ASTUnit.h revision 727d93ef49e18147149354fadd10e86b13bc4ab0
1//===--- ASTUnit.h - ASTUnit utility ----------------------------*- 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// ASTUnit utility class. 11// 12//===----------------------------------------------------------------------===// 13 14#ifndef LLVM_CLANG_FRONTEND_ASTUNIT_H 15#define LLVM_CLANG_FRONTEND_ASTUNIT_H 16 17#include "clang/Index/ASTLocation.h" 18#include "clang/Frontend/PCHBitCodes.h" 19#include "clang/Sema/Sema.h" 20#include "clang/Lex/PreprocessingRecord.h" 21#include "clang/Basic/SourceManager.h" 22#include "clang/Basic/FileManager.h" 23#include "clang-c/Index.h" 24#include "llvm/ADT/IntrusiveRefCntPtr.h" 25#include "llvm/ADT/OwningPtr.h" 26#include "llvm/ADT/SmallVector.h" 27#include "llvm/ADT/StringMap.h" 28#include "llvm/System/Path.h" 29#include "llvm/Support/Timer.h" 30#include <map> 31#include <string> 32#include <vector> 33#include <cassert> 34#include <utility> 35#include <sys/types.h> 36 37namespace llvm { 38 class MemoryBuffer; 39} 40 41namespace clang { 42class ASTContext; 43class CodeCompleteConsumer; 44class CompilerInvocation; 45class Decl; 46class Diagnostic; 47class FileEntry; 48class FileManager; 49class HeaderSearch; 50class Preprocessor; 51class SourceManager; 52class TargetInfo; 53 54using namespace idx; 55 56/// \brief Utility class for loading a ASTContext from a PCH file. 57/// 58class ASTUnit { 59public: 60 typedef std::map<FileID, std::vector<PreprocessedEntity *> > 61 PreprocessedEntitiesByFileMap; 62private: 63 llvm::IntrusiveRefCntPtr<Diagnostic> Diagnostics; 64 llvm::OwningPtr<FileManager> FileMgr; 65 llvm::OwningPtr<SourceManager> SourceMgr; 66 llvm::OwningPtr<HeaderSearch> HeaderInfo; 67 llvm::OwningPtr<TargetInfo> Target; 68 llvm::OwningPtr<Preprocessor> PP; 69 llvm::OwningPtr<ASTContext> Ctx; 70 71 /// \brief The AST consumer that received information about the translation 72 /// unit as it was parsed or loaded. 73 llvm::OwningPtr<ASTConsumer> Consumer; 74 75 /// \brief The semantic analysis object used to type-check the translation 76 /// unit. 77 llvm::OwningPtr<Sema> TheSema; 78 79 /// Optional owned invocation, just used to make the invocation used in 80 /// LoadFromCommandLine available. 81 llvm::OwningPtr<CompilerInvocation> Invocation; 82 83 // OnlyLocalDecls - when true, walking this AST should only visit declarations 84 // that come from the AST itself, not from included precompiled headers. 85 // FIXME: This is temporary; eventually, CIndex will always do this. 86 bool OnlyLocalDecls; 87 88 /// \brief Whether to capture any diagnostics produced. 89 bool CaptureDiagnostics; 90 91 /// \brief Track whether the main file was loaded from an AST or not. 92 bool MainFileIsAST; 93 94 /// \brief Whether this AST represents a complete translation unit. 95 bool CompleteTranslationUnit; 96 97 /// Track the top-level decls which appeared in an ASTUnit which was loaded 98 /// from a source file. 99 // 100 // FIXME: This is just an optimization hack to avoid deserializing large parts 101 // of a PCH file when using the Index library on an ASTUnit loaded from 102 // source. In the long term we should make the Index library use efficient and 103 // more scalable search mechanisms. 104 std::vector<Decl*> TopLevelDecls; 105 106 /// The name of the original source file used to generate this ASTUnit. 107 std::string OriginalSourceFile; 108 109 // Critical optimization when using clang_getCursor(). 110 ASTLocation LastLoc; 111 112 /// \brief The set of diagnostics produced when creating this 113 /// translation unit. 114 llvm::SmallVector<StoredDiagnostic, 4> StoredDiagnostics; 115 116 /// \brief Temporary files that should be removed when the ASTUnit is 117 /// destroyed. 118 llvm::SmallVector<llvm::sys::Path, 4> TemporaryFiles; 119 120 /// \brief A mapping from file IDs to the set of preprocessed entities 121 /// stored in that file. 122 /// 123 /// FIXME: This is just an optimization hack to avoid searching through 124 /// many preprocessed entities during cursor traversal in the CIndex library. 125 /// Ideally, we would just be able to perform a binary search within the 126 /// list of preprocessed entities. 127 PreprocessedEntitiesByFileMap PreprocessedEntitiesByFile; 128 129 /// \brief Simple hack to allow us to assert that ASTUnit is not being 130 /// used concurrently, which is not supported. 131 /// 132 /// Clients should create instances of the ConcurrencyCheck class whenever 133 /// using the ASTUnit in a way that isn't intended to be concurrent, which is 134 /// just about any usage. 135 unsigned int ConcurrencyCheckValue; 136 static const unsigned int CheckLocked = 28573289; 137 static const unsigned int CheckUnlocked = 9803453; 138 139 /// \brief Counter that determines when we want to try building a 140 /// precompiled preamble. 141 /// 142 /// If zero, we will never build a precompiled preamble. Otherwise, 143 /// it's treated as a counter that decrements each time we reparse 144 /// without the benefit of a precompiled preamble. When it hits 1, 145 /// we'll attempt to rebuild the precompiled header. This way, if 146 /// building the precompiled preamble fails, we won't try again for 147 /// some number of calls. 148 unsigned PreambleRebuildCounter; 149 150 /// \brief The file in which the precompiled preamble is stored. 151 std::string PreambleFile; 152 153 /// \brief The contents of the preamble that has been precompiled to 154 /// \c PreambleFile. 155 std::vector<char> Preamble; 156 157 /// \brief Whether the preamble ends at the start of a new line. 158 /// 159 /// Used to inform the lexer as to whether it's starting at the beginning of 160 /// a line after skipping the preamble. 161 bool PreambleEndsAtStartOfLine; 162 163 /// \brief The size of the source buffer that we've reserved for the main 164 /// file within the precompiled preamble. 165 unsigned PreambleReservedSize; 166 167 /// \brief Keeps track of the files that were used when computing the 168 /// preamble, with both their buffer size and their modification time. 169 /// 170 /// If any of the files have changed from one compile to the next, 171 /// the preamble must be thrown away. 172 llvm::StringMap<std::pair<off_t, time_t> > FilesInPreamble; 173 174 /// \brief When non-NULL, this is the buffer used to store the contents of 175 /// the main file when it has been padded for use with the precompiled 176 /// preamble. 177 llvm::MemoryBuffer *SavedMainFileBuffer; 178 179 /// \brief The number of warnings that occurred while parsing the preamble. 180 /// 181 /// This value will be used to restore the state of the \c Diagnostic object 182 /// when re-using the precompiled preamble. Note that only the 183 /// number of warnings matters, since we will not save the preamble 184 /// when any errors are present. 185 unsigned NumWarningsInPreamble; 186 187 /// \brief The number of diagnostics that were stored when parsing 188 /// the precompiled preamble. 189 /// 190 /// This value is used to determine how many of the stored 191 /// diagnostics should be retained when reparsing in the presence of 192 /// a precompiled preamble. 193 unsigned NumStoredDiagnosticsInPreamble; 194 195 /// \brief The group of timers associated with this translation unit. 196 llvm::OwningPtr<llvm::TimerGroup> TimerGroup; 197 198 /// \brief A list of the PCH ID numbers for each of the top-level 199 /// declarations parsed within the precompiled preamble. 200 std::vector<pch::DeclID> TopLevelDeclsInPreamble; 201 202 /// 203 /// \defgroup CodeCompleteCaching Code-completion caching 204 /// 205 /// \{ 206 /// 207 208 /// \brief Whether we should be caching code-completion results. 209 bool ShouldCacheCodeCompletionResults; 210 211public: 212 /// \brief A cached code-completion result, which may be introduced in one of 213 /// many different contexts. 214 struct CachedCodeCompletionResult { 215 /// \brief The code-completion string corresponding to this completion 216 /// result. 217 CodeCompletionString *Completion; 218 219 /// \brief A bitmask that indicates which code-completion contexts should 220 /// contain this completion result. 221 /// 222 /// The bits in the bitmask correspond to the values of 223 /// CodeCompleteContext::Kind. To map from a completion context kind to a 224 /// bit, subtract one from the completion context kind and shift 1 by that 225 /// number of bits. Many completions can occur in several different 226 /// contexts. 227 unsigned ShowInContexts; 228 229 /// \brief The priority given to this code-completion result. 230 unsigned Priority; 231 232 /// \brief The libclang cursor kind corresponding to this code-completion 233 /// result. 234 CXCursorKind Kind; 235 236 /// \brief The simplified type class for a non-macro completion result. 237 SimplifiedTypeClass TypeClass; 238 239 /// \brief The type of a non-macro completion result, stored as a unique 240 /// integer used by the string map of cached completion types. 241 /// 242 /// This value will be zero if the type is not known, or a unique value 243 /// determined by the formatted type string. Se \c CachedCompletionTypes 244 /// for more information. 245 unsigned Type; 246 }; 247 248 /// \brief Retrieve the mapping from formatted type names to unique type 249 /// identifiers. 250 llvm::StringMap<unsigned> &getCachedCompletionTypes() { 251 return CachedCompletionTypes; 252 } 253 254private: 255 /// \brief The set of cached code-completion results. 256 std::vector<CachedCodeCompletionResult> CachedCompletionResults; 257 258 /// \brief A mapping from the formatted type name to a unique number for that 259 /// type, which is used for type equality comparisons. 260 llvm::StringMap<unsigned> CachedCompletionTypes; 261 262 /// \brief The number of top-level declarations present the last time we 263 /// cached code-completion results. 264 /// 265 /// The value is used to help detect when we should repopulate the global 266 /// completion cache. 267 unsigned NumTopLevelDeclsAtLastCompletionCache; 268 269 /// \brief The number of reparses left until we'll consider updating the 270 /// code-completion cache. 271 /// 272 /// This is meant to avoid thrashing during reparsing, by not allowing the 273 /// code-completion cache to be updated on every reparse. 274 unsigned CacheCodeCompletionCoolDown; 275 276 /// \brief Cache any "global" code-completion results, so that we can avoid 277 /// recomputing them with each completion. 278 void CacheCodeCompletionResults(); 279 280 /// \brief Clear out and deallocate 281 void ClearCachedCompletionResults(); 282 283 /// 284 /// \} 285 /// 286 287 /// \brief The timers we've created from the various parses, reparses, etc. 288 /// involved in this translation unit. 289 std::vector<llvm::Timer *> Timers; 290 291 ASTUnit(const ASTUnit&); // DO NOT IMPLEMENT 292 ASTUnit &operator=(const ASTUnit &); // DO NOT IMPLEMENT 293 294 explicit ASTUnit(bool MainFileIsAST); 295 296 void CleanTemporaryFiles(); 297 bool Parse(llvm::MemoryBuffer *OverrideMainBuffer); 298 299 std::pair<llvm::MemoryBuffer *, std::pair<unsigned, bool> > 300 ComputePreamble(CompilerInvocation &Invocation, 301 unsigned MaxLines, bool &CreatedBuffer); 302 303 llvm::MemoryBuffer *getMainBufferWithPrecompiledPreamble( 304 bool AllowRebuild = true, 305 unsigned MaxLines = 0); 306 void RealizeTopLevelDeclsFromPreamble(); 307 308public: 309 class ConcurrencyCheck { 310 volatile ASTUnit &Self; 311 312 public: 313 explicit ConcurrencyCheck(ASTUnit &Self) 314 : Self(Self) 315 { 316 assert(Self.ConcurrencyCheckValue == CheckUnlocked && 317 "Concurrent access to ASTUnit!"); 318 Self.ConcurrencyCheckValue = CheckLocked; 319 } 320 321 ~ConcurrencyCheck() { 322 Self.ConcurrencyCheckValue = CheckUnlocked; 323 } 324 }; 325 friend class ConcurrencyCheck; 326 327 ~ASTUnit(); 328 329 bool isMainFileAST() const { return MainFileIsAST; } 330 331 const Diagnostic &getDiagnostics() const { return *Diagnostics; } 332 Diagnostic &getDiagnostics() { return *Diagnostics; } 333 334 const SourceManager &getSourceManager() const { return *SourceMgr; } 335 SourceManager &getSourceManager() { return *SourceMgr; } 336 337 const Preprocessor &getPreprocessor() const { return *PP.get(); } 338 Preprocessor &getPreprocessor() { return *PP.get(); } 339 340 const ASTContext &getASTContext() const { return *Ctx.get(); } 341 ASTContext &getASTContext() { return *Ctx.get(); } 342 343 bool hasSema() const { return TheSema; } 344 Sema &getSema() const { 345 assert(TheSema && "ASTUnit does not have a Sema object!"); 346 return *TheSema; 347 } 348 349 const FileManager &getFileManager() const { return *FileMgr; } 350 FileManager &getFileManager() { return *FileMgr; } 351 352 const std::string &getOriginalSourceFileName(); 353 const std::string &getPCHFileName(); 354 355 /// \brief Add a temporary file that the ASTUnit depends on. 356 /// 357 /// This file will be erased when the ASTUnit is destroyed. 358 void addTemporaryFile(const llvm::sys::Path &TempFile) { 359 TemporaryFiles.push_back(TempFile); 360 } 361 362 bool getOnlyLocalDecls() const { return OnlyLocalDecls; } 363 364 /// \brief Retrieve the maximum PCH level of declarations that a 365 /// traversal of the translation unit should consider. 366 unsigned getMaxPCHLevel() const; 367 368 void setLastASTLocation(ASTLocation ALoc) { LastLoc = ALoc; } 369 ASTLocation getLastASTLocation() const { return LastLoc; } 370 371 typedef std::vector<Decl *>::iterator top_level_iterator; 372 373 top_level_iterator top_level_begin() { 374 assert(!isMainFileAST() && "Invalid call for AST based ASTUnit!"); 375 if (!TopLevelDeclsInPreamble.empty()) 376 RealizeTopLevelDeclsFromPreamble(); 377 return TopLevelDecls.begin(); 378 } 379 380 top_level_iterator top_level_end() { 381 assert(!isMainFileAST() && "Invalid call for AST based ASTUnit!"); 382 if (!TopLevelDeclsInPreamble.empty()) 383 RealizeTopLevelDeclsFromPreamble(); 384 return TopLevelDecls.end(); 385 } 386 387 std::size_t top_level_size() const { 388 assert(!isMainFileAST() && "Invalid call for AST based ASTUnit!"); 389 return TopLevelDeclsInPreamble.size() + TopLevelDecls.size(); 390 } 391 392 bool top_level_empty() const { 393 assert(!isMainFileAST() && "Invalid call for AST based ASTUnit!"); 394 return TopLevelDeclsInPreamble.empty() && TopLevelDecls.empty(); 395 } 396 397 /// \brief Add a new top-level declaration. 398 void addTopLevelDecl(Decl *D) { 399 TopLevelDecls.push_back(D); 400 } 401 402 /// \brief Add a new top-level declaration, identified by its ID in 403 /// the precompiled preamble. 404 void addTopLevelDeclFromPreamble(pch::DeclID D) { 405 TopLevelDeclsInPreamble.push_back(D); 406 } 407 408 /// \brief Retrieve the mapping from File IDs to the preprocessed entities 409 /// within that file. 410 PreprocessedEntitiesByFileMap &getPreprocessedEntitiesByFile() { 411 return PreprocessedEntitiesByFile; 412 } 413 414 // Retrieve the diagnostics associated with this AST 415 typedef const StoredDiagnostic *stored_diag_iterator; 416 stored_diag_iterator stored_diag_begin() const { 417 return StoredDiagnostics.begin(); 418 } 419 stored_diag_iterator stored_diag_end() const { 420 return StoredDiagnostics.end(); 421 } 422 unsigned stored_diag_size() const { return StoredDiagnostics.size(); } 423 424 llvm::SmallVector<StoredDiagnostic, 4> &getStoredDiagnostics() { 425 return StoredDiagnostics; 426 } 427 428 typedef std::vector<CachedCodeCompletionResult>::iterator 429 cached_completion_iterator; 430 431 cached_completion_iterator cached_completion_begin() { 432 return CachedCompletionResults.begin(); 433 } 434 435 cached_completion_iterator cached_completion_end() { 436 return CachedCompletionResults.end(); 437 } 438 439 unsigned cached_completion_size() const { 440 return CachedCompletionResults.size(); 441 } 442 443 /// \brief Whether this AST represents a complete translation unit. 444 /// 445 /// If false, this AST is only a partial translation unit, e.g., one 446 /// that might still be used as a precompiled header or preamble. 447 bool isCompleteTranslationUnit() const { return CompleteTranslationUnit; } 448 449 /// \brief A mapping from a file name to the memory buffer that stores the 450 /// remapped contents of that file. 451 typedef std::pair<std::string, const llvm::MemoryBuffer *> RemappedFile; 452 453 /// \brief Create a ASTUnit from a PCH file. 454 /// 455 /// \param Filename - The PCH file to load. 456 /// 457 /// \param Diags - The diagnostics engine to use for reporting errors; its 458 /// lifetime is expected to extend past that of the returned ASTUnit. 459 /// 460 /// \returns - The initialized ASTUnit or null if the PCH failed to load. 461 static ASTUnit *LoadFromPCHFile(const std::string &Filename, 462 llvm::IntrusiveRefCntPtr<Diagnostic> Diags, 463 bool OnlyLocalDecls = false, 464 RemappedFile *RemappedFiles = 0, 465 unsigned NumRemappedFiles = 0, 466 bool CaptureDiagnostics = false); 467 468 /// LoadFromCompilerInvocation - Create an ASTUnit from a source file, via a 469 /// CompilerInvocation object. 470 /// 471 /// \param CI - The compiler invocation to use; it must have exactly one input 472 /// source file. The ASTUnit takes ownership of the CompilerInvocation object. 473 /// 474 /// \param Diags - The diagnostics engine to use for reporting errors; its 475 /// lifetime is expected to extend past that of the returned ASTUnit. 476 // 477 // FIXME: Move OnlyLocalDecls, UseBumpAllocator to setters on the ASTUnit, we 478 // shouldn't need to specify them at construction time. 479 static ASTUnit *LoadFromCompilerInvocation(CompilerInvocation *CI, 480 llvm::IntrusiveRefCntPtr<Diagnostic> Diags, 481 bool OnlyLocalDecls = false, 482 bool CaptureDiagnostics = false, 483 bool PrecompilePreamble = false, 484 bool CompleteTranslationUnit = true, 485 bool CacheCodeCompletionResults = false); 486 487 /// LoadFromCommandLine - Create an ASTUnit from a vector of command line 488 /// arguments, which must specify exactly one source file. 489 /// 490 /// \param ArgBegin - The beginning of the argument vector. 491 /// 492 /// \param ArgEnd - The end of the argument vector. 493 /// 494 /// \param Diags - The diagnostics engine to use for reporting errors; its 495 /// lifetime is expected to extend past that of the returned ASTUnit. 496 /// 497 /// \param ResourceFilesPath - The path to the compiler resource files. 498 // 499 // FIXME: Move OnlyLocalDecls, UseBumpAllocator to setters on the ASTUnit, we 500 // shouldn't need to specify them at construction time. 501 static ASTUnit *LoadFromCommandLine(const char **ArgBegin, 502 const char **ArgEnd, 503 llvm::IntrusiveRefCntPtr<Diagnostic> Diags, 504 llvm::StringRef ResourceFilesPath, 505 bool OnlyLocalDecls = false, 506 RemappedFile *RemappedFiles = 0, 507 unsigned NumRemappedFiles = 0, 508 bool CaptureDiagnostics = false, 509 bool PrecompilePreamble = false, 510 bool CompleteTranslationUnit = true, 511 bool CacheCodeCompletionResults = false); 512 513 /// \brief Reparse the source files using the same command-line options that 514 /// were originally used to produce this translation unit. 515 /// 516 /// \returns True if a failure occurred that causes the ASTUnit not to 517 /// contain any translation-unit information, false otherwise. 518 bool Reparse(RemappedFile *RemappedFiles = 0, 519 unsigned NumRemappedFiles = 0); 520 521 /// \brief Perform code completion at the given file, line, and 522 /// column within this translation unit. 523 /// 524 /// \param File The file in which code completion will occur. 525 /// 526 /// \param Line The line at which code completion will occur. 527 /// 528 /// \param Column The column at which code completion will occur. 529 /// 530 /// \param IncludeMacros Whether to include macros in the code-completion 531 /// results. 532 /// 533 /// \param IncludeCodePatterns Whether to include code patterns (such as a 534 /// for loop) in the code-completion results. 535 /// 536 /// FIXME: The Diag, LangOpts, SourceMgr, FileMgr, and 537 /// StoredDiagnostics parameters are all disgusting hacks. They will 538 /// go away. 539 void CodeComplete(llvm::StringRef File, unsigned Line, unsigned Column, 540 RemappedFile *RemappedFiles, unsigned NumRemappedFiles, 541 bool IncludeMacros, bool IncludeCodePatterns, 542 CodeCompleteConsumer &Consumer, 543 Diagnostic &Diag, LangOptions &LangOpts, 544 SourceManager &SourceMgr, FileManager &FileMgr, 545 llvm::SmallVectorImpl<StoredDiagnostic> &StoredDiagnostics); 546 547 /// \brief Save this translation unit to a file with the given name. 548 /// 549 /// \returns True if an error occurred, false otherwise. 550 bool Save(llvm::StringRef File); 551}; 552 553} // namespace clang 554 555#endif 556