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/OwningPtr.h" 18#include "llvm/ADT/SmallString.h" 19#include "llvm/ADT/Twine.h" 20#include "llvm/Support/Locale.h" 21#include "llvm/Support/MemoryBuffer.h" 22#include "llvm/Support/raw_ostream.h" 23#include "llvm/Support/system_error.h" 24using namespace llvm; 25 26static const size_t TabStop = 8; 27 28namespace { 29 struct LineNoCacheTy { 30 int LastQueryBufferID; 31 const char *LastQuery; 32 unsigned LineNoOfQuery; 33 }; 34} 35 36static LineNoCacheTy *getCache(void *Ptr) { 37 return (LineNoCacheTy*)Ptr; 38} 39 40 41SourceMgr::~SourceMgr() { 42 // Delete the line # cache if allocated. 43 if (LineNoCacheTy *Cache = getCache(LineNoCache)) 44 delete Cache; 45 46 while (!Buffers.empty()) { 47 delete Buffers.back().Buffer; 48 Buffers.pop_back(); 49 } 50} 51 52/// AddIncludeFile - Search for a file with the specified name in the current 53/// directory or in one of the IncludeDirs. If no file is found, this returns 54/// ~0, otherwise it returns the buffer ID of the stacked file. 55size_t SourceMgr::AddIncludeFile(const std::string &Filename, 56 SMLoc IncludeLoc, 57 std::string &IncludedFile) { 58 OwningPtr<MemoryBuffer> NewBuf; 59 IncludedFile = Filename; 60 MemoryBuffer::getFile(IncludedFile.c_str(), NewBuf); 61 62 // If the file didn't exist directly, see if it's in an include path. 63 for (unsigned i = 0, e = IncludeDirectories.size(); i != e && !NewBuf; ++i) { 64 IncludedFile = IncludeDirectories[i] + "/" + Filename; 65 MemoryBuffer::getFile(IncludedFile.c_str(), NewBuf); 66 } 67 68 if (!NewBuf) return ~0U; 69 70 return AddNewSourceBuffer(NewBuf.take(), IncludeLoc); 71} 72 73 74/// FindBufferContainingLoc - Return the ID of the buffer containing the 75/// specified location, returning -1 if not found. 76int SourceMgr::FindBufferContainingLoc(SMLoc Loc) const { 77 for (unsigned i = 0, e = Buffers.size(); i != e; ++i) 78 if (Loc.getPointer() >= Buffers[i].Buffer->getBufferStart() && 79 // Use <= here so that a pointer to the null at the end of the buffer 80 // is included as part of the buffer. 81 Loc.getPointer() <= Buffers[i].Buffer->getBufferEnd()) 82 return i; 83 return -1; 84} 85 86/// getLineAndColumn - Find the line and column number for the specified 87/// location in the specified file. This is not a fast method. 88std::pair<unsigned, unsigned> 89SourceMgr::getLineAndColumn(SMLoc Loc, int BufferID) const { 90 if (BufferID == -1) BufferID = FindBufferContainingLoc(Loc); 91 assert(BufferID != -1 && "Invalid Location!"); 92 93 MemoryBuffer *Buff = getBufferInfo(BufferID).Buffer; 94 95 // Count the number of \n's between the start of the file and the specified 96 // location. 97 unsigned LineNo = 1; 98 99 const char *BufStart = Buff->getBufferStart(); 100 const char *Ptr = BufStart; 101 102 // If we have a line number cache, and if the query is to a later point in the 103 // same file, start searching from the last query location. This optimizes 104 // for the case when multiple diagnostics come out of one file in order. 105 if (LineNoCacheTy *Cache = getCache(LineNoCache)) 106 if (Cache->LastQueryBufferID == BufferID && 107 Cache->LastQuery <= Loc.getPointer()) { 108 Ptr = Cache->LastQuery; 109 LineNo = Cache->LineNoOfQuery; 110 } 111 112 // Scan for the location being queried, keeping track of the number of lines 113 // we see. 114 for (; SMLoc::getFromPointer(Ptr) != Loc; ++Ptr) 115 if (*Ptr == '\n') ++LineNo; 116 117 // Allocate the line number cache if it doesn't exist. 118 if (LineNoCache == 0) 119 LineNoCache = new LineNoCacheTy(); 120 121 // Update the line # cache. 122 LineNoCacheTy &Cache = *getCache(LineNoCache); 123 Cache.LastQueryBufferID = BufferID; 124 Cache.LastQuery = Ptr; 125 Cache.LineNoOfQuery = LineNo; 126 127 size_t NewlineOffs = StringRef(BufStart, Ptr-BufStart).find_last_of("\n\r"); 128 if (NewlineOffs == StringRef::npos) NewlineOffs = ~(size_t)0; 129 return std::make_pair(LineNo, Ptr-BufStart-NewlineOffs); 130} 131 132void SourceMgr::PrintIncludeStack(SMLoc IncludeLoc, raw_ostream &OS) const { 133 if (IncludeLoc == SMLoc()) return; // Top of stack. 134 135 int CurBuf = FindBufferContainingLoc(IncludeLoc); 136 assert(CurBuf != -1 && "Invalid or unspecified location!"); 137 138 PrintIncludeStack(getBufferInfo(CurBuf).IncludeLoc, OS); 139 140 OS << "Included from " 141 << getBufferInfo(CurBuf).Buffer->getBufferIdentifier() 142 << ":" << FindLineNumber(IncludeLoc, CurBuf) << ":\n"; 143} 144 145 146/// GetMessage - Return an SMDiagnostic at the specified location with the 147/// specified string. 148/// 149/// @param Type - If non-null, the kind of message (e.g., "error") which is 150/// prefixed to the message. 151SMDiagnostic SourceMgr::GetMessage(SMLoc Loc, SourceMgr::DiagKind Kind, 152 const Twine &Msg, 153 ArrayRef<SMRange> Ranges, 154 ArrayRef<SMFixIt> FixIts) const { 155 156 // First thing to do: find the current buffer containing the specified 157 // location to pull out the source line. 158 SmallVector<std::pair<unsigned, unsigned>, 4> ColRanges; 159 std::pair<unsigned, unsigned> LineAndCol; 160 const char *BufferID = "<unknown>"; 161 std::string LineStr; 162 163 if (Loc.isValid()) { 164 int CurBuf = FindBufferContainingLoc(Loc); 165 assert(CurBuf != -1 && "Invalid or unspecified location!"); 166 167 MemoryBuffer *CurMB = getBufferInfo(CurBuf).Buffer; 168 BufferID = CurMB->getBufferIdentifier(); 169 170 // Scan backward to find the start of the line. 171 const char *LineStart = Loc.getPointer(); 172 const char *BufStart = CurMB->getBufferStart(); 173 while (LineStart != BufStart && LineStart[-1] != '\n' && 174 LineStart[-1] != '\r') 175 --LineStart; 176 177 // Get the end of the line. 178 const char *LineEnd = Loc.getPointer(); 179 const char *BufEnd = CurMB->getBufferEnd(); 180 while (LineEnd != BufEnd && LineEnd[0] != '\n' && LineEnd[0] != '\r') 181 ++LineEnd; 182 LineStr = std::string(LineStart, LineEnd); 183 184 // Convert any ranges to column ranges that only intersect the line of the 185 // location. 186 for (unsigned i = 0, e = Ranges.size(); i != e; ++i) { 187 SMRange R = Ranges[i]; 188 if (!R.isValid()) continue; 189 190 // If the line doesn't contain any part of the range, then ignore it. 191 if (R.Start.getPointer() > LineEnd || R.End.getPointer() < LineStart) 192 continue; 193 194 // Ignore pieces of the range that go onto other lines. 195 if (R.Start.getPointer() < LineStart) 196 R.Start = SMLoc::getFromPointer(LineStart); 197 if (R.End.getPointer() > LineEnd) 198 R.End = SMLoc::getFromPointer(LineEnd); 199 200 // Translate from SMLoc ranges to column ranges. 201 // FIXME: Handle multibyte characters. 202 ColRanges.push_back(std::make_pair(R.Start.getPointer()-LineStart, 203 R.End.getPointer()-LineStart)); 204 } 205 206 LineAndCol = getLineAndColumn(Loc, CurBuf); 207 } 208 209 return SMDiagnostic(*this, Loc, BufferID, LineAndCol.first, 210 LineAndCol.second-1, Kind, Msg.str(), 211 LineStr, ColRanges, FixIts); 212} 213 214void SourceMgr::PrintMessage(raw_ostream &OS, SMLoc Loc, 215 SourceMgr::DiagKind Kind, 216 const Twine &Msg, ArrayRef<SMRange> Ranges, 217 ArrayRef<SMFixIt> FixIts, bool ShowColors) const { 218 SMDiagnostic Diagnostic = GetMessage(Loc, Kind, Msg, Ranges, FixIts); 219 220 // Report the message with the diagnostic handler if present. 221 if (DiagHandler) { 222 DiagHandler(Diagnostic, DiagContext); 223 return; 224 } 225 226 if (Loc != SMLoc()) { 227 int CurBuf = FindBufferContainingLoc(Loc); 228 assert(CurBuf != -1 && "Invalid or unspecified location!"); 229 PrintIncludeStack(getBufferInfo(CurBuf).IncludeLoc, OS); 230 } 231 232 Diagnostic.print(0, OS, ShowColors); 233} 234 235void SourceMgr::PrintMessage(SMLoc Loc, SourceMgr::DiagKind Kind, 236 const Twine &Msg, ArrayRef<SMRange> Ranges, 237 ArrayRef<SMFixIt> FixIts, bool ShowColors) const { 238 PrintMessage(llvm::errs(), Loc, Kind, Msg, Ranges, FixIts, ShowColors); 239} 240 241//===----------------------------------------------------------------------===// 242// SMDiagnostic Implementation 243//===----------------------------------------------------------------------===// 244 245SMDiagnostic::SMDiagnostic(const SourceMgr &sm, SMLoc L, StringRef FN, 246 int Line, int Col, SourceMgr::DiagKind Kind, 247 StringRef Msg, StringRef LineStr, 248 ArrayRef<std::pair<unsigned,unsigned> > Ranges, 249 ArrayRef<SMFixIt> Hints) 250 : SM(&sm), Loc(L), Filename(FN), LineNo(Line), ColumnNo(Col), Kind(Kind), 251 Message(Msg), LineContents(LineStr), Ranges(Ranges.vec()), 252 FixIts(Hints.begin(), Hints.end()) { 253 std::sort(FixIts.begin(), FixIts.end()); 254} 255 256static void buildFixItLine(std::string &CaretLine, std::string &FixItLine, 257 ArrayRef<SMFixIt> FixIts, ArrayRef<char> SourceLine){ 258 if (FixIts.empty()) 259 return; 260 261 const char *LineStart = SourceLine.begin(); 262 const char *LineEnd = SourceLine.end(); 263 264 size_t PrevHintEndCol = 0; 265 266 for (ArrayRef<SMFixIt>::iterator I = FixIts.begin(), E = FixIts.end(); 267 I != E; ++I) { 268 // If the fixit contains a newline or tab, ignore it. 269 if (I->getText().find_first_of("\n\r\t") != StringRef::npos) 270 continue; 271 272 SMRange R = I->getRange(); 273 274 // If the line doesn't contain any part of the range, then ignore it. 275 if (R.Start.getPointer() > LineEnd || R.End.getPointer() < LineStart) 276 continue; 277 278 // Translate from SMLoc to column. 279 // Ignore pieces of the range that go onto other lines. 280 // FIXME: Handle multibyte characters in the source line. 281 unsigned FirstCol; 282 if (R.Start.getPointer() < LineStart) 283 FirstCol = 0; 284 else 285 FirstCol = R.Start.getPointer() - LineStart; 286 287 // If we inserted a long previous hint, push this one forwards, and add 288 // an extra space to show that this is not part of the previous 289 // completion. This is sort of the best we can do when two hints appear 290 // to overlap. 291 // 292 // Note that if this hint is located immediately after the previous 293 // hint, no space will be added, since the location is more important. 294 unsigned HintCol = FirstCol; 295 if (HintCol < PrevHintEndCol) 296 HintCol = PrevHintEndCol + 1; 297 298 // FIXME: This assertion is intended to catch unintended use of multibyte 299 // characters in fixits. If we decide to do this, we'll have to track 300 // separate byte widths for the source and fixit lines. 301 assert((size_t)llvm::sys::locale::columnWidth(I->getText()) == 302 I->getText().size()); 303 304 // This relies on one byte per column in our fixit hints. 305 unsigned LastColumnModified = HintCol + I->getText().size(); 306 if (LastColumnModified > FixItLine.size()) 307 FixItLine.resize(LastColumnModified, ' '); 308 309 std::copy(I->getText().begin(), I->getText().end(), 310 FixItLine.begin() + HintCol); 311 312 PrevHintEndCol = LastColumnModified; 313 314 // For replacements, mark the removal range with '~'. 315 // FIXME: Handle multibyte characters in the source line. 316 unsigned LastCol; 317 if (R.End.getPointer() >= LineEnd) 318 LastCol = LineEnd - LineStart; 319 else 320 LastCol = R.End.getPointer() - LineStart; 321 322 std::fill(&CaretLine[FirstCol], &CaretLine[LastCol], '~'); 323 } 324} 325 326static void printSourceLine(raw_ostream &S, StringRef LineContents) { 327 // Print out the source line one character at a time, so we can expand tabs. 328 for (unsigned i = 0, e = LineContents.size(), OutCol = 0; i != e; ++i) { 329 if (LineContents[i] != '\t') { 330 S << LineContents[i]; 331 ++OutCol; 332 continue; 333 } 334 335 // If we have a tab, emit at least one space, then round up to 8 columns. 336 do { 337 S << ' '; 338 ++OutCol; 339 } while ((OutCol % TabStop) != 0); 340 } 341 S << '\n'; 342} 343 344static bool isNonASCII(char c) { 345 return c & 0x80; 346} 347 348void SMDiagnostic::print(const char *ProgName, raw_ostream &S, 349 bool ShowColors) const { 350 // Display colors only if OS supports colors. 351 ShowColors &= S.has_colors(); 352 353 if (ShowColors) 354 S.changeColor(raw_ostream::SAVEDCOLOR, true); 355 356 if (ProgName && ProgName[0]) 357 S << ProgName << ": "; 358 359 if (!Filename.empty()) { 360 if (Filename == "-") 361 S << "<stdin>"; 362 else 363 S << Filename; 364 365 if (LineNo != -1) { 366 S << ':' << LineNo; 367 if (ColumnNo != -1) 368 S << ':' << (ColumnNo+1); 369 } 370 S << ": "; 371 } 372 373 switch (Kind) { 374 case SourceMgr::DK_Error: 375 if (ShowColors) 376 S.changeColor(raw_ostream::RED, true); 377 S << "error: "; 378 break; 379 case SourceMgr::DK_Warning: 380 if (ShowColors) 381 S.changeColor(raw_ostream::MAGENTA, true); 382 S << "warning: "; 383 break; 384 case SourceMgr::DK_Note: 385 if (ShowColors) 386 S.changeColor(raw_ostream::BLACK, true); 387 S << "note: "; 388 break; 389 } 390 391 if (ShowColors) { 392 S.resetColor(); 393 S.changeColor(raw_ostream::SAVEDCOLOR, true); 394 } 395 396 S << Message << '\n'; 397 398 if (ShowColors) 399 S.resetColor(); 400 401 if (LineNo == -1 || ColumnNo == -1) 402 return; 403 404 // FIXME: If there are multibyte or multi-column characters in the source, all 405 // our ranges will be wrong. To do this properly, we'll need a byte-to-column 406 // map like Clang's TextDiagnostic. For now, we'll just handle tabs by 407 // expanding them later, and bail out rather than show incorrect ranges and 408 // misaligned fixits for any other odd characters. 409 if (std::find_if(LineContents.begin(), LineContents.end(), isNonASCII) != 410 LineContents.end()) { 411 printSourceLine(S, LineContents); 412 return; 413 } 414 size_t NumColumns = LineContents.size(); 415 416 // Build the line with the caret and ranges. 417 std::string CaretLine(NumColumns+1, ' '); 418 419 // Expand any ranges. 420 for (unsigned r = 0, e = Ranges.size(); r != e; ++r) { 421 std::pair<unsigned, unsigned> R = Ranges[r]; 422 std::fill(&CaretLine[R.first], 423 &CaretLine[std::min((size_t)R.second, CaretLine.size())], 424 '~'); 425 } 426 427 // Add any fix-its. 428 // FIXME: Find the beginning of the line properly for multibyte characters. 429 std::string FixItInsertionLine; 430 buildFixItLine(CaretLine, FixItInsertionLine, FixIts, 431 makeArrayRef(Loc.getPointer() - ColumnNo, 432 LineContents.size())); 433 434 // Finally, plop on the caret. 435 if (unsigned(ColumnNo) <= NumColumns) 436 CaretLine[ColumnNo] = '^'; 437 else 438 CaretLine[NumColumns] = '^'; 439 440 // ... and remove trailing whitespace so the output doesn't wrap for it. We 441 // know that the line isn't completely empty because it has the caret in it at 442 // least. 443 CaretLine.erase(CaretLine.find_last_not_of(' ')+1); 444 445 printSourceLine(S, LineContents); 446 447 if (ShowColors) 448 S.changeColor(raw_ostream::GREEN, true); 449 450 // Print out the caret line, matching tabs in the source line. 451 for (unsigned i = 0, e = CaretLine.size(), OutCol = 0; i != e; ++i) { 452 if (i >= LineContents.size() || LineContents[i] != '\t') { 453 S << CaretLine[i]; 454 ++OutCol; 455 continue; 456 } 457 458 // Okay, we have a tab. Insert the appropriate number of characters. 459 do { 460 S << CaretLine[i]; 461 ++OutCol; 462 } while ((OutCol % TabStop) != 0); 463 } 464 S << '\n'; 465 466 if (ShowColors) 467 S.resetColor(); 468 469 // Print out the replacement line, matching tabs in the source line. 470 if (FixItInsertionLine.empty()) 471 return; 472 473 for (size_t i = 0, e = FixItInsertionLine.size(), OutCol = 0; i < e; ++i) { 474 if (i >= LineContents.size() || LineContents[i] != '\t') { 475 S << FixItInsertionLine[i]; 476 ++OutCol; 477 continue; 478 } 479 480 // Okay, we have a tab. Insert the appropriate number of characters. 481 do { 482 S << FixItInsertionLine[i]; 483 // FIXME: This is trying not to break up replacements, but then to re-sync 484 // with the tabs between replacements. This will fail, though, if two 485 // fix-it replacements are exactly adjacent, or if a fix-it contains a 486 // space. Really we should be precomputing column widths, which we'll 487 // need anyway for multibyte chars. 488 if (FixItInsertionLine[i] != ' ') 489 ++i; 490 ++OutCol; 491 } while (((OutCol % TabStop) != 0) && i != e); 492 } 493 S << '\n'; 494} 495