1 | //===- ASTWriter.h - AST File Writer ----------------------------*- C++ -*-===// |
2 | // |
3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
4 | // See https://llvm.org/LICENSE.txt for license information. |
5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
6 | // |
7 | //===----------------------------------------------------------------------===// |
8 | // |
9 | // This file defines the ASTWriter class, which writes an AST file |
10 | // containing a serialized representation of a translation unit. |
11 | // |
12 | //===----------------------------------------------------------------------===// |
13 | |
14 | #ifndef LLVM_CLANG_SERIALIZATION_ASTWRITER_H |
15 | #define LLVM_CLANG_SERIALIZATION_ASTWRITER_H |
16 | |
17 | #include "clang/AST/ASTMutationListener.h" |
18 | #include "clang/AST/Decl.h" |
19 | #include "clang/AST/Type.h" |
20 | #include "clang/Basic/LLVM.h" |
21 | #include "clang/Basic/Module.h" |
22 | #include "clang/Basic/SourceLocation.h" |
23 | #include "clang/Sema/Sema.h" |
24 | #include "clang/Sema/SemaConsumer.h" |
25 | #include "clang/Serialization/ASTBitCodes.h" |
26 | #include "clang/Serialization/ASTDeserializationListener.h" |
27 | #include "clang/Serialization/PCHContainerOperations.h" |
28 | #include "clang/Serialization/SourceLocationEncoding.h" |
29 | #include "llvm/ADT/ArrayRef.h" |
30 | #include "llvm/ADT/DenseMap.h" |
31 | #include "llvm/ADT/DenseSet.h" |
32 | #include "llvm/ADT/MapVector.h" |
33 | #include "llvm/ADT/STLExtras.h" |
34 | #include "llvm/ADT/SetVector.h" |
35 | #include "llvm/ADT/SmallVector.h" |
36 | #include "llvm/ADT/StringRef.h" |
37 | #include "llvm/Bitstream/BitstreamWriter.h" |
38 | #include <cassert> |
39 | #include <cstddef> |
40 | #include <cstdint> |
41 | #include <ctime> |
42 | #include <memory> |
43 | #include <queue> |
44 | #include <string> |
45 | #include <utility> |
46 | #include <vector> |
47 | |
48 | namespace clang { |
49 | |
50 | class ASTContext; |
51 | class ASTReader; |
52 | class Attr; |
53 | class CXXRecordDecl; |
54 | class FileEntry; |
55 | class FPOptionsOverride; |
56 | class FunctionDecl; |
57 | class ; |
58 | class ; |
59 | class IdentifierResolver; |
60 | class LangOptions; |
61 | class MacroDefinitionRecord; |
62 | class MacroInfo; |
63 | class Module; |
64 | class InMemoryModuleCache; |
65 | class ModuleFileExtension; |
66 | class ModuleFileExtensionWriter; |
67 | class NamedDecl; |
68 | class ObjCInterfaceDecl; |
69 | class PreprocessingRecord; |
70 | class Preprocessor; |
71 | class RecordDecl; |
72 | class Sema; |
73 | class SourceManager; |
74 | class Stmt; |
75 | class StoredDeclsList; |
76 | class SwitchCase; |
77 | class Token; |
78 | |
79 | /// Writes an AST file containing the contents of a translation unit. |
80 | /// |
81 | /// The ASTWriter class produces a bitstream containing the serialized |
82 | /// representation of a given abstract syntax tree and its supporting |
83 | /// data structures. This bitstream can be de-serialized via an |
84 | /// instance of the ASTReader class. |
85 | class ASTWriter : public ASTDeserializationListener, |
86 | public ASTMutationListener { |
87 | public: |
88 | friend class ASTDeclWriter; |
89 | friend class ASTRecordWriter; |
90 | |
91 | using RecordData = SmallVector<uint64_t, 64>; |
92 | using RecordDataImpl = SmallVectorImpl<uint64_t>; |
93 | using RecordDataRef = ArrayRef<uint64_t>; |
94 | |
95 | private: |
96 | /// Map that provides the ID numbers of each type within the |
97 | /// output stream, plus those deserialized from a chained PCH. |
98 | /// |
99 | /// The ID numbers of types are consecutive (in order of discovery) |
100 | /// and start at 1. 0 is reserved for NULL. When types are actually |
101 | /// stored in the stream, the ID number is shifted by 2 bits to |
102 | /// allow for the const/volatile qualifiers. |
103 | /// |
104 | /// Keys in the map never have const/volatile qualifiers. |
105 | using TypeIdxMap = llvm::DenseMap<QualType, serialization::TypeIdx, |
106 | serialization::UnsafeQualTypeDenseMapInfo>; |
107 | |
108 | using LocSeq = SourceLocationSequence; |
109 | |
110 | /// The bitstream writer used to emit this precompiled header. |
111 | llvm::BitstreamWriter &Stream; |
112 | |
113 | /// The buffer associated with the bitstream. |
114 | const SmallVectorImpl<char> &Buffer; |
115 | |
116 | /// The PCM manager which manages memory buffers for pcm files. |
117 | InMemoryModuleCache &ModuleCache; |
118 | |
119 | /// The ASTContext we're writing. |
120 | ASTContext *Context = nullptr; |
121 | |
122 | /// The preprocessor we're writing. |
123 | Preprocessor *PP = nullptr; |
124 | |
125 | /// The reader of existing AST files, if we're chaining. |
126 | ASTReader *Chain = nullptr; |
127 | |
128 | /// The module we're currently writing, if any. |
129 | Module *WritingModule = nullptr; |
130 | |
131 | /// The byte range representing all the UNHASHED_CONTROL_BLOCK. |
132 | std::pair<uint64_t, uint64_t> UnhashedControlBlockRange; |
133 | /// The bit offset of the AST block hash blob. |
134 | uint64_t ASTBlockHashOffset = 0; |
135 | /// The bit offset of the signature blob. |
136 | uint64_t SignatureOffset = 0; |
137 | |
138 | /// The bit offset of the first bit inside the AST_BLOCK. |
139 | uint64_t ASTBlockStartOffset = 0; |
140 | |
141 | /// The byte range representing all the AST_BLOCK. |
142 | std::pair<uint64_t, uint64_t> ASTBlockRange; |
143 | |
144 | /// The base directory for any relative paths we emit. |
145 | std::string BaseDirectory; |
146 | |
147 | /// Indicates whether timestamps should be written to the produced |
148 | /// module file. This is the case for files implicitly written to the |
149 | /// module cache, where we need the timestamps to determine if the module |
150 | /// file is up to date, but not otherwise. |
151 | bool IncludeTimestamps; |
152 | |
153 | /// Indicates whether the AST file being written is an implicit module. |
154 | /// If that's the case, we may be able to skip writing some information that |
155 | /// are guaranteed to be the same in the importer by the context hash. |
156 | bool BuildingImplicitModule = false; |
157 | |
158 | /// Indicates when the AST writing is actively performing |
159 | /// serialization, rather than just queueing updates. |
160 | bool WritingAST = false; |
161 | |
162 | /// Indicates that we are done serializing the collection of decls |
163 | /// and types to emit. |
164 | bool DoneWritingDeclsAndTypes = false; |
165 | |
166 | /// Indicates that the AST contained compiler errors. |
167 | bool ASTHasCompilerErrors = false; |
168 | |
169 | /// Mapping from input file entries to the index into the |
170 | /// offset table where information about that input file is stored. |
171 | llvm::DenseMap<const FileEntry *, uint32_t> InputFileIDs; |
172 | |
173 | /// Stores a declaration or a type to be written to the AST file. |
174 | class DeclOrType { |
175 | public: |
176 | DeclOrType(Decl *D) : Stored(D), IsType(false) {} |
177 | DeclOrType(QualType T) : Stored(T.getAsOpaquePtr()), IsType(true) {} |
178 | |
179 | bool isType() const { return IsType; } |
180 | bool isDecl() const { return !IsType; } |
181 | |
182 | QualType getType() const { |
183 | assert(isType() && "Not a type!" ); |
184 | return QualType::getFromOpaquePtr(Ptr: Stored); |
185 | } |
186 | |
187 | Decl *getDecl() const { |
188 | assert(isDecl() && "Not a decl!" ); |
189 | return static_cast<Decl *>(Stored); |
190 | } |
191 | |
192 | private: |
193 | void *Stored; |
194 | bool IsType; |
195 | }; |
196 | |
197 | /// The declarations and types to emit. |
198 | std::queue<DeclOrType> DeclTypesToEmit; |
199 | |
200 | /// The first ID number we can use for our own declarations. |
201 | serialization::DeclID FirstDeclID = serialization::NUM_PREDEF_DECL_IDS; |
202 | |
203 | /// The decl ID that will be assigned to the next new decl. |
204 | serialization::DeclID NextDeclID = FirstDeclID; |
205 | |
206 | /// Map that provides the ID numbers of each declaration within |
207 | /// the output stream, as well as those deserialized from a chained PCH. |
208 | /// |
209 | /// The ID numbers of declarations are consecutive (in order of |
210 | /// discovery) and start at 2. 1 is reserved for the translation |
211 | /// unit, while 0 is reserved for NULL. |
212 | llvm::DenseMap<const Decl *, serialization::DeclID> DeclIDs; |
213 | |
214 | /// Offset of each declaration in the bitstream, indexed by |
215 | /// the declaration's ID. |
216 | std::vector<serialization::DeclOffset> DeclOffsets; |
217 | |
218 | /// The offset of the DECLTYPES_BLOCK. The offsets in DeclOffsets |
219 | /// are relative to this value. |
220 | uint64_t DeclTypesBlockStartOffset = 0; |
221 | |
222 | /// Sorted (by file offset) vector of pairs of file offset/DeclID. |
223 | using LocDeclIDsTy = |
224 | SmallVector<std::pair<unsigned, serialization::DeclID>, 64>; |
225 | struct DeclIDInFileInfo { |
226 | LocDeclIDsTy DeclIDs; |
227 | |
228 | /// Set when the DeclIDs vectors from all files are joined, this |
229 | /// indicates the index that this particular vector has in the global one. |
230 | unsigned FirstDeclIndex; |
231 | }; |
232 | using FileDeclIDsTy = |
233 | llvm::DenseMap<FileID, std::unique_ptr<DeclIDInFileInfo>>; |
234 | |
235 | /// Map from file SLocEntries to info about the file-level declarations |
236 | /// that it contains. |
237 | FileDeclIDsTy FileDeclIDs; |
238 | |
239 | void associateDeclWithFile(const Decl *D, serialization::DeclID); |
240 | |
241 | /// The first ID number we can use for our own types. |
242 | serialization::TypeID FirstTypeID = serialization::NUM_PREDEF_TYPE_IDS; |
243 | |
244 | /// The type ID that will be assigned to the next new type. |
245 | serialization::TypeID NextTypeID = FirstTypeID; |
246 | |
247 | /// Map that provides the ID numbers of each type within the |
248 | /// output stream, plus those deserialized from a chained PCH. |
249 | /// |
250 | /// The ID numbers of types are consecutive (in order of discovery) |
251 | /// and start at 1. 0 is reserved for NULL. When types are actually |
252 | /// stored in the stream, the ID number is shifted by 2 bits to |
253 | /// allow for the const/volatile qualifiers. |
254 | /// |
255 | /// Keys in the map never have const/volatile qualifiers. |
256 | TypeIdxMap TypeIdxs; |
257 | |
258 | /// Offset of each type in the bitstream, indexed by |
259 | /// the type's ID. |
260 | std::vector<serialization::UnderalignedInt64> TypeOffsets; |
261 | |
262 | /// The first ID number we can use for our own identifiers. |
263 | serialization::IdentID FirstIdentID = serialization::NUM_PREDEF_IDENT_IDS; |
264 | |
265 | /// The identifier ID that will be assigned to the next new identifier. |
266 | serialization::IdentID NextIdentID = FirstIdentID; |
267 | |
268 | /// Map that provides the ID numbers of each identifier in |
269 | /// the output stream. |
270 | /// |
271 | /// The ID numbers for identifiers are consecutive (in order of |
272 | /// discovery), starting at 1. An ID of zero refers to a NULL |
273 | /// IdentifierInfo. |
274 | llvm::MapVector<const IdentifierInfo *, serialization::IdentID> IdentifierIDs; |
275 | |
276 | /// The first ID number we can use for our own macros. |
277 | serialization::MacroID FirstMacroID = serialization::NUM_PREDEF_MACRO_IDS; |
278 | |
279 | /// The identifier ID that will be assigned to the next new identifier. |
280 | serialization::MacroID NextMacroID = FirstMacroID; |
281 | |
282 | /// Map that provides the ID numbers of each macro. |
283 | llvm::DenseMap<MacroInfo *, serialization::MacroID> MacroIDs; |
284 | |
285 | struct MacroInfoToEmitData { |
286 | const IdentifierInfo *Name; |
287 | MacroInfo *MI; |
288 | serialization::MacroID ID; |
289 | }; |
290 | |
291 | /// The macro infos to emit. |
292 | std::vector<MacroInfoToEmitData> MacroInfosToEmit; |
293 | |
294 | llvm::DenseMap<const IdentifierInfo *, uint32_t> |
295 | IdentMacroDirectivesOffsetMap; |
296 | |
297 | /// @name FlushStmt Caches |
298 | /// @{ |
299 | |
300 | /// Set of parent Stmts for the currently serializing sub-stmt. |
301 | llvm::DenseSet<Stmt *> ParentStmts; |
302 | |
303 | /// Offsets of sub-stmts already serialized. The offset points |
304 | /// just after the stmt record. |
305 | llvm::DenseMap<Stmt *, uint64_t> SubStmtEntries; |
306 | |
307 | /// @} |
308 | |
309 | /// Offsets of each of the identifier IDs into the identifier |
310 | /// table. |
311 | std::vector<uint32_t> IdentifierOffsets; |
312 | |
313 | /// The first ID number we can use for our own submodules. |
314 | serialization::SubmoduleID FirstSubmoduleID = |
315 | serialization::NUM_PREDEF_SUBMODULE_IDS; |
316 | |
317 | /// The submodule ID that will be assigned to the next new submodule. |
318 | serialization::SubmoduleID NextSubmoduleID = FirstSubmoduleID; |
319 | |
320 | /// The first ID number we can use for our own selectors. |
321 | serialization::SelectorID FirstSelectorID = |
322 | serialization::NUM_PREDEF_SELECTOR_IDS; |
323 | |
324 | /// The selector ID that will be assigned to the next new selector. |
325 | serialization::SelectorID NextSelectorID = FirstSelectorID; |
326 | |
327 | /// Map that provides the ID numbers of each Selector. |
328 | llvm::MapVector<Selector, serialization::SelectorID> SelectorIDs; |
329 | |
330 | /// Offset of each selector within the method pool/selector |
331 | /// table, indexed by the Selector ID (-1). |
332 | std::vector<uint32_t> SelectorOffsets; |
333 | |
334 | /// Mapping from macro definitions (as they occur in the preprocessing |
335 | /// record) to the macro IDs. |
336 | llvm::DenseMap<const MacroDefinitionRecord *, |
337 | serialization::PreprocessedEntityID> MacroDefinitions; |
338 | |
339 | /// Cache of indices of anonymous declarations within their lexical |
340 | /// contexts. |
341 | llvm::DenseMap<const Decl *, unsigned> AnonymousDeclarationNumbers; |
342 | |
343 | /// An update to a Decl. |
344 | class DeclUpdate { |
345 | /// A DeclUpdateKind. |
346 | unsigned Kind; |
347 | union { |
348 | const Decl *Dcl; |
349 | void *Type; |
350 | SourceLocation::UIntTy Loc; |
351 | unsigned Val; |
352 | Module *Mod; |
353 | const Attr *Attribute; |
354 | }; |
355 | |
356 | public: |
357 | DeclUpdate(unsigned Kind) : Kind(Kind), Dcl(nullptr) {} |
358 | DeclUpdate(unsigned Kind, const Decl *Dcl) : Kind(Kind), Dcl(Dcl) {} |
359 | DeclUpdate(unsigned Kind, QualType Type) |
360 | : Kind(Kind), Type(Type.getAsOpaquePtr()) {} |
361 | DeclUpdate(unsigned Kind, SourceLocation Loc) |
362 | : Kind(Kind), Loc(Loc.getRawEncoding()) {} |
363 | DeclUpdate(unsigned Kind, unsigned Val) : Kind(Kind), Val(Val) {} |
364 | DeclUpdate(unsigned Kind, Module *M) : Kind(Kind), Mod(M) {} |
365 | DeclUpdate(unsigned Kind, const Attr *Attribute) |
366 | : Kind(Kind), Attribute(Attribute) {} |
367 | |
368 | unsigned getKind() const { return Kind; } |
369 | const Decl *getDecl() const { return Dcl; } |
370 | QualType getType() const { return QualType::getFromOpaquePtr(Ptr: Type); } |
371 | |
372 | SourceLocation getLoc() const { |
373 | return SourceLocation::getFromRawEncoding(Encoding: Loc); |
374 | } |
375 | |
376 | unsigned getNumber() const { return Val; } |
377 | Module *getModule() const { return Mod; } |
378 | const Attr *getAttr() const { return Attribute; } |
379 | }; |
380 | |
381 | using UpdateRecord = SmallVector<DeclUpdate, 1>; |
382 | using DeclUpdateMap = llvm::MapVector<const Decl *, UpdateRecord>; |
383 | |
384 | /// Mapping from declarations that came from a chained PCH to the |
385 | /// record containing modifications to them. |
386 | DeclUpdateMap DeclUpdates; |
387 | |
388 | using FirstLatestDeclMap = llvm::DenseMap<Decl *, Decl *>; |
389 | |
390 | /// Map of first declarations from a chained PCH that point to the |
391 | /// most recent declarations in another PCH. |
392 | FirstLatestDeclMap FirstLatestDecls; |
393 | |
394 | /// Declarations encountered that might be external |
395 | /// definitions. |
396 | /// |
397 | /// We keep track of external definitions and other 'interesting' declarations |
398 | /// as we are emitting declarations to the AST file. The AST file contains a |
399 | /// separate record for these declarations, which are provided to the AST |
400 | /// consumer by the AST reader. This is behavior is required to properly cope with, |
401 | /// e.g., tentative variable definitions that occur within |
402 | /// headers. The declarations themselves are stored as declaration |
403 | /// IDs, since they will be written out to an EAGERLY_DESERIALIZED_DECLS |
404 | /// record. |
405 | SmallVector<serialization::DeclID, 16> EagerlyDeserializedDecls; |
406 | SmallVector<serialization::DeclID, 16> ModularCodegenDecls; |
407 | |
408 | /// DeclContexts that have received extensions since their serialized |
409 | /// form. |
410 | /// |
411 | /// For namespaces, when we're chaining and encountering a namespace, we check |
412 | /// if its primary namespace comes from the chain. If it does, we add the |
413 | /// primary to this set, so that we can write out lexical content updates for |
414 | /// it. |
415 | llvm::SmallSetVector<const DeclContext *, 16> UpdatedDeclContexts; |
416 | |
417 | /// Keeps track of declarations that we must emit, even though we're |
418 | /// not guaranteed to be able to find them by walking the AST starting at the |
419 | /// translation unit. |
420 | SmallVector<const Decl *, 16> DeclsToEmitEvenIfUnreferenced; |
421 | |
422 | /// The set of Objective-C class that have categories we |
423 | /// should serialize. |
424 | llvm::SetVector<ObjCInterfaceDecl *> ObjCClassesWithCategories; |
425 | |
426 | /// The set of declarations that may have redeclaration chains that |
427 | /// need to be serialized. |
428 | llvm::SmallVector<const Decl *, 16> Redeclarations; |
429 | |
430 | /// A cache of the first local declaration for "interesting" |
431 | /// redeclaration chains. |
432 | llvm::DenseMap<const Decl *, const Decl *> FirstLocalDeclCache; |
433 | |
434 | /// Mapping from SwitchCase statements to IDs. |
435 | llvm::DenseMap<SwitchCase *, unsigned> SwitchCaseIDs; |
436 | |
437 | /// The number of statements written to the AST file. |
438 | unsigned NumStatements = 0; |
439 | |
440 | /// The number of macros written to the AST file. |
441 | unsigned NumMacros = 0; |
442 | |
443 | /// The number of lexical declcontexts written to the AST |
444 | /// file. |
445 | unsigned NumLexicalDeclContexts = 0; |
446 | |
447 | /// The number of visible declcontexts written to the AST |
448 | /// file. |
449 | unsigned NumVisibleDeclContexts = 0; |
450 | |
451 | /// A mapping from each known submodule to its ID number, which will |
452 | /// be a positive integer. |
453 | llvm::DenseMap<const Module *, unsigned> SubmoduleIDs; |
454 | |
455 | /// A list of the module file extension writers. |
456 | std::vector<std::unique_ptr<ModuleFileExtensionWriter>> |
457 | ModuleFileExtensionWriters; |
458 | |
459 | /// Mapping from a source location entry to whether it is affecting or not. |
460 | llvm::BitVector IsSLocAffecting; |
461 | |
462 | /// Mapping from \c FileID to an index into the FileID adjustment table. |
463 | std::vector<FileID> NonAffectingFileIDs; |
464 | std::vector<unsigned> NonAffectingFileIDAdjustments; |
465 | |
466 | /// Mapping from an offset to an index into the offset adjustment table. |
467 | std::vector<SourceRange> NonAffectingRanges; |
468 | std::vector<SourceLocation::UIntTy> NonAffectingOffsetAdjustments; |
469 | |
470 | /// Computes input files that didn't affect compilation of the current module, |
471 | /// and initializes data structures necessary for leaving those files out |
472 | /// during \c SourceManager serialization. |
473 | void computeNonAffectingInputFiles(); |
474 | |
475 | /// Returns an adjusted \c FileID, accounting for any non-affecting input |
476 | /// files. |
477 | FileID getAdjustedFileID(FileID FID) const; |
478 | /// Returns an adjusted number of \c FileIDs created within the specified \c |
479 | /// FileID, accounting for any non-affecting input files. |
480 | unsigned getAdjustedNumCreatedFIDs(FileID FID) const; |
481 | /// Returns an adjusted \c SourceLocation, accounting for any non-affecting |
482 | /// input files. |
483 | SourceLocation getAdjustedLocation(SourceLocation Loc) const; |
484 | /// Returns an adjusted \c SourceRange, accounting for any non-affecting input |
485 | /// files. |
486 | SourceRange getAdjustedRange(SourceRange Range) const; |
487 | /// Returns an adjusted \c SourceLocation offset, accounting for any |
488 | /// non-affecting input files. |
489 | SourceLocation::UIntTy getAdjustedOffset(SourceLocation::UIntTy Offset) const; |
490 | /// Returns an adjustment for offset into SourceManager, accounting for any |
491 | /// non-affecting input files. |
492 | SourceLocation::UIntTy getAdjustment(SourceLocation::UIntTy Offset) const; |
493 | |
494 | /// Retrieve or create a submodule ID for this module. |
495 | unsigned getSubmoduleID(Module *Mod); |
496 | |
497 | /// Write the given subexpression to the bitstream. |
498 | void WriteSubStmt(Stmt *S); |
499 | |
500 | void WriteBlockInfoBlock(); |
501 | void WriteControlBlock(Preprocessor &PP, ASTContext &Context, |
502 | StringRef isysroot); |
503 | |
504 | /// Write out the signature and diagnostic options, and return the signature. |
505 | void writeUnhashedControlBlock(Preprocessor &PP, ASTContext &Context); |
506 | ASTFileSignature backpatchSignature(); |
507 | |
508 | /// Calculate hash of the pcm content. |
509 | std::pair<ASTFileSignature, ASTFileSignature> createSignature() const; |
510 | |
511 | void (SourceManager &SourceMgr, HeaderSearchOptions &HSOpts); |
512 | void WriteSourceManagerBlock(SourceManager &SourceMgr, |
513 | const Preprocessor &PP); |
514 | void WritePreprocessor(const Preprocessor &PP, bool IsModule); |
515 | void (const HeaderSearch &HS); |
516 | void WritePreprocessorDetail(PreprocessingRecord &PPRec, |
517 | uint64_t MacroOffsetsBase); |
518 | void WriteSubmodules(Module *WritingModule); |
519 | |
520 | void WritePragmaDiagnosticMappings(const DiagnosticsEngine &Diag, |
521 | bool isModule); |
522 | |
523 | unsigned TypeExtQualAbbrev = 0; |
524 | void WriteTypeAbbrevs(); |
525 | void WriteType(QualType T); |
526 | |
527 | bool isLookupResultExternal(StoredDeclsList &Result, DeclContext *DC); |
528 | bool isLookupResultEntirelyExternal(StoredDeclsList &Result, DeclContext *DC); |
529 | |
530 | void GenerateNameLookupTable(const DeclContext *DC, |
531 | llvm::SmallVectorImpl<char> &LookupTable); |
532 | uint64_t WriteDeclContextLexicalBlock(ASTContext &Context, DeclContext *DC); |
533 | uint64_t WriteDeclContextVisibleBlock(ASTContext &Context, DeclContext *DC); |
534 | void WriteTypeDeclOffsets(); |
535 | void WriteFileDeclIDsMap(); |
536 | void (); |
537 | void WriteSelectors(Sema &SemaRef); |
538 | void WriteReferencedSelectorsPool(Sema &SemaRef); |
539 | void WriteIdentifierTable(Preprocessor &PP, IdentifierResolver &IdResolver, |
540 | bool IsModule); |
541 | void WriteDeclUpdatesBlocks(RecordDataImpl &OffsetsRecord); |
542 | void WriteDeclContextVisibleUpdate(const DeclContext *DC); |
543 | void WriteFPPragmaOptions(const FPOptionsOverride &Opts); |
544 | void WriteOpenCLExtensions(Sema &SemaRef); |
545 | void WriteCUDAPragmas(Sema &SemaRef); |
546 | void WriteObjCCategories(); |
547 | void WriteLateParsedTemplates(Sema &SemaRef); |
548 | void WriteOptimizePragmaOptions(Sema &SemaRef); |
549 | void WriteMSStructPragmaOptions(Sema &SemaRef); |
550 | void WriteMSPointersToMembersPragmaOptions(Sema &SemaRef); |
551 | void WritePackPragmaOptions(Sema &SemaRef); |
552 | void WriteFloatControlPragmaOptions(Sema &SemaRef); |
553 | void WriteModuleFileExtension(Sema &SemaRef, |
554 | ModuleFileExtensionWriter &Writer); |
555 | |
556 | unsigned DeclParmVarAbbrev = 0; |
557 | unsigned DeclContextLexicalAbbrev = 0; |
558 | unsigned DeclContextVisibleLookupAbbrev = 0; |
559 | unsigned UpdateVisibleAbbrev = 0; |
560 | unsigned DeclRecordAbbrev = 0; |
561 | unsigned DeclTypedefAbbrev = 0; |
562 | unsigned DeclVarAbbrev = 0; |
563 | unsigned DeclFieldAbbrev = 0; |
564 | unsigned DeclEnumAbbrev = 0; |
565 | unsigned DeclObjCIvarAbbrev = 0; |
566 | unsigned DeclCXXMethodAbbrev = 0; |
567 | unsigned DeclDependentNonTemplateCXXMethodAbbrev = 0; |
568 | unsigned DeclTemplateCXXMethodAbbrev = 0; |
569 | unsigned DeclMemberSpecializedCXXMethodAbbrev = 0; |
570 | unsigned DeclTemplateSpecializedCXXMethodAbbrev = 0; |
571 | unsigned DeclDependentSpecializationCXXMethodAbbrev = 0; |
572 | unsigned DeclTemplateTypeParmAbbrev = 0; |
573 | unsigned DeclUsingShadowAbbrev = 0; |
574 | |
575 | unsigned DeclRefExprAbbrev = 0; |
576 | unsigned CharacterLiteralAbbrev = 0; |
577 | unsigned IntegerLiteralAbbrev = 0; |
578 | unsigned ExprImplicitCastAbbrev = 0; |
579 | unsigned BinaryOperatorAbbrev = 0; |
580 | unsigned CompoundAssignOperatorAbbrev = 0; |
581 | unsigned CallExprAbbrev = 0; |
582 | unsigned CXXOperatorCallExprAbbrev = 0; |
583 | unsigned CXXMemberCallExprAbbrev = 0; |
584 | |
585 | unsigned CompoundStmtAbbrev = 0; |
586 | |
587 | void WriteDeclAbbrevs(); |
588 | void WriteDecl(ASTContext &Context, Decl *D); |
589 | |
590 | ASTFileSignature WriteASTCore(Sema &SemaRef, StringRef isysroot, |
591 | Module *WritingModule); |
592 | |
593 | public: |
594 | /// Create a new precompiled header writer that outputs to |
595 | /// the given bitstream. |
596 | ASTWriter(llvm::BitstreamWriter &Stream, SmallVectorImpl<char> &Buffer, |
597 | InMemoryModuleCache &ModuleCache, |
598 | ArrayRef<std::shared_ptr<ModuleFileExtension>> Extensions, |
599 | bool IncludeTimestamps = true, bool BuildingImplicitModule = false); |
600 | ~ASTWriter() override; |
601 | |
602 | ASTContext &getASTContext() const { |
603 | assert(Context && "requested AST context when not writing AST" ); |
604 | return *Context; |
605 | } |
606 | |
607 | const LangOptions &getLangOpts() const; |
608 | |
609 | /// Get a timestamp for output into the AST file. The actual timestamp |
610 | /// of the specified file may be ignored if we have been instructed to not |
611 | /// include timestamps in the output file. |
612 | time_t getTimestampForOutput(const FileEntry *E) const; |
613 | |
614 | /// Write a precompiled header for the given semantic analysis. |
615 | /// |
616 | /// \param SemaRef a reference to the semantic analysis object that processed |
617 | /// the AST to be written into the precompiled header. |
618 | /// |
619 | /// \param WritingModule The module that we are writing. If null, we are |
620 | /// writing a precompiled header. |
621 | /// |
622 | /// \param isysroot if non-empty, write a relocatable file whose headers |
623 | /// are relative to the given system root. If we're writing a module, its |
624 | /// build directory will be used in preference to this if both are available. |
625 | /// |
626 | /// \return the module signature, which eventually will be a hash of |
627 | /// the module but currently is merely a random 32-bit number. |
628 | ASTFileSignature WriteAST(Sema &SemaRef, StringRef OutputFile, |
629 | Module *WritingModule, StringRef isysroot, |
630 | bool ShouldCacheASTInMemory = false); |
631 | |
632 | /// Emit a token. |
633 | void AddToken(const Token &Tok, RecordDataImpl &Record); |
634 | |
635 | /// Emit a AlignPackInfo. |
636 | void AddAlignPackInfo(const Sema::AlignPackInfo &Info, |
637 | RecordDataImpl &Record); |
638 | |
639 | /// Emit a FileID. |
640 | void AddFileID(FileID FID, RecordDataImpl &Record); |
641 | |
642 | /// Emit a source location. |
643 | void AddSourceLocation(SourceLocation Loc, RecordDataImpl &Record, |
644 | LocSeq *Seq = nullptr); |
645 | |
646 | /// Emit a source range. |
647 | void AddSourceRange(SourceRange Range, RecordDataImpl &Record, |
648 | LocSeq *Seq = nullptr); |
649 | |
650 | /// Emit a reference to an identifier. |
651 | void AddIdentifierRef(const IdentifierInfo *II, RecordDataImpl &Record); |
652 | |
653 | /// Get the unique number used to refer to the given selector. |
654 | serialization::SelectorID getSelectorRef(Selector Sel); |
655 | |
656 | /// Get the unique number used to refer to the given identifier. |
657 | serialization::IdentID getIdentifierRef(const IdentifierInfo *II); |
658 | |
659 | /// Get the unique number used to refer to the given macro. |
660 | serialization::MacroID getMacroRef(MacroInfo *MI, const IdentifierInfo *Name); |
661 | |
662 | /// Determine the ID of an already-emitted macro. |
663 | serialization::MacroID getMacroID(MacroInfo *MI); |
664 | |
665 | uint32_t getMacroDirectivesOffset(const IdentifierInfo *Name); |
666 | |
667 | /// Emit a reference to a type. |
668 | void AddTypeRef(QualType T, RecordDataImpl &Record); |
669 | |
670 | /// Force a type to be emitted and get its ID. |
671 | serialization::TypeID GetOrCreateTypeID(QualType T); |
672 | |
673 | /// Determine the type ID of an already-emitted type. |
674 | serialization::TypeID getTypeID(QualType T) const; |
675 | |
676 | /// Find the first local declaration of a given local redeclarable |
677 | /// decl. |
678 | const Decl *getFirstLocalDecl(const Decl *D); |
679 | |
680 | /// Is this a local declaration (that is, one that will be written to |
681 | /// our AST file)? This is the case for declarations that are neither imported |
682 | /// from another AST file nor predefined. |
683 | bool IsLocalDecl(const Decl *D) { |
684 | if (D->isFromASTFile()) |
685 | return false; |
686 | auto I = DeclIDs.find(Val: D); |
687 | return (I == DeclIDs.end() || |
688 | I->second >= serialization::NUM_PREDEF_DECL_IDS); |
689 | }; |
690 | |
691 | /// Emit a reference to a declaration. |
692 | void AddDeclRef(const Decl *D, RecordDataImpl &Record); |
693 | |
694 | /// Force a declaration to be emitted and get its ID. |
695 | serialization::DeclID GetDeclRef(const Decl *D); |
696 | |
697 | /// Determine the declaration ID of an already-emitted |
698 | /// declaration. |
699 | serialization::DeclID getDeclID(const Decl *D); |
700 | |
701 | unsigned getAnonymousDeclarationNumber(const NamedDecl *D); |
702 | |
703 | /// Add a string to the given record. |
704 | void AddString(StringRef Str, RecordDataImpl &Record); |
705 | |
706 | /// Convert a path from this build process into one that is appropriate |
707 | /// for emission in the module file. |
708 | bool PreparePathForOutput(SmallVectorImpl<char> &Path); |
709 | |
710 | /// Add a path to the given record. |
711 | void AddPath(StringRef Path, RecordDataImpl &Record); |
712 | |
713 | /// Emit the current record with the given path as a blob. |
714 | void EmitRecordWithPath(unsigned Abbrev, RecordDataRef Record, |
715 | StringRef Path); |
716 | |
717 | /// Add a version tuple to the given record |
718 | void AddVersionTuple(const VersionTuple &Version, RecordDataImpl &Record); |
719 | |
720 | /// Retrieve or create a submodule ID for this module, or return 0 if |
721 | /// the submodule is neither local (a submodle of the currently-written module) |
722 | /// nor from an imported module. |
723 | unsigned getLocalOrImportedSubmoduleID(const Module *Mod); |
724 | |
725 | /// Note that the identifier II occurs at the given offset |
726 | /// within the identifier table. |
727 | void SetIdentifierOffset(const IdentifierInfo *II, uint32_t Offset); |
728 | |
729 | /// Note that the selector Sel occurs at the given offset |
730 | /// within the method pool/selector table. |
731 | void SetSelectorOffset(Selector Sel, uint32_t Offset); |
732 | |
733 | /// Record an ID for the given switch-case statement. |
734 | unsigned RecordSwitchCaseID(SwitchCase *S); |
735 | |
736 | /// Retrieve the ID for the given switch-case statement. |
737 | unsigned getSwitchCaseID(SwitchCase *S); |
738 | |
739 | void ClearSwitchCaseIDs(); |
740 | |
741 | unsigned getTypeExtQualAbbrev() const { |
742 | return TypeExtQualAbbrev; |
743 | } |
744 | |
745 | unsigned getDeclParmVarAbbrev() const { return DeclParmVarAbbrev; } |
746 | unsigned getDeclRecordAbbrev() const { return DeclRecordAbbrev; } |
747 | unsigned getDeclTypedefAbbrev() const { return DeclTypedefAbbrev; } |
748 | unsigned getDeclVarAbbrev() const { return DeclVarAbbrev; } |
749 | unsigned getDeclFieldAbbrev() const { return DeclFieldAbbrev; } |
750 | unsigned getDeclEnumAbbrev() const { return DeclEnumAbbrev; } |
751 | unsigned getDeclObjCIvarAbbrev() const { return DeclObjCIvarAbbrev; } |
752 | unsigned getDeclCXXMethodAbbrev(FunctionDecl::TemplatedKind Kind) const { |
753 | switch (Kind) { |
754 | case FunctionDecl::TK_NonTemplate: |
755 | return DeclCXXMethodAbbrev; |
756 | case FunctionDecl::TK_FunctionTemplate: |
757 | return DeclTemplateCXXMethodAbbrev; |
758 | case FunctionDecl::TK_MemberSpecialization: |
759 | return DeclMemberSpecializedCXXMethodAbbrev; |
760 | case FunctionDecl::TK_FunctionTemplateSpecialization: |
761 | return DeclTemplateSpecializedCXXMethodAbbrev; |
762 | case FunctionDecl::TK_DependentNonTemplate: |
763 | return DeclDependentNonTemplateCXXMethodAbbrev; |
764 | case FunctionDecl::TK_DependentFunctionTemplateSpecialization: |
765 | return DeclDependentSpecializationCXXMethodAbbrev; |
766 | } |
767 | llvm_unreachable("Unknwon Template Kind!" ); |
768 | } |
769 | unsigned getDeclTemplateTypeParmAbbrev() const { |
770 | return DeclTemplateTypeParmAbbrev; |
771 | } |
772 | unsigned getDeclUsingShadowAbbrev() const { return DeclUsingShadowAbbrev; } |
773 | |
774 | unsigned getDeclRefExprAbbrev() const { return DeclRefExprAbbrev; } |
775 | unsigned getCharacterLiteralAbbrev() const { return CharacterLiteralAbbrev; } |
776 | unsigned getIntegerLiteralAbbrev() const { return IntegerLiteralAbbrev; } |
777 | unsigned getExprImplicitCastAbbrev() const { return ExprImplicitCastAbbrev; } |
778 | unsigned getBinaryOperatorAbbrev() const { return BinaryOperatorAbbrev; } |
779 | unsigned getCompoundAssignOperatorAbbrev() const { |
780 | return CompoundAssignOperatorAbbrev; |
781 | } |
782 | unsigned getCallExprAbbrev() const { return CallExprAbbrev; } |
783 | unsigned getCXXOperatorCallExprAbbrev() { return CXXOperatorCallExprAbbrev; } |
784 | unsigned getCXXMemberCallExprAbbrev() { return CXXMemberCallExprAbbrev; } |
785 | |
786 | unsigned getCompoundStmtAbbrev() const { return CompoundStmtAbbrev; } |
787 | |
788 | bool hasChain() const { return Chain; } |
789 | ASTReader *getChain() const { return Chain; } |
790 | |
791 | bool isWritingStdCXXNamedModules() const { |
792 | return WritingModule && WritingModule->isNamedModule(); |
793 | } |
794 | |
795 | private: |
796 | // ASTDeserializationListener implementation |
797 | void ReaderInitialized(ASTReader *Reader) override; |
798 | void IdentifierRead(serialization::IdentID ID, IdentifierInfo *II) override; |
799 | void MacroRead(serialization::MacroID ID, MacroInfo *MI) override; |
800 | void TypeRead(serialization::TypeIdx Idx, QualType T) override; |
801 | void SelectorRead(serialization::SelectorID ID, Selector Sel) override; |
802 | void MacroDefinitionRead(serialization::PreprocessedEntityID ID, |
803 | MacroDefinitionRecord *MD) override; |
804 | void ModuleRead(serialization::SubmoduleID ID, Module *Mod) override; |
805 | |
806 | // ASTMutationListener implementation. |
807 | void CompletedTagDefinition(const TagDecl *D) override; |
808 | void AddedVisibleDecl(const DeclContext *DC, const Decl *D) override; |
809 | void AddedCXXImplicitMember(const CXXRecordDecl *RD, const Decl *D) override; |
810 | void AddedCXXTemplateSpecialization( |
811 | const ClassTemplateDecl *TD, |
812 | const ClassTemplateSpecializationDecl *D) override; |
813 | void AddedCXXTemplateSpecialization( |
814 | const VarTemplateDecl *TD, |
815 | const VarTemplateSpecializationDecl *D) override; |
816 | void AddedCXXTemplateSpecialization(const FunctionTemplateDecl *TD, |
817 | const FunctionDecl *D) override; |
818 | void ResolvedExceptionSpec(const FunctionDecl *FD) override; |
819 | void DeducedReturnType(const FunctionDecl *FD, QualType ReturnType) override; |
820 | void ResolvedOperatorDelete(const CXXDestructorDecl *DD, |
821 | const FunctionDecl *Delete, |
822 | Expr *ThisArg) override; |
823 | void CompletedImplicitDefinition(const FunctionDecl *D) override; |
824 | void InstantiationRequested(const ValueDecl *D) override; |
825 | void VariableDefinitionInstantiated(const VarDecl *D) override; |
826 | void FunctionDefinitionInstantiated(const FunctionDecl *D) override; |
827 | void DefaultArgumentInstantiated(const ParmVarDecl *D) override; |
828 | void DefaultMemberInitializerInstantiated(const FieldDecl *D) override; |
829 | void AddedObjCCategoryToInterface(const ObjCCategoryDecl *CatD, |
830 | const ObjCInterfaceDecl *IFD) override; |
831 | void DeclarationMarkedUsed(const Decl *D) override; |
832 | void DeclarationMarkedOpenMPThreadPrivate(const Decl *D) override; |
833 | void DeclarationMarkedOpenMPDeclareTarget(const Decl *D, |
834 | const Attr *Attr) override; |
835 | void DeclarationMarkedOpenMPAllocate(const Decl *D, const Attr *A) override; |
836 | void RedefinedHiddenDefinition(const NamedDecl *D, Module *M) override; |
837 | void AddedAttributeToRecord(const Attr *Attr, |
838 | const RecordDecl *Record) override; |
839 | }; |
840 | |
841 | /// AST and semantic-analysis consumer that generates a |
842 | /// precompiled header from the parsed source code. |
843 | class PCHGenerator : public SemaConsumer { |
844 | const Preprocessor &PP; |
845 | std::string OutputFile; |
846 | std::string isysroot; |
847 | Sema *SemaPtr; |
848 | std::shared_ptr<PCHBuffer> Buffer; |
849 | llvm::BitstreamWriter Stream; |
850 | ASTWriter Writer; |
851 | bool AllowASTWithErrors; |
852 | bool ShouldCacheASTInMemory; |
853 | |
854 | protected: |
855 | ASTWriter &getWriter() { return Writer; } |
856 | const ASTWriter &getWriter() const { return Writer; } |
857 | SmallVectorImpl<char> &getPCH() const { return Buffer->Data; } |
858 | |
859 | public: |
860 | PCHGenerator(const Preprocessor &PP, InMemoryModuleCache &ModuleCache, |
861 | StringRef OutputFile, StringRef isysroot, |
862 | std::shared_ptr<PCHBuffer> Buffer, |
863 | ArrayRef<std::shared_ptr<ModuleFileExtension>> Extensions, |
864 | bool AllowASTWithErrors = false, bool IncludeTimestamps = true, |
865 | bool BuildingImplicitModule = false, |
866 | bool ShouldCacheASTInMemory = false); |
867 | ~PCHGenerator() override; |
868 | |
869 | void InitializeSema(Sema &S) override { SemaPtr = &S; } |
870 | void HandleTranslationUnit(ASTContext &Ctx) override; |
871 | ASTMutationListener *GetASTMutationListener() override; |
872 | ASTDeserializationListener *GetASTDeserializationListener() override; |
873 | bool hasEmittedPCH() const { return Buffer->IsComplete; } |
874 | }; |
875 | |
876 | /// A simple helper class to pack several bits in order into (a) 32 bit |
877 | /// integer(s). |
878 | class BitsPacker { |
879 | constexpr static uint32_t BitIndexUpbound = 32u; |
880 | |
881 | public: |
882 | BitsPacker() = default; |
883 | BitsPacker(const BitsPacker &) = delete; |
884 | BitsPacker(BitsPacker &&) = delete; |
885 | BitsPacker operator=(const BitsPacker &) = delete; |
886 | BitsPacker operator=(BitsPacker &&) = delete; |
887 | ~BitsPacker() = default; |
888 | |
889 | bool canWriteNextNBits(uint32_t BitsWidth) const { |
890 | return CurrentBitIndex + BitsWidth < BitIndexUpbound; |
891 | } |
892 | |
893 | void reset(uint32_t Value) { |
894 | UnderlyingValue = Value; |
895 | CurrentBitIndex = 0; |
896 | } |
897 | |
898 | void addBit(bool Value) { addBits(Value, BitsWidth: 1); } |
899 | void addBits(uint32_t Value, uint32_t BitsWidth) { |
900 | assert(BitsWidth < BitIndexUpbound); |
901 | assert((Value < (1u << BitsWidth)) && "Passing narrower bit width!" ); |
902 | assert(canWriteNextNBits(BitsWidth) && |
903 | "Inserting too much bits into a value!" ); |
904 | |
905 | UnderlyingValue |= Value << CurrentBitIndex; |
906 | CurrentBitIndex += BitsWidth; |
907 | } |
908 | |
909 | operator uint32_t() { return UnderlyingValue; } |
910 | |
911 | private: |
912 | uint32_t UnderlyingValue = 0; |
913 | uint32_t CurrentBitIndex = 0; |
914 | }; |
915 | |
916 | } // namespace clang |
917 | |
918 | #endif // LLVM_CLANG_SERIALIZATION_ASTWRITER_H |
919 | |