1 //===- SourceMgr.cpp - Manager for Simple Source Buffers & Diagnostics ----===//
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 // This file implements the SourceMgr class.  This class is used as a simple
11 // substrate for diagnostics, #include handling, and other low level things for
12 // simple parsers.
13 //
14 //===----------------------------------------------------------------------===//
15 
16 #include "llvm/Support/SourceMgr.h"
17 #include "llvm/ADT/Twine.h"
18 #include "llvm/Support/Locale.h"
19 #include "llvm/Support/MemoryBuffer.h"
20 #include "llvm/Support/Path.h"
21 #include "llvm/Support/raw_ostream.h"
22 using namespace llvm;
23 
24 static const size_t TabStop = 8;
25 
26 namespace {
27   struct LineNoCacheTy {
28     unsigned LastQueryBufferID;
29     const char *LastQuery;
30     unsigned LineNoOfQuery;
31   };
32 }
33 
getCache(void * Ptr)34 static LineNoCacheTy *getCache(void *Ptr) {
35   return (LineNoCacheTy*)Ptr;
36 }
37 
38 
~SourceMgr()39 SourceMgr::~SourceMgr() {
40   // Delete the line # cache if allocated.
41   if (LineNoCacheTy *Cache = getCache(LineNoCache))
42     delete Cache;
43 }
44 
AddIncludeFile(const std::string & Filename,SMLoc IncludeLoc,std::string & IncludedFile)45 unsigned SourceMgr::AddIncludeFile(const std::string &Filename,
46                                    SMLoc IncludeLoc,
47                                    std::string &IncludedFile) {
48   IncludedFile = Filename;
49   ErrorOr<std::unique_ptr<MemoryBuffer>> NewBufOrErr =
50     MemoryBuffer::getFile(IncludedFile);
51 
52   // If the file didn't exist directly, see if it's in an include path.
53   for (unsigned i = 0, e = IncludeDirectories.size(); i != e && !NewBufOrErr;
54        ++i) {
55     IncludedFile =
56         IncludeDirectories[i] + sys::path::get_separator().data() + Filename;
57     NewBufOrErr = MemoryBuffer::getFile(IncludedFile);
58   }
59 
60   if (!NewBufOrErr)
61     return 0;
62 
63   return AddNewSourceBuffer(std::move(*NewBufOrErr), IncludeLoc);
64 }
65 
FindBufferContainingLoc(SMLoc Loc) const66 unsigned SourceMgr::FindBufferContainingLoc(SMLoc Loc) const {
67   for (unsigned i = 0, e = Buffers.size(); i != e; ++i)
68     if (Loc.getPointer() >= Buffers[i].Buffer->getBufferStart() &&
69         // Use <= here so that a pointer to the null at the end of the buffer
70         // is included as part of the buffer.
71         Loc.getPointer() <= Buffers[i].Buffer->getBufferEnd())
72       return i + 1;
73   return 0;
74 }
75 
76 std::pair<unsigned, unsigned>
getLineAndColumn(SMLoc Loc,unsigned BufferID) const77 SourceMgr::getLineAndColumn(SMLoc Loc, unsigned BufferID) const {
78   if (!BufferID)
79     BufferID = FindBufferContainingLoc(Loc);
80   assert(BufferID && "Invalid Location!");
81 
82   const MemoryBuffer *Buff = getMemoryBuffer(BufferID);
83 
84   // Count the number of \n's between the start of the file and the specified
85   // location.
86   unsigned LineNo = 1;
87 
88   const char *BufStart = Buff->getBufferStart();
89   const char *Ptr = BufStart;
90 
91   // If we have a line number cache, and if the query is to a later point in the
92   // same file, start searching from the last query location.  This optimizes
93   // for the case when multiple diagnostics come out of one file in order.
94   if (LineNoCacheTy *Cache = getCache(LineNoCache))
95     if (Cache->LastQueryBufferID == BufferID &&
96         Cache->LastQuery <= Loc.getPointer()) {
97       Ptr = Cache->LastQuery;
98       LineNo = Cache->LineNoOfQuery;
99     }
100 
101   // Scan for the location being queried, keeping track of the number of lines
102   // we see.
103   for (; SMLoc::getFromPointer(Ptr) != Loc; ++Ptr)
104     if (*Ptr == '\n') ++LineNo;
105 
106   // Allocate the line number cache if it doesn't exist.
107   if (!LineNoCache)
108     LineNoCache = new LineNoCacheTy();
109 
110   // Update the line # cache.
111   LineNoCacheTy &Cache = *getCache(LineNoCache);
112   Cache.LastQueryBufferID = BufferID;
113   Cache.LastQuery = Ptr;
114   Cache.LineNoOfQuery = LineNo;
115 
116   size_t NewlineOffs = StringRef(BufStart, Ptr-BufStart).find_last_of("\n\r");
117   if (NewlineOffs == StringRef::npos) NewlineOffs = ~(size_t)0;
118   return std::make_pair(LineNo, Ptr-BufStart-NewlineOffs);
119 }
120 
PrintIncludeStack(SMLoc IncludeLoc,raw_ostream & OS) const121 void SourceMgr::PrintIncludeStack(SMLoc IncludeLoc, raw_ostream &OS) const {
122   if (IncludeLoc == SMLoc()) return;  // Top of stack.
123 
124   unsigned CurBuf = FindBufferContainingLoc(IncludeLoc);
125   assert(CurBuf && "Invalid or unspecified location!");
126 
127   PrintIncludeStack(getBufferInfo(CurBuf).IncludeLoc, OS);
128 
129   OS << "Included from "
130      << getBufferInfo(CurBuf).Buffer->getBufferIdentifier()
131      << ":" << FindLineNumber(IncludeLoc, CurBuf) << ":\n";
132 }
133 
134 
GetMessage(SMLoc Loc,SourceMgr::DiagKind Kind,const Twine & Msg,ArrayRef<SMRange> Ranges,ArrayRef<SMFixIt> FixIts) const135 SMDiagnostic SourceMgr::GetMessage(SMLoc Loc, SourceMgr::DiagKind Kind,
136                                    const Twine &Msg,
137                                    ArrayRef<SMRange> Ranges,
138                                    ArrayRef<SMFixIt> FixIts) const {
139 
140   // First thing to do: find the current buffer containing the specified
141   // location to pull out the source line.
142   SmallVector<std::pair<unsigned, unsigned>, 4> ColRanges;
143   std::pair<unsigned, unsigned> LineAndCol;
144   const char *BufferID = "<unknown>";
145   std::string LineStr;
146 
147   if (Loc.isValid()) {
148     unsigned CurBuf = FindBufferContainingLoc(Loc);
149     assert(CurBuf && "Invalid or unspecified location!");
150 
151     const MemoryBuffer *CurMB = getMemoryBuffer(CurBuf);
152     BufferID = CurMB->getBufferIdentifier();
153 
154     // Scan backward to find the start of the line.
155     const char *LineStart = Loc.getPointer();
156     const char *BufStart = CurMB->getBufferStart();
157     while (LineStart != BufStart && LineStart[-1] != '\n' &&
158            LineStart[-1] != '\r')
159       --LineStart;
160 
161     // Get the end of the line.
162     const char *LineEnd = Loc.getPointer();
163     const char *BufEnd = CurMB->getBufferEnd();
164     while (LineEnd != BufEnd && LineEnd[0] != '\n' && LineEnd[0] != '\r')
165       ++LineEnd;
166     LineStr = std::string(LineStart, LineEnd);
167 
168     // Convert any ranges to column ranges that only intersect the line of the
169     // location.
170     for (unsigned i = 0, e = Ranges.size(); i != e; ++i) {
171       SMRange R = Ranges[i];
172       if (!R.isValid()) continue;
173 
174       // If the line doesn't contain any part of the range, then ignore it.
175       if (R.Start.getPointer() > LineEnd || R.End.getPointer() < LineStart)
176         continue;
177 
178       // Ignore pieces of the range that go onto other lines.
179       if (R.Start.getPointer() < LineStart)
180         R.Start = SMLoc::getFromPointer(LineStart);
181       if (R.End.getPointer() > LineEnd)
182         R.End = SMLoc::getFromPointer(LineEnd);
183 
184       // Translate from SMLoc ranges to column ranges.
185       // FIXME: Handle multibyte characters.
186       ColRanges.push_back(std::make_pair(R.Start.getPointer()-LineStart,
187                                          R.End.getPointer()-LineStart));
188     }
189 
190     LineAndCol = getLineAndColumn(Loc, CurBuf);
191   }
192 
193   return SMDiagnostic(*this, Loc, BufferID, LineAndCol.first,
194                       LineAndCol.second-1, Kind, Msg.str(),
195                       LineStr, ColRanges, FixIts);
196 }
197 
PrintMessage(raw_ostream & OS,const SMDiagnostic & Diagnostic,bool ShowColors) const198 void SourceMgr::PrintMessage(raw_ostream &OS, const SMDiagnostic &Diagnostic,
199                              bool ShowColors) const {
200   // Report the message with the diagnostic handler if present.
201   if (DiagHandler) {
202     DiagHandler(Diagnostic, DiagContext);
203     return;
204   }
205 
206   if (Diagnostic.getLoc().isValid()) {
207     unsigned CurBuf = FindBufferContainingLoc(Diagnostic.getLoc());
208     assert(CurBuf && "Invalid or unspecified location!");
209     PrintIncludeStack(getBufferInfo(CurBuf).IncludeLoc, OS);
210   }
211 
212   Diagnostic.print(nullptr, OS, ShowColors);
213 }
214 
PrintMessage(raw_ostream & OS,SMLoc Loc,SourceMgr::DiagKind Kind,const Twine & Msg,ArrayRef<SMRange> Ranges,ArrayRef<SMFixIt> FixIts,bool ShowColors) const215 void SourceMgr::PrintMessage(raw_ostream &OS, SMLoc Loc,
216                              SourceMgr::DiagKind Kind,
217                              const Twine &Msg, ArrayRef<SMRange> Ranges,
218                              ArrayRef<SMFixIt> FixIts, bool ShowColors) const {
219   PrintMessage(OS, GetMessage(Loc, Kind, Msg, Ranges, FixIts), ShowColors);
220 }
221 
PrintMessage(SMLoc Loc,SourceMgr::DiagKind Kind,const Twine & Msg,ArrayRef<SMRange> Ranges,ArrayRef<SMFixIt> FixIts,bool ShowColors) const222 void SourceMgr::PrintMessage(SMLoc Loc, SourceMgr::DiagKind Kind,
223                              const Twine &Msg, ArrayRef<SMRange> Ranges,
224                              ArrayRef<SMFixIt> FixIts, bool ShowColors) const {
225   PrintMessage(llvm::errs(), Loc, Kind, Msg, Ranges, FixIts, ShowColors);
226 }
227 
228 //===----------------------------------------------------------------------===//
229 // SMDiagnostic Implementation
230 //===----------------------------------------------------------------------===//
231 
SMDiagnostic(const SourceMgr & sm,SMLoc L,StringRef FN,int Line,int Col,SourceMgr::DiagKind Kind,StringRef Msg,StringRef LineStr,ArrayRef<std::pair<unsigned,unsigned>> Ranges,ArrayRef<SMFixIt> Hints)232 SMDiagnostic::SMDiagnostic(const SourceMgr &sm, SMLoc L, StringRef FN,
233                            int Line, int Col, SourceMgr::DiagKind Kind,
234                            StringRef Msg, StringRef LineStr,
235                            ArrayRef<std::pair<unsigned,unsigned> > Ranges,
236                            ArrayRef<SMFixIt> Hints)
237   : SM(&sm), Loc(L), Filename(FN), LineNo(Line), ColumnNo(Col), Kind(Kind),
238     Message(Msg), LineContents(LineStr), Ranges(Ranges.vec()),
239     FixIts(Hints.begin(), Hints.end()) {
240   std::sort(FixIts.begin(), FixIts.end());
241 }
242 
buildFixItLine(std::string & CaretLine,std::string & FixItLine,ArrayRef<SMFixIt> FixIts,ArrayRef<char> SourceLine)243 static void buildFixItLine(std::string &CaretLine, std::string &FixItLine,
244                            ArrayRef<SMFixIt> FixIts, ArrayRef<char> SourceLine){
245   if (FixIts.empty())
246     return;
247 
248   const char *LineStart = SourceLine.begin();
249   const char *LineEnd = SourceLine.end();
250 
251   size_t PrevHintEndCol = 0;
252 
253   for (ArrayRef<SMFixIt>::iterator I = FixIts.begin(), E = FixIts.end();
254        I != E; ++I) {
255     // If the fixit contains a newline or tab, ignore it.
256     if (I->getText().find_first_of("\n\r\t") != StringRef::npos)
257       continue;
258 
259     SMRange R = I->getRange();
260 
261     // If the line doesn't contain any part of the range, then ignore it.
262     if (R.Start.getPointer() > LineEnd || R.End.getPointer() < LineStart)
263       continue;
264 
265     // Translate from SMLoc to column.
266     // Ignore pieces of the range that go onto other lines.
267     // FIXME: Handle multibyte characters in the source line.
268     unsigned FirstCol;
269     if (R.Start.getPointer() < LineStart)
270       FirstCol = 0;
271     else
272       FirstCol = R.Start.getPointer() - LineStart;
273 
274     // If we inserted a long previous hint, push this one forwards, and add
275     // an extra space to show that this is not part of the previous
276     // completion. This is sort of the best we can do when two hints appear
277     // to overlap.
278     //
279     // Note that if this hint is located immediately after the previous
280     // hint, no space will be added, since the location is more important.
281     unsigned HintCol = FirstCol;
282     if (HintCol < PrevHintEndCol)
283       HintCol = PrevHintEndCol + 1;
284 
285     // FIXME: This assertion is intended to catch unintended use of multibyte
286     // characters in fixits. If we decide to do this, we'll have to track
287     // separate byte widths for the source and fixit lines.
288     assert((size_t)llvm::sys::locale::columnWidth(I->getText()) ==
289            I->getText().size());
290 
291     // This relies on one byte per column in our fixit hints.
292     unsigned LastColumnModified = HintCol + I->getText().size();
293     if (LastColumnModified > FixItLine.size())
294       FixItLine.resize(LastColumnModified, ' ');
295 
296     std::copy(I->getText().begin(), I->getText().end(),
297               FixItLine.begin() + HintCol);
298 
299     PrevHintEndCol = LastColumnModified;
300 
301     // For replacements, mark the removal range with '~'.
302     // FIXME: Handle multibyte characters in the source line.
303     unsigned LastCol;
304     if (R.End.getPointer() >= LineEnd)
305       LastCol = LineEnd - LineStart;
306     else
307       LastCol = R.End.getPointer() - LineStart;
308 
309     std::fill(&CaretLine[FirstCol], &CaretLine[LastCol], '~');
310   }
311 }
312 
printSourceLine(raw_ostream & S,StringRef LineContents)313 static void printSourceLine(raw_ostream &S, StringRef LineContents) {
314   // Print out the source line one character at a time, so we can expand tabs.
315   for (unsigned i = 0, e = LineContents.size(), OutCol = 0; i != e; ++i) {
316     if (LineContents[i] != '\t') {
317       S << LineContents[i];
318       ++OutCol;
319       continue;
320     }
321 
322     // If we have a tab, emit at least one space, then round up to 8 columns.
323     do {
324       S << ' ';
325       ++OutCol;
326     } while ((OutCol % TabStop) != 0);
327   }
328   S << '\n';
329 }
330 
isNonASCII(char c)331 static bool isNonASCII(char c) {
332   return c & 0x80;
333 }
334 
print(const char * ProgName,raw_ostream & S,bool ShowColors) const335 void SMDiagnostic::print(const char *ProgName, raw_ostream &S,
336                          bool ShowColors) const {
337   // Display colors only if OS supports colors.
338   ShowColors &= S.has_colors();
339 
340   if (ShowColors)
341     S.changeColor(raw_ostream::SAVEDCOLOR, true);
342 
343   if (ProgName && ProgName[0])
344     S << ProgName << ": ";
345 
346   if (!Filename.empty()) {
347     if (Filename == "-")
348       S << "<stdin>";
349     else
350       S << Filename;
351 
352     if (LineNo != -1) {
353       S << ':' << LineNo;
354       if (ColumnNo != -1)
355         S << ':' << (ColumnNo+1);
356     }
357     S << ": ";
358   }
359 
360   switch (Kind) {
361   case SourceMgr::DK_Error:
362     if (ShowColors)
363       S.changeColor(raw_ostream::RED, true);
364     S << "error: ";
365     break;
366   case SourceMgr::DK_Warning:
367     if (ShowColors)
368       S.changeColor(raw_ostream::MAGENTA, true);
369     S << "warning: ";
370     break;
371   case SourceMgr::DK_Note:
372     if (ShowColors)
373       S.changeColor(raw_ostream::BLACK, true);
374     S << "note: ";
375     break;
376   }
377 
378   if (ShowColors) {
379     S.resetColor();
380     S.changeColor(raw_ostream::SAVEDCOLOR, true);
381   }
382 
383   S << Message << '\n';
384 
385   if (ShowColors)
386     S.resetColor();
387 
388   if (LineNo == -1 || ColumnNo == -1)
389     return;
390 
391   // FIXME: If there are multibyte or multi-column characters in the source, all
392   // our ranges will be wrong. To do this properly, we'll need a byte-to-column
393   // map like Clang's TextDiagnostic. For now, we'll just handle tabs by
394   // expanding them later, and bail out rather than show incorrect ranges and
395   // misaligned fixits for any other odd characters.
396   if (std::find_if(LineContents.begin(), LineContents.end(), isNonASCII) !=
397       LineContents.end()) {
398     printSourceLine(S, LineContents);
399     return;
400   }
401   size_t NumColumns = LineContents.size();
402 
403   // Build the line with the caret and ranges.
404   std::string CaretLine(NumColumns+1, ' ');
405 
406   // Expand any ranges.
407   for (unsigned r = 0, e = Ranges.size(); r != e; ++r) {
408     std::pair<unsigned, unsigned> R = Ranges[r];
409     std::fill(&CaretLine[R.first],
410               &CaretLine[std::min((size_t)R.second, CaretLine.size())],
411               '~');
412   }
413 
414   // Add any fix-its.
415   // FIXME: Find the beginning of the line properly for multibyte characters.
416   std::string FixItInsertionLine;
417   buildFixItLine(CaretLine, FixItInsertionLine, FixIts,
418                  makeArrayRef(Loc.getPointer() - ColumnNo,
419                               LineContents.size()));
420 
421   // Finally, plop on the caret.
422   if (unsigned(ColumnNo) <= NumColumns)
423     CaretLine[ColumnNo] = '^';
424   else
425     CaretLine[NumColumns] = '^';
426 
427   // ... and remove trailing whitespace so the output doesn't wrap for it.  We
428   // know that the line isn't completely empty because it has the caret in it at
429   // least.
430   CaretLine.erase(CaretLine.find_last_not_of(' ')+1);
431 
432   printSourceLine(S, LineContents);
433 
434   if (ShowColors)
435     S.changeColor(raw_ostream::GREEN, true);
436 
437   // Print out the caret line, matching tabs in the source line.
438   for (unsigned i = 0, e = CaretLine.size(), OutCol = 0; i != e; ++i) {
439     if (i >= LineContents.size() || LineContents[i] != '\t') {
440       S << CaretLine[i];
441       ++OutCol;
442       continue;
443     }
444 
445     // Okay, we have a tab.  Insert the appropriate number of characters.
446     do {
447       S << CaretLine[i];
448       ++OutCol;
449     } while ((OutCol % TabStop) != 0);
450   }
451   S << '\n';
452 
453   if (ShowColors)
454     S.resetColor();
455 
456   // Print out the replacement line, matching tabs in the source line.
457   if (FixItInsertionLine.empty())
458     return;
459 
460   for (size_t i = 0, e = FixItInsertionLine.size(), OutCol = 0; i < e; ++i) {
461     if (i >= LineContents.size() || LineContents[i] != '\t') {
462       S << FixItInsertionLine[i];
463       ++OutCol;
464       continue;
465     }
466 
467     // Okay, we have a tab.  Insert the appropriate number of characters.
468     do {
469       S << FixItInsertionLine[i];
470       // FIXME: This is trying not to break up replacements, but then to re-sync
471       // with the tabs between replacements. This will fail, though, if two
472       // fix-it replacements are exactly adjacent, or if a fix-it contains a
473       // space. Really we should be precomputing column widths, which we'll
474       // need anyway for multibyte chars.
475       if (FixItInsertionLine[i] != ' ')
476         ++i;
477       ++OutCol;
478     } while (((OutCol % TabStop) != 0) && i != e);
479   }
480   S << '\n';
481 }
482