SourceMgr.h revision 360784 
1//===- SourceMgr.h - Manager for Source Buffers & Diagnostics ---*- C++ -*-===//
2//
3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4// See https://llvm.org/LICENSE.txt for license information.
5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6//
7//===----------------------------------------------------------------------===//
8//
9// This file declares the SMDiagnostic and SourceMgr classes.  This
10// provides a simple substrate for diagnostics, #include handling, and other low
11// level things for simple parsers.
12//
13//===----------------------------------------------------------------------===//
14
15#ifndef LLVM_SUPPORT_SOURCEMGR_H
16#define LLVM_SUPPORT_SOURCEMGR_H
17
18#include "llvm/ADT/ArrayRef.h"
19#include "llvm/ADT/None.h"
20#include "llvm/ADT/PointerUnion.h"
21#include "llvm/ADT/SmallVector.h"
22#include "llvm/ADT/StringRef.h"
23#include "llvm/ADT/Twine.h"
24#include "llvm/Support/MemoryBuffer.h"
25#include "llvm/Support/SMLoc.h"
26#include <algorithm>
27#include <cassert>
28#include <memory>
29#include <string>
30#include <utility>
31#include <vector>
32
33namespace llvm {
34
35class raw_ostream;
36class SMDiagnostic;
37class SMFixIt;
38
39/// This owns the files read by a parser, handles include stacks,
40/// and handles diagnostic wrangling.
41class SourceMgr {
42public:
43  enum DiagKind {
44    DK_Error,
45    DK_Warning,
46    DK_Remark,
47    DK_Note,
48  };
49
50  /// Clients that want to handle their own diagnostics in a custom way can
51  /// register a function pointer+context as a diagnostic handler.
52  /// It gets called each time PrintMessage is invoked.
53  using DiagHandlerTy = void (*)(const SMDiagnostic &, void *Context);
54
55private:
56  struct SrcBuffer {
57    /// The memory buffer for the file.
58    std::unique_ptr<MemoryBuffer> Buffer;
59
60    /// Helper type for OffsetCache below: since we're storing many offsets
61    /// into relatively small files (often smaller than 2^8 or 2^16 bytes),
62    /// we select the offset vector element type dynamically based on the
63    /// size of Buffer.
64    using VariableSizeOffsets = PointerUnion<std::vector<uint8_t> *,
65                                             std::vector<uint16_t> *,
66                                             std::vector<uint32_t> *,
67                                             std::vector<uint64_t> *>;
68
69    /// Vector of offsets into Buffer at which there are line-endings
70    /// (lazily populated). Once populated, the '\n' that marks the end of
71    /// line number N from [1..] is at Buffer[OffsetCache[N-1]]. Since
72    /// these offsets are in sorted (ascending) order, they can be
73    /// binary-searched for the first one after any given offset (eg. an
74    /// offset corresponding to a particular SMLoc).
75    mutable VariableSizeOffsets OffsetCache;
76
77    /// Populate \c OffsetCache and look up a given \p Ptr in it, assuming
78    /// it points somewhere into \c Buffer. The static type parameter \p T
79    /// must be an unsigned integer type from uint{8,16,32,64}_t large
80    /// enough to store offsets inside \c Buffer.
81    template<typename T>
82    unsigned getLineNumber(const char *Ptr) const;
83
84    /// This is the location of the parent include, or null if at the top level.
85    SMLoc IncludeLoc;
86
87    SrcBuffer() = default;
88    SrcBuffer(SrcBuffer &&);
89    SrcBuffer(const SrcBuffer &) = delete;
90    SrcBuffer &operator=(const SrcBuffer &) = delete;
91    ~SrcBuffer();
92  };
93
94  /// This is all of the buffers that we are reading from.
95  std::vector<SrcBuffer> Buffers;
96
97  // This is the list of directories we should search for include files in.
98  std::vector<std::string> IncludeDirectories;
99
100  DiagHandlerTy DiagHandler = nullptr;
101  void *DiagContext = nullptr;
102
103  bool isValidBufferID(unsigned i) const { return i && i <= Buffers.size(); }
104
105public:
106  SourceMgr() = default;
107  SourceMgr(const SourceMgr &) = delete;
108  SourceMgr &operator=(const SourceMgr &) = delete;
109  SourceMgr(SourceMgr &&) = default;
110  SourceMgr &operator=(SourceMgr &&) = default;
111  ~SourceMgr() = default;
112
113  void setIncludeDirs(const std::vector<std::string> &Dirs) {
114    IncludeDirectories = Dirs;
115  }
116
117  /// Specify a diagnostic handler to be invoked every time PrintMessage is
118  /// called. \p Ctx is passed into the handler when it is invoked.
119  void setDiagHandler(DiagHandlerTy DH, void *Ctx = nullptr) {
120    DiagHandler = DH;
121    DiagContext = Ctx;
122  }
123
124  DiagHandlerTy getDiagHandler() const { return DiagHandler; }
125  void *getDiagContext() const { return DiagContext; }
126
127  const SrcBuffer &getBufferInfo(unsigned i) const {
128    assert(isValidBufferID(i));
129    return Buffers[i - 1];
130  }
131
132  const MemoryBuffer *getMemoryBuffer(unsigned i) const {
133    assert(isValidBufferID(i));
134    return Buffers[i - 1].Buffer.get();
135  }
136
137  unsigned getNumBuffers() const {
138    return Buffers.size();
139  }
140
141  unsigned getMainFileID() const {
142    assert(getNumBuffers());
143    return 1;
144  }
145
146  SMLoc getParentIncludeLoc(unsigned i) const {
147    assert(isValidBufferID(i));
148    return Buffers[i - 1].IncludeLoc;
149  }
150
151  /// Add a new source buffer to this source manager. This takes ownership of
152  /// the memory buffer.
153  unsigned AddNewSourceBuffer(std::unique_ptr<MemoryBuffer> F,
154                              SMLoc IncludeLoc) {
155    SrcBuffer NB;
156    NB.Buffer = std::move(F);
157    NB.IncludeLoc = IncludeLoc;
158    Buffers.push_back(std::move(NB));
159    return Buffers.size();
160  }
161
162  /// Search for a file with the specified name in the current directory or in
163  /// one of the IncludeDirs.
164  ///
165  /// If no file is found, this returns 0, otherwise it returns the buffer ID
166  /// of the stacked file. The full path to the included file can be found in
167  /// \p IncludedFile.
168  unsigned AddIncludeFile(const std::string &Filename, SMLoc IncludeLoc,
169                          std::string &IncludedFile);
170
171  /// Return the ID of the buffer containing the specified location.
172  ///
173  /// 0 is returned if the buffer is not found.
174  unsigned FindBufferContainingLoc(SMLoc Loc) const;
175
176  /// Find the line number for the specified location in the specified file.
177  /// This is not a fast method.
178  unsigned FindLineNumber(SMLoc Loc, unsigned BufferID = 0) const {
179    return getLineAndColumn(Loc, BufferID).first;
180  }
181
182  /// Find the line and column number for the specified location in the
183  /// specified file. This is not a fast method.
184  std::pair<unsigned, unsigned> getLineAndColumn(SMLoc Loc,
185                                                 unsigned BufferID = 0) const;
186
187  /// Emit a message about the specified location with the specified string.
188  ///
189  /// \param ShowColors Display colored messages if output is a terminal and
190  /// the default error handler is used.
191  void PrintMessage(raw_ostream &OS, SMLoc Loc, DiagKind Kind,
192                    const Twine &Msg,
193                    ArrayRef<SMRange> Ranges = None,
194                    ArrayRef<SMFixIt> FixIts = None,
195                    bool ShowColors = true) const;
196
197  /// Emits a diagnostic to llvm::errs().
198  void PrintMessage(SMLoc Loc, DiagKind Kind, const Twine &Msg,
199                    ArrayRef<SMRange> Ranges = None,
200                    ArrayRef<SMFixIt> FixIts = None,
201                    bool ShowColors = true) const;
202
203  /// Emits a manually-constructed diagnostic to the given output stream.
204  ///
205  /// \param ShowColors Display colored messages if output is a terminal and
206  /// the default error handler is used.
207  void PrintMessage(raw_ostream &OS, const SMDiagnostic &Diagnostic,
208                    bool ShowColors = true) const;
209
210  /// Return an SMDiagnostic at the specified location with the specified
211  /// string.
212  ///
213  /// \param Msg If non-null, the kind of message (e.g., "error") which is
214  /// prefixed to the message.
215  SMDiagnostic GetMessage(SMLoc Loc, DiagKind Kind, const Twine &Msg,
216                          ArrayRef<SMRange> Ranges = None,
217                          ArrayRef<SMFixIt> FixIts = None) const;
218
219  /// Prints the names of included files and the line of the file they were
220  /// included from. A diagnostic handler can use this before printing its
221  /// custom formatted message.
222  ///
223  /// \param IncludeLoc The location of the include.
224  /// \param OS the raw_ostream to print on.
225  void PrintIncludeStack(SMLoc IncludeLoc, raw_ostream &OS) const;
226};
227
228/// Represents a single fixit, a replacement of one range of text with another.
229class SMFixIt {
230  SMRange Range;
231
232  std::string Text;
233
234public:
235  // FIXME: Twine.str() is not very efficient.
236  SMFixIt(SMLoc Loc, const Twine &Insertion)
237    : Range(Loc, Loc), Text(Insertion.str()) {
238    assert(Loc.isValid());
239  }
240
241  // FIXME: Twine.str() is not very efficient.
242  SMFixIt(SMRange R, const Twine &Replacement)
243    : Range(R), Text(Replacement.str()) {
244    assert(R.isValid());
245  }
246
247  StringRef getText() const { return Text; }
248  SMRange getRange() const { return Range; }
249
250  bool operator<(const SMFixIt &Other) const {
251    if (Range.Start.getPointer() != Other.Range.Start.getPointer())
252      return Range.Start.getPointer() < Other.Range.Start.getPointer();
253    if (Range.End.getPointer() != Other.Range.End.getPointer())
254      return Range.End.getPointer() < Other.Range.End.getPointer();
255    return Text < Other.Text;
256  }
257};
258
259/// Instances of this class encapsulate one diagnostic report, allowing
260/// printing to a raw_ostream as a caret diagnostic.
261class SMDiagnostic {
262  const SourceMgr *SM = nullptr;
263  SMLoc Loc;
264  std::string Filename;
265  int LineNo = 0;
266  int ColumnNo = 0;
267  SourceMgr::DiagKind Kind = SourceMgr::DK_Error;
268  std::string Message, LineContents;
269  std::vector<std::pair<unsigned, unsigned>> Ranges;
270  SmallVector<SMFixIt, 4> FixIts;
271
272public:
273  // Null diagnostic.
274  SMDiagnostic() = default;
275  // Diagnostic with no location (e.g. file not found, command line arg error).
276  SMDiagnostic(StringRef filename, SourceMgr::DiagKind Knd, StringRef Msg)
277    : Filename(filename), LineNo(-1), ColumnNo(-1), Kind(Knd), Message(Msg) {}
278
279  // Diagnostic with a location.
280  SMDiagnostic(const SourceMgr &sm, SMLoc L, StringRef FN,
281               int Line, int Col, SourceMgr::DiagKind Kind,
282               StringRef Msg, StringRef LineStr,
283               ArrayRef<std::pair<unsigned,unsigned>> Ranges,
284               ArrayRef<SMFixIt> FixIts = None);
285
286  const SourceMgr *getSourceMgr() const { return SM; }
287  SMLoc getLoc() const { return Loc; }
288  StringRef getFilename() const { return Filename; }
289  int getLineNo() const { return LineNo; }
290  int getColumnNo() const { return ColumnNo; }
291  SourceMgr::DiagKind getKind() const { return Kind; }
292  StringRef getMessage() const { return Message; }
293  StringRef getLineContents() const { return LineContents; }
294  ArrayRef<std::pair<unsigned, unsigned>> getRanges() const { return Ranges; }
295
296  void addFixIt(const SMFixIt &Hint) {
297    FixIts.push_back(Hint);
298  }
299
300  ArrayRef<SMFixIt> getFixIts() const {
301    return FixIts;
302  }
303
304  void print(const char *ProgName, raw_ostream &S, bool ShowColors = true,
305             bool ShowKindLabel = true) const;
306};
307
308} // end namespace llvm
309
310#endif // LLVM_SUPPORT_SOURCEMGR_H
311