ASTReader.h revision 95f4292cc526c629fead321c7fcfd4fe0f3bc66e
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 number of macro definitions in this file. 281 unsigned LocalNumMacroDefinitions; 282 283 /// \brief Offsets of all of the macro definitions in the preprocessing 284 /// record in the AST file. 285 const uint32_t *MacroDefinitionOffsets; 286 287 // === Selectors === 288 289 /// \brief The number of selectors new to this file. 290 /// 291 /// This is the number of entries in SelectorOffsets. 292 unsigned LocalNumSelectors; 293 294 /// \brief Offsets into the selector lookup table's data array 295 /// where each selector resides. 296 const uint32_t *SelectorOffsets; 297 298 /// \brief A pointer to the character data that comprises the selector table 299 /// 300 /// The SelectorOffsets table refers into this memory. 301 const unsigned char *SelectorLookupTableData; 302 303 /// \brief A pointer to an on-disk hash table of opaque type 304 /// ASTSelectorLookupTable. 305 /// 306 /// This hash table provides the IDs of all selectors, and the associated 307 /// instance and factory methods. 308 void *SelectorLookupTable; 309 310 /// \brief Method selectors used in a @selector expression. Used for 311 /// implementation of -Wselector. 312 llvm::SmallVector<uint64_t, 64> ReferencedSelectorsData; 313 314 // === Declarations === 315 316 /// DeclsCursor - This is a cursor to the start of the DECLS_BLOCK block. It 317 /// has read all the abbreviations at the start of the block and is ready to 318 /// jump around with these in context. 319 llvm::BitstreamCursor DeclsCursor; 320 321 /// \brief The number of declarations in this AST file. 322 unsigned LocalNumDecls; 323 324 /// \brief Offset of each declaration within the bitstream, indexed 325 /// by the declaration ID (-1). 326 const uint32_t *DeclOffsets; 327 328 /// \brief A snapshot of the pending instantiations in the chain. 329 /// 330 /// This record tracks the instantiations that Sema has to perform at the 331 /// end of the TU. It consists of a pair of values for every pending 332 /// instantiation where the first value is the ID of the decl and the second 333 /// is the instantiation location. 334 llvm::SmallVector<uint64_t, 64> PendingInstantiations; 335 336 // === Types === 337 338 /// \brief The number of types in this AST file. 339 unsigned LocalNumTypes; 340 341 /// \brief Offset of each type within the bitstream, indexed by the 342 /// type ID, or the representation of a Type*. 343 const uint32_t *TypeOffsets; 344 345 // === Miscellaneous === 346 347 /// \brief The AST stat cache installed for this file, if any. 348 /// 349 /// The dynamic type of this stat cache is always ASTStatCache 350 void *StatCache; 351 352 /// \brief The number of preallocated preprocessing entities in the 353 /// preprocessing record. 354 unsigned NumPreallocatedPreprocessingEntities; 355 356 /// \brief The next module in source order. 357 PerFileData *NextInSource; 358 359 /// \brief All the modules that loaded this one. Can contain NULL for 360 /// directly loaded modules. 361 llvm::SmallVector<PerFileData *, 1> Loaders; 362 }; 363 364 /// \brief All loaded modules, indexed by name. 365 llvm::StringMap<PerFileData*> Modules; 366 367 /// \brief The first module in source order. 368 PerFileData *FirstInSource; 369 370 /// \brief The chain of AST files. The first entry is the one named by the 371 /// user, the last one is the one that doesn't depend on anything further. 372 /// That is, the entry I was created with -include-pch I+1. 373 llvm::SmallVector<PerFileData*, 2> Chain; 374 375 /// \brief SLocEntries that we're going to preload. 376 llvm::SmallVector<uint64_t, 64> PreloadSLocEntries; 377 378 /// \brief Types that have already been loaded from the chain. 379 /// 380 /// When the pointer at index I is non-NULL, the type with 381 /// ID = (I + 1) << FastQual::Width has already been loaded 382 std::vector<QualType> TypesLoaded; 383 384 /// \brief Map that provides the ID numbers of each type within the 385 /// output stream, plus those deserialized from a chained PCH. 386 /// 387 /// The ID numbers of types are consecutive (in order of discovery) 388 /// and start at 1. 0 is reserved for NULL. When types are actually 389 /// stored in the stream, the ID number is shifted by 2 bits to 390 /// allow for the const/volatile qualifiers. 391 /// 392 /// Keys in the map never have const/volatile qualifiers. 393 serialization::TypeIdxMap TypeIdxs; 394 395 /// \brief Declarations that have already been loaded from the chain. 396 /// 397 /// When the pointer at index I is non-NULL, the declaration with ID 398 /// = I + 1 has already been loaded. 399 std::vector<Decl *> DeclsLoaded; 400 401 typedef llvm::DenseMap<serialization::DeclID, 402 std::pair<PerFileData *, uint64_t> > 403 DeclReplacementMap; 404 /// \brief Declarations that have been replaced in a later file in the chain. 405 DeclReplacementMap ReplacedDecls; 406 407 /// \brief Information about the contents of a DeclContext. 408 struct DeclContextInfo { 409 void *NameLookupTableData; // a ASTDeclContextNameLookupTable. 410 const serialization::KindDeclIDPair *LexicalDecls; 411 unsigned NumLexicalDecls; 412 }; 413 // In a full chain, there could be multiple updates to every decl context, 414 // so this is a vector. However, typically a chain is only two elements long, 415 // with only one file containing updates, so there will be only one update 416 // per decl context. 417 typedef llvm::SmallVector<DeclContextInfo, 1> DeclContextInfos; 418 typedef llvm::DenseMap<const DeclContext *, DeclContextInfos> 419 DeclContextOffsetsMap; 420 // Updates for visible decls can occur for other contexts than just the 421 // TU, and when we read those update records, the actual context will not 422 // be available yet (unless it's the TU), so have this pending map using the 423 // ID as a key. It will be realized when the context is actually loaded. 424 typedef llvm::SmallVector<void *, 1> DeclContextVisibleUpdates; 425 typedef llvm::DenseMap<serialization::DeclID, DeclContextVisibleUpdates> 426 DeclContextVisibleUpdatesPending; 427 428 /// \brief Offsets of the lexical and visible declarations for each 429 /// DeclContext. 430 DeclContextOffsetsMap DeclContextOffsets; 431 432 /// \brief Updates to the visible declarations of declaration contexts that 433 /// haven't been loaded yet. 434 DeclContextVisibleUpdatesPending PendingVisibleUpdates; 435 436 typedef llvm::DenseMap<serialization::DeclID, serialization::DeclID> 437 FirstLatestDeclIDMap; 438 /// \brief Map of first declarations from a chained PCH that point to the 439 /// most recent declarations in another AST file. 440 FirstLatestDeclIDMap FirstLatestDeclIDs; 441 442 typedef llvm::SmallVector<serialization::DeclID, 4> 443 AdditionalTemplateSpecializations; 444 typedef llvm::DenseMap<serialization::DeclID, 445 AdditionalTemplateSpecializations> 446 AdditionalTemplateSpecializationsMap; 447 448 /// \brief Additional specializations (including partial) of templates that 449 /// were introduced after the template was serialized. 450 AdditionalTemplateSpecializationsMap AdditionalTemplateSpecializationsPending; 451 452 /// \brief Read the records that describe the contents of declcontexts. 453 bool ReadDeclContextStorage(llvm::BitstreamCursor &Cursor, 454 const std::pair<uint64_t, uint64_t> &Offsets, 455 DeclContextInfo &Info); 456 457 /// \brief A vector containing identifiers that have already been 458 /// loaded. 459 /// 460 /// If the pointer at index I is non-NULL, then it refers to the 461 /// IdentifierInfo for the identifier with ID=I+1 that has already 462 /// been loaded. 463 std::vector<IdentifierInfo *> IdentifiersLoaded; 464 465 /// \brief A vector containing selectors that have already been loaded. 466 /// 467 /// This vector is indexed by the Selector ID (-1). NULL selector 468 /// entries indicate that the particular selector ID has not yet 469 /// been loaded. 470 llvm::SmallVector<Selector, 16> SelectorsLoaded; 471 472 /// \brief The macro definitions we have already loaded. 473 llvm::SmallVector<MacroDefinition *, 16> MacroDefinitionsLoaded; 474 475 /// \name CodeGen-relevant special data 476 /// \brief Fields containing data that is relevant to CodeGen. 477 //@{ 478 479 /// \brief The IDs of all declarations that fulfill the criteria of 480 /// "interesting" decls. 481 /// 482 /// This contains the data loaded from all EXTERNAL_DEFINITIONS blocks in the 483 /// chain. The referenced declarations are deserialized and passed to the 484 /// consumer eagerly. 485 llvm::SmallVector<uint64_t, 16> ExternalDefinitions; 486 487 /// \brief The IDs of all tentative definitions stored in the the chain. 488 /// 489 /// Sema keeps track of all tentative definitions in a TU because it has to 490 /// complete them and pass them on to CodeGen. Thus, tentative definitions in 491 /// the PCH chain must be eagerly deserialized. 492 llvm::SmallVector<uint64_t, 16> TentativeDefinitions; 493 494 /// \brief The IDs of all CXXRecordDecls stored in the chain whose VTables are 495 /// used. 496 /// 497 /// CodeGen has to emit VTables for these records, so they have to be eagerly 498 /// deserialized. 499 llvm::SmallVector<uint64_t, 64> VTableUses; 500 501 //@} 502 503 /// \name Diagnostic-relevant special data 504 /// \brief Fields containing data that is used for generating diagnostics 505 //@{ 506 507 /// \brief A snapshot of Sema's unused file-scoped variable tracking, for 508 /// generating warnings. 509 llvm::SmallVector<uint64_t, 16> UnusedFileScopedDecls; 510 511 /// \brief A snapshot of Sema's weak undeclared identifier tracking, for 512 /// generating warnings. 513 llvm::SmallVector<uint64_t, 64> WeakUndeclaredIdentifiers; 514 515 /// \brief The IDs of type aliases for ext_vectors that exist in the chain. 516 /// 517 /// Used by Sema for finding sugared names for ext_vectors in diagnostics. 518 llvm::SmallVector<uint64_t, 4> ExtVectorDecls; 519 520 //@} 521 522 /// \name Sema-relevant special data 523 /// \brief Fields containing data that is used for semantic analysis 524 //@{ 525 526 /// \brief The IDs of all locally scoped external decls in the chain. 527 /// 528 /// Sema tracks these to validate that the types are consistent across all 529 /// local external declarations. 530 llvm::SmallVector<uint64_t, 16> LocallyScopedExternalDecls; 531 532 /// \brief The IDs of all dynamic class declarations in the chain. 533 /// 534 /// Sema tracks these because it checks for the key functions being defined 535 /// at the end of the TU, in which case it directs CodeGen to emit the VTable. 536 llvm::SmallVector<uint64_t, 16> DynamicClasses; 537 538 /// \brief The IDs of the declarations Sema stores directly. 539 /// 540 /// Sema tracks a few important decls, such as namespace std, directly. 541 llvm::SmallVector<uint64_t, 4> SemaDeclRefs; 542 543 /// \brief The IDs of the types ASTContext stores directly. 544 /// 545 /// The AST context tracks a few important types, such as va_list, directly. 546 llvm::SmallVector<uint64_t, 16> SpecialTypes; 547 548 //@} 549 550 /// \brief The original file name that was used to build the primary AST file, 551 /// which may have been modified for relocatable-pch support. 552 std::string OriginalFileName; 553 554 /// \brief The actual original file name that was used to build the primary 555 /// AST file. 556 std::string ActualOriginalFileName; 557 558 /// \brief Whether this precompiled header is a relocatable PCH file. 559 bool RelocatablePCH; 560 561 /// \brief The system include root to be used when loading the 562 /// precompiled header. 563 const char *isysroot; 564 565 /// \brief Whether to disable the normal validation performed on precompiled 566 /// headers when they are loaded. 567 bool DisableValidation; 568 569 /// \brief Mapping from switch-case IDs in the chain to switch-case statements 570 /// 571 /// Statements usually don't have IDs, but switch cases need them, so that the 572 /// switch statement can refer to them. 573 std::map<unsigned, SwitchCase *> SwitchCaseStmts; 574 575 /// \brief Mapping from label statement IDs in the chain to label statements. 576 /// 577 /// Statements usually don't have IDs, but labeled statements need them, so 578 /// that goto statements and address-of-label expressions can refer to them. 579 std::map<unsigned, LabelStmt *> LabelStmts; 580 581 /// \brief Mapping from label IDs to the set of "goto" statements 582 /// that point to that label before the label itself has been 583 /// de-serialized. 584 std::multimap<unsigned, GotoStmt *> UnresolvedGotoStmts; 585 586 /// \brief Mapping from label IDs to the set of address label 587 /// expressions that point to that label before the label itself has 588 /// been de-serialized. 589 std::multimap<unsigned, AddrLabelExpr *> UnresolvedAddrLabelExprs; 590 591 /// \brief The number of stat() calls that hit/missed the stat 592 /// cache. 593 unsigned NumStatHits, NumStatMisses; 594 595 /// \brief The number of source location entries de-serialized from 596 /// the PCH file. 597 unsigned NumSLocEntriesRead; 598 599 /// \brief The number of source location entries in the chain. 600 unsigned TotalNumSLocEntries; 601 602 /// \brief The next offset for a SLocEntry after everything in this reader. 603 unsigned NextSLocOffset; 604 605 /// \brief The number of statements (and expressions) de-serialized 606 /// from the chain. 607 unsigned NumStatementsRead; 608 609 /// \brief The total number of statements (and expressions) stored 610 /// in the chain. 611 unsigned TotalNumStatements; 612 613 /// \brief The number of macros de-serialized from the chain. 614 unsigned NumMacrosRead; 615 616 /// \brief The total number of macros stored in the chain. 617 unsigned TotalNumMacros; 618 619 /// \brief The number of selectors that have been read. 620 unsigned NumSelectorsRead; 621 622 /// \brief The number of method pool entries that have been read. 623 unsigned NumMethodPoolEntriesRead; 624 625 /// \brief The number of times we have looked up a selector in the method 626 /// pool and not found anything interesting. 627 unsigned NumMethodPoolMisses; 628 629 /// \brief The total number of method pool entries in the selector table. 630 unsigned TotalNumMethodPoolEntries; 631 632 /// Number of lexical decl contexts read/total. 633 unsigned NumLexicalDeclContextsRead, TotalLexicalDeclContexts; 634 635 /// Number of visible decl contexts read/total. 636 unsigned NumVisibleDeclContextsRead, TotalVisibleDeclContexts; 637 638 /// \brief Number of Decl/types that are currently deserializing. 639 unsigned NumCurrentElementsDeserializing; 640 641 /// \brief An IdentifierInfo that has been loaded but whose top-level 642 /// declarations of the same name have not (yet) been loaded. 643 struct PendingIdentifierInfo { 644 IdentifierInfo *II; 645 llvm::SmallVector<uint32_t, 4> DeclIDs; 646 }; 647 648 /// \brief The set of identifiers that were read while the AST reader was 649 /// (recursively) loading declarations. 650 /// 651 /// The declarations on the identifier chain for these identifiers will be 652 /// loaded once the recursive loading has completed. 653 std::deque<PendingIdentifierInfo> PendingIdentifierInfos; 654 655 /// \brief Contains declarations and definitions that will be 656 /// "interesting" to the ASTConsumer, when we get that AST consumer. 657 /// 658 /// "Interesting" declarations are those that have data that may 659 /// need to be emitted, such as inline function definitions or 660 /// Objective-C protocols. 661 std::deque<Decl *> InterestingDecls; 662 663 /// \brief When reading a Stmt tree, Stmt operands are placed in this stack. 664 llvm::SmallVector<Stmt *, 16> StmtStack; 665 666 /// \brief What kind of records we are reading. 667 enum ReadingKind { 668 Read_Decl, Read_Type, Read_Stmt 669 }; 670 671 /// \brief What kind of records we are reading. 672 ReadingKind ReadingKind; 673 674 /// \brief RAII object to change the reading kind. 675 class ReadingKindTracker { 676 ASTReader &Reader; 677 enum ReadingKind PrevKind; 678 679 ReadingKindTracker(const ReadingKindTracker&); // do not implement 680 ReadingKindTracker &operator=(const ReadingKindTracker&);// do not implement 681 682 public: 683 ReadingKindTracker(enum ReadingKind newKind, ASTReader &reader) 684 : Reader(reader), PrevKind(Reader.ReadingKind) { 685 Reader.ReadingKind = newKind; 686 } 687 688 ~ReadingKindTracker() { Reader.ReadingKind = PrevKind; } 689 }; 690 691 /// \brief All predefines buffers in the chain, to be treated as if 692 /// concatenated. 693 PCHPredefinesBlocks PCHPredefinesBuffers; 694 695 /// \brief Suggested contents of the predefines buffer, after this 696 /// PCH file has been processed. 697 /// 698 /// In most cases, this string will be empty, because the predefines 699 /// buffer computed to build the PCH file will be identical to the 700 /// predefines buffer computed from the command line. However, when 701 /// there are differences that the PCH reader can work around, this 702 /// predefines buffer may contain additional definitions. 703 std::string SuggestedPredefines; 704 705 /// \brief Reads a statement from the specified cursor. 706 Stmt *ReadStmtFromStream(PerFileData &F); 707 708 void MaybeAddSystemRootToFilename(std::string &Filename); 709 710 ASTReadResult ReadASTCore(llvm::StringRef FileName, ASTFileType Type); 711 ASTReadResult ReadASTBlock(PerFileData &F); 712 bool CheckPredefinesBuffers(); 713 bool ParseLineTable(PerFileData &F, llvm::SmallVectorImpl<uint64_t> &Record); 714 ASTReadResult ReadSourceManagerBlock(PerFileData &F); 715 ASTReadResult ReadSLocEntryRecord(unsigned ID); 716 PerFileData *SLocCursorForID(unsigned ID); 717 SourceLocation getImportLocation(PerFileData *F); 718 bool ParseLanguageOptions(const llvm::SmallVectorImpl<uint64_t> &Record); 719 720 struct RecordLocation { 721 RecordLocation(PerFileData *M, uint64_t O) 722 : F(M), Offset(O) {} 723 PerFileData *F; 724 uint64_t Offset; 725 }; 726 727 QualType ReadTypeRecord(unsigned Index); 728 RecordLocation TypeCursorForIndex(unsigned Index); 729 void LoadedDecl(unsigned Index, Decl *D); 730 Decl *ReadDeclRecord(unsigned Index, serialization::DeclID ID); 731 RecordLocation DeclCursorForIndex(unsigned Index, serialization::DeclID ID); 732 733 void PassInterestingDeclsToConsumer(); 734 735 /// \brief Produce an error diagnostic and return true. 736 /// 737 /// This routine should only be used for fatal errors that have to 738 /// do with non-routine failures (e.g., corrupted AST file). 739 void Error(const char *Msg); 740 741 ASTReader(const ASTReader&); // do not implement 742 ASTReader &operator=(const ASTReader &); // do not implement 743public: 744 typedef llvm::SmallVector<uint64_t, 64> RecordData; 745 746 /// \brief Load the AST file and validate its contents against the given 747 /// Preprocessor. 748 /// 749 /// \param PP the preprocessor associated with the context in which this 750 /// precompiled header will be loaded. 751 /// 752 /// \param Context the AST context that this precompiled header will be 753 /// loaded into. 754 /// 755 /// \param isysroot If non-NULL, the system include path specified by the 756 /// user. This is only used with relocatable PCH files. If non-NULL, 757 /// a relocatable PCH file will use the default path "/". 758 /// 759 /// \param DisableValidation If true, the AST reader will suppress most 760 /// of its regular consistency checking, allowing the use of precompiled 761 /// headers that cannot be determined to be compatible. 762 ASTReader(Preprocessor &PP, ASTContext *Context, const char *isysroot = 0, 763 bool DisableValidation = false); 764 765 /// \brief Load the AST file without using any pre-initialized Preprocessor. 766 /// 767 /// The necessary information to initialize a Preprocessor later can be 768 /// obtained by setting a ASTReaderListener. 769 /// 770 /// \param SourceMgr the source manager into which the AST file will be loaded 771 /// 772 /// \param FileMgr the file manager into which the AST file will be loaded. 773 /// 774 /// \param Diags the diagnostics system to use for reporting errors and 775 /// warnings relevant to loading the AST file. 776 /// 777 /// \param isysroot If non-NULL, the system include path specified by the 778 /// user. This is only used with relocatable PCH files. If non-NULL, 779 /// a relocatable PCH file will use the default path "/". 780 /// 781 /// \param DisableValidation If true, the AST reader will suppress most 782 /// of its regular consistency checking, allowing the use of precompiled 783 /// headers that cannot be determined to be compatible. 784 ASTReader(SourceManager &SourceMgr, FileManager &FileMgr, 785 Diagnostic &Diags, const char *isysroot = 0, 786 bool DisableValidation = false); 787 ~ASTReader(); 788 789 /// \brief Load the precompiled header designated by the given file 790 /// name. 791 ASTReadResult ReadAST(const std::string &FileName, ASTFileType Type); 792 793 /// \brief Set the AST callbacks listener. 794 void setListener(ASTReaderListener *listener) { 795 Listener.reset(listener); 796 } 797 798 /// \brief Set the AST deserialization listener. 799 void setDeserializationListener(ASTDeserializationListener *Listener); 800 801 /// \brief Set the Preprocessor to use. 802 void setPreprocessor(Preprocessor &pp); 803 804 /// \brief Sets and initializes the given Context. 805 void InitializeContext(ASTContext &Context); 806 807 /// \brief Retrieve the name of the named (primary) AST file 808 const std::string &getFileName() const { return Chain[0]->FileName; } 809 810 /// \brief Retrieve the name of the original source file name 811 const std::string &getOriginalSourceFile() { return OriginalFileName; } 812 813 /// \brief Retrieve the name of the original source file name directly from 814 /// the AST file, without actually loading the AST file. 815 static std::string getOriginalSourceFile(const std::string &ASTFileName, 816 Diagnostic &Diags); 817 818 /// \brief Returns the suggested contents of the predefines buffer, 819 /// which contains a (typically-empty) subset of the predefines 820 /// build prior to including the precompiled header. 821 const std::string &getSuggestedPredefines() { return SuggestedPredefines; } 822 823 /// \brief Read preprocessed entities into the 824 virtual void ReadPreprocessedEntities(); 825 826 /// \brief Returns the number of source locations found in the chain. 827 unsigned getTotalNumSLocs() const { 828 return TotalNumSLocEntries; 829 } 830 831 /// \brief Returns the next SLocEntry offset after the chain. 832 unsigned getNextSLocOffset() const { 833 return NextSLocOffset; 834 } 835 836 /// \brief Returns the number of identifiers found in the chain. 837 unsigned getTotalNumIdentifiers() const { 838 return static_cast<unsigned>(IdentifiersLoaded.size()); 839 } 840 841 /// \brief Returns the number of types found in the chain. 842 unsigned getTotalNumTypes() const { 843 return static_cast<unsigned>(TypesLoaded.size()); 844 } 845 846 /// \brief Returns the number of declarations found in the chain. 847 unsigned getTotalNumDecls() const { 848 return static_cast<unsigned>(DeclsLoaded.size()); 849 } 850 851 /// \brief Returns the number of selectors found in the chain. 852 unsigned getTotalNumSelectors() const { 853 return static_cast<unsigned>(SelectorsLoaded.size()); 854 } 855 856 /// \brief Returns the number of macro definitions found in the chain. 857 unsigned getTotalNumMacroDefinitions() const { 858 return static_cast<unsigned>(MacroDefinitionsLoaded.size()); 859 } 860 861 /// \brief Reads a TemplateArgumentLocInfo appropriate for the 862 /// given TemplateArgument kind. 863 TemplateArgumentLocInfo 864 GetTemplateArgumentLocInfo(PerFileData &F, TemplateArgument::ArgKind Kind, 865 const RecordData &Record, unsigned &Idx); 866 867 /// \brief Reads a TemplateArgumentLoc. 868 TemplateArgumentLoc 869 ReadTemplateArgumentLoc(PerFileData &F, 870 const RecordData &Record, unsigned &Idx); 871 872 /// \brief Reads a declarator info from the given record. 873 TypeSourceInfo *GetTypeSourceInfo(PerFileData &F, 874 const RecordData &Record, unsigned &Idx); 875 876 /// \brief Resolve and return the translation unit declaration. 877 TranslationUnitDecl *GetTranslationUnitDecl(); 878 879 /// \brief Resolve a type ID into a type, potentially building a new 880 /// type. 881 QualType GetType(serialization::TypeID ID); 882 883 /// \brief Returns the type ID associated with the given type. 884 /// If the type didn't come from the AST file the ID that is returned is 885 /// marked as "doesn't exist in AST". 886 serialization::TypeID GetTypeID(QualType T) const; 887 888 /// \brief Returns the type index associated with the given type. 889 /// If the type didn't come from the AST file the index that is returned is 890 /// marked as "doesn't exist in AST". 891 serialization::TypeIdx GetTypeIdx(QualType T) const; 892 893 /// \brief Resolve a declaration ID into a declaration, potentially 894 /// building a new declaration. 895 Decl *GetDecl(serialization::DeclID ID); 896 virtual Decl *GetExternalDecl(uint32_t ID); 897 898 /// \brief Resolve the offset of a statement into a statement. 899 /// 900 /// This operation will read a new statement from the external 901 /// source each time it is called, and is meant to be used via a 902 /// LazyOffsetPtr (which is used by Decls for the body of functions, etc). 903 virtual Stmt *GetExternalDeclStmt(uint64_t Offset); 904 905 /// ReadBlockAbbrevs - Enter a subblock of the specified BlockID with the 906 /// specified cursor. Read the abbreviations that are at the top of the block 907 /// and then leave the cursor pointing into the block. 908 bool ReadBlockAbbrevs(llvm::BitstreamCursor &Cursor, unsigned BlockID); 909 910 /// \brief Finds all the visible declarations with a given name. 911 /// The current implementation of this method just loads the entire 912 /// lookup table as unmaterialized references. 913 virtual DeclContext::lookup_result 914 FindExternalVisibleDeclsByName(const DeclContext *DC, 915 DeclarationName Name); 916 917 virtual void MaterializeVisibleDecls(const DeclContext *DC); 918 919 /// \brief Read all of the declarations lexically stored in a 920 /// declaration context. 921 /// 922 /// \param DC The declaration context whose declarations will be 923 /// read. 924 /// 925 /// \param Decls Vector that will contain the declarations loaded 926 /// from the external source. The caller is responsible for merging 927 /// these declarations with any declarations already stored in the 928 /// declaration context. 929 /// 930 /// \returns true if there was an error while reading the 931 /// declarations for this declaration context. 932 virtual bool FindExternalLexicalDecls(const DeclContext *DC, 933 bool (*isKindWeWant)(Decl::Kind), 934 llvm::SmallVectorImpl<Decl*> &Decls); 935 936 /// \brief Notify ASTReader that we started deserialization of 937 /// a decl or type so until FinishedDeserializing is called there may be 938 /// decls that are initializing. Must be paired with FinishedDeserializing. 939 virtual void StartedDeserializing() { ++NumCurrentElementsDeserializing; } 940 941 /// \brief Notify ASTReader that we finished the deserialization of 942 /// a decl or type. Must be paired with StartedDeserializing. 943 virtual void FinishedDeserializing(); 944 945 /// \brief Function that will be invoked when we begin parsing a new 946 /// translation unit involving this external AST source. 947 /// 948 /// This function will provide all of the external definitions to 949 /// the ASTConsumer. 950 virtual void StartTranslationUnit(ASTConsumer *Consumer); 951 952 /// \brief Print some statistics about AST usage. 953 virtual void PrintStats(); 954 955 /// \brief Initialize the semantic source with the Sema instance 956 /// being used to perform semantic analysis on the abstract syntax 957 /// tree. 958 virtual void InitializeSema(Sema &S); 959 960 /// \brief Inform the semantic consumer that Sema is no longer available. 961 virtual void ForgetSema() { SemaObj = 0; } 962 963 /// \brief Retrieve the IdentifierInfo for the named identifier. 964 /// 965 /// This routine builds a new IdentifierInfo for the given identifier. If any 966 /// declarations with this name are visible from translation unit scope, their 967 /// declarations will be deserialized and introduced into the declaration 968 /// chain of the identifier. 969 virtual IdentifierInfo *get(const char *NameStart, const char *NameEnd); 970 IdentifierInfo *get(llvm::StringRef Name) { 971 return get(Name.begin(), Name.end()); 972 } 973 974 /// \brief Retrieve an iterator into the set of all identifiers 975 /// in all loaded AST files. 976 virtual IdentifierIterator *getIdentifiers() const; 977 978 /// \brief Load the contents of the global method pool for a given 979 /// selector. 980 /// 981 /// \returns a pair of Objective-C methods lists containing the 982 /// instance and factory methods, respectively, with this selector. 983 virtual std::pair<ObjCMethodList, ObjCMethodList> 984 ReadMethodPool(Selector Sel); 985 986 /// \brief Load a selector from disk, registering its ID if it exists. 987 void LoadSelector(Selector Sel); 988 989 void SetIdentifierInfo(unsigned ID, IdentifierInfo *II); 990 void SetGloballyVisibleDecls(IdentifierInfo *II, 991 const llvm::SmallVectorImpl<uint32_t> &DeclIDs, 992 bool Nonrecursive = false); 993 994 /// \brief Report a diagnostic. 995 DiagnosticBuilder Diag(unsigned DiagID); 996 997 /// \brief Report a diagnostic. 998 DiagnosticBuilder Diag(SourceLocation Loc, unsigned DiagID); 999 1000 IdentifierInfo *DecodeIdentifierInfo(unsigned Idx); 1001 1002 IdentifierInfo *GetIdentifierInfo(const RecordData &Record, unsigned &Idx) { 1003 return DecodeIdentifierInfo(Record[Idx++]); 1004 } 1005 1006 virtual IdentifierInfo *GetIdentifier(unsigned ID) { 1007 return DecodeIdentifierInfo(ID); 1008 } 1009 1010 /// \brief Read the source location entry with index ID. 1011 virtual void ReadSLocEntry(unsigned ID); 1012 1013 Selector DecodeSelector(unsigned Idx); 1014 1015 virtual Selector GetExternalSelector(uint32_t ID); 1016 uint32_t GetNumExternalSelectors(); 1017 1018 Selector GetSelector(const RecordData &Record, unsigned &Idx) { 1019 return DecodeSelector(Record[Idx++]); 1020 } 1021 1022 /// \brief Read a declaration name. 1023 DeclarationName ReadDeclarationName(const RecordData &Record, unsigned &Idx); 1024 1025 NestedNameSpecifier *ReadNestedNameSpecifier(const RecordData &Record, 1026 unsigned &Idx); 1027 1028 /// \brief Read a template name. 1029 TemplateName ReadTemplateName(const RecordData &Record, unsigned &Idx); 1030 1031 /// \brief Read a template argument. 1032 TemplateArgument ReadTemplateArgument(PerFileData &F, 1033 const RecordData &Record,unsigned &Idx); 1034 1035 /// \brief Read a template parameter list. 1036 TemplateParameterList *ReadTemplateParameterList(PerFileData &F, 1037 const RecordData &Record, 1038 unsigned &Idx); 1039 1040 /// \brief Read a template argument array. 1041 void 1042 ReadTemplateArgumentList(llvm::SmallVector<TemplateArgument, 8> &TemplArgs, 1043 PerFileData &F, const RecordData &Record, 1044 unsigned &Idx); 1045 1046 /// \brief Read a UnresolvedSet structure. 1047 void ReadUnresolvedSet(UnresolvedSetImpl &Set, 1048 const RecordData &Record, unsigned &Idx); 1049 1050 /// \brief Read a C++ base specifier. 1051 CXXBaseSpecifier ReadCXXBaseSpecifier(PerFileData &F, 1052 const RecordData &Record,unsigned &Idx); 1053 1054 /// \brief Read a CXXBaseOrMemberInitializer array. 1055 std::pair<CXXBaseOrMemberInitializer **, unsigned> 1056 ReadCXXBaseOrMemberInitializers(PerFileData &F, 1057 const RecordData &Record, unsigned &Idx); 1058 1059 /// \brief Read a source location from raw form. 1060 SourceLocation ReadSourceLocation(PerFileData &Module, unsigned Raw) { 1061 (void)Module; // No remapping yet 1062 return SourceLocation::getFromRawEncoding(Raw); 1063 } 1064 1065 /// \brief Read a source location. 1066 SourceLocation ReadSourceLocation(PerFileData &Module, 1067 const RecordData &Record, unsigned& Idx) { 1068 return ReadSourceLocation(Module, Record[Idx++]); 1069 } 1070 1071 /// \brief Read a source range. 1072 SourceRange ReadSourceRange(PerFileData &F, 1073 const RecordData &Record, unsigned& Idx); 1074 1075 /// \brief Read an integral value 1076 llvm::APInt ReadAPInt(const RecordData &Record, unsigned &Idx); 1077 1078 /// \brief Read a signed integral value 1079 llvm::APSInt ReadAPSInt(const RecordData &Record, unsigned &Idx); 1080 1081 /// \brief Read a floating-point value 1082 llvm::APFloat ReadAPFloat(const RecordData &Record, unsigned &Idx); 1083 1084 // \brief Read a string 1085 std::string ReadString(const RecordData &Record, unsigned &Idx); 1086 1087 CXXTemporary *ReadCXXTemporary(const RecordData &Record, unsigned &Idx); 1088 1089 /// \brief Reads attributes from the current stream position. 1090 void ReadAttributes(PerFileData &F, AttrVec &Attrs); 1091 1092 /// \brief Reads a statement. 1093 Stmt *ReadStmt(PerFileData &F); 1094 1095 /// \brief Reads an expression. 1096 Expr *ReadExpr(PerFileData &F); 1097 1098 /// \brief Reads a sub-statement operand during statement reading. 1099 Stmt *ReadSubStmt() { 1100 assert(ReadingKind == Read_Stmt && 1101 "Should be called only during statement reading!"); 1102 // Subexpressions are stored from last to first, so the next Stmt we need 1103 // is at the back of the stack. 1104 assert(!StmtStack.empty() && "Read too many sub statements!"); 1105 return StmtStack.pop_back_val(); 1106 } 1107 1108 /// \brief Reads a sub-expression operand during statement reading. 1109 Expr *ReadSubExpr(); 1110 1111 /// \brief Reads the macro record located at the given offset. 1112 void ReadMacroRecord(PerFileData &F, uint64_t Offset); 1113 1114 /// \brief Read the set of macros defined by this external macro source. 1115 virtual void ReadDefinedMacros(); 1116 1117 /// \brief Retrieve the macro definition with the given ID. 1118 MacroDefinition *getMacroDefinition(serialization::MacroID ID); 1119 1120 /// \brief Retrieve the AST context that this AST reader supplements. 1121 ASTContext *getContext() { return Context; } 1122 1123 // \brief Contains declarations that were loaded before we have 1124 // access to a Sema object. 1125 llvm::SmallVector<NamedDecl *, 16> PreloadedDecls; 1126 1127 /// \brief Retrieve the semantic analysis object used to analyze the 1128 /// translation unit in which the precompiled header is being 1129 /// imported. 1130 Sema *getSema() { return SemaObj; } 1131 1132 /// \brief Retrieve the identifier table associated with the 1133 /// preprocessor. 1134 IdentifierTable &getIdentifierTable(); 1135 1136 /// \brief Record that the given ID maps to the given switch-case 1137 /// statement. 1138 void RecordSwitchCaseID(SwitchCase *SC, unsigned ID); 1139 1140 /// \brief Retrieve the switch-case statement with the given ID. 1141 SwitchCase *getSwitchCaseWithID(unsigned ID); 1142 1143 /// \brief Record that the given label statement has been 1144 /// deserialized and has the given ID. 1145 void RecordLabelStmt(LabelStmt *S, unsigned ID); 1146 1147 /// \brief Set the label of the given statement to the label 1148 /// identified by ID. 1149 /// 1150 /// Depending on the order in which the label and other statements 1151 /// referencing that label occur, this operation may complete 1152 /// immediately (updating the statement) or it may queue the 1153 /// statement to be back-patched later. 1154 void SetLabelOf(GotoStmt *S, unsigned ID); 1155 1156 /// \brief Set the label of the given expression to the label 1157 /// identified by ID. 1158 /// 1159 /// Depending on the order in which the label and other statements 1160 /// referencing that label occur, this operation may complete 1161 /// immediately (updating the statement) or it may queue the 1162 /// statement to be back-patched later. 1163 void SetLabelOf(AddrLabelExpr *S, unsigned ID); 1164}; 1165 1166/// \brief Helper class that saves the current stream position and 1167/// then restores it when destroyed. 1168struct SavedStreamPosition { 1169 explicit SavedStreamPosition(llvm::BitstreamCursor &Cursor) 1170 : Cursor(Cursor), Offset(Cursor.GetCurrentBitNo()) { } 1171 1172 ~SavedStreamPosition() { 1173 Cursor.JumpToBit(Offset); 1174 } 1175 1176private: 1177 llvm::BitstreamCursor &Cursor; 1178 uint64_t Offset; 1179}; 1180 1181inline void PCHValidator::Error(const char *Msg) { 1182 Reader.Error(Msg); 1183} 1184 1185} // end namespace clang 1186 1187#endif 1188