1 //===- SourceMgr.cpp - Manager for Simple Source Buffers & Diagnostics ----===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
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
14 //===----------------------------------------------------------------------===//
16 #include "llvm/Support/SourceMgr.h"
17 #include "llvm/ADT/ArrayRef.h"
18 #include "llvm/ADT/STLExtras.h"
19 #include "llvm/ADT/SmallVector.h"
20 #include "llvm/ADT/StringRef.h"
21 #include "llvm/ADT/Twine.h"
22 #include "llvm/Support/ErrorOr.h"
23 #include "llvm/Support/Locale.h"
24 #include "llvm/Support/MemoryBuffer.h"
25 #include "llvm/Support/Path.h"
26 #include "llvm/Support/SMLoc.h"
27 #include "llvm/Support/raw_ostream.h"
38 static const size_t TabStop = 8;
40 unsigned SourceMgr::AddIncludeFile(const std::string &Filename,
42 std::string &IncludedFile) {
43 IncludedFile = Filename;
44 ErrorOr<std::unique_ptr<MemoryBuffer>> NewBufOrErr =
45 MemoryBuffer::getFile(IncludedFile);
47 // If the file didn't exist directly, see if it's in an include path.
48 for (unsigned i = 0, e = IncludeDirectories.size(); i != e && !NewBufOrErr;
51 IncludeDirectories[i] + sys::path::get_separator().data() + Filename;
52 NewBufOrErr = MemoryBuffer::getFile(IncludedFile);
58 return AddNewSourceBuffer(std::move(*NewBufOrErr), IncludeLoc);
61 unsigned SourceMgr::FindBufferContainingLoc(SMLoc Loc) const {
62 for (unsigned i = 0, e = Buffers.size(); i != e; ++i)
63 if (Loc.getPointer() >= Buffers[i].Buffer->getBufferStart() &&
64 // Use <= here so that a pointer to the null at the end of the buffer
65 // is included as part of the buffer.
66 Loc.getPointer() <= Buffers[i].Buffer->getBufferEnd())
72 unsigned SourceMgr::SrcBuffer::getLineNumber(const char *Ptr) const {
74 // Ensure OffsetCache is allocated and populated with offsets of all the
76 std::vector<T> *Offsets = nullptr;
77 if (OffsetCache.isNull()) {
78 Offsets = new std::vector<T>();
79 OffsetCache = Offsets;
80 size_t Sz = Buffer->getBufferSize();
81 assert(Sz <= std::numeric_limits<T>::max());
82 StringRef S = Buffer->getBuffer();
83 for (size_t N = 0; N < Sz; ++N) {
85 Offsets->push_back(static_cast<T>(N));
89 Offsets = OffsetCache.get<std::vector<T> *>();
92 const char *BufStart = Buffer->getBufferStart();
93 assert(Ptr >= BufStart && Ptr <= Buffer->getBufferEnd());
94 ptrdiff_t PtrDiff = Ptr - BufStart;
95 assert(PtrDiff >= 0 && static_cast<size_t>(PtrDiff) <= std::numeric_limits<T>::max());
96 T PtrOffset = static_cast<T>(PtrDiff);
98 // std::lower_bound returns the first EOL offset that's not-less-than
99 // PtrOffset, meaning the EOL that _ends the line_ that PtrOffset is on
100 // (including if PtrOffset refers to the EOL itself). If there's no such
101 // EOL, returns end().
102 auto EOL = std::lower_bound(Offsets->begin(), Offsets->end(), PtrOffset);
104 // Lines count from 1, so add 1 to the distance from the 0th line.
105 return (1 + (EOL - Offsets->begin()));
108 SourceMgr::SrcBuffer::SrcBuffer(SourceMgr::SrcBuffer &&Other)
109 : Buffer(std::move(Other.Buffer)),
110 OffsetCache(Other.OffsetCache),
111 IncludeLoc(Other.IncludeLoc) {
112 Other.OffsetCache = nullptr;
115 SourceMgr::SrcBuffer::~SrcBuffer() {
116 if (!OffsetCache.isNull()) {
117 if (OffsetCache.is<std::vector<uint8_t>*>())
118 delete OffsetCache.get<std::vector<uint8_t>*>();
119 else if (OffsetCache.is<std::vector<uint16_t>*>())
120 delete OffsetCache.get<std::vector<uint16_t>*>();
121 else if (OffsetCache.is<std::vector<uint32_t>*>())
122 delete OffsetCache.get<std::vector<uint32_t>*>();
124 delete OffsetCache.get<std::vector<uint64_t>*>();
125 OffsetCache = nullptr;
129 std::pair<unsigned, unsigned>
130 SourceMgr::getLineAndColumn(SMLoc Loc, unsigned BufferID) const {
132 BufferID = FindBufferContainingLoc(Loc);
133 assert(BufferID && "Invalid Location!");
135 auto &SB = getBufferInfo(BufferID);
136 const char *Ptr = Loc.getPointer();
138 size_t Sz = SB.Buffer->getBufferSize();
140 if (Sz <= std::numeric_limits<uint8_t>::max())
141 LineNo = SB.getLineNumber<uint8_t>(Ptr);
142 else if (Sz <= std::numeric_limits<uint16_t>::max())
143 LineNo = SB.getLineNumber<uint16_t>(Ptr);
144 else if (Sz <= std::numeric_limits<uint32_t>::max())
145 LineNo = SB.getLineNumber<uint32_t>(Ptr);
147 LineNo = SB.getLineNumber<uint64_t>(Ptr);
149 const char *BufStart = SB.Buffer->getBufferStart();
150 size_t NewlineOffs = StringRef(BufStart, Ptr-BufStart).find_last_of("\n\r");
151 if (NewlineOffs == StringRef::npos) NewlineOffs = ~(size_t)0;
152 return std::make_pair(LineNo, Ptr-BufStart-NewlineOffs);
155 void SourceMgr::PrintIncludeStack(SMLoc IncludeLoc, raw_ostream &OS) const {
156 if (IncludeLoc == SMLoc()) return; // Top of stack.
158 unsigned CurBuf = FindBufferContainingLoc(IncludeLoc);
159 assert(CurBuf && "Invalid or unspecified location!");
161 PrintIncludeStack(getBufferInfo(CurBuf).IncludeLoc, OS);
163 OS << "Included from "
164 << getBufferInfo(CurBuf).Buffer->getBufferIdentifier()
165 << ":" << FindLineNumber(IncludeLoc, CurBuf) << ":\n";
168 SMDiagnostic SourceMgr::GetMessage(SMLoc Loc, SourceMgr::DiagKind Kind,
170 ArrayRef<SMRange> Ranges,
171 ArrayRef<SMFixIt> FixIts) const {
172 // First thing to do: find the current buffer containing the specified
173 // location to pull out the source line.
174 SmallVector<std::pair<unsigned, unsigned>, 4> ColRanges;
175 std::pair<unsigned, unsigned> LineAndCol;
176 StringRef BufferID = "<unknown>";
180 unsigned CurBuf = FindBufferContainingLoc(Loc);
181 assert(CurBuf && "Invalid or unspecified location!");
183 const MemoryBuffer *CurMB = getMemoryBuffer(CurBuf);
184 BufferID = CurMB->getBufferIdentifier();
186 // Scan backward to find the start of the line.
187 const char *LineStart = Loc.getPointer();
188 const char *BufStart = CurMB->getBufferStart();
189 while (LineStart != BufStart && LineStart[-1] != '\n' &&
190 LineStart[-1] != '\r')
193 // Get the end of the line.
194 const char *LineEnd = Loc.getPointer();
195 const char *BufEnd = CurMB->getBufferEnd();
196 while (LineEnd != BufEnd && LineEnd[0] != '\n' && LineEnd[0] != '\r')
198 LineStr = std::string(LineStart, LineEnd);
200 // Convert any ranges to column ranges that only intersect the line of the
202 for (unsigned i = 0, e = Ranges.size(); i != e; ++i) {
203 SMRange R = Ranges[i];
204 if (!R.isValid()) continue;
206 // If the line doesn't contain any part of the range, then ignore it.
207 if (R.Start.getPointer() > LineEnd || R.End.getPointer() < LineStart)
210 // Ignore pieces of the range that go onto other lines.
211 if (R.Start.getPointer() < LineStart)
212 R.Start = SMLoc::getFromPointer(LineStart);
213 if (R.End.getPointer() > LineEnd)
214 R.End = SMLoc::getFromPointer(LineEnd);
216 // Translate from SMLoc ranges to column ranges.
217 // FIXME: Handle multibyte characters.
218 ColRanges.push_back(std::make_pair(R.Start.getPointer()-LineStart,
219 R.End.getPointer()-LineStart));
222 LineAndCol = getLineAndColumn(Loc, CurBuf);
225 return SMDiagnostic(*this, Loc, BufferID, LineAndCol.first,
226 LineAndCol.second-1, Kind, Msg.str(),
227 LineStr, ColRanges, FixIts);
230 void SourceMgr::PrintMessage(raw_ostream &OS, const SMDiagnostic &Diagnostic,
231 bool ShowColors) const {
232 // Report the message with the diagnostic handler if present.
234 DiagHandler(Diagnostic, DiagContext);
238 if (Diagnostic.getLoc().isValid()) {
239 unsigned CurBuf = FindBufferContainingLoc(Diagnostic.getLoc());
240 assert(CurBuf && "Invalid or unspecified location!");
241 PrintIncludeStack(getBufferInfo(CurBuf).IncludeLoc, OS);
244 Diagnostic.print(nullptr, OS, ShowColors);
247 void SourceMgr::PrintMessage(raw_ostream &OS, SMLoc Loc,
248 SourceMgr::DiagKind Kind,
249 const Twine &Msg, ArrayRef<SMRange> Ranges,
250 ArrayRef<SMFixIt> FixIts, bool ShowColors) const {
251 PrintMessage(OS, GetMessage(Loc, Kind, Msg, Ranges, FixIts), ShowColors);
254 void SourceMgr::PrintMessage(SMLoc Loc, SourceMgr::DiagKind Kind,
255 const Twine &Msg, ArrayRef<SMRange> Ranges,
256 ArrayRef<SMFixIt> FixIts, bool ShowColors) const {
257 PrintMessage(errs(), Loc, Kind, Msg, Ranges, FixIts, ShowColors);
260 //===----------------------------------------------------------------------===//
261 // SMDiagnostic Implementation
262 //===----------------------------------------------------------------------===//
264 SMDiagnostic::SMDiagnostic(const SourceMgr &sm, SMLoc L, StringRef FN,
265 int Line, int Col, SourceMgr::DiagKind Kind,
266 StringRef Msg, StringRef LineStr,
267 ArrayRef<std::pair<unsigned,unsigned>> Ranges,
268 ArrayRef<SMFixIt> Hints)
269 : SM(&sm), Loc(L), Filename(FN), LineNo(Line), ColumnNo(Col), Kind(Kind),
270 Message(Msg), LineContents(LineStr), Ranges(Ranges.vec()),
271 FixIts(Hints.begin(), Hints.end()) {
272 llvm::sort(FixIts.begin(), FixIts.end());
275 static void buildFixItLine(std::string &CaretLine, std::string &FixItLine,
276 ArrayRef<SMFixIt> FixIts, ArrayRef<char> SourceLine){
280 const char *LineStart = SourceLine.begin();
281 const char *LineEnd = SourceLine.end();
283 size_t PrevHintEndCol = 0;
285 for (ArrayRef<SMFixIt>::iterator I = FixIts.begin(), E = FixIts.end();
287 // If the fixit contains a newline or tab, ignore it.
288 if (I->getText().find_first_of("\n\r\t") != StringRef::npos)
291 SMRange R = I->getRange();
293 // If the line doesn't contain any part of the range, then ignore it.
294 if (R.Start.getPointer() > LineEnd || R.End.getPointer() < LineStart)
297 // Translate from SMLoc to column.
298 // Ignore pieces of the range that go onto other lines.
299 // FIXME: Handle multibyte characters in the source line.
301 if (R.Start.getPointer() < LineStart)
304 FirstCol = R.Start.getPointer() - LineStart;
306 // If we inserted a long previous hint, push this one forwards, and add
307 // an extra space to show that this is not part of the previous
308 // completion. This is sort of the best we can do when two hints appear
311 // Note that if this hint is located immediately after the previous
312 // hint, no space will be added, since the location is more important.
313 unsigned HintCol = FirstCol;
314 if (HintCol < PrevHintEndCol)
315 HintCol = PrevHintEndCol + 1;
317 // FIXME: This assertion is intended to catch unintended use of multibyte
318 // characters in fixits. If we decide to do this, we'll have to track
319 // separate byte widths for the source and fixit lines.
320 assert((size_t)sys::locale::columnWidth(I->getText()) ==
321 I->getText().size());
323 // This relies on one byte per column in our fixit hints.
324 unsigned LastColumnModified = HintCol + I->getText().size();
325 if (LastColumnModified > FixItLine.size())
326 FixItLine.resize(LastColumnModified, ' ');
328 std::copy(I->getText().begin(), I->getText().end(),
329 FixItLine.begin() + HintCol);
331 PrevHintEndCol = LastColumnModified;
333 // For replacements, mark the removal range with '~'.
334 // FIXME: Handle multibyte characters in the source line.
336 if (R.End.getPointer() >= LineEnd)
337 LastCol = LineEnd - LineStart;
339 LastCol = R.End.getPointer() - LineStart;
341 std::fill(&CaretLine[FirstCol], &CaretLine[LastCol], '~');
345 static void printSourceLine(raw_ostream &S, StringRef LineContents) {
346 // Print out the source line one character at a time, so we can expand tabs.
347 for (unsigned i = 0, e = LineContents.size(), OutCol = 0; i != e; ++i) {
348 if (LineContents[i] != '\t') {
349 S << LineContents[i];
354 // If we have a tab, emit at least one space, then round up to 8 columns.
358 } while ((OutCol % TabStop) != 0);
363 static bool isNonASCII(char c) {
367 void SMDiagnostic::print(const char *ProgName, raw_ostream &S, bool ShowColors,
368 bool ShowKindLabel) const {
369 // Display colors only if OS supports colors.
370 ShowColors &= S.has_colors();
373 S.changeColor(raw_ostream::SAVEDCOLOR, true);
375 if (ProgName && ProgName[0])
376 S << ProgName << ": ";
378 if (!Filename.empty()) {
387 S << ':' << (ColumnNo+1);
394 case SourceMgr::DK_Error:
396 S.changeColor(raw_ostream::RED, true);
399 case SourceMgr::DK_Warning:
401 S.changeColor(raw_ostream::MAGENTA, true);
404 case SourceMgr::DK_Note:
406 S.changeColor(raw_ostream::BLACK, true);
409 case SourceMgr::DK_Remark:
411 S.changeColor(raw_ostream::BLUE, true);
418 S.changeColor(raw_ostream::SAVEDCOLOR, true);
422 S << Message << '\n';
427 if (LineNo == -1 || ColumnNo == -1)
430 // FIXME: If there are multibyte or multi-column characters in the source, all
431 // our ranges will be wrong. To do this properly, we'll need a byte-to-column
432 // map like Clang's TextDiagnostic. For now, we'll just handle tabs by
433 // expanding them later, and bail out rather than show incorrect ranges and
434 // misaligned fixits for any other odd characters.
435 if (find_if(LineContents, isNonASCII) != LineContents.end()) {
436 printSourceLine(S, LineContents);
439 size_t NumColumns = LineContents.size();
441 // Build the line with the caret and ranges.
442 std::string CaretLine(NumColumns+1, ' ');
444 // Expand any ranges.
445 for (unsigned r = 0, e = Ranges.size(); r != e; ++r) {
446 std::pair<unsigned, unsigned> R = Ranges[r];
447 std::fill(&CaretLine[R.first],
448 &CaretLine[std::min((size_t)R.second, CaretLine.size())],
453 // FIXME: Find the beginning of the line properly for multibyte characters.
454 std::string FixItInsertionLine;
455 buildFixItLine(CaretLine, FixItInsertionLine, FixIts,
456 makeArrayRef(Loc.getPointer() - ColumnNo,
457 LineContents.size()));
459 // Finally, plop on the caret.
460 if (unsigned(ColumnNo) <= NumColumns)
461 CaretLine[ColumnNo] = '^';
463 CaretLine[NumColumns] = '^';
465 // ... and remove trailing whitespace so the output doesn't wrap for it. We
466 // know that the line isn't completely empty because it has the caret in it at
468 CaretLine.erase(CaretLine.find_last_not_of(' ')+1);
470 printSourceLine(S, LineContents);
473 S.changeColor(raw_ostream::GREEN, true);
475 // Print out the caret line, matching tabs in the source line.
476 for (unsigned i = 0, e = CaretLine.size(), OutCol = 0; i != e; ++i) {
477 if (i >= LineContents.size() || LineContents[i] != '\t') {
483 // Okay, we have a tab. Insert the appropriate number of characters.
487 } while ((OutCol % TabStop) != 0);
494 // Print out the replacement line, matching tabs in the source line.
495 if (FixItInsertionLine.empty())
498 for (size_t i = 0, e = FixItInsertionLine.size(), OutCol = 0; i < e; ++i) {
499 if (i >= LineContents.size() || LineContents[i] != '\t') {
500 S << FixItInsertionLine[i];
505 // Okay, we have a tab. Insert the appropriate number of characters.
507 S << FixItInsertionLine[i];
508 // FIXME: This is trying not to break up replacements, but then to re-sync
509 // with the tabs between replacements. This will fail, though, if two
510 // fix-it replacements are exactly adjacent, or if a fix-it contains a
511 // space. Really we should be precomputing column widths, which we'll
512 // need anyway for multibyte chars.
513 if (FixItInsertionLine[i] != ' ')
516 } while (((OutCol % TabStop) != 0) && i != e);