Home | History | Annotate | Download | only in Support
      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"
     24 using namespace llvm;
     25 
     26 static const size_t TabStop = 8;
     27 
     28 namespace {
     29   struct LineNoCacheTy {
     30     int LastQueryBufferID;
     31     const char *LastQuery;
     32     unsigned LineNoOfQuery;
     33   };
     34 }
     35 
     36 static LineNoCacheTy *getCache(void *Ptr) {
     37   return (LineNoCacheTy*)Ptr;
     38 }
     39 
     40 
     41 SourceMgr::~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.
     55 size_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.
     76 int 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.
     88 std::pair<unsigned, unsigned>
     89 SourceMgr::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 
    132 void 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.
    151 SMDiagnostic 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 
    214 void SourceMgr::PrintMessage(SMLoc Loc, SourceMgr::DiagKind Kind,
    215                              const Twine &Msg, ArrayRef<SMRange> Ranges,
    216                              ArrayRef<SMFixIt> FixIts, bool ShowColors) const {
    217   SMDiagnostic Diagnostic = GetMessage(Loc, Kind, Msg, Ranges, FixIts);
    218 
    219   // Report the message with the diagnostic handler if present.
    220   if (DiagHandler) {
    221     DiagHandler(Diagnostic, DiagContext);
    222     return;
    223   }
    224 
    225   raw_ostream &OS = errs();
    226 
    227   if (Loc != SMLoc()) {
    228     int CurBuf = FindBufferContainingLoc(Loc);
    229     assert(CurBuf != -1 && "Invalid or unspecified location!");
    230     PrintIncludeStack(getBufferInfo(CurBuf).IncludeLoc, OS);
    231   }
    232 
    233   Diagnostic.print(0, OS, ShowColors);
    234 }
    235 
    236 //===----------------------------------------------------------------------===//
    237 // SMDiagnostic Implementation
    238 //===----------------------------------------------------------------------===//
    239 
    240 SMDiagnostic::SMDiagnostic(const SourceMgr &sm, SMLoc L, StringRef FN,
    241                            int Line, int Col, SourceMgr::DiagKind Kind,
    242                            StringRef Msg, StringRef LineStr,
    243                            ArrayRef<std::pair<unsigned,unsigned> > Ranges,
    244                            ArrayRef<SMFixIt> Hints)
    245   : SM(&sm), Loc(L), Filename(FN), LineNo(Line), ColumnNo(Col), Kind(Kind),
    246     Message(Msg), LineContents(LineStr), Ranges(Ranges.vec()),
    247     FixIts(Hints.begin(), Hints.end()) {
    248   std::sort(FixIts.begin(), FixIts.end());
    249 }
    250 
    251 static void buildFixItLine(std::string &CaretLine, std::string &FixItLine,
    252                            ArrayRef<SMFixIt> FixIts, ArrayRef<char> SourceLine){
    253   if (FixIts.empty())
    254     return;
    255 
    256   const char *LineStart = SourceLine.begin();
    257   const char *LineEnd = SourceLine.end();
    258 
    259   size_t PrevHintEndCol = 0;
    260 
    261   for (ArrayRef<SMFixIt>::iterator I = FixIts.begin(), E = FixIts.end();
    262        I != E; ++I) {
    263     // If the fixit contains a newline or tab, ignore it.
    264     if (I->getText().find_first_of("\n\r\t") != StringRef::npos)
    265       continue;
    266 
    267     SMRange R = I->getRange();
    268 
    269     // If the line doesn't contain any part of the range, then ignore it.
    270     if (R.Start.getPointer() > LineEnd || R.End.getPointer() < LineStart)
    271       continue;
    272 
    273     // Translate from SMLoc to column.
    274     // Ignore pieces of the range that go onto other lines.
    275     // FIXME: Handle multibyte characters in the source line.
    276     unsigned FirstCol;
    277     if (R.Start.getPointer() < LineStart)
    278       FirstCol = 0;
    279     else
    280       FirstCol = R.Start.getPointer() - LineStart;
    281 
    282     // If we inserted a long previous hint, push this one forwards, and add
    283     // an extra space to show that this is not part of the previous
    284     // completion. This is sort of the best we can do when two hints appear
    285     // to overlap.
    286     //
    287     // Note that if this hint is located immediately after the previous
    288     // hint, no space will be added, since the location is more important.
    289     unsigned HintCol = FirstCol;
    290     if (HintCol < PrevHintEndCol)
    291       HintCol = PrevHintEndCol + 1;
    292 
    293     // FIXME: This assertion is intended to catch unintended use of multibyte
    294     // characters in fixits. If we decide to do this, we'll have to track
    295     // separate byte widths for the source and fixit lines.
    296     assert((size_t)llvm::sys::locale::columnWidth(I->getText()) ==
    297            I->getText().size());
    298 
    299     // This relies on one byte per column in our fixit hints.
    300     unsigned LastColumnModified = HintCol + I->getText().size();
    301     if (LastColumnModified > FixItLine.size())
    302       FixItLine.resize(LastColumnModified, ' ');
    303 
    304     std::copy(I->getText().begin(), I->getText().end(),
    305               FixItLine.begin() + HintCol);
    306 
    307     PrevHintEndCol = LastColumnModified;
    308 
    309     // For replacements, mark the removal range with '~'.
    310     // FIXME: Handle multibyte characters in the source line.
    311     unsigned LastCol;
    312     if (R.End.getPointer() >= LineEnd)
    313       LastCol = LineEnd - LineStart;
    314     else
    315       LastCol = R.End.getPointer() - LineStart;
    316 
    317     std::fill(&CaretLine[FirstCol], &CaretLine[LastCol], '~');
    318   }
    319 }
    320 
    321 static void printSourceLine(raw_ostream &S, StringRef LineContents) {
    322   // Print out the source line one character at a time, so we can expand tabs.
    323   for (unsigned i = 0, e = LineContents.size(), OutCol = 0; i != e; ++i) {
    324     if (LineContents[i] != '\t') {
    325       S << LineContents[i];
    326       ++OutCol;
    327       continue;
    328     }
    329 
    330     // If we have a tab, emit at least one space, then round up to 8 columns.
    331     do {
    332       S << ' ';
    333       ++OutCol;
    334     } while ((OutCol % TabStop) != 0);
    335   }
    336   S << '\n';
    337 }
    338 
    339 static bool isNonASCII(char c) {
    340   return c & 0x80;
    341 }
    342 
    343 void SMDiagnostic::print(const char *ProgName, raw_ostream &S,
    344                          bool ShowColors) const {
    345   // Display colors only if OS supports colors.
    346   ShowColors &= S.has_colors();
    347 
    348   if (ShowColors)
    349     S.changeColor(raw_ostream::SAVEDCOLOR, true);
    350 
    351   if (ProgName && ProgName[0])
    352     S << ProgName << ": ";
    353 
    354   if (!Filename.empty()) {
    355     if (Filename == "-")
    356       S << "<stdin>";
    357     else
    358       S << Filename;
    359 
    360     if (LineNo != -1) {
    361       S << ':' << LineNo;
    362       if (ColumnNo != -1)
    363         S << ':' << (ColumnNo+1);
    364     }
    365     S << ": ";
    366   }
    367 
    368   switch (Kind) {
    369   case SourceMgr::DK_Error:
    370     if (ShowColors)
    371       S.changeColor(raw_ostream::RED, true);
    372     S << "error: ";
    373     break;
    374   case SourceMgr::DK_Warning:
    375     if (ShowColors)
    376       S.changeColor(raw_ostream::MAGENTA, true);
    377     S << "warning: ";
    378     break;
    379   case SourceMgr::DK_Note:
    380     if (ShowColors)
    381       S.changeColor(raw_ostream::BLACK, true);
    382     S << "note: ";
    383     break;
    384   }
    385 
    386   if (ShowColors) {
    387     S.resetColor();
    388     S.changeColor(raw_ostream::SAVEDCOLOR, true);
    389   }
    390 
    391   S << Message << '\n';
    392 
    393   if (ShowColors)
    394     S.resetColor();
    395 
    396   if (LineNo == -1 || ColumnNo == -1)
    397     return;
    398 
    399   // FIXME: If there are multibyte or multi-column characters in the source, all
    400   // our ranges will be wrong. To do this properly, we'll need a byte-to-column
    401   // map like Clang's TextDiagnostic. For now, we'll just handle tabs by
    402   // expanding them later, and bail out rather than show incorrect ranges and
    403   // misaligned fixits for any other odd characters.
    404   if (std::find_if(LineContents.begin(), LineContents.end(), isNonASCII) !=
    405       LineContents.end()) {
    406     printSourceLine(S, LineContents);
    407     return;
    408   }
    409   size_t NumColumns = LineContents.size();
    410 
    411   // Build the line with the caret and ranges.
    412   std::string CaretLine(NumColumns+1, ' ');
    413 
    414   // Expand any ranges.
    415   for (unsigned r = 0, e = Ranges.size(); r != e; ++r) {
    416     std::pair<unsigned, unsigned> R = Ranges[r];
    417     std::fill(&CaretLine[R.first],
    418               &CaretLine[std::min((size_t)R.second, CaretLine.size())],
    419               '~');
    420   }
    421 
    422   // Add any fix-its.
    423   // FIXME: Find the beginning of the line properly for multibyte characters.
    424   std::string FixItInsertionLine;
    425   buildFixItLine(CaretLine, FixItInsertionLine, FixIts,
    426                  makeArrayRef(Loc.getPointer() - ColumnNo,
    427                               LineContents.size()));
    428 
    429   // Finally, plop on the caret.
    430   if (unsigned(ColumnNo) <= NumColumns)
    431     CaretLine[ColumnNo] = '^';
    432   else
    433     CaretLine[NumColumns] = '^';
    434 
    435   // ... and remove trailing whitespace so the output doesn't wrap for it.  We
    436   // know that the line isn't completely empty because it has the caret in it at
    437   // least.
    438   CaretLine.erase(CaretLine.find_last_not_of(' ')+1);
    439 
    440   printSourceLine(S, LineContents);
    441 
    442   if (ShowColors)
    443     S.changeColor(raw_ostream::GREEN, true);
    444 
    445   // Print out the caret line, matching tabs in the source line.
    446   for (unsigned i = 0, e = CaretLine.size(), OutCol = 0; i != e; ++i) {
    447     if (i >= LineContents.size() || LineContents[i] != '\t') {
    448       S << CaretLine[i];
    449       ++OutCol;
    450       continue;
    451     }
    452 
    453     // Okay, we have a tab.  Insert the appropriate number of characters.
    454     do {
    455       S << CaretLine[i];
    456       ++OutCol;
    457     } while ((OutCol % TabStop) != 0);
    458   }
    459   S << '\n';
    460 
    461   if (ShowColors)
    462     S.resetColor();
    463 
    464   // Print out the replacement line, matching tabs in the source line.
    465   if (FixItInsertionLine.empty())
    466     return;
    467 
    468   for (size_t i = 0, e = FixItInsertionLine.size(), OutCol = 0; i != e; ++i) {
    469     if (i >= LineContents.size() || LineContents[i] != '\t') {
    470       S << FixItInsertionLine[i];
    471       ++OutCol;
    472       continue;
    473     }
    474 
    475     // Okay, we have a tab.  Insert the appropriate number of characters.
    476     do {
    477       S << FixItInsertionLine[i];
    478       // FIXME: This is trying not to break up replacements, but then to re-sync
    479       // with the tabs between replacements. This will fail, though, if two
    480       // fix-it replacements are exactly adjacent, or if a fix-it contains a
    481       // space. Really we should be precomputing column widths, which we'll
    482       // need anyway for multibyte chars.
    483       if (FixItInsertionLine[i] != ' ')
    484         ++i;
    485       ++OutCol;
    486     } while (((OutCol % TabStop) != 0) && i != e);
    487   }
    488   S << '\n';
    489 }
    490