1//===- SourceLocation.h - Compact identifier for Source Files ---*- 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/// \file
10/// Defines the clang::SourceLocation class and associated facilities.
11//
12//===----------------------------------------------------------------------===//
13
14#ifndef LLVM_CLANG_BASIC_SOURCELOCATION_H
15#define LLVM_CLANG_BASIC_SOURCELOCATION_H
16
17#include "clang/Basic/LLVM.h"
18#include "llvm/ADT/StringRef.h"
19#include "llvm/Support/PointerLikeTypeTraits.h"
20#include <cassert>
21#include <cstdint>
22#include <string>
23#include <utility>
24
25namespace llvm {
26
27template <typename T> struct DenseMapInfo;
28
29} // namespace llvm
30
31namespace clang {
32
33class SourceManager;
34
35/// An opaque identifier used by SourceManager which refers to a
36/// source file (MemoryBuffer) along with its \#include path and \#line data.
37///
38class FileID {
39 /// A mostly-opaque identifier, where 0 is "invalid", >0 is
40 /// this module, and <-1 is something loaded from another module.
41 int ID = 0;
42
43public:
44 bool isValid() const { return ID != 0; }
45 bool isInvalid() const { return ID == 0; }
46
47 bool operator==(const FileID &RHS) const { return ID == RHS.ID; }
48 bool operator<(const FileID &RHS) const { return ID < RHS.ID; }
49 bool operator<=(const FileID &RHS) const { return ID <= RHS.ID; }
50 bool operator!=(const FileID &RHS) const { return !(*this == RHS); }
51 bool operator>(const FileID &RHS) const { return RHS < *this; }
52 bool operator>=(const FileID &RHS) const { return RHS <= *this; }
53
54 static FileID getSentinel() { return get(-1); }
55 unsigned getHashValue() const { return static_cast<unsigned>(ID); }
56
57private:
58 friend class ASTWriter;
59 friend class ASTReader;
60 friend class SourceManager;
61
62 static FileID get(int V) {
63 FileID F;
64 F.ID = V;
65 return F;
66 }
67
68 int getOpaqueValue() const { return ID; }
69};
70
71/// Encodes a location in the source. The SourceManager can decode this
72/// to get at the full include stack, line and column information.
73///
74/// Technically, a source location is simply an offset into the manager's view
75/// of the input source, which is all input buffers (including macro
76/// expansions) concatenated in an effectively arbitrary order. The manager
77/// actually maintains two blocks of input buffers. One, starting at offset
78/// 0 and growing upwards, contains all buffers from this module. The other,
79/// starting at the highest possible offset and growing downwards, contains
80/// buffers of loaded modules.
81///
82/// In addition, one bit of SourceLocation is used for quick access to the
83/// information whether the location is in a file or a macro expansion.
84///
85/// It is important that this type remains small. It is currently 32 bits wide.
86class SourceLocation {
87 friend class ASTReader;
88 friend class ASTWriter;
89 friend class SourceManager;
90
91 unsigned ID = 0;
92
93 enum : unsigned {
94 MacroIDBit = 1U << 31
95 };
96
97public:
98 bool isFileID() const { return (ID & MacroIDBit) == 0; }
99 bool isMacroID() const { return (ID & MacroIDBit) != 0; }
100
101 /// Return true if this is a valid SourceLocation object.
102 ///
103 /// Invalid SourceLocations are often used when events have no corresponding
104 /// location in the source (e.g. a diagnostic is required for a command line
105 /// option).
106 bool isValid() const { return ID != 0; }
107 bool isInvalid() const { return ID == 0; }
108
109private:
110 /// Return the offset into the manager's global input view.
111 unsigned getOffset() const {
112 return ID & ~MacroIDBit;
113 }
114
115 static SourceLocation getFileLoc(unsigned ID) {
116 assert((ID & MacroIDBit) == 0 && "Ran out of source locations!");
117 SourceLocation L;
118 L.ID = ID;
119 return L;
120 }
121
122 static SourceLocation getMacroLoc(unsigned ID) {
123 assert((ID & MacroIDBit) == 0 && "Ran out of source locations!");
124 SourceLocation L;
125 L.ID = MacroIDBit | ID;
126 return L;
127 }
128
129public:
130 /// Return a source location with the specified offset from this
131 /// SourceLocation.
132 SourceLocation getLocWithOffset(int Offset) const {
133 assert(((getOffset()+Offset) & MacroIDBit) == 0 && "offset overflow");
134 SourceLocation L;
135 L.ID = ID+Offset;
136 return L;
137 }
138
139 /// When a SourceLocation itself cannot be used, this returns
140 /// an (opaque) 32-bit integer encoding for it.
141 ///
142 /// This should only be passed to SourceLocation::getFromRawEncoding, it
143 /// should not be inspected directly.
144 unsigned getRawEncoding() const { return ID; }
145
146 /// Turn a raw encoding of a SourceLocation object into
147 /// a real SourceLocation.
148 ///
149 /// \see getRawEncoding.
150 static SourceLocation getFromRawEncoding(unsigned Encoding) {
151 SourceLocation X;
152 X.ID = Encoding;
153 return X;
154 }
155
156 /// When a SourceLocation itself cannot be used, this returns
157 /// an (opaque) pointer encoding for it.
158 ///
159 /// This should only be passed to SourceLocation::getFromPtrEncoding, it
160 /// should not be inspected directly.
161 void* getPtrEncoding() const {
162 // Double cast to avoid a warning "cast to pointer from integer of different
163 // size".
164 return (void*)(uintptr_t)getRawEncoding();
165 }
166
167 /// Turn a pointer encoding of a SourceLocation object back
168 /// into a real SourceLocation.
169 static SourceLocation getFromPtrEncoding(const void *Encoding) {
170 return getFromRawEncoding((unsigned)(uintptr_t)Encoding);
171 }
172
173 static bool isPairOfFileLocations(SourceLocation Start, SourceLocation End) {
174 return Start.isValid() && Start.isFileID() && End.isValid() &&
175 End.isFileID();
176 }
177
178 void print(raw_ostream &OS, const SourceManager &SM) const;
179 std::string printToString(const SourceManager &SM) const;
180 void dump(const SourceManager &SM) const;
181};
182
183inline bool operator==(const SourceLocation &LHS, const SourceLocation &RHS) {
184 return LHS.getRawEncoding() == RHS.getRawEncoding();
185}
186
187inline bool operator!=(const SourceLocation &LHS, const SourceLocation &RHS) {
188 return !(LHS == RHS);
189}
190
191inline bool operator<(const SourceLocation &LHS, const SourceLocation &RHS) {
192 return LHS.getRawEncoding() < RHS.getRawEncoding();
193}
194
195/// A trivial tuple used to represent a source range.
196class SourceRange {
197 SourceLocation B;
198 SourceLocation E;
199
200public:
201 SourceRange() = default;
202 SourceRange(SourceLocation loc) : B(loc), E(loc) {}
203 SourceRange(SourceLocation begin, SourceLocation end) : B(begin), E(end) {}
204
205 SourceLocation getBegin() const { return B; }
206 SourceLocation getEnd() const { return E; }
207
208 void setBegin(SourceLocation b) { B = b; }
209 void setEnd(SourceLocation e) { E = e; }
210
211 bool isValid() const { return B.isValid() && E.isValid(); }
212 bool isInvalid() const { return !isValid(); }
213
214 bool operator==(const SourceRange &X) const {
215 return B == X.B && E == X.E;
216 }
217
218 bool operator!=(const SourceRange &X) const {
219 return B != X.B || E != X.E;
220 }
221
222 void print(raw_ostream &OS, const SourceManager &SM) const;
223 std::string printToString(const SourceManager &SM) const;
224 void dump(const SourceManager &SM) const;
225};
226
227/// Represents a character-granular source range.
228///
229/// The underlying SourceRange can either specify the starting/ending character
230/// of the range, or it can specify the start of the range and the start of the
231/// last token of the range (a "token range"). In the token range case, the
232/// size of the last token must be measured to determine the actual end of the
233/// range.
234class CharSourceRange {
235 SourceRange Range;
236 bool IsTokenRange = false;
237
238public:
239 CharSourceRange() = default;
240 CharSourceRange(SourceRange R, bool ITR) : Range(R), IsTokenRange(ITR) {}
241
242 static CharSourceRange getTokenRange(SourceRange R) {
243 return CharSourceRange(R, true);
244 }
245
246 static CharSourceRange getCharRange(SourceRange R) {
247 return CharSourceRange(R, false);
248 }
249
250 static CharSourceRange getTokenRange(SourceLocation B, SourceLocation E) {
251 return getTokenRange(SourceRange(B, E));
252 }
253
254 static CharSourceRange getCharRange(SourceLocation B, SourceLocation E) {
255 return getCharRange(SourceRange(B, E));
256 }
257
258 /// Return true if the end of this range specifies the start of
259 /// the last token. Return false if the end of this range specifies the last
260 /// character in the range.
261 bool isTokenRange() const { return IsTokenRange; }
262 bool isCharRange() const { return !IsTokenRange; }
263
264 SourceLocation getBegin() const { return Range.getBegin(); }
265 SourceLocation getEnd() const { return Range.getEnd(); }
266 SourceRange getAsRange() const { return Range; }
267
268 void setBegin(SourceLocation b) { Range.setBegin(b); }
269 void setEnd(SourceLocation e) { Range.setEnd(e); }
270 void setTokenRange(bool TR) { IsTokenRange = TR; }
271
272 bool isValid() const { return Range.isValid(); }
273 bool isInvalid() const { return !isValid(); }
274};
275
276/// Represents an unpacked "presumed" location which can be presented
277/// to the user.
278///
279/// A 'presumed' location can be modified by \#line and GNU line marker
280/// directives and is always the expansion point of a normal location.
281///
282/// You can get a PresumedLoc from a SourceLocation with SourceManager.
283class PresumedLoc {
284 const char *Filename = nullptr;
285 unsigned Line, Col;
286 SourceLocation IncludeLoc;
287
288public:
289 PresumedLoc() = default;
290 PresumedLoc(const char *FN, unsigned Ln, unsigned Co, SourceLocation IL)
291 : Filename(FN), Line(Ln), Col(Co), IncludeLoc(IL) {}
292
293 /// Return true if this object is invalid or uninitialized.
294 ///
295 /// This occurs when created with invalid source locations or when walking
296 /// off the top of a \#include stack.
297 bool isInvalid() const { return Filename == nullptr; }
298 bool isValid() const { return Filename != nullptr; }
299
300 /// Return the presumed filename of this location.
301 ///
302 /// This can be affected by \#line etc.
303 const char *getFilename() const {
304 assert(isValid());
305 return Filename;
306 }
307
308 /// Return the presumed line number of this location.
309 ///
310 /// This can be affected by \#line etc.
311 unsigned getLine() const {
312 assert(isValid());
313 return Line;
314 }
315
316 /// Return the presumed column number of this location.
317 ///
318 /// This cannot be affected by \#line, but is packaged here for convenience.
319 unsigned getColumn() const {
320 assert(isValid());
321 return Col;
322 }
323
324 /// Return the presumed include location of this location.
325 ///
326 /// This can be affected by GNU linemarker directives.
327 SourceLocation getIncludeLoc() const {
328 assert(isValid());
329 return IncludeLoc;
330 }
331};
332
333class FileEntry;
334
335/// A SourceLocation and its associated SourceManager.
336///
337/// This is useful for argument passing to functions that expect both objects.
338class FullSourceLoc : public SourceLocation {
339 const SourceManager *SrcMgr = nullptr;
340
341public:
342 /// Creates a FullSourceLoc where isValid() returns \c false.
343 FullSourceLoc() = default;
344
345 explicit FullSourceLoc(SourceLocation Loc, const SourceManager &SM)
346 : SourceLocation(Loc), SrcMgr(&SM) {}
347
348 bool hasManager() const {
349 bool hasSrcMgr = SrcMgr != nullptr;
350 assert(hasSrcMgr == isValid() && "FullSourceLoc has location but no manager");
351 return hasSrcMgr;
352 }
353
354 /// \pre This FullSourceLoc has an associated SourceManager.
355 const SourceManager &getManager() const {
356 assert(SrcMgr && "SourceManager is NULL.");
357 return *SrcMgr;
358 }
359
360 FileID getFileID() const;
361
362 FullSourceLoc getExpansionLoc() const;
363 FullSourceLoc getSpellingLoc() const;
364 FullSourceLoc getFileLoc() const;
365 PresumedLoc getPresumedLoc(bool UseLineDirectives = true) const;
366 bool isMacroArgExpansion(FullSourceLoc *StartLoc = nullptr) const;
367 FullSourceLoc getImmediateMacroCallerLoc() const;
368 std::pair<FullSourceLoc, StringRef> getModuleImportLoc() const;
369 unsigned getFileOffset() const;
370
371 unsigned getExpansionLineNumber(bool *Invalid = nullptr) const;
372 unsigned getExpansionColumnNumber(bool *Invalid = nullptr) const;
373
374 unsigned getSpellingLineNumber(bool *Invalid = nullptr) const;
375 unsigned getSpellingColumnNumber(bool *Invalid = nullptr) const;
376
377 const char *getCharacterData(bool *Invalid = nullptr) const;
378
379 unsigned getLineNumber(bool *Invalid = nullptr) const;
380 unsigned getColumnNumber(bool *Invalid = nullptr) const;
381
382 const FileEntry *getFileEntry() const;
383
384 /// Return a StringRef to the source buffer data for the
385 /// specified FileID.
386 StringRef getBufferData(bool *Invalid = nullptr) const;
387
388 /// Decompose the specified location into a raw FileID + Offset pair.
389 ///
390 /// The first element is the FileID, the second is the offset from the
391 /// start of the buffer of the location.
392 std::pair<FileID, unsigned> getDecomposedLoc() const;
393
394 bool isInSystemHeader() const;
395
396 /// Determines the order of 2 source locations in the translation unit.
397 ///
398 /// \returns true if this source location comes before 'Loc', false otherwise.
399 bool isBeforeInTranslationUnitThan(SourceLocation Loc) const;
400
401 /// Determines the order of 2 source locations in the translation unit.
402 ///
403 /// \returns true if this source location comes before 'Loc', false otherwise.
404 bool isBeforeInTranslationUnitThan(FullSourceLoc Loc) const {
405 assert(Loc.isValid());
406 assert(SrcMgr == Loc.SrcMgr && "Loc comes from another SourceManager!");
407 return isBeforeInTranslationUnitThan((SourceLocation)Loc);
408 }
409
410 /// Comparison function class, useful for sorting FullSourceLocs.
411 struct BeforeThanCompare {
412 bool operator()(const FullSourceLoc& lhs, const FullSourceLoc& rhs) const {
413 return lhs.isBeforeInTranslationUnitThan(rhs);
414 }
415 };
416
417 /// Prints information about this FullSourceLoc to stderr.
418 ///
419 /// This is useful for debugging.
420 void dump() const;
421
422 friend bool
423 operator==(const FullSourceLoc &LHS, const FullSourceLoc &RHS) {
424 return LHS.getRawEncoding() == RHS.getRawEncoding() &&
425 LHS.SrcMgr == RHS.SrcMgr;
426 }
427
428 friend bool
429 operator!=(const FullSourceLoc &LHS, const FullSourceLoc &RHS) {
430 return !(LHS == RHS);
431 }
432};
433
434} // namespace clang
435
436namespace llvm {
437
438 /// Define DenseMapInfo so that FileID's can be used as keys in DenseMap and
439 /// DenseSets.
440 template <>
441 struct DenseMapInfo<clang::FileID> {
442 static clang::FileID getEmptyKey() {
443 return {};
444 }
445
446 static clang::FileID getTombstoneKey() {
447 return clang::FileID::getSentinel();
448 }
449
450 static unsigned getHashValue(clang::FileID S) {
451 return S.getHashValue();
452 }
453
454 static bool isEqual(clang::FileID LHS, clang::FileID RHS) {
455 return LHS == RHS;
456 }
457 };
458
459 // Teach SmallPtrSet how to handle SourceLocation.
460 template<>
461 struct PointerLikeTypeTraits<clang::SourceLocation> {
462 enum { NumLowBitsAvailable = 0 };
463
464 static void *getAsVoidPointer(clang::SourceLocation L) {
465 return L.getPtrEncoding();
466 }
467
468 static clang::SourceLocation getFromVoidPointer(void *P) {
469 return clang::SourceLocation::getFromRawEncoding((unsigned)(uintptr_t)P);
470 }
471 };
472
473} // namespace llvm
474
475#endif // LLVM_CLANG_BASIC_SOURCELOCATION_H
476