1//===--- SourceManager.h - Track and cache source files ---------*- 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/// \file
11/// \brief Defines the SourceManager interface.
12///
13/// There are three different types of locations in a %file: a spelling
14/// location, an expansion location, and a presumed location.
15///
16/// Given an example of:
17/// \code
18/// #define min(x, y) x < y ? x : y
19/// \endcode
20///
21/// and then later on a use of min:
22/// \code
23/// #line 17
24/// return min(a, b);
25/// \endcode
26///
27/// The expansion location is the line in the source code where the macro
28/// was expanded (the return statement), the spelling location is the
29/// location in the source where the macro was originally defined,
30/// and the presumed location is where the line directive states that
31/// the line is 17, or any other line.
32///
33//===----------------------------------------------------------------------===//
34
35#ifndef LLVM_CLANG_BASIC_SOURCEMANAGER_H
36#define LLVM_CLANG_BASIC_SOURCEMANAGER_H
37
38#include "clang/Basic/FileManager.h"
39#include "clang/Basic/LLVM.h"
40#include "clang/Basic/SourceLocation.h"
41#include "llvm/ADT/ArrayRef.h"
42#include "llvm/ADT/BitVector.h"
43#include "llvm/ADT/DenseMap.h"
44#include "llvm/ADT/DenseSet.h"
45#include "llvm/ADT/IntrusiveRefCntPtr.h"
46#include "llvm/ADT/PointerIntPair.h"
47#include "llvm/ADT/SmallVector.h"
48#include "llvm/ADT/StringRef.h"
49#include "llvm/Support/Allocator.h"
50#include "llvm/Support/Compiler.h"
51#include "llvm/Support/MemoryBuffer.h"
52#include <algorithm>
53#include <cassert>
54#include <cstddef>
55#include <cstdint>
56#include <map>
57#include <memory>
58#include <string>
59#include <utility>
60#include <vector>
61
62namespace clang {
63
64class ASTReader;
65class ASTWriter;
66class DiagnosticsEngine;
67class LineTableInfo;
68class SourceManager;
69
70/// \brief Public enums and private classes that are part of the
71/// SourceManager implementation.
72///
73namespace SrcMgr {
74
75 /// \brief Indicates whether a file or directory holds normal user code,
76 /// system code, or system code which is implicitly 'extern "C"' in C++ mode.
77 ///
78 /// Entire directories can be tagged with this (this is maintained by
79 /// DirectoryLookup and friends) as can specific FileInfos when a \#pragma
80 /// system_header is seen or in various other cases.
81 ///
82 enum CharacteristicKind {
83 C_User, C_System, C_ExternCSystem
84 };
85
86 /// \brief One instance of this struct is kept for every file loaded or used.
87 ///
88 /// This object owns the MemoryBuffer object.
89 class LLVM_ALIGNAS(8) ContentCache {
90 enum CCFlags {
91 /// \brief Whether the buffer is invalid.
92 InvalidFlag = 0x01,
93 /// \brief Whether the buffer should not be freed on destruction.
94 DoNotFreeFlag = 0x02
95 };
96
97 /// \brief The actual buffer containing the characters from the input
98 /// file.
99 ///
100 /// This is owned by the ContentCache object. The bits indicate
101 /// whether the buffer is invalid.
102 mutable llvm::PointerIntPair<llvm::MemoryBuffer *, 2> Buffer;
103
104 public:
105 /// \brief Reference to the file entry representing this ContentCache.
106 ///
107 /// This reference does not own the FileEntry object.
108 ///
109 /// It is possible for this to be NULL if the ContentCache encapsulates
110 /// an imaginary text buffer.
111 const FileEntry *OrigEntry;
112
113 /// \brief References the file which the contents were actually loaded from.
114 ///
115 /// Can be different from 'Entry' if we overridden the contents of one file
116 /// with the contents of another file.
117 const FileEntry *ContentsEntry;
118
119 /// \brief A bump pointer allocated array of offsets for each source line.
120 ///
121 /// This is lazily computed. This is owned by the SourceManager
122 /// BumpPointerAllocator object.
123 unsigned *SourceLineCache;
124
125 /// \brief The number of lines in this ContentCache.
126 ///
127 /// This is only valid if SourceLineCache is non-null.
128 unsigned NumLines;
129
130 /// \brief Indicates whether the buffer itself was provided to override
131 /// the actual file contents.
132 ///
133 /// When true, the original entry may be a virtual file that does not
134 /// exist.
135 unsigned BufferOverridden : 1;
136
137 /// \brief True if this content cache was initially created for a source
138 /// file considered as a system one.
139 unsigned IsSystemFile : 1;
140
141 /// \brief True if this file may be transient, that is, if it might not
142 /// exist at some later point in time when this content entry is used,
143 /// after serialization and deserialization.
144 unsigned IsTransient : 1;
145
146 ContentCache(const FileEntry *Ent = nullptr) : ContentCache(Ent, Ent) {}
147
148 ContentCache(const FileEntry *Ent, const FileEntry *contentEnt)
149 : Buffer(nullptr, false), OrigEntry(Ent), ContentsEntry(contentEnt),
150 SourceLineCache(nullptr), NumLines(0), BufferOverridden(false),
151 IsSystemFile(false), IsTransient(false) {}
152
153 /// The copy ctor does not allow copies where source object has either
154 /// a non-NULL Buffer or SourceLineCache. Ownership of allocated memory
155 /// is not transferred, so this is a logical error.
156 ContentCache(const ContentCache &RHS)
157 : Buffer(nullptr, false), SourceLineCache(nullptr),
158 BufferOverridden(false), IsSystemFile(false), IsTransient(false) {
159 OrigEntry = RHS.OrigEntry;
160 ContentsEntry = RHS.ContentsEntry;
161
162 assert(RHS.Buffer.getPointer() == nullptr &&
163 RHS.SourceLineCache == nullptr &&
164 "Passed ContentCache object cannot own a buffer.");
165
166 NumLines = RHS.NumLines;
167 }
168
169 ContentCache &operator=(const ContentCache& RHS) = delete;
170
171 ~ContentCache();
172
173 /// \brief Returns the memory buffer for the associated content.
174 ///
175 /// \param Diag Object through which diagnostics will be emitted if the
176 /// buffer cannot be retrieved.
177 ///
178 /// \param Loc If specified, is the location that invalid file diagnostics
179 /// will be emitted at.
180 ///
181 /// \param Invalid If non-NULL, will be set \c true if an error occurred.
182 llvm::MemoryBuffer *getBuffer(DiagnosticsEngine &Diag,
183 const SourceManager &SM,
184 SourceLocation Loc = SourceLocation(),
185 bool *Invalid = nullptr) const;
186
187 /// \brief Returns the size of the content encapsulated by this
188 /// ContentCache.
189 ///
190 /// This can be the size of the source file or the size of an
191 /// arbitrary scratch buffer. If the ContentCache encapsulates a source
192 /// file this size is retrieved from the file's FileEntry.
193 unsigned getSize() const;
194
195 /// \brief Returns the number of bytes actually mapped for this
196 /// ContentCache.
197 ///
198 /// This can be 0 if the MemBuffer was not actually expanded.
199 unsigned getSizeBytesMapped() const;
200
201 /// Returns the kind of memory used to back the memory buffer for
202 /// this content cache. This is used for performance analysis.
203 llvm::MemoryBuffer::BufferKind getMemoryBufferKind() const;
204
205 void setBuffer(std::unique_ptr<llvm::MemoryBuffer> B) {
206 assert(!Buffer.getPointer() && "MemoryBuffer already set.");
207 Buffer.setPointer(B.release());
208 Buffer.setInt(0);
209 }
210
211 /// \brief Get the underlying buffer, returning NULL if the buffer is not
212 /// yet available.
213 llvm::MemoryBuffer *getRawBuffer() const { return Buffer.getPointer(); }
214
215 /// \brief Replace the existing buffer (which will be deleted)
216 /// with the given buffer.
217 void replaceBuffer(llvm::MemoryBuffer *B, bool DoNotFree = false);
218
219 /// \brief Determine whether the buffer itself is invalid.
220 bool isBufferInvalid() const {
221 return Buffer.getInt() & InvalidFlag;
222 }
223
224 /// \brief Determine whether the buffer should be freed.
225 bool shouldFreeBuffer() const {
226 return (Buffer.getInt() & DoNotFreeFlag) == 0;
227 }
228 };
229
230 // Assert that the \c ContentCache objects will always be 8-byte aligned so
231 // that we can pack 3 bits of integer into pointers to such objects.
232 static_assert(alignof(ContentCache) >= 8,
233 "ContentCache must be 8-byte aligned.");
234
235 /// \brief Information about a FileID, basically just the logical file
236 /// that it represents and include stack information.
237 ///
238 /// Each FileInfo has include stack information, indicating where it came
239 /// from. This information encodes the \#include chain that a token was
240 /// expanded from. The main include file has an invalid IncludeLoc.
241 ///
242 /// FileInfos contain a "ContentCache *", with the contents of the file.
243 ///
244 class FileInfo {
245 /// \brief The location of the \#include that brought in this file.
246 ///
247 /// This is an invalid SLOC for the main file (top of the \#include chain).
248 unsigned IncludeLoc; // Really a SourceLocation
249
250 /// \brief Number of FileIDs (files and macros) that were created during
251 /// preprocessing of this \#include, including this SLocEntry.
252 ///
253 /// Zero means the preprocessor didn't provide such info for this SLocEntry.
254 unsigned NumCreatedFIDs;
255
256 /// \brief Contains the ContentCache* and the bits indicating the
257 /// characteristic of the file and whether it has \#line info, all
258 /// bitmangled together.
259 uintptr_t Data;
260
261 friend class clang::SourceManager;
262 friend class clang::ASTWriter;
263 friend class clang::ASTReader;
264
265 public:
266 /// \brief Return a FileInfo object.
267 static FileInfo get(SourceLocation IL, const ContentCache *Con,
268 CharacteristicKind FileCharacter) {
269 FileInfo X;
270 X.IncludeLoc = IL.getRawEncoding();
271 X.NumCreatedFIDs = 0;
272 X.Data = (uintptr_t)Con;
273 assert((X.Data & 7) == 0 &&"ContentCache pointer insufficiently aligned");
274 assert((unsigned)FileCharacter < 4 && "invalid file character");
275 X.Data |= (unsigned)FileCharacter;
276 return X;
277 }
278
279 SourceLocation getIncludeLoc() const {
280 return SourceLocation::getFromRawEncoding(IncludeLoc);
281 }
282
283 const ContentCache* getContentCache() const {
284 return reinterpret_cast<const ContentCache*>(Data & ~uintptr_t(7));
285 }
286
287 /// \brief Return whether this is a system header or not.
288 CharacteristicKind getFileCharacteristic() const {
289 return (CharacteristicKind)(Data & 3);
290 }
291
292 /// \brief Return true if this FileID has \#line directives in it.
293 bool hasLineDirectives() const { return (Data & 4) != 0; }
294
295 /// \brief Set the flag that indicates that this FileID has
296 /// line table entries associated with it.
297 void setHasLineDirectives() {
298 Data |= 4;
299 }
300 };
301
302 /// \brief Each ExpansionInfo encodes the expansion location - where
303 /// the token was ultimately expanded, and the SpellingLoc - where the actual
304 /// character data for the token came from.
305 class ExpansionInfo {
306 // Really these are all SourceLocations.
307
308 /// \brief Where the spelling for the token can be found.
309 unsigned SpellingLoc;
310
311 /// In a macro expansion, ExpansionLocStart and ExpansionLocEnd
312 /// indicate the start and end of the expansion. In object-like macros,
313 /// they will be the same. In a function-like macro expansion, the start
314 /// will be the identifier and the end will be the ')'. Finally, in
315 /// macro-argument instantiations, the end will be 'SourceLocation()', an
316 /// invalid location.
317 unsigned ExpansionLocStart, ExpansionLocEnd;
318
319 public:
320 SourceLocation getSpellingLoc() const {
321 return SourceLocation::getFromRawEncoding(SpellingLoc);
322 }
323
324 SourceLocation getExpansionLocStart() const {
325 return SourceLocation::getFromRawEncoding(ExpansionLocStart);
326 }
327
328 SourceLocation getExpansionLocEnd() const {
329 SourceLocation EndLoc =
330 SourceLocation::getFromRawEncoding(ExpansionLocEnd);
331 return EndLoc.isInvalid() ? getExpansionLocStart() : EndLoc;
332 }
333
334 std::pair<SourceLocation,SourceLocation> getExpansionLocRange() const {
335 return std::make_pair(getExpansionLocStart(), getExpansionLocEnd());
336 }
337
338 bool isMacroArgExpansion() const {
339 // Note that this needs to return false for default constructed objects.
340 return getExpansionLocStart().isValid() &&
341 SourceLocation::getFromRawEncoding(ExpansionLocEnd).isInvalid();
342 }
343
344 bool isMacroBodyExpansion() const {
345 return getExpansionLocStart().isValid() &&
346 SourceLocation::getFromRawEncoding(ExpansionLocEnd).isValid();
347 }
348
349 bool isFunctionMacroExpansion() const {
350 return getExpansionLocStart().isValid() &&
351 getExpansionLocStart() != getExpansionLocEnd();
352 }
353
354 /// \brief Return a ExpansionInfo for an expansion.
355 ///
356 /// Start and End specify the expansion range (where the macro is
357 /// expanded), and SpellingLoc specifies the spelling location (where
358 /// the characters from the token come from). All three can refer to
359 /// normal File SLocs or expansion locations.
360 static ExpansionInfo create(SourceLocation SpellingLoc,
361 SourceLocation Start, SourceLocation End) {
362 ExpansionInfo X;
363 X.SpellingLoc = SpellingLoc.getRawEncoding();
364 X.ExpansionLocStart = Start.getRawEncoding();
365 X.ExpansionLocEnd = End.getRawEncoding();
366 return X;
367 }
368
369 /// \brief Return a special ExpansionInfo for the expansion of
370 /// a macro argument into a function-like macro's body.
371 ///
372 /// ExpansionLoc specifies the expansion location (where the macro is
373 /// expanded). This doesn't need to be a range because a macro is always
374 /// expanded at a macro parameter reference, and macro parameters are
375 /// always exactly one token. SpellingLoc specifies the spelling location
376 /// (where the characters from the token come from). ExpansionLoc and
377 /// SpellingLoc can both refer to normal File SLocs or expansion locations.
378 ///
379 /// Given the code:
380 /// \code
381 /// #define F(x) f(x)
382 /// F(42);
383 /// \endcode
384 ///
385 /// When expanding '\c F(42)', the '\c x' would call this with an
386 /// SpellingLoc pointing at '\c 42' and an ExpansionLoc pointing at its
387 /// location in the definition of '\c F'.
388 static ExpansionInfo createForMacroArg(SourceLocation SpellingLoc,
389 SourceLocation ExpansionLoc) {
390 // We store an intentionally invalid source location for the end of the
391 // expansion range to mark that this is a macro argument ion rather than
392 // a normal one.
393 return create(SpellingLoc, ExpansionLoc, SourceLocation());
394 }
395 };
396
397 /// \brief This is a discriminated union of FileInfo and ExpansionInfo.
398 ///
399 /// SourceManager keeps an array of these objects, and they are uniquely
400 /// identified by the FileID datatype.
401 class SLocEntry {
402 unsigned Offset : 31;
403 unsigned IsExpansion : 1;
404 union {
405 FileInfo File;
406 ExpansionInfo Expansion;
407 };
408
409 public:
410 unsigned getOffset() const { return Offset; }
411
412 bool isExpansion() const { return IsExpansion; }
413 bool isFile() const { return !isExpansion(); }
414
415 const FileInfo &getFile() const {
416 assert(isFile() && "Not a file SLocEntry!");
417 return File;
418 }
419
420 const ExpansionInfo &getExpansion() const {
421 assert(isExpansion() && "Not a macro expansion SLocEntry!");
422 return Expansion;
423 }
424
425 static SLocEntry get(unsigned Offset, const FileInfo &FI) {
426 assert(!(Offset & (1 << 31)) && "Offset is too large");
427 SLocEntry E;
428 E.Offset = Offset;
429 E.IsExpansion = false;
430 E.File = FI;
431 return E;
432 }
433
434 static SLocEntry get(unsigned Offset, const ExpansionInfo &Expansion) {
435 assert(!(Offset & (1 << 31)) && "Offset is too large");
436 SLocEntry E;
437 E.Offset = Offset;
438 E.IsExpansion = true;
439 E.Expansion = Expansion;
440 return E;
441 }
442 };
443
444} // end SrcMgr namespace.
445
446/// \brief External source of source location entries.
447class ExternalSLocEntrySource {
448public:
449 virtual ~ExternalSLocEntrySource();
450
451 /// \brief Read the source location entry with index ID, which will always be
452 /// less than -1.
453 ///
454 /// \returns true if an error occurred that prevented the source-location
455 /// entry from being loaded.
456 virtual bool ReadSLocEntry(int ID) = 0;
457
458 /// \brief Retrieve the module import location and name for the given ID, if
459 /// in fact it was loaded from a module (rather than, say, a precompiled
460 /// header).
461 virtual std::pair<SourceLocation, StringRef> getModuleImportLoc(int ID) = 0;
462};
463
464/// \brief Holds the cache used by isBeforeInTranslationUnit.
465///
466/// The cache structure is complex enough to be worth breaking out of
467/// SourceManager.
468class InBeforeInTUCacheEntry {
469 /// \brief The FileID's of the cached query.
470 ///
471 /// If these match up with a subsequent query, the result can be reused.
472 FileID LQueryFID, RQueryFID;
473
474 /// \brief True if LQueryFID was created before RQueryFID.
475 ///
476 /// This is used to compare macro expansion locations.
477 bool IsLQFIDBeforeRQFID;
478
479 /// \brief The file found in common between the two \#include traces, i.e.,
480 /// the nearest common ancestor of the \#include tree.
481 FileID CommonFID;
482
483 /// \brief The offset of the previous query in CommonFID.
484 ///
485 /// Usually, this represents the location of the \#include for QueryFID, but
486 /// if LQueryFID is a parent of RQueryFID (or vice versa) then these can be a
487 /// random token in the parent.
488 unsigned LCommonOffset, RCommonOffset;
489
490public:
491 /// \brief Return true if the currently cached values match up with
492 /// the specified LHS/RHS query.
493 ///
494 /// If not, we can't use the cache.
495 bool isCacheValid(FileID LHS, FileID RHS) const {
496 return LQueryFID == LHS && RQueryFID == RHS;
497 }
498
499 /// \brief If the cache is valid, compute the result given the
500 /// specified offsets in the LHS/RHS FileID's.
501 bool getCachedResult(unsigned LOffset, unsigned ROffset) const {
502 // If one of the query files is the common file, use the offset. Otherwise,
503 // use the #include loc in the common file.
504 if (LQueryFID != CommonFID) LOffset = LCommonOffset;
505 if (RQueryFID != CommonFID) ROffset = RCommonOffset;
506
507 // It is common for multiple macro expansions to be "included" from the same
508 // location (expansion location), in which case use the order of the FileIDs
509 // to determine which came first. This will also take care the case where
510 // one of the locations points at the inclusion/expansion point of the other
511 // in which case its FileID will come before the other.
512 if (LOffset == ROffset)
513 return IsLQFIDBeforeRQFID;
514
515 return LOffset < ROffset;
516 }
517
518 /// \brief Set up a new query.
519 void setQueryFIDs(FileID LHS, FileID RHS, bool isLFIDBeforeRFID) {
520 assert(LHS != RHS);
521 LQueryFID = LHS;
522 RQueryFID = RHS;
523 IsLQFIDBeforeRQFID = isLFIDBeforeRFID;
524 }
525
526 void clear() {
527 LQueryFID = RQueryFID = FileID();
528 IsLQFIDBeforeRQFID = false;
529 }
530
531 void setCommonLoc(FileID commonFID, unsigned lCommonOffset,
532 unsigned rCommonOffset) {
533 CommonFID = commonFID;
534 LCommonOffset = lCommonOffset;
535 RCommonOffset = rCommonOffset;
536 }
537};
538
539/// \brief The stack used when building modules on demand, which is used
540/// to provide a link between the source managers of the different compiler
541/// instances.
542typedef ArrayRef<std::pair<std::string, FullSourceLoc>> ModuleBuildStack;
543
544/// \brief This class handles loading and caching of source files into memory.
545///
546/// This object owns the MemoryBuffer objects for all of the loaded
547/// files and assigns unique FileID's for each unique \#include chain.
548///
549/// The SourceManager can be queried for information about SourceLocation
550/// objects, turning them into either spelling or expansion locations. Spelling
551/// locations represent where the bytes corresponding to a token came from and
552/// expansion locations represent where the location is in the user's view. In
553/// the case of a macro expansion, for example, the spelling location indicates
554/// where the expanded token came from and the expansion location specifies
555/// where it was expanded.
556class SourceManager : public RefCountedBase<SourceManager> {
557 /// \brief DiagnosticsEngine object.
558 DiagnosticsEngine &Diag;
559
560 FileManager &FileMgr;
561
562 mutable llvm::BumpPtrAllocator ContentCacheAlloc;
563
564 /// \brief Memoized information about all of the files tracked by this
565 /// SourceManager.
566 ///
567 /// This map allows us to merge ContentCache entries based
568 /// on their FileEntry*. All ContentCache objects will thus have unique,
569 /// non-null, FileEntry pointers.
570 llvm::DenseMap<const FileEntry*, SrcMgr::ContentCache*> FileInfos;
571
572 /// \brief True if the ContentCache for files that are overridden by other
573 /// files, should report the original file name. Defaults to true.
574 bool OverridenFilesKeepOriginalName;
575
576 /// \brief True if non-system source files should be treated as volatile
577 /// (likely to change while trying to use them). Defaults to false.
578 bool UserFilesAreVolatile;
579
580 /// \brief True if all files read during this compilation should be treated
581 /// as transient (may not be present in later compilations using a module
582 /// file created from this compilation). Defaults to false.
583 bool FilesAreTransient;
584
585 struct OverriddenFilesInfoTy {
586 /// \brief Files that have been overridden with the contents from another
587 /// file.
588 llvm::DenseMap<const FileEntry *, const FileEntry *> OverriddenFiles;
589 /// \brief Files that were overridden with a memory buffer.
590 llvm::DenseSet<const FileEntry *> OverriddenFilesWithBuffer;
591 };
592
593 /// \brief Lazily create the object keeping overridden files info, since
594 /// it is uncommonly used.
595 std::unique_ptr<OverriddenFilesInfoTy> OverriddenFilesInfo;
596
597 OverriddenFilesInfoTy &getOverriddenFilesInfo() {
598 if (!OverriddenFilesInfo)
599 OverriddenFilesInfo.reset(new OverriddenFilesInfoTy);
600 return *OverriddenFilesInfo;
601 }
602
603 /// \brief Information about various memory buffers that we have read in.
604 ///
605 /// All FileEntry* within the stored ContentCache objects are NULL,
606 /// as they do not refer to a file.
607 std::vector<SrcMgr::ContentCache*> MemBufferInfos;
608
609 /// \brief The table of SLocEntries that are local to this module.
610 ///
611 /// Positive FileIDs are indexes into this table. Entry 0 indicates an invalid
612 /// expansion.
613 SmallVector<SrcMgr::SLocEntry, 0> LocalSLocEntryTable;
614
615 /// \brief The table of SLocEntries that are loaded from other modules.
616 ///
617 /// Negative FileIDs are indexes into this table. To get from ID to an index,
618 /// use (-ID - 2).
619 mutable SmallVector<SrcMgr::SLocEntry, 0> LoadedSLocEntryTable;
620
621 /// \brief The starting offset of the next local SLocEntry.
622 ///
623 /// This is LocalSLocEntryTable.back().Offset + the size of that entry.
624 unsigned NextLocalOffset;
625
626 /// \brief The starting offset of the latest batch of loaded SLocEntries.
627 ///
628 /// This is LoadedSLocEntryTable.back().Offset, except that that entry might
629 /// not have been loaded, so that value would be unknown.
630 unsigned CurrentLoadedOffset;
631
632 /// \brief The highest possible offset is 2^31-1, so CurrentLoadedOffset
633 /// starts at 2^31.
634 static const unsigned MaxLoadedOffset = 1U << 31U;
635
636 /// \brief A bitmap that indicates whether the entries of LoadedSLocEntryTable
637 /// have already been loaded from the external source.
638 ///
639 /// Same indexing as LoadedSLocEntryTable.
640 llvm::BitVector SLocEntryLoaded;
641
642 /// \brief An external source for source location entries.
643 ExternalSLocEntrySource *ExternalSLocEntries;
644
645 /// \brief A one-entry cache to speed up getFileID.
646 ///
647 /// LastFileIDLookup records the last FileID looked up or created, because it
648 /// is very common to look up many tokens from the same file.
649 mutable FileID LastFileIDLookup;
650
651 /// \brief Holds information for \#line directives.
652 ///
653 /// This is referenced by indices from SLocEntryTable.
654 LineTableInfo *LineTable;
655
656 /// \brief These ivars serve as a cache used in the getLineNumber
657 /// method which is used to speedup getLineNumber calls to nearby locations.
658 mutable FileID LastLineNoFileIDQuery;
659 mutable SrcMgr::ContentCache *LastLineNoContentCache;
660 mutable unsigned LastLineNoFilePos;
661 mutable unsigned LastLineNoResult;
662
663 /// \brief The file ID for the main source file of the translation unit.
664 FileID MainFileID;
665
666 /// \brief The file ID for the precompiled preamble there is one.
667 FileID PreambleFileID;
668
669 // Statistics for -print-stats.
670 mutable unsigned NumLinearScans, NumBinaryProbes;
671
672 /// \brief Associates a FileID with its "included/expanded in" decomposed
673 /// location.
674 ///
675 /// Used to cache results from and speed-up \c getDecomposedIncludedLoc
676 /// function.
677 mutable llvm::DenseMap<FileID, std::pair<FileID, unsigned>> IncludedLocMap;
678
679 /// The key value into the IsBeforeInTUCache table.
680 typedef std::pair<FileID, FileID> IsBeforeInTUCacheKey;
681
682 /// The IsBeforeInTranslationUnitCache is a mapping from FileID pairs
683 /// to cache results.
684 typedef llvm::DenseMap<IsBeforeInTUCacheKey, InBeforeInTUCacheEntry>
685 InBeforeInTUCache;
686
687 /// Cache results for the isBeforeInTranslationUnit method.
688 mutable InBeforeInTUCache IBTUCache;
689 mutable InBeforeInTUCacheEntry IBTUCacheOverflow;
690
691 /// Return the cache entry for comparing the given file IDs
692 /// for isBeforeInTranslationUnit.
693 InBeforeInTUCacheEntry &getInBeforeInTUCache(FileID LFID, FileID RFID) const;
694
695 // Cache for the "fake" buffer used for error-recovery purposes.
696 mutable std::unique_ptr<llvm::MemoryBuffer> FakeBufferForRecovery;
697
698 mutable std::unique_ptr<SrcMgr::ContentCache> FakeContentCacheForRecovery;
699
700 /// \brief Lazily computed map of macro argument chunks to their expanded
701 /// source location.
702 typedef std::map<unsigned, SourceLocation> MacroArgsMap;
703
704 mutable llvm::DenseMap<FileID, std::unique_ptr<MacroArgsMap>>
705 MacroArgsCacheMap;
706
707 /// \brief The stack of modules being built, which is used to detect
708 /// cycles in the module dependency graph as modules are being built, as
709 /// well as to describe why we're rebuilding a particular module.
710 ///
711 /// There is no way to set this value from the command line. If we ever need
712 /// to do so (e.g., if on-demand module construction moves out-of-process),
713 /// we can add a cc1-level option to do so.
714 SmallVector<std::pair<std::string, FullSourceLoc>, 2> StoredModuleBuildStack;
715
716public:
717 SourceManager(DiagnosticsEngine &Diag, FileManager &FileMgr,
718 bool UserFilesAreVolatile = false);
719 explicit SourceManager(const SourceManager &) = delete;
720 SourceManager &operator=(const SourceManager &) = delete;
721 ~SourceManager();
722
723 void clearIDTables();
724
725 DiagnosticsEngine &getDiagnostics() const { return Diag; }
726
727 FileManager &getFileManager() const { return FileMgr; }
728
729 /// \brief Set true if the SourceManager should report the original file name
730 /// for contents of files that were overridden by other files. Defaults to
731 /// true.
732 void setOverridenFilesKeepOriginalName(bool value) {
733 OverridenFilesKeepOriginalName = value;
734 }
735
736 /// \brief True if non-system source files should be treated as volatile
737 /// (likely to change while trying to use them).
738 bool userFilesAreVolatile() const { return UserFilesAreVolatile; }
739
740 /// \brief Retrieve the module build stack.
741 ModuleBuildStack getModuleBuildStack() const {
742 return StoredModuleBuildStack;
743 }
744
745 /// \brief Set the module build stack.
746 void setModuleBuildStack(ModuleBuildStack stack) {
747 StoredModuleBuildStack.clear();
748 StoredModuleBuildStack.append(stack.begin(), stack.end());
749 }
750
751 /// \brief Push an entry to the module build stack.
752 void pushModuleBuildStack(StringRef moduleName, FullSourceLoc importLoc) {
753 StoredModuleBuildStack.push_back(std::make_pair(moduleName.str(),importLoc));
754 }
755
756 //===--------------------------------------------------------------------===//
757 // MainFileID creation and querying methods.
758 //===--------------------------------------------------------------------===//
759
760 /// \brief Returns the FileID of the main source file.
761 FileID getMainFileID() const { return MainFileID; }
762
763 /// \brief Set the file ID for the main source file.
764 void setMainFileID(FileID FID) {
765 MainFileID = FID;
766 }
767
768 /// \brief Set the file ID for the precompiled preamble.
769 void setPreambleFileID(FileID Preamble) {
770 assert(PreambleFileID.isInvalid() && "PreambleFileID already set!");
771 PreambleFileID = Preamble;
772 }
773
774 /// \brief Get the file ID for the precompiled preamble if there is one.
775 FileID getPreambleFileID() const { return PreambleFileID; }
776
777 //===--------------------------------------------------------------------===//
778 // Methods to create new FileID's and macro expansions.
779 //===--------------------------------------------------------------------===//
780
781 /// \brief Create a new FileID that represents the specified file
782 /// being \#included from the specified IncludePosition.
783 ///
784 /// This translates NULL into standard input.
785 FileID createFileID(const FileEntry *SourceFile, SourceLocation IncludePos,
786 SrcMgr::CharacteristicKind FileCharacter,
787 int LoadedID = 0, unsigned LoadedOffset = 0) {
788 const SrcMgr::ContentCache *
789 IR = getOrCreateContentCache(SourceFile,
790 /*isSystemFile=*/FileCharacter != SrcMgr::C_User);
791 assert(IR && "getOrCreateContentCache() cannot return NULL");
792 return createFileID(IR, IncludePos, FileCharacter, LoadedID, LoadedOffset);
793 }
794
795 /// \brief Create a new FileID that represents the specified memory buffer.
796 ///
797 /// This does no caching of the buffer and takes ownership of the
798 /// MemoryBuffer, so only pass a MemoryBuffer to this once.
799 FileID createFileID(std::unique_ptr<llvm::MemoryBuffer> Buffer,
800 SrcMgr::CharacteristicKind FileCharacter = SrcMgr::C_User,
801 int LoadedID = 0, unsigned LoadedOffset = 0,
802 SourceLocation IncludeLoc = SourceLocation()) {
803 return createFileID(createMemBufferContentCache(std::move(Buffer)),
804 IncludeLoc, FileCharacter, LoadedID, LoadedOffset);
805 }
806
807 /// \brief Get the FileID for \p SourceFile if it exists. Otherwise, create a
808 /// new FileID for the \p SourceFile.
809 FileID getOrCreateFileID(const FileEntry *SourceFile,
810 SrcMgr::CharacteristicKind FileCharacter) {
811 FileID ID = translateFile(SourceFile);
812 return ID.isValid() ? ID : createFileID(SourceFile, SourceLocation(),
813 FileCharacter);
814 }
815
816 /// \brief Return a new SourceLocation that encodes the
817 /// fact that a token from SpellingLoc should actually be referenced from
818 /// ExpansionLoc, and that it represents the expansion of a macro argument
819 /// into the function-like macro body.
820 SourceLocation createMacroArgExpansionLoc(SourceLocation Loc,
821 SourceLocation ExpansionLoc,
822 unsigned TokLength);
823
824 /// \brief Return a new SourceLocation that encodes the fact
825 /// that a token from SpellingLoc should actually be referenced from
826 /// ExpansionLoc.
827 SourceLocation createExpansionLoc(SourceLocation Loc,
828 SourceLocation ExpansionLocStart,
829 SourceLocation ExpansionLocEnd,
830 unsigned TokLength,
831 int LoadedID = 0,
832 unsigned LoadedOffset = 0);
833
834 /// \brief Retrieve the memory buffer associated with the given file.
835 ///
836 /// \param Invalid If non-NULL, will be set \c true if an error
837 /// occurs while retrieving the memory buffer.
838 llvm::MemoryBuffer *getMemoryBufferForFile(const FileEntry *File,
839 bool *Invalid = nullptr);
840
841 /// \brief Override the contents of the given source file by providing an
842 /// already-allocated buffer.
843 ///
844 /// \param SourceFile the source file whose contents will be overridden.
845 ///
846 /// \param Buffer the memory buffer whose contents will be used as the
847 /// data in the given source file.
848 ///
849 /// \param DoNotFree If true, then the buffer will not be freed when the
850 /// source manager is destroyed.
851 void overrideFileContents(const FileEntry *SourceFile,
852 llvm::MemoryBuffer *Buffer, bool DoNotFree);
853 void overrideFileContents(const FileEntry *SourceFile,
854 std::unique_ptr<llvm::MemoryBuffer> Buffer) {
855 overrideFileContents(SourceFile, Buffer.release(), /*DoNotFree*/ false);
856 }
857
858 /// \brief Override the given source file with another one.
859 ///
860 /// \param SourceFile the source file which will be overridden.
861 ///
862 /// \param NewFile the file whose contents will be used as the
863 /// data instead of the contents of the given source file.
864 void overrideFileContents(const FileEntry *SourceFile,
865 const FileEntry *NewFile);
866
867 /// \brief Returns true if the file contents have been overridden.
868 bool isFileOverridden(const FileEntry *File) const {
869 if (OverriddenFilesInfo) {
870 if (OverriddenFilesInfo->OverriddenFilesWithBuffer.count(File))
871 return true;
872 if (OverriddenFilesInfo->OverriddenFiles.find(File) !=
873 OverriddenFilesInfo->OverriddenFiles.end())
874 return true;
875 }
876 return false;
877 }
878
879 /// \brief Disable overridding the contents of a file, previously enabled
880 /// with #overrideFileContents.
881 ///
882 /// This should be called before parsing has begun.
883 void disableFileContentsOverride(const FileEntry *File);
884
885 /// \brief Specify that a file is transient.
886 void setFileIsTransient(const FileEntry *SourceFile);
887
888 /// \brief Specify that all files that are read during this compilation are
889 /// transient.
890 void setAllFilesAreTransient(bool Transient) {
891 FilesAreTransient = Transient;
892 }
893
894 //===--------------------------------------------------------------------===//
895 // FileID manipulation methods.
896 //===--------------------------------------------------------------------===//
897
898 /// \brief Return the buffer for the specified FileID.
899 ///
900 /// If there is an error opening this buffer the first time, this
901 /// manufactures a temporary buffer and returns a non-empty error string.
902 llvm::MemoryBuffer *getBuffer(FileID FID, SourceLocation Loc,
903 bool *Invalid = nullptr) const {
904 bool MyInvalid = false;
905 const SrcMgr::SLocEntry &Entry = getSLocEntry(FID, &MyInvalid);
906 if (MyInvalid || !Entry.isFile()) {
907 if (Invalid)
908 *Invalid = true;
909
910 return getFakeBufferForRecovery();
911 }
912
913 return Entry.getFile().getContentCache()->getBuffer(Diag, *this, Loc,
914 Invalid);
915 }
916
917 llvm::MemoryBuffer *getBuffer(FileID FID, bool *Invalid = nullptr) const {
918 bool MyInvalid = false;
919 const SrcMgr::SLocEntry &Entry = getSLocEntry(FID, &MyInvalid);
920 if (MyInvalid || !Entry.isFile()) {
921 if (Invalid)
922 *Invalid = true;
923
924 return getFakeBufferForRecovery();
925 }
926
927 return Entry.getFile().getContentCache()->getBuffer(Diag, *this,
928 SourceLocation(),
929 Invalid);
930 }
931
932 /// \brief Returns the FileEntry record for the provided FileID.
933 const FileEntry *getFileEntryForID(FileID FID) const {
934 bool MyInvalid = false;
935 const SrcMgr::SLocEntry &Entry = getSLocEntry(FID, &MyInvalid);
936 if (MyInvalid || !Entry.isFile())
937 return nullptr;
938
939 const SrcMgr::ContentCache *Content = Entry.getFile().getContentCache();
940 if (!Content)
941 return nullptr;
942 return Content->OrigEntry;
943 }
944
945 /// \brief Returns the FileEntry record for the provided SLocEntry.
946 const FileEntry *getFileEntryForSLocEntry(const SrcMgr::SLocEntry &sloc) const
947 {
948 const SrcMgr::ContentCache *Content = sloc.getFile().getContentCache();
949 if (!Content)
950 return nullptr;
951 return Content->OrigEntry;
952 }
953
954 /// \brief Return a StringRef to the source buffer data for the
955 /// specified FileID.
956 ///
957 /// \param FID The file ID whose contents will be returned.
958 /// \param Invalid If non-NULL, will be set true if an error occurred.
959 StringRef getBufferData(FileID FID, bool *Invalid = nullptr) const;
960
961 /// \brief Get the number of FileIDs (files and macros) that were created
962 /// during preprocessing of \p FID, including it.
963 unsigned getNumCreatedFIDsForFileID(FileID FID) const {
964 bool Invalid = false;
965 const SrcMgr::SLocEntry &Entry = getSLocEntry(FID, &Invalid);
966 if (Invalid || !Entry.isFile())
967 return 0;
968
969 return Entry.getFile().NumCreatedFIDs;
970 }
971
972 /// \brief Set the number of FileIDs (files and macros) that were created
973 /// during preprocessing of \p FID, including it.
974 void setNumCreatedFIDsForFileID(FileID FID, unsigned NumFIDs) const {
975 bool Invalid = false;
976 const SrcMgr::SLocEntry &Entry = getSLocEntry(FID, &Invalid);
977 if (Invalid || !Entry.isFile())
978 return;
979
980 assert(Entry.getFile().NumCreatedFIDs == 0 && "Already set!");
981 const_cast<SrcMgr::FileInfo &>(Entry.getFile()).NumCreatedFIDs = NumFIDs;
982 }
983
984 //===--------------------------------------------------------------------===//
985 // SourceLocation manipulation methods.
986 //===--------------------------------------------------------------------===//
987
988 /// \brief Return the FileID for a SourceLocation.
989 ///
990 /// This is a very hot method that is used for all SourceManager queries
991 /// that start with a SourceLocation object. It is responsible for finding
992 /// the entry in SLocEntryTable which contains the specified location.
993 ///
994 FileID getFileID(SourceLocation SpellingLoc) const {
995 unsigned SLocOffset = SpellingLoc.getOffset();
996
997 // If our one-entry cache covers this offset, just return it.
998 if (isOffsetInFileID(LastFileIDLookup, SLocOffset))
999 return LastFileIDLookup;
1000
1001 return getFileIDSlow(SLocOffset);
1002 }
1003
1004 /// \brief Return the filename of the file containing a SourceLocation.
1005 StringRef getFilename(SourceLocation SpellingLoc) const {
1006 if (const FileEntry *F = getFileEntryForID(getFileID(SpellingLoc)))
1007 return F->getName();
1008 return StringRef();
1009 }
1010
1011 /// \brief Return the source location corresponding to the first byte of
1012 /// the specified file.
1013 SourceLocation getLocForStartOfFile(FileID FID) const {
1014 bool Invalid = false;
1015 const SrcMgr::SLocEntry &Entry = getSLocEntry(FID, &Invalid);
1016 if (Invalid || !Entry.isFile())
1017 return SourceLocation();
1018
1019 unsigned FileOffset = Entry.getOffset();
1020 return SourceLocation::getFileLoc(FileOffset);
1021 }
1022
1023 /// \brief Return the source location corresponding to the last byte of the
1024 /// specified file.
1025 SourceLocation getLocForEndOfFile(FileID FID) const {
1026 bool Invalid = false;
1027 const SrcMgr::SLocEntry &Entry = getSLocEntry(FID, &Invalid);
1028 if (Invalid || !Entry.isFile())
1029 return SourceLocation();
1030
1031 unsigned FileOffset = Entry.getOffset();
1032 return SourceLocation::getFileLoc(FileOffset + getFileIDSize(FID));
1033 }
1034
1035 /// \brief Returns the include location if \p FID is a \#include'd file
1036 /// otherwise it returns an invalid location.
1037 SourceLocation getIncludeLoc(FileID FID) const {
1038 bool Invalid = false;
1039 const SrcMgr::SLocEntry &Entry = getSLocEntry(FID, &Invalid);
1040 if (Invalid || !Entry.isFile())
1041 return SourceLocation();
1042
1043 return Entry.getFile().getIncludeLoc();
1044 }
1045
1046 // \brief Returns the import location if the given source location is
1047 // located within a module, or an invalid location if the source location
1048 // is within the current translation unit.
1049 std::pair<SourceLocation, StringRef>
1050 getModuleImportLoc(SourceLocation Loc) const {
1051 FileID FID = getFileID(Loc);
1052
1053 // Positive file IDs are in the current translation unit, and -1 is a
1054 // placeholder.
1055 if (FID.ID >= -1)
1056 return std::make_pair(SourceLocation(), "");
1057
1058 return ExternalSLocEntries->getModuleImportLoc(FID.ID);
1059 }
1060
1061 /// \brief Given a SourceLocation object \p Loc, return the expansion
1062 /// location referenced by the ID.
1063 SourceLocation getExpansionLoc(SourceLocation Loc) const {
1064 // Handle the non-mapped case inline, defer to out of line code to handle
1065 // expansions.
1066 if (Loc.isFileID()) return Loc;
1067 return getExpansionLocSlowCase(Loc);
1068 }
1069
1070 /// \brief Given \p Loc, if it is a macro location return the expansion
1071 /// location or the spelling location, depending on if it comes from a
1072 /// macro argument or not.
1073 SourceLocation getFileLoc(SourceLocation Loc) const {
1074 if (Loc.isFileID()) return Loc;
1075 return getFileLocSlowCase(Loc);
1076 }
1077
1078 /// \brief Return the start/end of the expansion information for an
1079 /// expansion location.
1080 ///
1081 /// \pre \p Loc is required to be an expansion location.
1082 std::pair<SourceLocation,SourceLocation>
1083 getImmediateExpansionRange(SourceLocation Loc) const;
1084
1085 /// \brief Given a SourceLocation object, return the range of
1086 /// tokens covered by the expansion in the ultimate file.
1087 std::pair<SourceLocation,SourceLocation>
1088 getExpansionRange(SourceLocation Loc) const;
1089
1090 /// \brief Given a SourceRange object, return the range of
1091 /// tokens covered by the expansion in the ultimate file.
1092 SourceRange getExpansionRange(SourceRange Range) const {
1093 return SourceRange(getExpansionRange(Range.getBegin()).first,
1094 getExpansionRange(Range.getEnd()).second);
1095 }
1096
1097 /// \brief Given a SourceLocation object, return the spelling
1098 /// location referenced by the ID.
1099 ///
1100 /// This is the place where the characters that make up the lexed token
1101 /// can be found.
1102 SourceLocation getSpellingLoc(SourceLocation Loc) const {
1103 // Handle the non-mapped case inline, defer to out of line code to handle
1104 // expansions.
1105 if (Loc.isFileID()) return Loc;
1106 return getSpellingLocSlowCase(Loc);
1107 }
1108
1109 /// \brief Given a SourceLocation object, return the spelling location
1110 /// referenced by the ID.
1111 ///
1112 /// This is the first level down towards the place where the characters
1113 /// that make up the lexed token can be found. This should not generally
1114 /// be used by clients.
1115 SourceLocation getImmediateSpellingLoc(SourceLocation Loc) const;
1116
1117 /// \brief Decompose the specified location into a raw FileID + Offset pair.
1118 ///
1119 /// The first element is the FileID, the second is the offset from the
1120 /// start of the buffer of the location.
1121 std::pair<FileID, unsigned> getDecomposedLoc(SourceLocation Loc) const {
1122 FileID FID = getFileID(Loc);
1123 bool Invalid = false;
1124 const SrcMgr::SLocEntry &E = getSLocEntry(FID, &Invalid);
1125 if (Invalid)
1126 return std::make_pair(FileID(), 0);
1127 return std::make_pair(FID, Loc.getOffset()-E.getOffset());
1128 }
1129
1130 /// \brief Decompose the specified location into a raw FileID + Offset pair.
1131 ///
1132 /// If the location is an expansion record, walk through it until we find
1133 /// the final location expanded.
1134 std::pair<FileID, unsigned>
1135 getDecomposedExpansionLoc(SourceLocation Loc) const {
1136 FileID FID = getFileID(Loc);
1137 bool Invalid = false;
1138 const SrcMgr::SLocEntry *E = &getSLocEntry(FID, &Invalid);
1139 if (Invalid)
1140 return std::make_pair(FileID(), 0);
1141
1142 unsigned Offset = Loc.getOffset()-E->getOffset();
1143 if (Loc.isFileID())
1144 return std::make_pair(FID, Offset);
1145
1146 return getDecomposedExpansionLocSlowCase(E);
1147 }
1148
1149 /// \brief Decompose the specified location into a raw FileID + Offset pair.
1150 ///
1151 /// If the location is an expansion record, walk through it until we find
1152 /// its spelling record.
1153 std::pair<FileID, unsigned>
1154 getDecomposedSpellingLoc(SourceLocation Loc) const {
1155 FileID FID = getFileID(Loc);
1156 bool Invalid = false;
1157 const SrcMgr::SLocEntry *E = &getSLocEntry(FID, &Invalid);
1158 if (Invalid)
1159 return std::make_pair(FileID(), 0);
1160
1161 unsigned Offset = Loc.getOffset()-E->getOffset();
1162 if (Loc.isFileID())
1163 return std::make_pair(FID, Offset);
1164 return getDecomposedSpellingLocSlowCase(E, Offset);
1165 }
1166
1167 /// \brief Returns the "included/expanded in" decomposed location of the given
1168 /// FileID.
1169 std::pair<FileID, unsigned> getDecomposedIncludedLoc(FileID FID) const;
1170
1171 /// \brief Returns the offset from the start of the file that the
1172 /// specified SourceLocation represents.
1173 ///
1174 /// This is not very meaningful for a macro ID.
1175 unsigned getFileOffset(SourceLocation SpellingLoc) const {
1176 return getDecomposedLoc(SpellingLoc).second;
1177 }
1178
1179 /// \brief Tests whether the given source location represents a macro
1180 /// argument's expansion into the function-like macro definition.
1181 ///
1182 /// \param StartLoc If non-null and function returns true, it is set to the
1183 /// start location of the macro argument expansion.
1184 ///
1185 /// Such source locations only appear inside of the expansion
1186 /// locations representing where a particular function-like macro was
1187 /// expanded.
1188 bool isMacroArgExpansion(SourceLocation Loc,
1189 SourceLocation *StartLoc = nullptr) const;
1190
1191 /// \brief Tests whether the given source location represents the expansion of
1192 /// a macro body.
1193 ///
1194 /// This is equivalent to testing whether the location is part of a macro
1195 /// expansion but not the expansion of an argument to a function-like macro.
1196 bool isMacroBodyExpansion(SourceLocation Loc) const;
1197
1198 /// \brief Returns true if the given MacroID location points at the beginning
1199 /// of the immediate macro expansion.
1200 ///
1201 /// \param MacroBegin If non-null and function returns true, it is set to the
1202 /// begin location of the immediate macro expansion.
1203 bool isAtStartOfImmediateMacroExpansion(SourceLocation Loc,
1204 SourceLocation *MacroBegin = nullptr) const;
1205
1206 /// \brief Returns true if the given MacroID location points at the character
1207 /// end of the immediate macro expansion.
1208 ///
1209 /// \param MacroEnd If non-null and function returns true, it is set to the
1210 /// character end location of the immediate macro expansion.
1211 bool
1212 isAtEndOfImmediateMacroExpansion(SourceLocation Loc,
1213 SourceLocation *MacroEnd = nullptr) const;
1214
1215 /// \brief Returns true if \p Loc is inside the [\p Start, +\p Length)
1216 /// chunk of the source location address space.
1217 ///
1218 /// If it's true and \p RelativeOffset is non-null, it will be set to the
1219 /// relative offset of \p Loc inside the chunk.
1220 bool isInSLocAddrSpace(SourceLocation Loc,
1221 SourceLocation Start, unsigned Length,
1222 unsigned *RelativeOffset = nullptr) const {
1223 assert(((Start.getOffset() < NextLocalOffset &&
1224 Start.getOffset()+Length <= NextLocalOffset) ||
1225 (Start.getOffset() >= CurrentLoadedOffset &&
1226 Start.getOffset()+Length < MaxLoadedOffset)) &&
1227 "Chunk is not valid SLoc address space");
1228 unsigned LocOffs = Loc.getOffset();
1229 unsigned BeginOffs = Start.getOffset();
1230 unsigned EndOffs = BeginOffs + Length;
1231 if (LocOffs >= BeginOffs && LocOffs < EndOffs) {
1232 if (RelativeOffset)
1233 *RelativeOffset = LocOffs - BeginOffs;
1234 return true;
1235 }
1236
1237 return false;
1238 }
1239
1240 /// \brief Return true if both \p LHS and \p RHS are in the local source
1241 /// location address space or the loaded one.
1242 ///
1243 /// If it's true and \p RelativeOffset is non-null, it will be set to the
1244 /// offset of \p RHS relative to \p LHS.
1245 bool isInSameSLocAddrSpace(SourceLocation LHS, SourceLocation RHS,
1246 int *RelativeOffset) const {
1247 unsigned LHSOffs = LHS.getOffset(), RHSOffs = RHS.getOffset();
1248 bool LHSLoaded = LHSOffs >= CurrentLoadedOffset;
1249 bool RHSLoaded = RHSOffs >= CurrentLoadedOffset;
1250
1251 if (LHSLoaded == RHSLoaded) {
1252 if (RelativeOffset)
1253 *RelativeOffset = RHSOffs - LHSOffs;
1254 return true;
1255 }
1256
1257 return false;
1258 }
1259
1260 //===--------------------------------------------------------------------===//
1261 // Queries about the code at a SourceLocation.
1262 //===--------------------------------------------------------------------===//
1263
1264 /// \brief Return a pointer to the start of the specified location
1265 /// in the appropriate spelling MemoryBuffer.
1266 ///
1267 /// \param Invalid If non-NULL, will be set \c true if an error occurs.
1268 const char *getCharacterData(SourceLocation SL,
1269 bool *Invalid = nullptr) const;
1270
1271 /// \brief Return the column # for the specified file position.
1272 ///
1273 /// This is significantly cheaper to compute than the line number. This
1274 /// returns zero if the column number isn't known. This may only be called
1275 /// on a file sloc, so you must choose a spelling or expansion location
1276 /// before calling this method.
1277 unsigned getColumnNumber(FileID FID, unsigned FilePos,
1278 bool *Invalid = nullptr) const;
1279 unsigned getSpellingColumnNumber(SourceLocation Loc,
1280 bool *Invalid = nullptr) const;
1281 unsigned getExpansionColumnNumber(SourceLocation Loc,
1282 bool *Invalid = nullptr) const;
1283 unsigned getPresumedColumnNumber(SourceLocation Loc,
1284 bool *Invalid = nullptr) const;
1285
1286 /// \brief Given a SourceLocation, return the spelling line number
1287 /// for the position indicated.
1288 ///
1289 /// This requires building and caching a table of line offsets for the
1290 /// MemoryBuffer, so this is not cheap: use only when about to emit a
1291 /// diagnostic.
1292 unsigned getLineNumber(FileID FID, unsigned FilePos, bool *Invalid = nullptr) const;
1293 unsigned getSpellingLineNumber(SourceLocation Loc, bool *Invalid = nullptr) const;
1294 unsigned getExpansionLineNumber(SourceLocation Loc, bool *Invalid = nullptr) const;
1295 unsigned getPresumedLineNumber(SourceLocation Loc, bool *Invalid = nullptr) const;
1296
1297 /// \brief Return the filename or buffer identifier of the buffer the
1298 /// location is in.
1299 ///
1300 /// Note that this name does not respect \#line directives. Use
1301 /// getPresumedLoc for normal clients.
1302 StringRef getBufferName(SourceLocation Loc, bool *Invalid = nullptr) const;
1303
1304 /// \brief Return the file characteristic of the specified source
1305 /// location, indicating whether this is a normal file, a system
1306 /// header, or an "implicit extern C" system header.
1307 ///
1308 /// This state can be modified with flags on GNU linemarker directives like:
1309 /// \code
1310 /// # 4 "foo.h" 3
1311 /// \endcode
1312 /// which changes all source locations in the current file after that to be
1313 /// considered to be from a system header.
1314 SrcMgr::CharacteristicKind getFileCharacteristic(SourceLocation Loc) const;
1315
1316 /// \brief Returns the "presumed" location of a SourceLocation specifies.
1317 ///
1318 /// A "presumed location" can be modified by \#line or GNU line marker
1319 /// directives. This provides a view on the data that a user should see
1320 /// in diagnostics, for example.
1321 ///
1322 /// Note that a presumed location is always given as the expansion point of
1323 /// an expansion location, not at the spelling location.
1324 ///
1325 /// \returns The presumed location of the specified SourceLocation. If the
1326 /// presumed location cannot be calculated (e.g., because \p Loc is invalid
1327 /// or the file containing \p Loc has changed on disk), returns an invalid
1328 /// presumed location.
1329 PresumedLoc getPresumedLoc(SourceLocation Loc,
1330 bool UseLineDirectives = true) const;
1331
1332 /// \brief Returns whether the PresumedLoc for a given SourceLocation is
1333 /// in the main file.
1334 ///
1335 /// This computes the "presumed" location for a SourceLocation, then checks
1336 /// whether it came from a file other than the main file. This is different
1337 /// from isWrittenInMainFile() because it takes line marker directives into
1338 /// account.
1339 bool isInMainFile(SourceLocation Loc) const;
1340
1341 /// \brief Returns true if the spelling locations for both SourceLocations
1342 /// are part of the same file buffer.
1343 ///
1344 /// This check ignores line marker directives.
1345 bool isWrittenInSameFile(SourceLocation Loc1, SourceLocation Loc2) const {
1346 return getFileID(Loc1) == getFileID(Loc2);
1347 }
1348
1349 /// \brief Returns true if the spelling location for the given location
1350 /// is in the main file buffer.
1351 ///
1352 /// This check ignores line marker directives.
1353 bool isWrittenInMainFile(SourceLocation Loc) const {
1354 return getFileID(Loc) == getMainFileID();
1355 }
1356
1357 /// \brief Returns if a SourceLocation is in a system header.
1358 bool isInSystemHeader(SourceLocation Loc) const {
1359 return getFileCharacteristic(Loc) != SrcMgr::C_User;
1360 }
1361
1362 /// \brief Returns if a SourceLocation is in an "extern C" system header.
1363 bool isInExternCSystemHeader(SourceLocation Loc) const {
1364 return getFileCharacteristic(Loc) == SrcMgr::C_ExternCSystem;
1365 }
1366
1367 /// \brief Returns whether \p Loc is expanded from a macro in a system header.
1368 bool isInSystemMacro(SourceLocation loc) const {
1369 return loc.isMacroID() && isInSystemHeader(getSpellingLoc(loc));
1370 }
1371
1372 /// \brief The size of the SLocEntry that \p FID represents.
1373 unsigned getFileIDSize(FileID FID) const;
1374
1375 /// \brief Given a specific FileID, returns true if \p Loc is inside that
1376 /// FileID chunk and sets relative offset (offset of \p Loc from beginning
1377 /// of FileID) to \p relativeOffset.
1378 bool isInFileID(SourceLocation Loc, FileID FID,
1379 unsigned *RelativeOffset = nullptr) const {
1380 unsigned Offs = Loc.getOffset();
1381 if (isOffsetInFileID(FID, Offs)) {
1382 if (RelativeOffset)
1383 *RelativeOffset = Offs - getSLocEntry(FID).getOffset();
1384 return true;
1385 }
1386
1387 return false;
1388 }
1389
1390 //===--------------------------------------------------------------------===//
1391 // Line Table Manipulation Routines
1392 //===--------------------------------------------------------------------===//
1393
1394 /// \brief Return the uniqued ID for the specified filename.
1395 ///
1396 unsigned getLineTableFilenameID(StringRef Str);
1397
1398 /// \brief Add a line note to the line table for the FileID and offset
1399 /// specified by Loc.
1400 ///
1401 /// If FilenameID is -1, it is considered to be unspecified.
1402 void AddLineNote(SourceLocation Loc, unsigned LineNo, int FilenameID);
1403 void AddLineNote(SourceLocation Loc, unsigned LineNo, int FilenameID,
1404 bool IsFileEntry, bool IsFileExit,
1405 bool IsSystemHeader, bool IsExternCHeader);
1406
1407 /// \brief Determine if the source manager has a line table.
1408 bool hasLineTable() const { return LineTable != nullptr; }
1409
1410 /// \brief Retrieve the stored line table.
1411 LineTableInfo &getLineTable();
1412
1413 //===--------------------------------------------------------------------===//
1414 // Queries for performance analysis.
1415 //===--------------------------------------------------------------------===//
1416
1417 /// \brief Return the total amount of physical memory allocated by the
1418 /// ContentCache allocator.
1419 size_t getContentCacheSize() const {
1420 return ContentCacheAlloc.getTotalMemory();
1421 }
1422
1423 struct MemoryBufferSizes {
1424 const size_t malloc_bytes;
1425 const size_t mmap_bytes;
1426
1427 MemoryBufferSizes(size_t malloc_bytes, size_t mmap_bytes)
1428 : malloc_bytes(malloc_bytes), mmap_bytes(mmap_bytes) {}
1429 };
1430
1431 /// \brief Return the amount of memory used by memory buffers, breaking down
1432 /// by heap-backed versus mmap'ed memory.
1433 MemoryBufferSizes getMemoryBufferSizes() const;
1434
1435 /// \brief Return the amount of memory used for various side tables and
1436 /// data structures in the SourceManager.
1437 size_t getDataStructureSizes() const;
1438
1439 //===--------------------------------------------------------------------===//
1440 // Other miscellaneous methods.
1441 //===--------------------------------------------------------------------===//
1442
1443 /// \brief Get the source location for the given file:line:col triplet.
1444 ///
1445 /// If the source file is included multiple times, the source location will
1446 /// be based upon the first inclusion.
1447 SourceLocation translateFileLineCol(const FileEntry *SourceFile,
1448 unsigned Line, unsigned Col) const;
1449
1450 /// \brief Get the FileID for the given file.
1451 ///
1452 /// If the source file is included multiple times, the FileID will be the
1453 /// first inclusion.
1454 FileID translateFile(const FileEntry *SourceFile) const;
1455
1456 /// \brief Get the source location in \p FID for the given line:col.
1457 /// Returns null location if \p FID is not a file SLocEntry.
1458 SourceLocation translateLineCol(FileID FID,
1459 unsigned Line, unsigned Col) const;
1460
1461 /// \brief If \p Loc points inside a function macro argument, the returned
1462 /// location will be the macro location in which the argument was expanded.
1463 /// If a macro argument is used multiple times, the expanded location will
1464 /// be at the first expansion of the argument.
1465 /// e.g.
1466 /// MY_MACRO(foo);
1467 /// ^
1468 /// Passing a file location pointing at 'foo', will yield a macro location
1469 /// where 'foo' was expanded into.
1470 SourceLocation getMacroArgExpandedLocation(SourceLocation Loc) const;
1471
1472 /// \brief Determines the order of 2 source locations in the translation unit.
1473 ///
1474 /// \returns true if LHS source location comes before RHS, false otherwise.
1475 bool isBeforeInTranslationUnit(SourceLocation LHS, SourceLocation RHS) const;
1476
1477 /// \brief Determines the order of 2 source locations in the "source location
1478 /// address space".
1479 bool isBeforeInSLocAddrSpace(SourceLocation LHS, SourceLocation RHS) const {
1480 return isBeforeInSLocAddrSpace(LHS, RHS.getOffset());
1481 }
1482
1483 /// \brief Determines the order of a source location and a source location
1484 /// offset in the "source location address space".
1485 ///
1486 /// Note that we always consider source locations loaded from
1487 bool isBeforeInSLocAddrSpace(SourceLocation LHS, unsigned RHS) const {
1488 unsigned LHSOffset = LHS.getOffset();
1489 bool LHSLoaded = LHSOffset >= CurrentLoadedOffset;
1490 bool RHSLoaded = RHS >= CurrentLoadedOffset;
1491 if (LHSLoaded == RHSLoaded)
1492 return LHSOffset < RHS;
1493
1494 return LHSLoaded;
1495 }
1496
1497 // Iterators over FileInfos.
1498 typedef llvm::DenseMap<const FileEntry*, SrcMgr::ContentCache*>
1499 ::const_iterator fileinfo_iterator;
1500 fileinfo_iterator fileinfo_begin() const { return FileInfos.begin(); }
1501 fileinfo_iterator fileinfo_end() const { return FileInfos.end(); }
1502 bool hasFileInfo(const FileEntry *File) const {
1503 return FileInfos.find(File) != FileInfos.end();
1504 }
1505
1506 /// \brief Print statistics to stderr.
1507 ///
1508 void PrintStats() const;
1509
1510 void dump() const;
1511
1512 /// \brief Get the number of local SLocEntries we have.
1513 unsigned local_sloc_entry_size() const { return LocalSLocEntryTable.size(); }
1514
1515 /// \brief Get a local SLocEntry. This is exposed for indexing.
1516 const SrcMgr::SLocEntry &getLocalSLocEntry(unsigned Index,
1517 bool *Invalid = nullptr) const {
1518 assert(Index < LocalSLocEntryTable.size() && "Invalid index");
1519 return LocalSLocEntryTable[Index];
1520 }
1521
1522 /// \brief Get the number of loaded SLocEntries we have.
1523 unsigned loaded_sloc_entry_size() const { return LoadedSLocEntryTable.size();}
1524
1525 /// \brief Get a loaded SLocEntry. This is exposed for indexing.
1526 const SrcMgr::SLocEntry &getLoadedSLocEntry(unsigned Index,
1527 bool *Invalid = nullptr) const {
1528 assert(Index < LoadedSLocEntryTable.size() && "Invalid index");
1529 if (SLocEntryLoaded[Index])
1530 return LoadedSLocEntryTable[Index];
1531 return loadSLocEntry(Index, Invalid);
1532 }
1533
1534 const SrcMgr::SLocEntry &getSLocEntry(FileID FID,
1535 bool *Invalid = nullptr) const {
1536 if (FID.ID == 0 || FID.ID == -1) {
1537 if (Invalid) *Invalid = true;
1538 return LocalSLocEntryTable[0];
1539 }
1540 return getSLocEntryByID(FID.ID, Invalid);
1541 }
1542
1543 unsigned getNextLocalOffset() const { return NextLocalOffset; }
1544
1545 void setExternalSLocEntrySource(ExternalSLocEntrySource *Source) {
1546 assert(LoadedSLocEntryTable.empty() &&
1547 "Invalidating existing loaded entries");
1548 ExternalSLocEntries = Source;
1549 }
1550
1551 /// \brief Allocate a number of loaded SLocEntries, which will be actually
1552 /// loaded on demand from the external source.
1553 ///
1554 /// NumSLocEntries will be allocated, which occupy a total of TotalSize space
1555 /// in the global source view. The lowest ID and the base offset of the
1556 /// entries will be returned.
1557 std::pair<int, unsigned>
1558 AllocateLoadedSLocEntries(unsigned NumSLocEntries, unsigned TotalSize);
1559
1560 /// \brief Returns true if \p Loc came from a PCH/Module.
1561 bool isLoadedSourceLocation(SourceLocation Loc) const {
1562 return Loc.getOffset() >= CurrentLoadedOffset;
1563 }
1564
1565 /// \brief Returns true if \p Loc did not come from a PCH/Module.
1566 bool isLocalSourceLocation(SourceLocation Loc) const {
1567 return Loc.getOffset() < NextLocalOffset;
1568 }
1569
1570 /// \brief Returns true if \p FID came from a PCH/Module.
1571 bool isLoadedFileID(FileID FID) const {
1572 assert(FID.ID != -1 && "Using FileID sentinel value");
1573 return FID.ID < 0;
1574 }
1575
1576 /// \brief Returns true if \p FID did not come from a PCH/Module.
1577 bool isLocalFileID(FileID FID) const {
1578 return !isLoadedFileID(FID);
1579 }
1580
1581 /// Gets the location of the immediate macro caller, one level up the stack
1582 /// toward the initial macro typed into the source.
1583 SourceLocation getImmediateMacroCallerLoc(SourceLocation Loc) const {
1584 if (!Loc.isMacroID()) return Loc;
1585
1586 // When we have the location of (part of) an expanded parameter, its
1587 // spelling location points to the argument as expanded in the macro call,
1588 // and therefore is used to locate the macro caller.
1589 if (isMacroArgExpansion(Loc))
1590 return getImmediateSpellingLoc(Loc);
1591
1592 // Otherwise, the caller of the macro is located where this macro is
1593 // expanded (while the spelling is part of the macro definition).
1594 return getImmediateExpansionRange(Loc).first;
1595 }
1596
1597private:
1598 llvm::MemoryBuffer *getFakeBufferForRecovery() const;
1599 const SrcMgr::ContentCache *getFakeContentCacheForRecovery() const;
1600
1601 const SrcMgr::SLocEntry &loadSLocEntry(unsigned Index, bool *Invalid) const;
1602
1603 /// \brief Get the entry with the given unwrapped FileID.
1604 const SrcMgr::SLocEntry &getSLocEntryByID(int ID,
1605 bool *Invalid = nullptr) const {
1606 assert(ID != -1 && "Using FileID sentinel value");
1607 if (ID < 0)
1608 return getLoadedSLocEntryByID(ID, Invalid);
1609 return getLocalSLocEntry(static_cast<unsigned>(ID), Invalid);
1610 }
1611
1612 const SrcMgr::SLocEntry &
1613 getLoadedSLocEntryByID(int ID, bool *Invalid = nullptr) const {
1614 return getLoadedSLocEntry(static_cast<unsigned>(-ID - 2), Invalid);
1615 }
1616
1617 /// Implements the common elements of storing an expansion info struct into
1618 /// the SLocEntry table and producing a source location that refers to it.
1619 SourceLocation createExpansionLocImpl(const SrcMgr::ExpansionInfo &Expansion,
1620 unsigned TokLength,
1621 int LoadedID = 0,
1622 unsigned LoadedOffset = 0);
1623
1624 /// \brief Return true if the specified FileID contains the
1625 /// specified SourceLocation offset. This is a very hot method.
1626 inline bool isOffsetInFileID(FileID FID, unsigned SLocOffset) const {
1627 const SrcMgr::SLocEntry &Entry = getSLocEntry(FID);
1628 // If the entry is after the offset, it can't contain it.
1629 if (SLocOffset < Entry.getOffset()) return false;
1630
1631 // If this is the very last entry then it does.
1632 if (FID.ID == -2)
1633 return true;
1634
1635 // If it is the last local entry, then it does if the location is local.
1636 if (FID.ID+1 == static_cast<int>(LocalSLocEntryTable.size()))
1637 return SLocOffset < NextLocalOffset;
1638
1639 // Otherwise, the entry after it has to not include it. This works for both
1640 // local and loaded entries.
1641 return SLocOffset < getSLocEntryByID(FID.ID+1).getOffset();
1642 }
1643
1644 /// \brief Returns the previous in-order FileID or an invalid FileID if there
1645 /// is no previous one.
1646 FileID getPreviousFileID(FileID FID) const;
1647
1648 /// \brief Returns the next in-order FileID or an invalid FileID if there is
1649 /// no next one.
1650 FileID getNextFileID(FileID FID) const;
1651
1652 /// \brief Create a new fileID for the specified ContentCache and
1653 /// include position.
1654 ///
1655 /// This works regardless of whether the ContentCache corresponds to a
1656 /// file or some other input source.
1657 FileID createFileID(const SrcMgr::ContentCache* File,
1658 SourceLocation IncludePos,
1659 SrcMgr::CharacteristicKind DirCharacter,
1660 int LoadedID, unsigned LoadedOffset);
1661
1662 const SrcMgr::ContentCache *
1663 getOrCreateContentCache(const FileEntry *SourceFile,
1664 bool isSystemFile = false);
1665
1666 /// \brief Create a new ContentCache for the specified memory buffer.
1667 const SrcMgr::ContentCache *
1668 createMemBufferContentCache(std::unique_ptr<llvm::MemoryBuffer> Buf);
1669
1670 FileID getFileIDSlow(unsigned SLocOffset) const;
1671 FileID getFileIDLocal(unsigned SLocOffset) const;
1672 FileID getFileIDLoaded(unsigned SLocOffset) const;
1673
1674 SourceLocation getExpansionLocSlowCase(SourceLocation Loc) const;
1675 SourceLocation getSpellingLocSlowCase(SourceLocation Loc) const;
1676 SourceLocation getFileLocSlowCase(SourceLocation Loc) const;
1677
1678 std::pair<FileID, unsigned>
1679 getDecomposedExpansionLocSlowCase(const SrcMgr::SLocEntry *E) const;
1680 std::pair<FileID, unsigned>
1681 getDecomposedSpellingLocSlowCase(const SrcMgr::SLocEntry *E,
1682 unsigned Offset) const;
1683 void computeMacroArgsCache(MacroArgsMap &MacroArgsCache, FileID FID) const;
1684 void associateFileChunkWithMacroArgExp(MacroArgsMap &MacroArgsCache,
1685 FileID FID,
1686 SourceLocation SpellLoc,
1687 SourceLocation ExpansionLoc,
1688 unsigned ExpansionLength) const;
1689 friend class ASTReader;
1690 friend class ASTWriter;
1691};
1692
1693/// \brief Comparison function object.
1694template<typename T>
1695class BeforeThanCompare;
1696
1697/// \brief Compare two source locations.
1698template<>
1699class BeforeThanCompare<SourceLocation> {
1700 SourceManager &SM;
1701
1702public:
1703 explicit BeforeThanCompare(SourceManager &SM) : SM(SM) { }
1704
1705 bool operator()(SourceLocation LHS, SourceLocation RHS) const {
1706 return SM.isBeforeInTranslationUnit(LHS, RHS);
1707 }
1708};
1709
1710/// \brief Compare two non-overlapping source ranges.
1711template<>
1712class BeforeThanCompare<SourceRange> {
1713 SourceManager &SM;
1714
1715public:
1716 explicit BeforeThanCompare(SourceManager &SM) : SM(SM) { }
1717
1718 bool operator()(SourceRange LHS, SourceRange RHS) const {
1719 return SM.isBeforeInTranslationUnit(LHS.getBegin(), RHS.getBegin());
1720 }
1721};
1722
1723} // end namespace clang
1724
1725#endif // LLVM_CLANG_BASIC_SOURCEMANAGER_H
1726