lib/Support/SourceMgr.cpp

   1 //===- SourceMgr.cpp - Manager for Simple Source Buffers & Diagnostics ----===//
   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 implements the SourceMgr class.  This class is used as a simple
  10 // substrate for diagnostics, #include handling, and other low level things for
  11 // simple parsers.
  12 //
  13 //===----------------------------------------------------------------------===//
  14
  15 #include "llvm/Support/SourceMgr.h"
  16 #include "llvm/ADT/ArrayRef.h"
  17 #include "llvm/ADT/STLExtras.h"
  18 #include "llvm/ADT/SmallVector.h"
  19 #include "llvm/ADT/StringRef.h"
  20 #include "llvm/ADT/Twine.h"
  21 #include "llvm/Support/ErrorOr.h"
  22 #include "llvm/Support/Locale.h"
  23 #include "llvm/Support/MemoryBuffer.h"
  24 #include "llvm/Support/Path.h"
  25 #include "llvm/Support/SMLoc.h"
  26 #include "llvm/Support/WithColor.h"
  27 #include "llvm/Support/raw_ostream.h"
  28 #include <algorithm>
  29 #include <cassert>
  30 #include <cstddef>
  31 #include <limits>
  32 #include <memory>
  33 #include <string>
  34 #include <utility>
  35
  36 using namespace llvm;
  37
  38 static const size_t TabStop = 8;
  39
  40 unsigned SourceMgr::AddIncludeFile(const std::string &Filename,
  41                                    SMLoc IncludeLoc,
  42                                    std::string &IncludedFile) {
  43   IncludedFile = Filename;
  44   ErrorOr<std::unique_ptr<MemoryBuffer>> NewBufOrErr =
  45     MemoryBuffer::getFile(IncludedFile);
  46
  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;
  49        ++i) {
  50     IncludedFile =
  51         IncludeDirectories[i] + sys::path::get_separator().data() + Filename;
  52     NewBufOrErr = MemoryBuffer::getFile(IncludedFile);
  53   }
  54
  55   if (!NewBufOrErr)
  56     return 0;
  57
  58   return AddNewSourceBuffer(std::move(*NewBufOrErr), IncludeLoc);
  59 }
  60
  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())
  67       return i + 1;
  68   return 0;
  69 }
  70
  71 template <typename T>
  72 unsigned SourceMgr::SrcBuffer::getLineNumber(const char *Ptr) const {
  73
  74   // Ensure OffsetCache is allocated and populated with offsets of all the
  75   // '\n' bytes.
  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) {
  84       if (S[N] == '\n') {
  85         Offsets->push_back(static_cast<T>(N));
  86       }
  87     }
  88   } else {
  89     Offsets = OffsetCache.get<std::vector<T> *>();
  90   }
  91
  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);
  97
  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);
 103
 104   // Lines count from 1, so add 1 to the distance from the 0th line.
 105   return (1 + (EOL - Offsets->begin()));
 106 }
 107
 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;
 113 }
 114
 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>*>();
 123     else
 124       delete OffsetCache.get<std::vector<uint64_t>*>();
 125     OffsetCache = nullptr;
 126   }
 127 }
 128
 129 std::pair<unsigned, unsigned>
 130 SourceMgr::getLineAndColumn(SMLoc Loc, unsigned BufferID) const {
 131   if (!BufferID)
 132     BufferID = FindBufferContainingLoc(Loc);
 133   assert(BufferID && "Invalid Location!");
 134
 135   auto &SB = getBufferInfo(BufferID);
 136   const char *Ptr = Loc.getPointer();
 137
 138   size_t Sz = SB.Buffer->getBufferSize();
 139   unsigned LineNo;
 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);
 146   else
 147     LineNo = SB.getLineNumber<uint64_t>(Ptr);
 148
 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);
 153 }
 154
 155 void SourceMgr::PrintIncludeStack(SMLoc IncludeLoc, raw_ostream &OS) const {
 156   if (IncludeLoc == SMLoc()) return;  // Top of stack.
 157
 158   unsigned CurBuf = FindBufferContainingLoc(IncludeLoc);
 159   assert(CurBuf && "Invalid or unspecified location!");
 160
 161   PrintIncludeStack(getBufferInfo(CurBuf).IncludeLoc, OS);
 162
 163   OS << "Included from "
 164      << getBufferInfo(CurBuf).Buffer->getBufferIdentifier()
 165      << ":" << FindLineNumber(IncludeLoc, CurBuf) << ":\n";
 166 }
 167
 168 SMDiagnostic SourceMgr::GetMessage(SMLoc Loc, SourceMgr::DiagKind Kind,
 169                                    const Twine &Msg,
 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>";
 177   std::string LineStr;
 178
 179   if (Loc.isValid()) {
 180     unsigned CurBuf = FindBufferContainingLoc(Loc);
 181     assert(CurBuf && "Invalid or unspecified location!");
 182
 183     const MemoryBuffer *CurMB = getMemoryBuffer(CurBuf);
 184     BufferID = CurMB->getBufferIdentifier();
 185
 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')
 191       --LineStart;
 192
 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')
 197       ++LineEnd;
 198     LineStr = std::string(LineStart, LineEnd);
 199
 200     // Convert any ranges to column ranges that only intersect the line of the
 201     // location.
 202     for (unsigned i = 0, e = Ranges.size(); i != e; ++i) {
 203       SMRange R = Ranges[i];
 204       if (!R.isValid()) continue;
 205
 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)
 208         continue;
 209
 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);
 215
 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));
 220     }
 221
 222     LineAndCol = getLineAndColumn(Loc, CurBuf);
 223   }
 224
 225   return SMDiagnostic(*this, Loc, BufferID, LineAndCol.first,
 226                       LineAndCol.second-1, Kind, Msg.str(),
 227                       LineStr, ColRanges, FixIts);
 228 }
 229
 230 void SourceMgr::PrintMessage(raw_ostream &OS, const SMDiagnostic &Diagnostic,
 231                              bool ShowColors) const {
 232   // Report the message with the diagnostic handler if present.
 233   if (DiagHandler) {
 234     DiagHandler(Diagnostic, DiagContext);
 235     return;
 236   }
 237
 238   if (Diagnostic.getLoc().isValid()) {
 239     unsigned CurBuf = FindBufferContainingLoc(Diagnostic.getLoc());
 240     assert(CurBuf && "Invalid or unspecified location!");
 241     PrintIncludeStack(getBufferInfo(CurBuf).IncludeLoc, OS);
 242   }
 243
 244   Diagnostic.print(nullptr, OS, ShowColors);
 245 }
 246
 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);
 252 }
 253
 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);
 258 }
 259
 260 //===----------------------------------------------------------------------===//
 261 // SMDiagnostic Implementation
 262 //===----------------------------------------------------------------------===//
 263
 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);
 273 }
 274
 275 static void buildFixItLine(std::string &CaretLine, std::string &FixItLine,
 276                            ArrayRef<SMFixIt> FixIts, ArrayRef<char> SourceLine){
 277   if (FixIts.empty())
 278     return;
 279
 280   const char *LineStart = SourceLine.begin();
 281   const char *LineEnd = SourceLine.end();
 282
 283   size_t PrevHintEndCol = 0;
 284
 285   for (ArrayRef<SMFixIt>::iterator I = FixIts.begin(), E = FixIts.end();
 286        I != E; ++I) {
 287     // If the fixit contains a newline or tab, ignore it.
 288     if (I->getText().find_first_of("\n\r\t") != StringRef::npos)
 289       continue;
 290
 291     SMRange R = I->getRange();
 292
 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)
 295       continue;
 296
 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.
 300     unsigned FirstCol;
 301     if (R.Start.getPointer() < LineStart)
 302       FirstCol = 0;
 303     else
 304       FirstCol = R.Start.getPointer() - LineStart;
 305
 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
 309     // to overlap.
 310     //
 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;
 316
 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());
 322
 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, ' ');
 327
 328     std::copy(I->getText().begin(), I->getText().end(),
 329               FixItLine.begin() + HintCol);
 330
 331     PrevHintEndCol = LastColumnModified;
 332
 333     // For replacements, mark the removal range with '~'.
 334     // FIXME: Handle multibyte characters in the source line.
 335     unsigned LastCol;
 336     if (R.End.getPointer() >= LineEnd)
 337       LastCol = LineEnd - LineStart;
 338     else
 339       LastCol = R.End.getPointer() - LineStart;
 340
 341     std::fill(&CaretLine[FirstCol], &CaretLine[LastCol], '~');
 342   }
 343 }
 344
 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     size_t NextTab = LineContents.find('\t', i);
 349     // If there were no tabs left, print the rest, we are done.
 350     if (NextTab == StringRef::npos) {
 351       S << LineContents.drop_front(i);
 352       break;
 353     }
 354
 355     // Otherwise, print from i to NextTab.
 356     S << LineContents.slice(i, NextTab);
 357     OutCol += NextTab - i;
 358     i = NextTab;
 359
 360     // If we have a tab, emit at least one space, then round up to 8 columns.
 361     do {
 362       S << ' ';
 363       ++OutCol;
 364     } while ((OutCol % TabStop) != 0);
 365   }
 366   S << '\n';
 367 }
 368
 369 static bool isNonASCII(char c) {
 370   return c & 0x80;
 371 }
 372
 373 void SMDiagnostic::print(const char *ProgName, raw_ostream &OS,
 374                          bool ShowColors, bool ShowKindLabel) const {
 375   {
 376     WithColor S(OS, raw_ostream::SAVEDCOLOR, true, false, !ShowColors);
 377
 378     if (ProgName && ProgName[0])
 379       S << ProgName << ": ";
 380
 381     if (!Filename.empty()) {
 382       if (Filename == "-")
 383         S << "<stdin>";
 384       else
 385         S << Filename;
 386
 387       if (LineNo != -1) {
 388         S << ':' << LineNo;
 389         if (ColumnNo != -1)
 390           S << ':' << (ColumnNo + 1);
 391       }
 392       S << ": ";
 393     }
 394   }
 395
 396   if (ShowKindLabel) {
 397     switch (Kind) {
 398     case SourceMgr::DK_Error:
 399       WithColor::error(OS, "", !ShowColors);
 400       break;
 401     case SourceMgr::DK_Warning:
 402       WithColor::warning(OS, "", !ShowColors);
 403       break;
 404     case SourceMgr::DK_Note:
 405       WithColor::note(OS, "", !ShowColors);
 406       break;
 407     case SourceMgr::DK_Remark:
 408       WithColor::remark(OS, "", !ShowColors);
 409       break;
 410     }
 411   }
 412
 413   WithColor(OS, raw_ostream::SAVEDCOLOR, true, false, !ShowColors)
 414       << Message << '\n';
 415
 416   if (LineNo == -1 || ColumnNo == -1)
 417     return;
 418
 419   // FIXME: If there are multibyte or multi-column characters in the source, all
 420   // our ranges will be wrong. To do this properly, we'll need a byte-to-column
 421   // map like Clang's TextDiagnostic. For now, we'll just handle tabs by
 422   // expanding them later, and bail out rather than show incorrect ranges and
 423   // misaligned fixits for any other odd characters.
 424   if (find_if(LineContents, isNonASCII) != LineContents.end()) {
 425     printSourceLine(OS, LineContents);
 426     return;
 427   }
 428   size_t NumColumns = LineContents.size();
 429
 430   // Build the line with the caret and ranges.
 431   std::string CaretLine(NumColumns+1, ' ');
 432
 433   // Expand any ranges.
 434   for (unsigned r = 0, e = Ranges.size(); r != e; ++r) {
 435     std::pair<unsigned, unsigned> R = Ranges[r];
 436     std::fill(&CaretLine[R.first],
 437               &CaretLine[std::min((size_t)R.second, CaretLine.size())],
 438               '~');
 439   }
 440
 441   // Add any fix-its.
 442   // FIXME: Find the beginning of the line properly for multibyte characters.
 443   std::string FixItInsertionLine;
 444   buildFixItLine(CaretLine, FixItInsertionLine, FixIts,
 445                  makeArrayRef(Loc.getPointer() - ColumnNo,
 446                               LineContents.size()));
 447
 448   // Finally, plop on the caret.
 449   if (unsigned(ColumnNo) <= NumColumns)
 450     CaretLine[ColumnNo] = '^';
 451   else
 452     CaretLine[NumColumns] = '^';
 453
 454   // ... and remove trailing whitespace so the output doesn't wrap for it.  We
 455   // know that the line isn't completely empty because it has the caret in it at
 456   // least.
 457   CaretLine.erase(CaretLine.find_last_not_of(' ')+1);
 458
 459   printSourceLine(OS, LineContents);
 460
 461   {
 462     WithColor S(OS, raw_ostream::GREEN, true, false, !ShowColors);
 463
 464     // Print out the caret line, matching tabs in the source line.
 465     for (unsigned i = 0, e = CaretLine.size(), OutCol = 0; i != e; ++i) {
 466       if (i >= LineContents.size() || LineContents[i] != '\t') {
 467         S << CaretLine[i];
 468         ++OutCol;
 469         continue;
 470       }
 471
 472       // Okay, we have a tab.  Insert the appropriate number of characters.
 473       do {
 474         S << CaretLine[i];
 475         ++OutCol;
 476       } while ((OutCol % TabStop) != 0);
 477     }
 478     S << '\n';
 479   }
 480
 481   // Print out the replacement line, matching tabs in the source line.
 482   if (FixItInsertionLine.empty())
 483     return;
 484
 485   for (size_t i = 0, e = FixItInsertionLine.size(), OutCol = 0; i < e; ++i) {
 486     if (i >= LineContents.size() || LineContents[i] != '\t') {
 487       OS << FixItInsertionLine[i];
 488       ++OutCol;
 489       continue;
 490     }
 491
 492     // Okay, we have a tab.  Insert the appropriate number of characters.
 493     do {
 494       OS << FixItInsertionLine[i];
 495       // FIXME: This is trying not to break up replacements, but then to re-sync
 496       // with the tabs between replacements. This will fail, though, if two
 497       // fix-it replacements are exactly adjacent, or if a fix-it contains a
 498       // space. Really we should be precomputing column widths, which we'll
 499       // need anyway for multibyte chars.
 500       if (FixItInsertionLine[i] != ' ')
 501         ++i;
 502       ++OutCol;
 503     } while (((OutCol % TabStop) != 0) && i != e);
 504   }
 505   OS << '\n';
 506 }