C++ SourceManager::getBuffer方法代码示例

TokenRewriter::TokenRewriter(FileID FID, SourceManager &SM,
                             const LangOptions &LangOpts) {
  ScratchBuf.reset(new ScratchBuffer(SM));

  // Create a lexer to lex all the tokens of the main file in raw mode.
  const llvm::MemoryBuffer *FromFile = SM.getBuffer(FID);
  Lexer RawLex(FID, FromFile, SM, LangOpts);

  // Return all comments and whitespace as tokens.
  RawLex.SetKeepWhitespaceMode(true);

  // Lex the file, populating our datastructures.
  Token RawTok;
  RawLex.LexFromRawLexer(RawTok);
  while (RawTok.isNot(tok::eof)) {
#if 0
    if (Tok.is(tok::raw_identifier)) {
      // Look up the identifier info for the token.  This should use
      // IdentifierTable directly instead of PP.
      PP.LookUpIdentifierInfo(Tok);
    }
#endif

    AddToken(RawTok, TokenList.end());
    RawLex.LexFromRawLexer(RawTok);
  }
}

/// \brief Lex the specified source file to determine whether it contains
/// any expected-* directives.  As a Lexer is used rather than a full-blown
/// Preprocessor, directives inside skipped #if blocks will still be found.
///
/// \return true if any directives were found.
static bool findDirectives(SourceManager &SM, FileID FID,
                           const LangOptions &LangOpts) {
  // Create a raw lexer to pull all the comments out of FID.
  if (FID.isInvalid())
    return false;

  // Create a lexer to lex all the tokens of the main file in raw mode.
  const llvm::MemoryBuffer *FromFile = SM.getBuffer(FID);
  Lexer RawLex(FID, FromFile, SM, LangOpts);

  // Return comments as tokens, this is how we find expected diagnostics.
  RawLex.SetCommentRetentionState(true);

  Token Tok;
  Tok.setKind(tok::comment);
  VerifyDiagnosticConsumer::DirectiveStatus Status =
    VerifyDiagnosticConsumer::HasNoDirectives;
  while (Tok.isNot(tok::eof)) {
    RawLex.Lex(Tok);
    if (!Tok.is(tok::comment)) continue;

    std::string Comment = RawLex.getSpelling(Tok, SM, LangOpts);
    if (Comment.empty()) continue;

    // Find first directive.
    if (ParseDirective(Comment, 0, SM, Tok.getLocation(),
                       SM.getDiagnostics(), Status))
      return true;
  }
  return false;
}

  static void PrintSourceForLocation(const SourceLocation &Loc,
                                     SourceManager &SM) {
    const char *LocData = SM.getCharacterData(Loc, /*Invalid=*/nullptr);
    unsigned LocColumn =
        SM.getSpellingColumnNumber(Loc, /*Invalid=*/nullptr) - 1;
    FileID FID = SM.getFileID(Loc);
    llvm::MemoryBuffer *Buffer = SM.getBuffer(FID, Loc, /*Invalid=*/nullptr);

    assert(LocData >= Buffer->getBufferStart() &&
           LocData < Buffer->getBufferEnd());

    const char *LineBegin = LocData - LocColumn;

    assert(LineBegin >= Buffer->getBufferStart());

    const char *LineEnd = nullptr;

    for (LineEnd = LineBegin; *LineEnd != '\n' && *LineEnd != '\r' &&
                              LineEnd < Buffer->getBufferEnd();
         ++LineEnd)
      ;

    llvm::StringRef LineString(LineBegin, LineEnd - LineBegin);

    llvm::errs() << LineString << '\n';
    std::string Space(LocColumn, ' ');
    llvm::errs() << Space.c_str() << '\n';
  }

// Checks if 'typedef' keyword can be removed - we do it only if
// it is the only declaration in a declaration chain.
static bool CheckRemoval(SourceManager &SM, const SourceLocation &LocStart,
                         const SourceLocation &LocEnd, ASTContext &Context,
                         SourceRange &ResultRange) {
  FileID FID = SM.getFileID(LocEnd);
  llvm::MemoryBuffer *Buffer = SM.getBuffer(FID, LocEnd);
  Lexer DeclLexer(SM.getLocForStartOfFile(FID), Context.getLangOpts(),
                  Buffer->getBufferStart(), SM.getCharacterData(LocStart),
                  Buffer->getBufferEnd());
  Token DeclToken;
  bool result = false;
  int parenthesisLevel = 0;

  while (!DeclLexer.LexFromRawLexer(DeclToken)) {
    if (DeclToken.getKind() == tok::TokenKind::l_paren)
      parenthesisLevel++;
    if (DeclToken.getKind() == tok::TokenKind::r_paren)
      parenthesisLevel--;
    if (DeclToken.getKind() == tok::TokenKind::semi)
      break;
    // if there is comma and we are not between open parenthesis then it is
    // two or more declatarions in this chain
    if (parenthesisLevel == 0 && DeclToken.getKind() == tok::TokenKind::comma)
      return false;

    if (DeclToken.isOneOf(tok::TokenKind::identifier,
                          tok::TokenKind::raw_identifier)) {
      auto TokenStr = DeclToken.getRawIdentifier().str();

      if (TokenStr == "typedef") {
        ResultRange =
            SourceRange(DeclToken.getLocation(), DeclToken.getEndLoc());
        result = true;
      }
    }
  }
  // assert if there was keyword 'typedef' in declaration
  assert(result && "No typedef found");

  return result;
}

void TextDiagnosticPrinter::EmitCaretDiagnostic(SourceLocation Loc,
                                                SourceRange *Ranges,
                                                unsigned NumRanges,
                                                SourceManager &SM,
                                          const CodeModificationHint *Hints,
                                                unsigned NumHints,
                                                unsigned Columns) {
  assert(!Loc.isInvalid() && "must have a valid source location here");

  // If this is a macro ID, first emit information about where this was
  // instantiated (recursively) then emit information about where. the token was
  // spelled from.
  if (!Loc.isFileID()) {
    SourceLocation OneLevelUp = SM.getImmediateInstantiationRange(Loc).first;
    // FIXME: Map ranges?
    EmitCaretDiagnostic(OneLevelUp, Ranges, NumRanges, SM, 0, 0, Columns);

    Loc = SM.getImmediateSpellingLoc(Loc);
    
    // Map the ranges.
    for (unsigned i = 0; i != NumRanges; ++i) {
      SourceLocation S = Ranges[i].getBegin(), E = Ranges[i].getEnd();
      if (S.isMacroID()) S = SM.getImmediateSpellingLoc(S);
      if (E.isMacroID()) E = SM.getImmediateSpellingLoc(E);
      Ranges[i] = SourceRange(S, E);
    }
    
    if (ShowLocation) {
      std::pair<FileID, unsigned> IInfo = SM.getDecomposedInstantiationLoc(Loc);
      
      // Emit the file/line/column that this expansion came from.
      OS << SM.getBuffer(IInfo.first)->getBufferIdentifier() << ':'
         << SM.getLineNumber(IInfo.first, IInfo.second) << ':';
      if (ShowColumn)
        OS << SM.getColumnNumber(IInfo.first, IInfo.second) << ':';
      OS << ' ';
    }
    OS << "note: instantiated from:\n";
    
    EmitCaretDiagnostic(Loc, Ranges, NumRanges, SM, Hints, NumHints, Columns);
    return;
  }
  
  // Decompose the location into a FID/Offset pair.
  std::pair<FileID, unsigned> LocInfo = SM.getDecomposedLoc(Loc);
  FileID FID = LocInfo.first;
  unsigned FileOffset = LocInfo.second;
  
  // Get information about the buffer it points into.
  std::pair<const char*, const char*> BufferInfo = SM.getBufferData(FID);
  const char *BufStart = BufferInfo.first;

  unsigned ColNo = SM.getColumnNumber(FID, FileOffset);
  unsigned CaretEndColNo 
    = ColNo + Lexer::MeasureTokenLength(Loc, SM, *LangOpts);

  // Rewind from the current position to the start of the line.
  const char *TokPtr = BufStart+FileOffset;
  const char *LineStart = TokPtr-ColNo+1; // Column # is 1-based.
  
  
  // Compute the line end.  Scan forward from the error position to the end of
  // the line.
  const char *LineEnd = TokPtr;
  while (*LineEnd != '\n' && *LineEnd != '\r' && *LineEnd != '\0')
    ++LineEnd;
  
  // Copy the line of code into an std::string for ease of manipulation.
  std::string SourceLine(LineStart, LineEnd);
  
  // Create a line for the caret that is filled with spaces that is the same
  // length as the line of source code.
  std::string CaretLine(LineEnd-LineStart, ' ');
  
  // Highlight all of the characters covered by Ranges with ~ characters.
  if (NumRanges) {
    unsigned LineNo = SM.getLineNumber(FID, FileOffset);
    
    for (unsigned i = 0, e = NumRanges; i != e; ++i)
      HighlightRange(Ranges[i], SM, LineNo, FID, CaretLine, SourceLine);
  }
  
  // Next, insert the caret itself.
  if (ColNo-1 < CaretLine.size())
    CaretLine[ColNo-1] = '^';
  else
    CaretLine.push_back('^');
  
  // Scan the source line, looking for tabs.  If we find any, manually expand
  // them to 8 characters and update the CaretLine to match.
  for (unsigned i = 0; i != SourceLine.size(); ++i) {
    if (SourceLine[i] != '\t') continue;
    
    // Replace this tab with at least one space.
    SourceLine[i] = ' ';
    
    // Compute the number of spaces we need to insert.
    unsigned NumSpaces = ((i+8)&~7) - (i+1);
    assert(NumSpaces < 8 && "Invalid computation of space amt");
    
//.........这里部分代码省略.........

bool
XdfObject::Read(SourceManager& sm, Diagnostic& diags)
{
    const llvm::MemoryBuffer& in = *sm.getBuffer(sm.getMainFileID());
    InputBuffer inbuf(in);
    inbuf.setLittleEndian();

    // Read object header
    if (inbuf.getReadableSize() < FILEHEAD_SIZE)
    {
        diags.Report(SourceLocation(), diag::err_object_header_unreadable);
        return false;
    }
    unsigned long magic = ReadU32(inbuf);
    if (magic != XDF_MAGIC)
    {
        diags.Report(SourceLocation(), diag::err_not_file_type) << "XDF";
        return false;
    }
    unsigned long scnum = ReadU32(inbuf);
    unsigned long symnum = ReadU32(inbuf);
    unsigned long headers_len = ReadU32(inbuf);
    if (inbuf.getReadableSize() < headers_len)
    {
        diags.Report(SourceLocation(), diag::err_xdf_headers_unreadable);
        return false;
    }

    unsigned long section_offset = FILEHEAD_SIZE;
    unsigned long symtab_offset = section_offset + SECTHEAD_SIZE*scnum;
    unsigned long strtab_offset = symtab_offset + SYMBOL_SIZE*symnum;
    ReadString read_string(in, strtab_offset,
                           FILEHEAD_SIZE+headers_len-strtab_offset,
                           diags);

    // Storage for nrelocs, indexed by section number
    std::vector<unsigned long> sects_nrelocs;
    sects_nrelocs.reserve(scnum);

    // Create sections
    for (unsigned long i=0; i<scnum; ++i)
    {
        // Start with symbol=0 as it's not created yet; updated later.
        std::auto_ptr<XdfSection> xsect(new XdfSection(SymbolRef(0)));
        unsigned long name_sym_index;
        IntNum lma, vma;
        unsigned long align;
        bool bss;
        unsigned long filepos;
        unsigned long nrelocs;
        xsect->Read(in, section_offset+SECTHEAD_SIZE*i, &name_sym_index, &lma,
                    &vma, &align, &bss, &filepos, &nrelocs);
        xsect->scnum = i;

        // get section name from section symbol entry
        inbuf.setPosition(symtab_offset+name_sym_index*SYMBOL_SIZE+8);
        llvm::StringRef sectname = read_string(ReadU32(inbuf));

        std::auto_ptr<Section> section(
            new Section(sectname, xsect->bits != 0, bss, SourceLocation()));

        section->setFilePos(filepos);
        section->setVMA(vma);
        section->setLMA(lma);

        if (bss)
        {
            Bytecode& gap =
                section->AppendGap(xsect->size, SourceLocation());
            Diagnostic nodiags(0);
            gap.CalcLen(0, nodiags);    // force length calculation of gap
        }
        else
        {
            // Read section data
            inbuf.setPosition(filepos);
            if (inbuf.getReadableSize() < xsect->size)
            {
                diags.Report(SourceLocation(),
                             diag::err_section_data_unreadable) << sectname;
                return false;
            }

            section->bytecodes_front().getFixed().Write(inbuf.Read(xsect->size),
                                                        xsect->size);
        }

        // Associate section data with section
        section->AddAssocData(xsect);

        // Add section to object
        m_object.AppendSection(section);

        sects_nrelocs.push_back(nrelocs);
    }

    // Create symbols
    inbuf.setPosition(symtab_offset);
    for (unsigned long i=0; i<symnum; ++i)
    {
//.........这里部分代码省略.........