedDeserializing is called there may be 947 /// decls that are initializing. Must be paired with FinishedDeserializing. 948 virtual void StartedDeserializing() { ++NumCurrentElementsDeserializing; } 949 950 /// \brief Notify ASTReader that we finished the deserialization of 951 /// a decl or type. Must be paired with StartedDeserializing. 952 virtual void FinishedDeserializing(); 953 954 /// \brief Function that will be invoked when we begin parsing a new 955 /// translation unit involving this external AST source. 956 /// 957 /// This function will provide all of the external definitions to 958 /// the ASTConsumer. 959 virtual void StartTranslationUnit(ASTConsumer *Consumer); 960 961 /// \brief Print some statistics about AST usage. 962 virtual void PrintStats(); 963 964 /// \brief Initialize the semantic source with the Sema instance 965 /// being used to perform semantic analysis on the abstract syntax 966 /// tree. 967 virtual void InitializeSema(Sema &S); 968 969 /// \brief Inform the semantic consumer that Sema is no longer available. 970 virtual void ForgetSema() { SemaObj = 0; } 971 972 /// \brief Retrieve the IdentifierInfo for the named identifier. 973 /// 974 /// This routine builds a new IdentifierInfo for the given identifier. If any 975 /// declarations with this name are visible from translation unit scope, their 976 /// declarations will be deserialized and introduced into the declaration 977 /// chain of the identifier. 978 virtual IdentifierInfo *get(const char *NameStart, const char *NameEnd); 979 IdentifierInfo *get(llvm::StringRef Name) { 980 return get(Name.begin(), Name.end()); 981 } 982 983 /// \brief Retrieve an iterator into the set of all identifiers 984 /// in all loaded AST files. 985 virtual IdentifierIterator *getIdentifiers() const; 986 987 /// \brief Load the contents of the global method pool for a given 988 /// selector. 989 /// 990 /// \returns a pair of Objective-C methods lists containing the 991 /// instance and factory methods, respectively, with this selector. 992 virtual std::pair<ObjCMethodList, ObjCMethodList> 993 ReadMethodPool(Selector Sel); 994 995 /// \brief Load a selector from disk, registering its ID if it exists. 996 void LoadSelector(Selector Sel); 997 998 void SetIdentifierInfo(unsigned ID, IdentifierInfo *II); 999 void SetGloballyVisibleDecls(IdentifierInfo *II, 1000 const llvm::SmallVectorImpl<uint32_t> &DeclIDs, 1001 bool Nonrecursive = false); 1002 1003 /// \brief Report a diagnostic. 1004 DiagnosticBuilder Diag(unsigned DiagID); 1005 1006 /// \brief Report a diagnostic. 1007 DiagnosticBuilder Diag(SourceLocation Loc, unsigned DiagID); 1008 1009 IdentifierInfo *DecodeIdentifierInfo(unsigned Idx); 1010 1011 IdentifierInfo *GetIdentifierInfo(const RecordData &Record, unsigned &Idx) { 1012 return DecodeIdentifierInfo(Record[Idx++]); 1013 } 1014 1015 virtual IdentifierInfo *GetIdentifier(unsigned ID) { 1016 return DecodeIdentifierInfo(ID); 1017 } 1018 1019 /// \brief Read the source location entry with index ID. 1020 virtual void ReadSLocEntry(unsigned ID); 1021 1022 Selector DecodeSelector(unsigned Idx); 1023 1024 virtual Selector GetExternalSelector(uint32_t ID); 1025 uint32_t GetNumExternalSelectors(); 1026 1027 Selector GetSelector(const RecordData &Record, unsigned &Idx) { 1028 return DecodeSelector(Record[Idx++]); 1029 } 1030 1031 /// \brief Read a declaration name. 1032 DeclarationName ReadDeclarationName(const RecordData &Record, unsigned &Idx); 1033 void ReadDeclarationNameLoc(PerFileData &F, 1034 DeclarationNameLoc &DNLoc, DeclarationName Name, 1035 const RecordData &Record, unsigned &Idx); 1036 void ReadDeclarationNameInfo(PerFileData &F, DeclarationNameInfo &NameInfo, 1037 const RecordData &Record, unsigned &Idx); 1038 1039 void ReadQualifierInfo(PerFileData &F, QualifierInfo &Info, 1040 const RecordData &Record, unsigned &Idx); 1041 1042 NestedNameSpecifier *ReadNestedNameSpecifier(const RecordData &Record, 1043 unsigned &Idx); 1044 1045 /// \brief Read a template name. 1046 TemplateName ReadTemplateName(const RecordData &Record, unsigned &Idx); 1047 1048 /// \brief Read a template argument. 1049 TemplateArgument ReadTemplateArgument(PerFileData &F, 1050 const RecordData &Record,unsigned &Idx); 1051 1052 /// \brief Read a template parameter list. 1053 TemplateParameterList *ReadTemplateParameterList(PerFileData &F, 1054 const RecordData &Record, 1055 unsigned &Idx); 1056 1057 /// \brief Read a template argument array. 1058 void 1059 ReadTemplateArgumentList(llvm::SmallVector<TemplateArgument, 8> &TemplArgs, 1060 PerFileData &F, const RecordData &Record, 1061 unsigned &Idx); 1062 1063 /// \brief Read a UnresolvedSet structure. 1064 void ReadUnresolvedSet(UnresolvedSetImpl &Set, 1065 const RecordData &Record, unsigned &Idx); 1066 1067 /// \brief Read a C++ base specifier. 1068 CXXBaseSpecifier ReadCXXBaseSpecifier(PerFileData &F, 1069 const RecordData &Record,unsigned &Idx); 1070 1071 /// \brief Read a CXXBaseOrMemberInitializer array. 1072 std::pair<CXXBaseOrMemberInitializer **, unsigned> 1073 ReadCXXBaseOrMemberInitializers(PerFileData &F, 1074 const RecordData &Record, unsigned &Idx); 1075 1076 /// \brief Read a source location from raw form. 1077 SourceLocation ReadSourceLocation(PerFileData &Module, unsigned Raw) { 1078 (void)Module; // No remapping yet 1079 return SourceLocation::getFromRawEncoding(Raw); 1080 } 1081 1082 /// \brief Read a source location. 1083 SourceLocation ReadSourceLocation(PerFileData &Module, 1084 const RecordData &Record, unsigned& Idx) { 1085 return ReadSourceLocation(Module, Record[Idx++]); 1086 } 1087 1088 /// \brief Read a source range. 1089 SourceRange ReadSourceRange(PerFileData &F, 1090 const RecordData &Record, unsigned& Idx); 1091 1092 /// \brief Read an integral value 1093 llvm::APInt ReadAPInt(const RecordData &Record, unsigned &Idx); 1094 1095 /// \brief Read a signed integral value 1096 llvm::APSInt ReadAPSInt(const RecordData &Record, unsigned &Idx); 1097 1098 /// \brief Read a floating-point value 1099 llvm::APFloat ReadAPFloat(const RecordData &Record, unsigned &Idx); 1100 1101 // \brief Read a string 1102 std::string ReadString(const RecordData &Record, unsigned &Idx); 1103 1104 CXXTemporary *ReadCXXTemporary(const RecordData &Record, unsigned &Idx); 1105 1106 /// \brief Reads attributes from the current stream position. 1107 void ReadAttributes(PerFileData &F, AttrVec &Attrs, 1108 const RecordData &Record, unsigned &Idx); 1109 1110 /// \brief Reads a statement. 1111 Stmt *ReadStmt(PerFileData &F); 1112 1113 /// \brief Reads an expression. 1114 Expr *ReadExpr(PerFileData &F); 1115 1116 /// \brief Reads a sub-statement operand during statement reading. 1117 Stmt *ReadSubStmt() { 1118 assert(ReadingKind == Read_Stmt && 1119 "Should be called only during statement reading!"); 1120 // Subexpressions are stored from last to first, so the next Stmt we need 1121 // is at the back of the stack. 1122 assert(!StmtStack.empty() && "Read too many sub statements!"); 1123 return StmtStack.pop_back_val(); 1124 } 1125 1126 /// \brief Reads a sub-expression operand during statement reading. 1127 Expr *ReadSubExpr(); 1128 1129 /// \brief Reads the macro record located at the given offset. 1130 void ReadMacroRecord(PerFileData &F, uint64_t Offset); 1131 1132 /// \brief Read the set of macros defined by this external macro source. 1133 virtual void ReadDefinedMacros(); 1134 1135 /// \brief Retrieve the macro definition with the given ID. 1136 MacroDefinition *getMacroDefinition(serialization::MacroID ID); 1137 1138 /// \brief Retrieve the AST context that this AST reader supplements. 1139 ASTContext *getContext() { return Context; } 1140 1141 // \brief Contains declarations that were loaded before we have 1142 // access to a Sema object. 1143 llvm::SmallVector<NamedDecl *, 16> PreloadedDecls; 1144 1145 /// \brief Retrieve the semantic analysis object used to analyze the 1146 /// translation unit in which the precompiled header is being 1147 /// imported. 1148 Sema *getSema() { return SemaObj; } 1149 1150 /// \brief Retrieve the identifier table associated with the 1151 /// preprocessor. 1152 IdentifierTable &getIdentifierTable(); 1153 1154 /// \brief Record that the given ID maps to the given switch-case 1155 /// statement. 1156 void RecordSwitchCaseID(SwitchCase *SC, unsigned ID); 1157 1158 /// \brief Retrieve the switch-case statement with the given ID. 1159 SwitchCase *getSwitchCaseWithID(unsigned ID); 1160 1161 /// \brief Record that the given label statement has been 1162 /// deserialized and has the given ID. 1163 void RecordLabelStmt(LabelStmt *S, unsigned ID); 1164 1165 /// \brief Set the label of the given statement to the label 1166 /// identified by ID. 1167 /// 1168 /// Depending on the order in which the label and other statements 1169 /// referencing that label occur, this operation may complete 1170 /// immediately (updating the statement) or it may queue the 1171 /// statement to be back-patched later. 1172 void SetLabelOf(GotoStmt *S, unsigned ID); 1173 1174 /// \brief Set the label of the given expression to the label 1175 /// identified by ID. 1176 /// 1177 /// Depending on the order in which the label and other statements 1178 /// referencing that label occur, this operation may complete 1179 /// immediately (updating the statement) or it may queue the 1180 /// statement to be back-patched later. 1181 void SetLabelOf(AddrLabelExpr *S, unsigned ID); 1182}; 1183 1184/// \brief Helper class that saves the current stream position and 1185/// then restores it when destroyed. 1186struct SavedStreamPosition { 1187 explicit SavedStreamPosition(llvm::BitstreamCursor &Cursor) 1188 : Cursor(Cursor), Offset(Cursor.GetCurrentBitNo()) { } 1189 1190 ~SavedStreamPosition() { 1191 Cursor.JumpToBit(Offset); 1192 } 1193 1194private: 1195 llvm::BitstreamCursor &Cursor; 1196 uint64_t Offset; 1197}; 1198 1199inline void PCHValidator::Error(const char *Msg) { 1200 Reader.Error(Msg); 1201} 1202 1203} // end namespace clang 1204 1205#endif 1206