ASTReader.h revision 1d9f1fe7173e3084325f43c78af812a36d8a2a7c
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 Attr; 51class Decl; 52class DeclContext; 53class NestedNameSpecifier; 54class CXXBaseSpecifier; 55class CXXBaseOrMemberInitializer; 56class GotoStmt; 57class LabelStmt; 58class MacroDefinition; 59class NamedDecl; 60class Preprocessor; 61class Sema; 62class SwitchCase; 63class ASTDeserializationListener; 64class ASTReader; 65class ASTDeclReader; 66class ASTStmtReader; 67class ASTIdentifierLookupTrait; 68class TypeLocReader; 69struct HeaderFileInfo; 70 71struct PCHPredefinesBlock { 72 /// \brief The file ID for this predefines buffer in a PCH file. 73 FileID BufferID; 74 75 /// \brief This predefines buffer in a PCH file. 76 llvm::StringRef Data; 77}; 78typedef llvm::SmallVector<PCHPredefinesBlock, 2> PCHPredefinesBlocks; 79 80/// \brief Abstract interface for callback invocations by the ASTReader. 81/// 82/// While reading an AST file, the ASTReader will call the methods of the 83/// listener to pass on specific information. Some of the listener methods can 84/// return true to indicate to the ASTReader that the information (and 85/// consequently the AST file) is invalid. 86class ASTReaderListener { 87public: 88 virtual ~ASTReaderListener(); 89 90 /// \brief Receives the language options. 91 /// 92 /// \returns true to indicate the options are invalid or false otherwise. 93 virtual bool ReadLanguageOptions(const LangOptions &LangOpts) { 94 return false; 95 } 96 97 /// \brief Receives the target triple. 98 /// 99 /// \returns true to indicate the target triple is invalid or false otherwise. 100 virtual bool ReadTargetTriple(llvm::StringRef Triple) { 101 return false; 102 } 103 104 /// \brief Receives the contents of the predefines buffer. 105 /// 106 /// \param Buffers Information about the predefines buffers. 107 /// 108 /// \param OriginalFileName The original file name for the AST file, which 109 /// will appear as an entry in the predefines buffer. 110 /// 111 /// \param SuggestedPredefines If necessary, additional definitions are added 112 /// here. 113 /// 114 /// \returns true to indicate the predefines are invalid or false otherwise. 115 virtual bool ReadPredefinesBuffer(const PCHPredefinesBlocks &Buffers, 116 llvm::StringRef OriginalFileName, 117 std::string &SuggestedPredefines) { 118 return false; 119 } 120 121 /// \brief Receives a HeaderFileInfo entry. 122 virtual void ReadHeaderFileInfo(const HeaderFileInfo &HFI, unsigned ID) {} 123 124 /// \brief Receives __COUNTER__ value. 125 virtual void ReadCounter(unsigned Value) {} 126}; 127 128/// \brief ASTReaderListener implementation to validate the information of 129/// the PCH file against an initialized Preprocessor. 130class PCHValidator : public ASTReaderListener { 131 Preprocessor &PP; 132 ASTReader &Reader; 133 134 unsigned NumHeaderInfos; 135 136public: 137 PCHValidator(Preprocessor &PP, ASTReader &Reader) 138 : PP(PP), Reader(Reader), NumHeaderInfos(0) {} 139 140 virtual bool ReadLanguageOptions(const LangOptions &LangOpts); 141 virtual bool ReadTargetTriple(llvm::StringRef Triple); 142 virtual bool ReadPredefinesBuffer(const PCHPredefinesBlocks &Buffers, 143 llvm::StringRef OriginalFileName, 144 std::string &SuggestedPredefines); 145 virtual void ReadHeaderFileInfo(const HeaderFileInfo &HFI, unsigned ID); 146 virtual void ReadCounter(unsigned Value); 147 148private: 149 void Error(const char *Msg); 150}; 151 152/// \brief Reads an AST files chain containing the contents of a translation 153/// unit. 154/// 155/// The ASTReader class reads bitstreams (produced by the ASTWriter 156/// class) containing the serialized representation of a given 157/// abstract syntax tree and its supporting data structures. An 158/// instance of the ASTReader can be attached to an ASTContext object, 159/// which will provide access to the contents of the AST files. 160/// 161/// The AST reader provides lazy de-serialization of declarations, as 162/// required when traversing the AST. Only those AST nodes that are 163/// actually required will be de-serialized. 164class ASTReader 165 : public ExternalPreprocessorSource, 166 public ExternalPreprocessingRecordSource, 167 public ExternalSemaSource, 168 public IdentifierInfoLookup, 169 public ExternalIdentifierLookup, 170 public ExternalSLocEntrySource { 171public: 172 enum ASTReadResult { Success, Failure, IgnorePCH }; 173 /// \brief Types of AST files. 174 enum ASTFileType { 175 Module, ///< File is a module proper. 176 PCH, ///< File is a PCH file treated as such. 177 Preamble, ///< File is a PCH file treated as the preamble. 178 MainFile ///< File is a PCH file treated as the actual main file. 179 }; 180 friend class PCHValidator; 181 friend class ASTDeclReader; 182 friend class ASTStmtReader; 183 friend class ASTIdentifierLookupTrait; 184 friend class TypeLocReader; 185private: 186 /// \brief The receiver of some callbacks invoked by ASTReader. 187 llvm::OwningPtr<ASTReaderListener> Listener; 188 189 /// \brief The receiver of deserialization events. 190 ASTDeserializationListener *DeserializationListener; 191 192 SourceManager &SourceMgr; 193 FileManager &FileMgr; 194 Diagnostic &Diags; 195 196 /// \brief The semantic analysis object that will be processing the 197 /// AST files and the translation unit that uses it. 198 Sema *SemaObj; 199 200 /// \brief The preprocessor that will be loading the source file. 201 Preprocessor *PP; 202 203 /// \brief The AST context into which we'll read the AST files. 204 ASTContext *Context; 205 206 /// \brief The AST consumer. 207 ASTConsumer *Consumer; 208 209 /// \brief Information that is needed for every module. 210 struct PerFileData { 211 PerFileData(ASTFileType Ty); 212 ~PerFileData(); 213 214 // === General information === 215 216 /// \brief The type of this AST file. 217 ASTFileType Type; 218 219 /// \brief The file name of the AST file. 220 std::string FileName; 221 222 /// \brief The memory buffer that stores the data associated with 223 /// this AST file. 224 llvm::OwningPtr<llvm::MemoryBuffer> Buffer; 225 226 /// \brief The size of this file, in bits. 227 uint64_t SizeInBits; 228 229 /// \brief The bitstream reader from which we'll read the AST file. 230 llvm::BitstreamReader StreamFile; 231 232 /// \brief The main bitstream cursor for the main block. 233 llvm::BitstreamCursor Stream; 234 235 // === Source Locations === 236 237 /// \brief Cursor used to read source location entries. 238 llvm::BitstreamCursor SLocEntryCursor; 239 240 /// \brief The number of source location entries in this AST file. 241 unsigned LocalNumSLocEntries; 242 243 /// \brief Offsets for all of the source location entries in the 244 /// AST file. 245 const uint32_t *SLocOffsets; 246 247 /// \brief The entire size of this module's source location offset range. 248 unsigned LocalSLocSize; 249 250 // === Identifiers === 251 252 /// \brief The number of identifiers in this AST file. 253 unsigned LocalNumIdentifiers; 254 255 /// \brief Offsets into the identifier table data. 256 /// 257 /// This array is indexed by the identifier ID (-1), and provides 258 /// the offset into IdentifierTableData where the string data is 259 /// stored. 260 const uint32_t *IdentifierOffsets; 261 262 /// \brief Actual data for the on-disk hash table. 263 /// 264 /// This pointer points into a memory buffer, where the on-disk hash 265 /// table for identifiers actually lives. 266 const char *IdentifierTableData; 267 268 /// \brief A pointer to an on-disk hash table of opaque type 269 /// IdentifierHashTable. 270 void *IdentifierLookupTable; 271 272 // === Macros === 273 274 /// \brief The cursor to the start of the preprocessor block, which stores 275 /// all of the macro definitions. 276 llvm::BitstreamCursor MacroCursor; 277 278 /// \brief The number of macro definitions in this file. 279 unsigned LocalNumMacroDefinitions; 280 281 /// \brief Offsets of all of the macro definitions in the preprocessing 282 /// record in the AST file. 283 const uint32_t *MacroDefinitionOffsets; 284 285 // === Selectors === 286 287 /// \brief The number of selectors new to this file. 288 /// 289 /// This is the number of entries in SelectorOffsets. 290 unsigned LocalNumSelectors; 291 292 /// \brief Offsets into the selector lookup table's data array 293 /// where each selector resides. 294 const uint32_t *SelectorOffsets; 295 296 /// \brief A pointer to the character data that comprises the selector table 297 /// 298 /// The SelectorOffsets table refers into this memory. 299 const unsigned char *SelectorLookupTableData; 300 301 /// \brief A pointer to an on-disk hash table of opaque type 302 /// ASTSelectorLookupTable. 303 /// 304 /// This hash table provides the IDs of all selectors, and the associated 305 /// instance and factory methods. 306 void *SelectorLookupTable; 307 308 /// \brief Method selectors used in a @selector expression. Used for 309 /// implementation of -Wselector. 310 llvm::SmallVector<uint64_t, 64> ReferencedSelectorsData; 311 312 // === Declarations === 313 314 /// DeclsCursor - This is a cursor to the start of the DECLS_BLOCK block. It 315 /// has read all the abbreviations at the start of the block and is ready to 316 /// jump around with these in context. 317 llvm::BitstreamCursor DeclsCursor; 318 319 /// \brief The number of declarations in this AST file. 320 unsigned LocalNumDecls; 321 322 /// \brief Offset of each declaration within the bitstream, indexed 323 /// by the declaration ID (-1). 324 const uint32_t *DeclOffsets; 325 326 /// \brief A snapshot of the pending instantiations in the chain. 327 /// 328 /// This record tracks the instantiations that Sema has to perform at the 329 /// end of the TU. It consists of a pair of values for every pending 330 /// instantiation where the first value is the ID of the decl and the second 331 /// is the instantiation location. 332 llvm::SmallVector<uint64_t, 64> PendingInstantiations; 333 334 // === Types === 335 336 /// \brief The number of types in this AST file. 337 unsigned LocalNumTypes; 338 339 /// \brief Offset of each type within the bitstream, indexed by the 340 /// type ID, or the representation of a Type*. 341 const uint32_t *TypeOffsets; 342 343 // === Miscellaneous === 344 345 /// \brief The AST stat cache installed for this file, if any. 346 /// 347 /// The dynamic type of this stat cache is always ASTStatCache 348 void *StatCache; 349 350 /// \brief The number of preallocated preprocessing entities in the 351 /// preprocessing record. 352 unsigned NumPreallocatedPreprocessingEntities; 353 354 /// \brief The next module in source order. 355 PerFileData *NextInSource; 356 357 /// \brief All the modules that loaded this one. Can contain NULL for 358 /// directly loaded modules. 359 llvm::SmallVector<PerFileData *, 1> Loaders; 360 }; 361 362 /// \brief All loaded modules, indexed by name. 363 llvm::StringMap<PerFileData*> Modules; 364 365 /// \brief The first module in source order. 366 PerFileData *FirstInSource; 367 368 /// \brief The chain of AST files. The first entry is the one named by the 369 /// user, the last one is the one that doesn't depend on anything further. 370 /// That is, the entry I was created with -include-pch I+1. 371 llvm::SmallVector<PerFileData*, 2> Chain; 372 373 /// \brief SLocEntries that we're going to preload. 374 llvm::SmallVector<uint64_t, 64> PreloadSLocEntries; 375 376 /// \brief Types that have already been loaded from the chain. 377 /// 378 /// When the pointer at index I is non-NULL, the type with 379 /// ID = (I + 1) << FastQual::Width has already been loaded 380 std::vector<QualType> TypesLoaded; 381 382 /// \brief Map that provides the ID numbers of each type within the 383 /// output stream, plus those deserialized from a chained PCH. 384 /// 385 /// The ID numbers of types are consecutive (in order of discovery) 386 /// and start at 1. 0 is reserved for NULL. When types are actually 387 /// stored in the stream, the ID number is shifted by 2 bits to 388 /// allow for the const/volatile qualifiers. 389 /// 390 /// Keys in the map never have const/volatile qualifiers. 391 serialization::TypeIdxMap TypeIdxs; 392 393 /// \brief Declarations that have already been loaded from the chain. 394 /// 395 /// When the pointer at index I is non-NULL, the declaration with ID 396 /// = I + 1 has already been loaded. 397 std::vector<Decl *> DeclsLoaded; 398 399 typedef llvm::DenseMap<serialization::DeclID, 400 std::pair<PerFileData *, uint64_t> > 401 DeclReplacementMap; 402 /// \brief Declarations that have been replaced in a later file in the chain. 403 DeclReplacementMap ReplacedDecls; 404 405 /// \brief Information about the contents of a DeclContext. 406 struct DeclContextInfo { 407 void *NameLookupTableData; // a ASTDeclContextNameLookupTable. 408 const serialization::DeclID *LexicalDecls; 409 unsigned NumLexicalDecls; 410 }; 411 // In a full chain, there could be multiple updates to every decl context, 412 // so this is a vector. However, typically a chain is only two elements long, 413 // with only one file containing updates, so there will be only one update 414 // per decl context. 415 typedef llvm::SmallVector<DeclContextInfo, 1> DeclContextInfos; 416 typedef llvm::DenseMap<const DeclContext *, DeclContextInfos> 417 DeclContextOffsetsMap; 418 // Updates for visible decls can occur for other contexts than just the 419 // TU, and when we read those update records, the actual context will not 420 // be available yet (unless it's the TU), so have this pending map using the 421 // ID as a key. It will be realized when the context is actually loaded. 422 typedef llvm::SmallVector<void *, 1> DeclContextVisibleUpdates; 423 typedef llvm::DenseMap<serialization::DeclID, DeclContextVisibleUpdates> 424 DeclContextVisibleUpdatesPending; 425 426 /// \brief Offsets of the lexical and visible declarations for each 427 /// DeclContext. 428 DeclContextOffsetsMap DeclContextOffsets; 429 430 /// \brief Updates to the visible declarations of declaration contexts that 431 /// haven't been loaded yet. 432 DeclContextVisibleUpdatesPending PendingVisibleUpdates; 433 434 typedef llvm::DenseMap<serialization::DeclID, serialization::DeclID> 435 FirstLatestDeclIDMap; 436 /// \brief Map of first declarations from a chained PCH that point to the 437 /// most recent declarations in another AST file. 438 FirstLatestDeclIDMap FirstLatestDeclIDs; 439 440 typedef llvm::SmallVector<serialization::DeclID, 4> 441 AdditionalTemplateSpecializations; 442 typedef llvm::DenseMap<serialization::DeclID, 443 AdditionalTemplateSpecializations> 444 AdditionalTemplateSpecializationsMap; 445 446 /// \brief Additional specializations (including partial) of templates that 447 /// were introduced after the template was serialized. 448 AdditionalTemplateSpecializationsMap AdditionalTemplateSpecializationsPending; 449 450 /// \brief Read the records that describe the contents of declcontexts. 451 bool ReadDeclContextStorage(llvm::BitstreamCursor &Cursor, 452 const std::pair<uint64_t, uint64_t> &Offsets, 453 DeclContextInfo &Info); 454 455 /// \brief A vector containing identifiers that have already been 456 /// loaded. 457 /// 458 /// If the pointer at index I is non-NULL, then it refers to the 459 /// IdentifierInfo for the identifier with ID=I+1 that has already 460 /// been loaded. 461 std::vector<IdentifierInfo *> IdentifiersLoaded; 462 463 /// \brief A vector containing selectors that have already been loaded. 464 /// 465 /// This vector is indexed by the Selector ID (-1). NULL selector 466 /// entries indicate that the particular selector ID has not yet 467 /// been loaded. 468 llvm::SmallVector<Selector, 16> SelectorsLoaded; 469 470 /// \brief The macro definitions we have already loaded. 471 llvm::SmallVector<MacroDefinition *, 16> MacroDefinitionsLoaded; 472 473 /// \name CodeGen-relevant special data 474 /// \brief Fields containing data that is relevant to CodeGen. 475 //@{ 476 477 /// \brief The IDs of all declarations that fulfill the criteria of 478 /// "interesting" decls. 479 /// 480 /// This contains the data loaded from all EXTERNAL_DEFINITIONS blocks in the 481 /// chain. The referenced declarations are deserialized and passed to the 482 /// consumer eagerly. 483 llvm::SmallVector<uint64_t, 16> ExternalDefinitions; 484 485 /// \brief The IDs of all tentative definitions stored in the the chain. 486 /// 487 /// Sema keeps track of all tentative definitions in a TU because it has to 488 /// complete them and pass them on to CodeGen. Thus, tentative definitions in 489 /// the PCH chain must be eagerly deserialized. 490 llvm::SmallVector<uint64_t, 16> TentativeDefinitions; 491 492 /// \brief The IDs of all CXXRecordDecls stored in the chain whose VTables are 493 /// used. 494 /// 495 /// CodeGen has to emit VTables for these records, so they have to be eagerly 496 /// deserialized. 497 llvm::SmallVector<uint64_t, 64> VTableUses; 498 499 //@} 500 501 /// \name Diagnostic-relevant special data 502 /// \brief Fields containing data that is used for generating diagnostics 503 //@{ 504 505 /// \brief A snapshot of Sema's unused file-scoped variable tracking, for 506 /// generating warnings. 507 llvm::SmallVector<uint64_t, 16> UnusedFileScopedDecls; 508 509 /// \brief A snapshot of Sema's weak undeclared identifier tracking, for 510 /// generating warnings. 511 llvm::SmallVector<uint64_t, 64> WeakUndeclaredIdentifiers; 512 513 /// \brief The IDs of type aliases for ext_vectors that exist in the chain. 514 /// 515 /// Used by Sema for finding sugared names for ext_vectors in diagnostics. 516 llvm::SmallVector<uint64_t, 4> ExtVectorDecls; 517 518 //@} 519 520 /// \name Sema-relevant special data 521 /// \brief Fields containing data that is used for semantic analysis 522 //@{ 523 524 /// \brief The IDs of all locally scoped external decls in the chain. 525 /// 526 /// Sema tracks these to validate that the types are consistent across all 527 /// local external declarations. 528 llvm::SmallVector<uint64_t, 16> LocallyScopedExternalDecls; 529 530 /// \brief The IDs of all dynamic class declarations in the chain. 531 /// 532 /// Sema tracks these because it checks for the key functions being defined 533 /// at the end of the TU, in which case it directs CodeGen to emit the VTable. 534 llvm::SmallVector<uint64_t, 16> DynamicClasses; 535 536 /// \brief The IDs of the declarations Sema stores directly. 537 /// 538 /// Sema tracks a few important decls, such as namespace std, directly. 539 llvm::SmallVector<uint64_t, 4> SemaDeclRefs; 540 541 /// \brief The IDs of the types ASTContext stores directly. 542 /// 543 /// The AST context tracks a few important types, such as va_list, directly. 544 llvm::SmallVector<uint64_t, 16> SpecialTypes; 545 546 //@} 547 548 /// \brief The original file name that was used to build the primary AST file, 549 /// which may have been modified for relocatable-pch support. 550 std::string OriginalFileName; 551 552 /// \brief The actual original file name that was used to build the primary 553 /// AST file. 554 std::string ActualOriginalFileName; 555 556 /// \brief Whether this precompiled header is a relocatable PCH file. 557 bool RelocatablePCH; 558 559 /// \brief The system include root to be used when loading the 560 /// precompiled header. 561 const char *isysroot; 562 563 /// \brief Whether to disable the normal validation performed on precompiled 564 /// headers when they are loaded. 565 bool DisableValidation; 566 567 /// \brief Mapping from switch-case IDs in the chain to switch-case statements 568 /// 569 /// Statements usually don't have IDs, but switch cases need them, so that the 570 /// switch statement can refer to them. 571 std::map<unsigned, SwitchCase *> SwitchCaseStmts; 572 573 /// \brief Mapping from label statement IDs in the chain to label statements. 574 /// 575 /// Statements usually don't have IDs, but labeled statements need them, so 576 /// that goto statements and address-of-label expressions can refer to them. 577 std::map<unsigned, LabelStmt *> LabelStmts; 578 579 /// \brief Mapping from label IDs to the set of "goto" statements 580 /// that point to that label before the label itself has been 581 /// de-serialized. 582 std::multimap<unsigned, GotoStmt *> UnresolvedGotoStmts; 583 584 /// \brief Mapping from label IDs to the set of address label 585 /// expressions that point to that label before the label itself has 586 /// been de-serialized. 587 std::multimap<unsigned, AddrLabelExpr *> UnresolvedAddrLabelExprs; 588 589 /// \brief The number of stat() calls that hit/missed the stat 590 /// cache. 591 unsigned NumStatHits, NumStatMisses; 592 593 /// \brief The number of source location entries de-serialized from 594 /// the PCH file. 595 unsigned NumSLocEntriesRead; 596 597 /// \brief The number of source location entries in the chain. 598 unsigned TotalNumSLocEntries; 599 600 /// \brief The next offset for a SLocEntry after everything in this reader. 601 unsigned NextSLocOffset; 602 603 /// \brief The number of statements (and expressions) de-serialized 604 /// from the chain. 605 unsigned NumStatementsRead; 606 607 /// \brief The total number of statements (and expressions) stored 608 /// in the chain. 609 unsigned TotalNumStatements; 610 611 /// \brief The number of macros de-serialized from the chain. 612 unsigned NumMacrosRead; 613 614 /// \brief The total number of macros stored in the chain. 615 unsigned TotalNumMacros; 616 617 /// \brief The number of selectors that have been read. 618 unsigned NumSelectorsRead; 619 620 /// \brief The number of method pool entries that have been read. 621 unsigned NumMethodPoolEntriesRead; 622 623 /// \brief The number of times we have looked up a selector in the method 624 /// pool and not found anything interesting. 625 unsigned NumMethodPoolMisses; 626 627 /// \brief The total number of method pool entries in the selector table. 628 unsigned TotalNumMethodPoolEntries; 629 630 /// Number of lexical decl contexts read/total. 631 unsigned NumLexicalDeclContextsRead, TotalLexicalDeclContexts; 632 633 /// Number of visible decl contexts read/total. 634 unsigned NumVisibleDeclContextsRead, TotalVisibleDeclContexts; 635 636 /// \brief Number of Decl/types that are currently deserializing. 637 unsigned NumCurrentElementsDeserializing; 638 639 /// \brief An IdentifierInfo that has been loaded but whose top-level 640 /// declarations of the same name have not (yet) been loaded. 641 struct PendingIdentifierInfo { 642 IdentifierInfo *II; 643 llvm::SmallVector<uint32_t, 4> DeclIDs; 644 }; 645 646 /// \brief The set of identifiers that were read while the AST reader was 647 /// (recursively) loading declarations. 648 /// 649 /// The declarations on the identifier chain for these identifiers will be 650 /// loaded once the recursive loading has completed. 651 std::deque<PendingIdentifierInfo> PendingIdentifierInfos; 652 653 /// \brief Contains declarations and definitions that will be 654 /// "interesting" to the ASTConsumer, when we get that AST consumer. 655 /// 656 /// "Interesting" declarations are those that have data that may 657 /// need to be emitted, such as inline function definitions or 658 /// Objective-C protocols. 659 std::deque<Decl *> InterestingDecls; 660 661 /// \brief When reading a Stmt tree, Stmt operands are placed in this stack. 662 llvm::SmallVector<Stmt *, 16> StmtStack; 663 664 /// \brief What kind of records we are reading. 665 enum ReadingKind { 666 Read_Decl, Read_Type, Read_Stmt 667 }; 668 669 /// \brief What kind of records we are reading. 670 ReadingKind ReadingKind; 671 672 /// \brief RAII object to change the reading kind. 673 class ReadingKindTracker { 674 ASTReader &Reader; 675 enum ReadingKind PrevKind; 676 677 ReadingKindTracker(const ReadingKindTracker&); // do not implement 678 ReadingKindTracker &operator=(const ReadingKindTracker&);// do not implement 679 680 public: 681 ReadingKindTracker(enum ReadingKind newKind, ASTReader &reader) 682 : Reader(reader), PrevKind(Reader.ReadingKind) { 683 Reader.ReadingKind = newKind; 684 } 685 686 ~ReadingKindTracker() { Reader.ReadingKind = PrevKind; } 687 }; 688 689 /// \brief All predefines buffers in the chain, to be treated as if 690 /// concatenated. 691 PCHPredefinesBlocks PCHPredefinesBuffers; 692 693 /// \brief Suggested contents of the predefines buffer, after this 694 /// PCH file has been processed. 695 /// 696 /// In most cases, this string will be empty, because the predefines 697 /// buffer computed to build the PCH file will be identical to the 698 /// predefines buffer computed from the command line. However, when 699 /// there are differences that the PCH reader can work around, this 700 /// predefines buffer may contain additional definitions. 701 std::string SuggestedPredefines; 702 703 /// \brief Reads a statement from the specified cursor. 704 Stmt *ReadStmtFromStream(PerFileData &F); 705 706 void MaybeAddSystemRootToFilename(std::string &Filename); 707 708 ASTReadResult ReadASTCore(llvm::StringRef FileName, ASTFileType Type); 709 ASTReadResult ReadASTBlock(PerFileData &F); 710 bool CheckPredefinesBuffers(); 711 bool ParseLineTable(PerFileData &F, llvm::SmallVectorImpl<uint64_t> &Record); 712 ASTReadResult ReadSourceManagerBlock(PerFileData &F); 713 ASTReadResult ReadSLocEntryRecord(unsigned ID); 714 PerFileData *SLocCursorForID(unsigned ID); 715 SourceLocation getImportLocation(PerFileData *F); 716 bool ParseLanguageOptions(const llvm::SmallVectorImpl<uint64_t> &Record); 717 718 struct RecordLocation { 719 RecordLocation(PerFileData *M, uint64_t O) 720 : F(M), Offset(O) {} 721 PerFileData *F; 722 uint64_t Offset; 723 }; 724 725 QualType ReadTypeRecord(unsigned Index); 726 RecordLocation TypeCursorForIndex(unsigned Index); 727 void LoadedDecl(unsigned Index, Decl *D); 728 Decl *ReadDeclRecord(unsigned Index, serialization::DeclID ID); 729 RecordLocation DeclCursorForIndex(unsigned Index, serialization::DeclID ID); 730 731 void PassInterestingDeclsToConsumer(); 732 733 /// \brief Produce an error diagnostic and return true. 734 /// 735 /// This routine should only be used for fatal errors that have to 736 /// do with non-routine failures (e.g., corrupted AST file). 737 void Error(const char *Msg); 738 739 ASTReader(const ASTReader&); // do not implement 740 ASTReader &operator=(const ASTReader &); // do not implement 741public: 742 typedef llvm::SmallVector<uint64_t, 64> RecordData; 743 744 /// \brief Load the AST file and validate its contents against the given 745 /// Preprocessor. 746 /// 747 /// \param PP the preprocessor associated with the context in which this 748 /// precompiled header will be loaded. 749 /// 750 /// \param Context the AST context that this precompiled header will be 751 /// loaded into. 752 /// 753 /// \param isysroot If non-NULL, the system include path specified by the 754 /// user. This is only used with relocatable PCH files. If non-NULL, 755 /// a relocatable PCH file will use the default path "/". 756 /// 757 /// \param DisableValidation If true, the AST reader will suppress most 758 /// of its regular consistency checking, allowing the use of precompiled 759 /// headers that cannot be determined to be compatible. 760 ASTReader(Preprocessor &PP, ASTContext *Context, const char *isysroot = 0, 761 bool DisableValidation = false); 762 763 /// \brief Load the AST file without using any pre-initialized Preprocessor. 764 /// 765 /// The necessary information to initialize a Preprocessor later can be 766 /// obtained by setting a ASTReaderListener. 767 /// 768 /// \param SourceMgr the source manager into which the AST file will be loaded 769 /// 770 /// \param FileMgr the file manager into which the AST file will be loaded. 771 /// 772 /// \param Diags the diagnostics system to use for reporting errors and 773 /// warnings relevant to loading the AST file. 774 /// 775 /// \param isysroot If non-NULL, the system include path specified by the 776 /// user. This is only used with relocatable PCH files. If non-NULL, 777 /// a relocatable PCH file will use the default path "/". 778 /// 779 /// \param DisableValidation If true, the AST reader will suppress most 780 /// of its regular consistency checking, allowing the use of precompiled 781 /// headers that cannot be determined to be compatible. 782 ASTReader(SourceManager &SourceMgr, FileManager &FileMgr, 783 Diagnostic &Diags, const char *isysroot = 0, 784 bool DisableValidation = false); 785 ~ASTReader(); 786 787 /// \brief Load the precompiled header designated by the given file 788 /// name. 789 ASTReadResult ReadAST(const std::string &FileName, ASTFileType Type); 790 791 /// \brief Set the AST callbacks listener. 792 void setListener(ASTReaderListener *listener) { 793 Listener.reset(listener); 794 } 795 796 /// \brief Set the AST deserialization listener. 797 void setDeserializationListener(ASTDeserializationListener *Listener); 798 799 /// \brief Set the Preprocessor to use. 800 void setPreprocessor(Preprocessor &pp); 801 802 /// \brief Sets and initializes the given Context. 803 void InitializeContext(ASTContext &Context); 804 805 /// \brief Retrieve the name of the named (primary) AST file 806 const std::string &getFileName() const { return Chain[0]->FileName; } 807 808 /// \brief Retrieve the name of the original source file name 809 const std::string &getOriginalSourceFile() { return OriginalFileName; } 810 811 /// \brief Retrieve the name of the original source file name directly from 812 /// the AST file, without actually loading the AST file. 813 static std::string getOriginalSourceFile(const std::string &ASTFileName, 814 Diagnostic &Diags); 815 816 /// \brief Returns the suggested contents of the predefines buffer, 817 /// which contains a (typically-empty) subset of the predefines 818 /// build prior to including the precompiled header. 819 const std::string &getSuggestedPredefines() { return SuggestedPredefines; } 820 821 /// \brief Read preprocessed entities into the 822 virtual void ReadPreprocessedEntities(); 823 824 /// \brief Returns the number of source locations found in the chain. 825 unsigned getTotalNumSLocs() const { 826 return TotalNumSLocEntries; 827 } 828 829 /// \brief Returns the next SLocEntry offset after the chain. 830 unsigned getNextSLocOffset() const { 831 return NextSLocOffset; 832 } 833 834 /// \brief Returns the number of identifiers found in the chain. 835 unsigned getTotalNumIdentifiers() const { 836 return static_cast<unsigned>(IdentifiersLoaded.size()); 837 } 838 839 /// \brief Returns the number of types found in the chain. 840 unsigned getTotalNumTypes() const { 841 return static_cast<unsigned>(TypesLoaded.size()); 842 } 843 844 /// \brief Returns the number of declarations found in the chain. 845 unsigned getTotalNumDecls() const { 846 return static_cast<unsigned>(DeclsLoaded.size()); 847 } 848 849 /// \brief Returns the number of selectors found in the chain. 850 unsigned getTotalNumSelectors() const { 851 return static_cast<unsigned>(SelectorsLoaded.size()); 852 } 853 854 /// \brief Returns the number of macro definitions found in the chain. 855 unsigned getTotalNumMacroDefinitions() const { 856 return static_cast<unsigned>(MacroDefinitionsLoaded.size()); 857 } 858 859 /// \brief Reads a TemplateArgumentLocInfo appropriate for the 860 /// given TemplateArgument kind. 861 TemplateArgumentLocInfo 862 GetTemplateArgumentLocInfo(PerFileData &F, TemplateArgument::ArgKind Kind, 863 const RecordData &Record, unsigned &Idx); 864 865 /// \brief Reads a TemplateArgumentLoc. 866 TemplateArgumentLoc 867 ReadTemplateArgumentLoc(PerFileData &F, 868 const RecordData &Record, unsigned &Idx); 869 870 /// \brief Reads a declarator info from the given record. 871 TypeSourceInfo *GetTypeSourceInfo(PerFileData &F, 872 const RecordData &Record, unsigned &Idx); 873 874 /// \brief Resolve and return the translation unit declaration. 875 TranslationUnitDecl *GetTranslationUnitDecl(); 876 877 /// \brief Resolve a type ID into a type, potentially building a new 878 /// type. 879 QualType GetType(serialization::TypeID ID); 880 881 /// \brief Returns the type ID associated with the given type. 882 /// If the type didn't come from the AST file the ID that is returned is 883 /// marked as "doesn't exist in AST". 884 serialization::TypeID GetTypeID(QualType T) const; 885 886 /// \brief Returns the type index associated with the given type. 887 /// If the type didn't come from the AST file the index that is returned is 888 /// marked as "doesn't exist in AST". 889 serialization::TypeIdx GetTypeIdx(QualType T) const; 890 891 /// \brief Resolve a declaration ID into a declaration, potentially 892 /// building a new declaration. 893 Decl *GetDecl(serialization::DeclID ID); 894 virtual Decl *GetExternalDecl(uint32_t ID); 895 896 /// \brief Resolve the offset of a statement into a statement. 897 /// 898 /// This operation will read a new statement from the external 899 /// source each time it is called, and is meant to be used via a 900 /// LazyOffsetPtr (which is used by Decls for the body of functions, etc). 901 virtual Stmt *GetExternalDeclStmt(uint64_t Offset); 902 903 /// ReadBlockAbbrevs - Enter a subblock of the specified BlockID with the 904 /// specified cursor. Read the abbreviations that are at the top of the block 905 /// and then leave the cursor pointing into the block. 906 bool ReadBlockAbbrevs(llvm::BitstreamCursor &Cursor, unsigned BlockID); 907 908 /// \brief Finds all the visible declarations with a given name. 909 /// The current implementation of this method just loads the entire 910 /// lookup table as unmaterialized references. 911 virtual DeclContext::lookup_result 912 FindExternalVisibleDeclsByName(const DeclContext *DC, 913 DeclarationName Name); 914 915 virtual void MaterializeVisibleDecls(const DeclContext *DC); 916 917 /// \brief Read all of the declarations lexically stored in a 918 /// declaration context. 919 /// 920 /// \param DC The declaration context whose declarations will be 921 /// read. 922 /// 923 /// \param Decls Vector that will contain the declarations loaded 924 /// from the external source. The caller is responsible for merging 925 /// these declarations with any declarations already stored in the 926 /// declaration context. 927 /// 928 /// \returns true if there was an error while reading the 929 /// declarations for this declaration context. 930 virtual bool FindExternalLexicalDecls(const DeclContext *DC, 931 llvm::SmallVectorImpl<Decl*> &Decls); 932 933 /// \brief Notify ASTReader that we started deserialization of 934 /// a decl or type so until FinishedDeserializing is called there may be 935 /// decls that are initializing. Must be paired with FinishedDeserializing. 936 virtual void StartedDeserializing() { ++NumCurrentElementsDeserializing; } 937 938 /// \brief Notify ASTReader that we finished the deserialization of 939 /// a decl or type. Must be paired with StartedDeserializing. 940 virtual void FinishedDeserializing(); 941 942 /// \brief Function that will be invoked when we begin parsing a new 943 /// translation unit involving this external AST source. 944 /// 945 /// This function will provide all of the external definitions to 946 /// the ASTConsumer. 947 virtual void StartTranslationUnit(ASTConsumer *Consumer); 948 949 /// \brief Print some statistics about AST usage. 950 virtual void PrintStats(); 951 952 /// \brief Initialize the semantic source with the Sema instance 953 /// being used to perform semantic analysis on the abstract syntax 954 /// tree. 955 virtual void InitializeSema(Sema &S); 956 957 /// \brief Inform the semantic consumer that Sema is no longer available. 958 virtual void ForgetSema() { SemaObj = 0; } 959 960 /// \brief Retrieve the IdentifierInfo for the named identifier. 961 /// 962 /// This routine builds a new IdentifierInfo for the given identifier. If any 963 /// declarations with this name are visible from translation unit scope, their 964 /// declarations will be deserialized and introduced into the declaration 965 /// chain of the identifier. 966 virtual IdentifierInfo *get(const char *NameStart, const char *NameEnd); 967 IdentifierInfo *get(llvm::StringRef Name) { 968 return get(Name.begin(), Name.end()); 969 } 970 971 /// \brief Load the contents of the global method pool for a given 972 /// selector. 973 /// 974 /// \returns a pair of Objective-C methods lists containing the 975 /// instance and factory methods, respectively, with this selector. 976 virtual std::pair<ObjCMethodList, ObjCMethodList> 977 ReadMethodPool(Selector Sel); 978 979 /// \brief Load a selector from disk, registering its ID if it exists. 980 void LoadSelector(Selector Sel); 981 982 void SetIdentifierInfo(unsigned ID, IdentifierInfo *II); 983 void SetGloballyVisibleDecls(IdentifierInfo *II, 984 const llvm::SmallVectorImpl<uint32_t> &DeclIDs, 985 bool Nonrecursive = false); 986 987 /// \brief Report a diagnostic. 988 DiagnosticBuilder Diag(unsigned DiagID); 989 990 /// \brief Report a diagnostic. 991 DiagnosticBuilder Diag(SourceLocation Loc, unsigned DiagID); 992 993 IdentifierInfo *DecodeIdentifierInfo(unsigned Idx); 994 995 IdentifierInfo *GetIdentifierInfo(const RecordData &Record, unsigned &Idx) { 996 return DecodeIdentifierInfo(Record[Idx++]); 997 } 998 999 virtual IdentifierInfo *GetIdentifier(unsigned ID) { 1000 return DecodeIdentifierInfo(ID); 1001 } 1002 1003 /// \brief Read the source location entry with index ID. 1004 virtual void ReadSLocEntry(unsigned ID); 1005 1006 Selector DecodeSelector(unsigned Idx); 1007 1008 virtual Selector GetExternalSelector(uint32_t ID); 1009 uint32_t GetNumExternalSelectors(); 1010 1011 Selector GetSelector(const RecordData &Record, unsigned &Idx) { 1012 return DecodeSelector(Record[Idx++]); 1013 } 1014 1015 /// \brief Read a declaration name. 1016 DeclarationName ReadDeclarationName(const RecordData &Record, unsigned &Idx); 1017 1018 NestedNameSpecifier *ReadNestedNameSpecifier(const RecordData &Record, 1019 unsigned &Idx); 1020 1021 /// \brief Read a template name. 1022 TemplateName ReadTemplateName(const RecordData &Record, unsigned &Idx); 1023 1024 /// \brief Read a template argument. 1025 TemplateArgument ReadTemplateArgument(PerFileData &F, 1026 const RecordData &Record,unsigned &Idx); 1027 1028 /// \brief Read a template parameter list. 1029 TemplateParameterList *ReadTemplateParameterList(PerFileData &F, 1030 const RecordData &Record, 1031 unsigned &Idx); 1032 1033 /// \brief Read a template argument array. 1034 void 1035 ReadTemplateArgumentList(llvm::SmallVector<TemplateArgument, 8> &TemplArgs, 1036 PerFileData &F, const RecordData &Record, 1037 unsigned &Idx); 1038 1039 /// \brief Read a UnresolvedSet structure. 1040 void ReadUnresolvedSet(UnresolvedSetImpl &Set, 1041 const RecordData &Record, unsigned &Idx); 1042 1043 /// \brief Read a C++ base specifier. 1044 CXXBaseSpecifier ReadCXXBaseSpecifier(PerFileData &F, 1045 const RecordData &Record,unsigned &Idx); 1046 1047 /// \brief Read a CXXBaseOrMemberInitializer array. 1048 std::pair<CXXBaseOrMemberInitializer **, unsigned> 1049 ReadCXXBaseOrMemberInitializers(PerFileData &F, 1050 const RecordData &Record, unsigned &Idx); 1051 1052 /// \brief Read a source location from raw form. 1053 SourceLocation ReadSourceLocation(PerFileData &Module, unsigned Raw) { 1054 (void)Module; // No remapping yet 1055 return SourceLocation::getFromRawEncoding(Raw); 1056 } 1057 1058 /// \brief Read a source location. 1059 SourceLocation ReadSourceLocation(PerFileData &Module, 1060 const RecordData &Record, unsigned& Idx) { 1061 return ReadSourceLocation(Module, Record[Idx++]); 1062 } 1063 1064 /// \brief Read a source range. 1065 SourceRange ReadSourceRange(PerFileData &F, 1066 const RecordData &Record, unsigned& Idx); 1067 1068 /// \brief Read an integral value 1069 llvm::APInt ReadAPInt(const RecordData &Record, unsigned &Idx); 1070 1071 /// \brief Read a signed integral value 1072 llvm::APSInt ReadAPSInt(const RecordData &Record, unsigned &Idx); 1073 1074 /// \brief Read a floating-point value 1075 llvm::APFloat ReadAPFloat(const RecordData &Record, unsigned &Idx); 1076 1077 // \brief Read a string 1078 std::string ReadString(const RecordData &Record, unsigned &Idx); 1079 1080 CXXTemporary *ReadCXXTemporary(const RecordData &Record, unsigned &Idx); 1081 1082 /// \brief Reads attributes from the current stream position. 1083 void ReadAttributes(PerFileData &F, AttrVec &Attrs); 1084 1085 /// \brief Reads a statement. 1086 Stmt *ReadStmt(PerFileData &F); 1087 1088 /// \brief Reads an expression. 1089 Expr *ReadExpr(PerFileData &F); 1090 1091 /// \brief Reads a sub-statement operand during statement reading. 1092 Stmt *ReadSubStmt() { 1093 assert(ReadingKind == Read_Stmt && 1094 "Should be called only during statement reading!"); 1095 // Subexpressions are stored from last to first, so the next Stmt we need 1096 // is at the back of the stack. 1097 assert(!StmtStack.empty() && "Read too many sub statements!"); 1098 return StmtStack.pop_back_val(); 1099 } 1100 1101 /// \brief Reads a sub-expression operand during statement reading. 1102 Expr *ReadSubExpr(); 1103 1104 /// \brief Reads the macro record located at the given offset. 1105 void ReadMacroRecord(PerFileData &F, uint64_t Offset); 1106 1107 /// \brief Read the set of macros defined by this external macro source. 1108 virtual void ReadDefinedMacros(); 1109 1110 /// \brief Retrieve the macro definition with the given ID. 1111 MacroDefinition *getMacroDefinition(serialization::MacroID ID); 1112 1113 /// \brief Retrieve the AST context that this AST reader supplements. 1114 ASTContext *getContext() { return Context; } 1115 1116 // \brief Contains declarations that were loaded before we have 1117 // access to a Sema object. 1118 llvm::SmallVector<NamedDecl *, 16> PreloadedDecls; 1119 1120 /// \brief Retrieve the semantic analysis object used to analyze the 1121 /// translation unit in which the precompiled header is being 1122 /// imported. 1123 Sema *getSema() { return SemaObj; } 1124 1125 /// \brief Retrieve the identifier table associated with the 1126 /// preprocessor. 1127 IdentifierTable &getIdentifierTable(); 1128 1129 /// \brief Record that the given ID maps to the given switch-case 1130 /// statement. 1131 void RecordSwitchCaseID(SwitchCase *SC, unsigned ID); 1132 1133 /// \brief Retrieve the switch-case statement with the given ID. 1134 SwitchCase *getSwitchCaseWithID(unsigned ID); 1135 1136 /// \brief Record that the given label statement has been 1137 /// deserialized and has the given ID. 1138 void RecordLabelStmt(LabelStmt *S, unsigned ID); 1139 1140 /// \brief Set the label of the given statement to the label 1141 /// identified by ID. 1142 /// 1143 /// Depending on the order in which the label and other statements 1144 /// referencing that label occur, this operation may complete 1145 /// immediately (updating the statement) or it may queue the 1146 /// statement to be back-patched later. 1147 void SetLabelOf(GotoStmt *S, unsigned ID); 1148 1149 /// \brief Set the label of the given expression to the label 1150 /// identified by ID. 1151 /// 1152 /// Depending on the order in which the label and other statements 1153 /// referencing that label occur, this operation may complete 1154 /// immediately (updating the statement) or it may queue the 1155 /// statement to be back-patched later. 1156 void SetLabelOf(AddrLabelExpr *S, unsigned ID); 1157}; 1158 1159/// \brief Helper class that saves the current stream position and 1160/// then restores it when destroyed. 1161struct SavedStreamPosition { 1162 explicit SavedStreamPosition(llvm::BitstreamCursor &Cursor) 1163 : Cursor(Cursor), Offset(Cursor.GetCurrentBitNo()) { } 1164 1165 ~SavedStreamPosition() { 1166 Cursor.JumpToBit(Offset); 1167 } 1168 1169private: 1170 llvm::BitstreamCursor &Cursor; 1171 uint64_t Offset; 1172}; 1173 1174inline void PCHValidator::Error(const char *Msg) { 1175 Reader.Error(Msg); 1176} 1177 1178} // end namespace clang 1179 1180#endif 1181