1 //===--- SourceLocation.h - Compact identifier for 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 clang::SourceLocation class and associated facilities.
12 ///
13 //===----------------------------------------------------------------------===//
14 
15 #ifndef LLVM_CLANG_BASIC_SOURCELOCATION_H
16 #define LLVM_CLANG_BASIC_SOURCELOCATION_H
17 
18 #include "clang/Basic/LLVM.h"
19 #include "llvm/Support/Compiler.h"
20 #include "llvm/Support/PointerLikeTypeTraits.h"
21 #include <cassert>
22 #include <functional>
23 #include <string>
24 #include <utility>
25 
26 namespace llvm {
27   class MemoryBuffer;
28   template <typename T> struct DenseMapInfo;
29   template <typename T> struct isPodLike;
30 }
31 
32 namespace clang {
33 
34 class SourceManager;
35 
36 /// \brief An opaque identifier used by SourceManager which refers to a
37 /// source file (MemoryBuffer) along with its \#include path and \#line data.
38 ///
39 class FileID {
40   /// \brief A mostly-opaque identifier, where 0 is "invalid", >0 is
41   /// this module, and <-1 is something loaded from another module.
42   int ID;
43 public:
FileID()44   FileID() : ID(0) {}
45 
isValid()46   bool isValid() const { return ID != 0; }
isInvalid()47   bool isInvalid() const { return ID == 0; }
48 
49   bool operator==(const FileID &RHS) const { return ID == RHS.ID; }
50   bool operator<(const FileID &RHS) const { return ID < RHS.ID; }
51   bool operator<=(const FileID &RHS) const { return ID <= RHS.ID; }
52   bool operator!=(const FileID &RHS) const { return !(*this == RHS); }
53   bool operator>(const FileID &RHS) const { return RHS < *this; }
54   bool operator>=(const FileID &RHS) const { return RHS <= *this; }
55 
getSentinel()56   static FileID getSentinel() { return get(-1); }
getHashValue()57   unsigned getHashValue() const { return static_cast<unsigned>(ID); }
58 
59 private:
60   friend class SourceManager;
61   friend class ASTWriter;
62   friend class ASTReader;
63 
get(int V)64   static FileID get(int V) {
65     FileID F;
66     F.ID = V;
67     return F;
68   }
getOpaqueValue()69   int getOpaqueValue() const { return ID; }
70 };
71 
72 
73 /// \brief Encodes a location in the source. The SourceManager can decode this
74 /// to get at the full include stack, line and column information.
75 ///
76 /// Technically, a source location is simply an offset into the manager's view
77 /// of the input source, which is all input buffers (including macro
78 /// expansions) concatenated in an effectively arbitrary order. The manager
79 /// actually maintains two blocks of input buffers. One, starting at offset
80 /// 0 and growing upwards, contains all buffers from this module. The other,
81 /// starting at the highest possible offset and growing downwards, contains
82 /// buffers of loaded modules.
83 ///
84 /// In addition, one bit of SourceLocation is used for quick access to the
85 /// information whether the location is in a file or a macro expansion.
86 ///
87 /// It is important that this type remains small. It is currently 32 bits wide.
88 class SourceLocation {
89   unsigned ID;
90   friend class SourceManager;
91   friend class ASTReader;
92   friend class ASTWriter;
93   enum : unsigned {
94     MacroIDBit = 1U << 31
95   };
96 public:
97 
SourceLocation()98   SourceLocation() : ID(0) {}
99 
isFileID()100   bool isFileID() const  { return (ID & MacroIDBit) == 0; }
isMacroID()101   bool isMacroID() const { return (ID & MacroIDBit) != 0; }
102 
103   /// \brief Return true if this is a valid SourceLocation object.
104   ///
105   /// Invalid SourceLocations are often used when events have no corresponding
106   /// location in the source (e.g. a diagnostic is required for a command line
107   /// option).
isValid()108   bool isValid() const { return ID != 0; }
isInvalid()109   bool isInvalid() const { return ID == 0; }
110 
111 private:
112   /// \brief Return the offset into the manager's global input view.
getOffset()113   unsigned getOffset() const {
114     return ID & ~MacroIDBit;
115   }
116 
getFileLoc(unsigned ID)117   static SourceLocation getFileLoc(unsigned ID) {
118     assert((ID & MacroIDBit) == 0 && "Ran out of source locations!");
119     SourceLocation L;
120     L.ID = ID;
121     return L;
122   }
123 
getMacroLoc(unsigned ID)124   static SourceLocation getMacroLoc(unsigned ID) {
125     assert((ID & MacroIDBit) == 0 && "Ran out of source locations!");
126     SourceLocation L;
127     L.ID = MacroIDBit | ID;
128     return L;
129   }
130 public:
131 
132   /// \brief Return a source location with the specified offset from this
133   /// SourceLocation.
getLocWithOffset(int Offset)134   SourceLocation getLocWithOffset(int Offset) const {
135     assert(((getOffset()+Offset) & MacroIDBit) == 0 && "offset overflow");
136     SourceLocation L;
137     L.ID = ID+Offset;
138     return L;
139   }
140 
141   /// \brief When a SourceLocation itself cannot be used, this returns
142   /// an (opaque) 32-bit integer encoding for it.
143   ///
144   /// This should only be passed to SourceLocation::getFromRawEncoding, it
145   /// should not be inspected directly.
getRawEncoding()146   unsigned getRawEncoding() const { return ID; }
147 
148   /// \brief Turn a raw encoding of a SourceLocation object into
149   /// a real SourceLocation.
150   ///
151   /// \see getRawEncoding.
getFromRawEncoding(unsigned Encoding)152   static SourceLocation getFromRawEncoding(unsigned Encoding) {
153     SourceLocation X;
154     X.ID = Encoding;
155     return X;
156   }
157 
158   /// \brief When a SourceLocation itself cannot be used, this returns
159   /// an (opaque) pointer encoding for it.
160   ///
161   /// This should only be passed to SourceLocation::getFromPtrEncoding, it
162   /// should not be inspected directly.
getPtrEncoding()163   void* getPtrEncoding() const {
164     // Double cast to avoid a warning "cast to pointer from integer of different
165     // size".
166     return (void*)(uintptr_t)getRawEncoding();
167   }
168 
169   /// \brief Turn a pointer encoding of a SourceLocation object back
170   /// into a real SourceLocation.
getFromPtrEncoding(const void * Encoding)171   static SourceLocation getFromPtrEncoding(const void *Encoding) {
172     return getFromRawEncoding((unsigned)(uintptr_t)Encoding);
173   }
174 
175   void print(raw_ostream &OS, const SourceManager &SM) const;
176   std::string printToString(const SourceManager &SM) const;
177   void dump(const SourceManager &SM) const;
178 };
179 
180 inline bool operator==(const SourceLocation &LHS, const SourceLocation &RHS) {
181   return LHS.getRawEncoding() == RHS.getRawEncoding();
182 }
183 
184 inline bool operator!=(const SourceLocation &LHS, const SourceLocation &RHS) {
185   return !(LHS == RHS);
186 }
187 
188 inline bool operator<(const SourceLocation &LHS, const SourceLocation &RHS) {
189   return LHS.getRawEncoding() < RHS.getRawEncoding();
190 }
191 
192 /// \brief A trivial tuple used to represent a source range.
193 class SourceRange {
194   SourceLocation B;
195   SourceLocation E;
196 public:
SourceRange()197   SourceRange(): B(SourceLocation()), E(SourceLocation()) {}
SourceRange(SourceLocation loc)198   SourceRange(SourceLocation loc) : B(loc), E(loc) {}
SourceRange(SourceLocation begin,SourceLocation end)199   SourceRange(SourceLocation begin, SourceLocation end) : B(begin), E(end) {}
200 
getBegin()201   SourceLocation getBegin() const { return B; }
getEnd()202   SourceLocation getEnd() const { return E; }
203 
setBegin(SourceLocation b)204   void setBegin(SourceLocation b) { B = b; }
setEnd(SourceLocation e)205   void setEnd(SourceLocation e) { E = e; }
206 
isValid()207   bool isValid() const { return B.isValid() && E.isValid(); }
isInvalid()208   bool isInvalid() const { return !isValid(); }
209 
210   bool operator==(const SourceRange &X) const {
211     return B == X.B && E == X.E;
212   }
213 
214   bool operator!=(const SourceRange &X) const {
215     return B != X.B || E != X.E;
216   }
217 };
218 
219 /// \brief Represents a character-granular source range.
220 ///
221 /// The underlying SourceRange can either specify the starting/ending character
222 /// of the range, or it can specify the start of the range and the start of the
223 /// last token of the range (a "token range").  In the token range case, the
224 /// size of the last token must be measured to determine the actual end of the
225 /// range.
226 class CharSourceRange {
227   SourceRange Range;
228   bool IsTokenRange;
229 public:
CharSourceRange()230   CharSourceRange() : IsTokenRange(false) {}
CharSourceRange(SourceRange R,bool ITR)231   CharSourceRange(SourceRange R, bool ITR) : Range(R), IsTokenRange(ITR) {}
232 
getTokenRange(SourceRange R)233   static CharSourceRange getTokenRange(SourceRange R) {
234     return CharSourceRange(R, true);
235   }
236 
getCharRange(SourceRange R)237   static CharSourceRange getCharRange(SourceRange R) {
238     return CharSourceRange(R, false);
239   }
240 
getTokenRange(SourceLocation B,SourceLocation E)241   static CharSourceRange getTokenRange(SourceLocation B, SourceLocation E) {
242     return getTokenRange(SourceRange(B, E));
243   }
getCharRange(SourceLocation B,SourceLocation E)244   static CharSourceRange getCharRange(SourceLocation B, SourceLocation E) {
245     return getCharRange(SourceRange(B, E));
246   }
247 
248   /// \brief Return true if the end of this range specifies the start of
249   /// the last token.  Return false if the end of this range specifies the last
250   /// character in the range.
isTokenRange()251   bool isTokenRange() const { return IsTokenRange; }
isCharRange()252   bool isCharRange() const { return !IsTokenRange; }
253 
getBegin()254   SourceLocation getBegin() const { return Range.getBegin(); }
getEnd()255   SourceLocation getEnd() const { return Range.getEnd(); }
getAsRange()256   SourceRange getAsRange() const { return Range; }
257 
setBegin(SourceLocation b)258   void setBegin(SourceLocation b) { Range.setBegin(b); }
setEnd(SourceLocation e)259   void setEnd(SourceLocation e) { Range.setEnd(e); }
260 
isValid()261   bool isValid() const { return Range.isValid(); }
isInvalid()262   bool isInvalid() const { return !isValid(); }
263 };
264 
265 /// \brief A SourceLocation and its associated SourceManager.
266 ///
267 /// This is useful for argument passing to functions that expect both objects.
268 class FullSourceLoc : public SourceLocation {
269   const SourceManager *SrcMgr;
270 public:
271   /// \brief Creates a FullSourceLoc where isValid() returns \c false.
FullSourceLoc()272   explicit FullSourceLoc() : SrcMgr(nullptr) {}
273 
FullSourceLoc(SourceLocation Loc,const SourceManager & SM)274   explicit FullSourceLoc(SourceLocation Loc, const SourceManager &SM)
275     : SourceLocation(Loc), SrcMgr(&SM) {}
276 
277   /// \pre This FullSourceLoc has an associated SourceManager.
getManager()278   const SourceManager &getManager() const {
279     assert(SrcMgr && "SourceManager is NULL.");
280     return *SrcMgr;
281   }
282 
283   FileID getFileID() const;
284 
285   FullSourceLoc getExpansionLoc() const;
286   FullSourceLoc getSpellingLoc() const;
287 
288   unsigned getExpansionLineNumber(bool *Invalid = nullptr) const;
289   unsigned getExpansionColumnNumber(bool *Invalid = nullptr) const;
290 
291   unsigned getSpellingLineNumber(bool *Invalid = nullptr) const;
292   unsigned getSpellingColumnNumber(bool *Invalid = nullptr) const;
293 
294   const char *getCharacterData(bool *Invalid = nullptr) const;
295 
296 
297   /// \brief Return a StringRef to the source buffer data for the
298   /// specified FileID.
299   StringRef getBufferData(bool *Invalid = nullptr) const;
300 
301   /// \brief Decompose the specified location into a raw FileID + Offset pair.
302   ///
303   /// The first element is the FileID, the second is the offset from the
304   /// start of the buffer of the location.
305   std::pair<FileID, unsigned> getDecomposedLoc() const;
306 
307   bool isInSystemHeader() const;
308 
309   /// \brief Determines the order of 2 source locations in the translation unit.
310   ///
311   /// \returns true if this source location comes before 'Loc', false otherwise.
312   bool isBeforeInTranslationUnitThan(SourceLocation Loc) const;
313 
314   /// \brief Determines the order of 2 source locations in the translation unit.
315   ///
316   /// \returns true if this source location comes before 'Loc', false otherwise.
isBeforeInTranslationUnitThan(FullSourceLoc Loc)317   bool isBeforeInTranslationUnitThan(FullSourceLoc Loc) const {
318     assert(Loc.isValid());
319     assert(SrcMgr == Loc.SrcMgr && "Loc comes from another SourceManager!");
320     return isBeforeInTranslationUnitThan((SourceLocation)Loc);
321   }
322 
323   /// \brief Comparison function class, useful for sorting FullSourceLocs.
324   struct BeforeThanCompare : public std::binary_function<FullSourceLoc,
325                                                          FullSourceLoc, bool> {
operatorBeforeThanCompare326     bool operator()(const FullSourceLoc& lhs, const FullSourceLoc& rhs) const {
327       return lhs.isBeforeInTranslationUnitThan(rhs);
328     }
329   };
330 
331   /// \brief Prints information about this FullSourceLoc to stderr.
332   ///
333   /// This is useful for debugging.
334   void dump() const;
335 
336   friend inline bool
337   operator==(const FullSourceLoc &LHS, const FullSourceLoc &RHS) {
338     return LHS.getRawEncoding() == RHS.getRawEncoding() &&
339           LHS.SrcMgr == RHS.SrcMgr;
340   }
341 
342   friend inline bool
343   operator!=(const FullSourceLoc &LHS, const FullSourceLoc &RHS) {
344     return !(LHS == RHS);
345   }
346 
347 };
348 
349 /// \brief Represents an unpacked "presumed" location which can be presented
350 /// to the user.
351 ///
352 /// A 'presumed' location can be modified by \#line and GNU line marker
353 /// directives and is always the expansion point of a normal location.
354 ///
355 /// You can get a PresumedLoc from a SourceLocation with SourceManager.
356 class PresumedLoc {
357   const char *Filename;
358   unsigned Line, Col;
359   SourceLocation IncludeLoc;
360 public:
PresumedLoc()361   PresumedLoc() : Filename(nullptr) {}
PresumedLoc(const char * FN,unsigned Ln,unsigned Co,SourceLocation IL)362   PresumedLoc(const char *FN, unsigned Ln, unsigned Co, SourceLocation IL)
363     : Filename(FN), Line(Ln), Col(Co), IncludeLoc(IL) {
364   }
365 
366   /// \brief Return true if this object is invalid or uninitialized.
367   ///
368   /// This occurs when created with invalid source locations or when walking
369   /// off the top of a \#include stack.
isInvalid()370   bool isInvalid() const { return Filename == nullptr; }
isValid()371   bool isValid() const { return Filename != nullptr; }
372 
373   /// \brief Return the presumed filename of this location.
374   ///
375   /// This can be affected by \#line etc.
getFilename()376   const char *getFilename() const { assert(isValid()); return Filename; }
377 
378   /// \brief Return the presumed line number of this location.
379   ///
380   /// This can be affected by \#line etc.
getLine()381   unsigned getLine() const { assert(isValid()); return Line; }
382 
383   /// \brief Return the presumed column number of this location.
384   ///
385   /// This cannot be affected by \#line, but is packaged here for convenience.
getColumn()386   unsigned getColumn() const { assert(isValid()); return Col; }
387 
388   /// \brief Return the presumed include location of this location.
389   ///
390   /// This can be affected by GNU linemarker directives.
getIncludeLoc()391   SourceLocation getIncludeLoc() const { assert(isValid()); return IncludeLoc; }
392 };
393 
394 
395 }  // end namespace clang
396 
397 namespace llvm {
398   /// Define DenseMapInfo so that FileID's can be used as keys in DenseMap and
399   /// DenseSets.
400   template <>
401   struct DenseMapInfo<clang::FileID> {
402     static inline clang::FileID getEmptyKey() {
403       return clang::FileID();
404     }
405     static inline clang::FileID getTombstoneKey() {
406       return clang::FileID::getSentinel();
407     }
408 
409     static unsigned getHashValue(clang::FileID S) {
410       return S.getHashValue();
411     }
412 
413     static bool isEqual(clang::FileID LHS, clang::FileID RHS) {
414       return LHS == RHS;
415     }
416   };
417 
418   template <>
419   struct isPodLike<clang::SourceLocation> { static const bool value = true; };
420   template <>
421   struct isPodLike<clang::FileID> { static const bool value = true; };
422 
423   // Teach SmallPtrSet how to handle SourceLocation.
424   template<>
425   class PointerLikeTypeTraits<clang::SourceLocation> {
426   public:
427     static inline void *getAsVoidPointer(clang::SourceLocation L) {
428       return L.getPtrEncoding();
429     }
430     static inline clang::SourceLocation getFromVoidPointer(void *P) {
431       return clang::SourceLocation::getFromRawEncoding((unsigned)(uintptr_t)P);
432     }
433     enum { NumLowBitsAvailable = 0 };
434   };
435 
436 }  // end namespace llvm
437 
438 #endif
439