Home | History | Annotate | Download | only in Frontend 
      1 //===--- PrintPreprocessedOutput.cpp - Implement the -E mode --------------===//
      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 code simply runs the preprocessor on the input file and prints out the
     11 // result.  This is the traditional behavior of the -E option.
     12 //
     13 //===----------------------------------------------------------------------===//
     14 
     15 #include "clang/Frontend/Utils.h"
     16 #include "clang/Basic/CharInfo.h"
     17 #include "clang/Basic/Diagnostic.h"
     18 #include "clang/Basic/SourceManager.h"
     19 #include "clang/Frontend/PreprocessorOutputOptions.h"
     20 #include "clang/Lex/MacroInfo.h"
     21 #include "clang/Lex/PPCallbacks.h"
     22 #include "clang/Lex/Pragma.h"
     23 #include "clang/Lex/Preprocessor.h"
     24 #include "clang/Lex/TokenConcatenation.h"
     25 #include "llvm/ADT/STLExtras.h"
     26 #include "llvm/ADT/SmallString.h"
     27 #include "llvm/ADT/StringRef.h"
     28 #include "llvm/Support/ErrorHandling.h"
     29 #include "llvm/Support/raw_ostream.h"
     30 #include <cstdio>
     31 using namespace clang;
     32 
     33 /// PrintMacroDefinition - Print a macro definition in a form that will be
     34 /// properly accepted back as a definition.
     35 static void PrintMacroDefinition(const IdentifierInfo &II, const MacroInfo &MI,
     36                                  Preprocessor &PP, raw_ostream &OS) {
     37   OS << "#define " << II.getName();
     38 
     39   if (MI.isFunctionLike()) {
     40     OS << '(';
     41     if (!MI.arg_empty()) {
     42       MacroInfo::arg_iterator AI = MI.arg_begin(), E = MI.arg_end();
     43       for (; AI+1 != E; ++AI) {
     44         OS << (*AI)->getName();
     45         OS << ',';
     46       }
     47 
     48       // Last argument.
     49       if ((*AI)->getName() == "__VA_ARGS__")
     50         OS << "...";
     51       else
     52         OS << (*AI)->getName();
     53     }
     54 
     55     if (MI.isGNUVarargs())
     56       OS << "...";  // #define foo(x...)
     57 
     58     OS << ')';
     59   }
     60 
     61   // GCC always emits a space, even if the macro body is empty.  However, do not
     62   // want to emit two spaces if the first token has a leading space.
     63   if (MI.tokens_empty() || !MI.tokens_begin()->hasLeadingSpace())
     64     OS << ' ';
     65 
     66   SmallString<128> SpellingBuffer;
     67   for (const auto &T : MI.tokens()) {
     68     if (T.hasLeadingSpace())
     69       OS << ' ';
     70 
     71     OS << PP.getSpelling(T, SpellingBuffer);
     72   }
     73 }
     74 
     75 //===----------------------------------------------------------------------===//
     76 // Preprocessed token printer
     77 //===----------------------------------------------------------------------===//
     78 
     79 namespace {
     80 class PrintPPOutputPPCallbacks : public PPCallbacks {
     81   Preprocessor &PP;
     82   SourceManager &SM;
     83   TokenConcatenation ConcatInfo;
     84 public:
     85   raw_ostream &OS;
     86 private:
     87   unsigned CurLine;
     88 
     89   bool EmittedTokensOnThisLine;
     90   bool EmittedDirectiveOnThisLine;
     91   SrcMgr::CharacteristicKind FileType;
     92   SmallString<512> CurFilename;
     93   bool Initialized;
     94   bool DisableLineMarkers;
     95   bool DumpDefines;
     96   bool UseLineDirectives;
     97   bool IsFirstFileEntered;
     98 public:
     99   PrintPPOutputPPCallbacks(Preprocessor &pp, raw_ostream &os, bool lineMarkers,
    100                            bool defines, bool UseLineDirectives)
    101       : PP(pp), SM(PP.getSourceManager()), ConcatInfo(PP), OS(os),
    102         DisableLineMarkers(lineMarkers), DumpDefines(defines),
    103         UseLineDirectives(UseLineDirectives) {
    104     CurLine = 0;
    105     CurFilename += "<uninit>";
    106     EmittedTokensOnThisLine = false;
    107     EmittedDirectiveOnThisLine = false;
    108     FileType = SrcMgr::C_User;
    109     Initialized = false;
    110     IsFirstFileEntered = false;
    111   }
    112 
    113   void setEmittedTokensOnThisLine() { EmittedTokensOnThisLine = true; }
    114   bool hasEmittedTokensOnThisLine() const { return EmittedTokensOnThisLine; }
    115 
    116   void setEmittedDirectiveOnThisLine() { EmittedDirectiveOnThisLine = true; }
    117   bool hasEmittedDirectiveOnThisLine() const {
    118     return EmittedDirectiveOnThisLine;
    119   }
    120 
    121   bool startNewLineIfNeeded(bool ShouldUpdateCurrentLine = true);
    122 
    123   void FileChanged(SourceLocation Loc, FileChangeReason Reason,
    124                    SrcMgr::CharacteristicKind FileType,
    125                    FileID PrevFID) override;
    126   void InclusionDirective(SourceLocation HashLoc, const Token &IncludeTok,
    127                           StringRef FileName, bool IsAngled,
    128                           CharSourceRange FilenameRange, const FileEntry *File,
    129                           StringRef SearchPath, StringRef RelativePath,
    130                           const Module *Imported) override;
    131   void Ident(SourceLocation Loc, StringRef str) override;
    132   void PragmaMessage(SourceLocation Loc, StringRef Namespace,
    133                      PragmaMessageKind Kind, StringRef Str) override;
    134   void PragmaDebug(SourceLocation Loc, StringRef DebugType) override;
    135   void PragmaDiagnosticPush(SourceLocation Loc, StringRef Namespace) override;
    136   void PragmaDiagnosticPop(SourceLocation Loc, StringRef Namespace) override;
    137   void PragmaDiagnostic(SourceLocation Loc, StringRef Namespace,
    138                         diag::Severity Map, StringRef Str) override;
    139   void PragmaWarning(SourceLocation Loc, StringRef WarningSpec,
    140                      ArrayRef<int> Ids) override;
    141   void PragmaWarningPush(SourceLocation Loc, int Level) override;
    142   void PragmaWarningPop(SourceLocation Loc) override;
    143 
    144   bool HandleFirstTokOnLine(Token &Tok);
    145 
    146   /// Move to the line of the provided source location. This will
    147   /// return true if the output stream required adjustment or if
    148   /// the requested location is on the first line.
    149   bool MoveToLine(SourceLocation Loc) {
    150     PresumedLoc PLoc = SM.getPresumedLoc(Loc);
    151     if (PLoc.isInvalid())
    152       return false;
    153     return MoveToLine(PLoc.getLine()) || (PLoc.getLine() == 1);
    154   }
    155   bool MoveToLine(unsigned LineNo);
    156 
    157   bool AvoidConcat(const Token &PrevPrevTok, const Token &PrevTok,
    158                    const Token &Tok) {
    159     return ConcatInfo.AvoidConcat(PrevPrevTok, PrevTok, Tok);
    160   }
    161   void WriteLineInfo(unsigned LineNo, const char *Extra=nullptr,
    162                      unsigned ExtraLen=0);
    163   bool LineMarkersAreDisabled() const { return DisableLineMarkers; }
    164   void HandleNewlinesInToken(const char *TokStr, unsigned Len);
    165 
    166   /// MacroDefined - This hook is called whenever a macro definition is seen.
    167   void MacroDefined(const Token &MacroNameTok,
    168                     const MacroDirective *MD) override;
    169 
    170   /// MacroUndefined - This hook is called whenever a macro #undef is seen.
    171   void MacroUndefined(const Token &MacroNameTok,
    172                       const MacroDefinition &MD) override;
    173 };
    174 }  // end anonymous namespace
    175 
    176 void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
    177                                              const char *Extra,
    178                                              unsigned ExtraLen) {
    179   startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
    180 
    181   // Emit #line directives or GNU line markers depending on what mode we're in.
    182   if (UseLineDirectives) {
    183     OS << "#line" << ' ' << LineNo << ' ' << '"';
    184     OS.write_escaped(CurFilename);
    185     OS << '"';
    186   } else {
    187     OS << '#' << ' ' << LineNo << ' ' << '"';
    188     OS.write_escaped(CurFilename);
    189     OS << '"';
    190 
    191     if (ExtraLen)
    192       OS.write(Extra, ExtraLen);
    193 
    194     if (FileType == SrcMgr::C_System)
    195       OS.write(" 3", 2);
    196     else if (FileType == SrcMgr::C_ExternCSystem)
    197       OS.write(" 3 4", 4);
    198   }
    199   OS << '\n';
    200 }
    201 
    202 /// MoveToLine - Move the output to the source line specified by the location
    203 /// object.  We can do this by emitting some number of \n's, or be emitting a
    204 /// #line directive.  This returns false if already at the specified line, true
    205 /// if some newlines were emitted.
    206 bool PrintPPOutputPPCallbacks::MoveToLine(unsigned LineNo) {
    207   // If this line is "close enough" to the original line, just print newlines,
    208   // otherwise print a #line directive.
    209   if (LineNo-CurLine <= 8) {
    210     if (LineNo-CurLine == 1)
    211       OS << '\n';
    212     else if (LineNo == CurLine)
    213       return false;    // Spelling line moved, but expansion line didn't.
    214     else {
    215       const char *NewLines = "\n\n\n\n\n\n\n\n";
    216       OS.write(NewLines, LineNo-CurLine);
    217     }
    218   } else if (!DisableLineMarkers) {
    219     // Emit a #line or line marker.
    220     WriteLineInfo(LineNo, nullptr, 0);
    221   } else {
    222     // Okay, we're in -P mode, which turns off line markers.  However, we still
    223     // need to emit a newline between tokens on different lines.
    224     startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
    225   }
    226 
    227   CurLine = LineNo;
    228   return true;
    229 }
    230 
    231 bool
    232 PrintPPOutputPPCallbacks::startNewLineIfNeeded(bool ShouldUpdateCurrentLine) {
    233   if (EmittedTokensOnThisLine || EmittedDirectiveOnThisLine) {
    234     OS << '\n';
    235     EmittedTokensOnThisLine = false;
    236     EmittedDirectiveOnThisLine = false;
    237     if (ShouldUpdateCurrentLine)
    238       ++CurLine;
    239     return true;
    240   }
    241 
    242   return false;
    243 }
    244 
    245 /// FileChanged - Whenever the preprocessor enters or exits a #include file
    246 /// it invokes this handler.  Update our conception of the current source
    247 /// position.
    248 void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
    249                                            FileChangeReason Reason,
    250                                        SrcMgr::CharacteristicKind NewFileType,
    251                                        FileID PrevFID) {
    252   // Unless we are exiting a #include, make sure to skip ahead to the line the
    253   // #include directive was at.
    254   SourceManager &SourceMgr = SM;
    255 
    256   PresumedLoc UserLoc = SourceMgr.getPresumedLoc(Loc);
    257   if (UserLoc.isInvalid())
    258     return;
    259 
    260   unsigned NewLine = UserLoc.getLine();
    261 
    262   if (Reason == PPCallbacks::EnterFile) {
    263     SourceLocation IncludeLoc = UserLoc.getIncludeLoc();
    264     if (IncludeLoc.isValid())
    265       MoveToLine(IncludeLoc);
    266   } else if (Reason == PPCallbacks::SystemHeaderPragma) {
    267     // GCC emits the # directive for this directive on the line AFTER the
    268     // directive and emits a bunch of spaces that aren't needed. This is because
    269     // otherwise we will emit a line marker for THIS line, which requires an
    270     // extra blank line after the directive to avoid making all following lines
    271     // off by one. We can do better by simply incrementing NewLine here.
    272     NewLine += 1;
    273   }
    274 
    275   CurLine = NewLine;
    276 
    277   CurFilename.clear();
    278   CurFilename += UserLoc.getFilename();
    279   FileType = NewFileType;
    280 
    281   if (DisableLineMarkers) {
    282     startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
    283     return;
    284   }
    285 
    286   if (!Initialized) {
    287     WriteLineInfo(CurLine);
    288     Initialized = true;
    289   }
    290 
    291   // Do not emit an enter marker for the main file (which we expect is the first
    292   // entered file). This matches gcc, and improves compatibility with some tools
    293   // which track the # line markers as a way to determine when the preprocessed
    294   // output is in the context of the main file.
    295   if (Reason == PPCallbacks::EnterFile && !IsFirstFileEntered) {
    296     IsFirstFileEntered = true;
    297     return;
    298   }
    299 
    300   switch (Reason) {
    301   case PPCallbacks::EnterFile:
    302     WriteLineInfo(CurLine, " 1", 2);
    303     break;
    304   case PPCallbacks::ExitFile:
    305     WriteLineInfo(CurLine, " 2", 2);
    306     break;
    307   case PPCallbacks::SystemHeaderPragma:
    308   case PPCallbacks::RenameFile:
    309     WriteLineInfo(CurLine);
    310     break;
    311   }
    312 }
    313 
    314 void PrintPPOutputPPCallbacks::InclusionDirective(SourceLocation HashLoc,
    315                                                   const Token &IncludeTok,
    316                                                   StringRef FileName,
    317                                                   bool IsAngled,
    318                                                   CharSourceRange FilenameRange,
    319                                                   const FileEntry *File,
    320                                                   StringRef SearchPath,
    321                                                   StringRef RelativePath,
    322                                                   const Module *Imported) {
    323   // When preprocessing, turn implicit imports into @imports.
    324   // FIXME: This is a stop-gap until a more comprehensive "preprocessing with
    325   // modules" solution is introduced.
    326   if (Imported) {
    327     startNewLineIfNeeded();
    328     MoveToLine(HashLoc);
    329     OS << "@import " << Imported->getFullModuleName() << ";"
    330        << " /* clang -E: implicit import for \"" << File->getName() << "\" */";
    331     // Since we want a newline after the @import, but not a #<line>, start a new
    332     // line immediately.
    333     EmittedTokensOnThisLine = true;
    334     startNewLineIfNeeded();
    335   }
    336 }
    337 
    338 /// Ident - Handle #ident directives when read by the preprocessor.
    339 ///
    340 void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, StringRef S) {
    341   MoveToLine(Loc);
    342 
    343   OS.write("#ident ", strlen("#ident "));
    344   OS.write(S.begin(), S.size());
    345   EmittedTokensOnThisLine = true;
    346 }
    347 
    348 /// MacroDefined - This hook is called whenever a macro definition is seen.
    349 void PrintPPOutputPPCallbacks::MacroDefined(const Token &MacroNameTok,
    350                                             const MacroDirective *MD) {
    351   const MacroInfo *MI = MD->getMacroInfo();
    352   // Only print out macro definitions in -dD mode.
    353   if (!DumpDefines ||
    354       // Ignore __FILE__ etc.
    355       MI->isBuiltinMacro()) return;
    356 
    357   MoveToLine(MI->getDefinitionLoc());
    358   PrintMacroDefinition(*MacroNameTok.getIdentifierInfo(), *MI, PP, OS);
    359   setEmittedDirectiveOnThisLine();
    360 }
    361 
    362 void PrintPPOutputPPCallbacks::MacroUndefined(const Token &MacroNameTok,
    363                                               const MacroDefinition &MD) {
    364   // Only print out macro definitions in -dD mode.
    365   if (!DumpDefines) return;
    366 
    367   MoveToLine(MacroNameTok.getLocation());
    368   OS << "#undef " << MacroNameTok.getIdentifierInfo()->getName();
    369   setEmittedDirectiveOnThisLine();
    370 }
    371 
    372 static void outputPrintable(llvm::raw_ostream& OS,
    373                                              const std::string &Str) {
    374     for (unsigned i = 0, e = Str.size(); i != e; ++i) {
    375       unsigned char Char = Str[i];
    376       if (isPrintable(Char) && Char != '\\' && Char != '"')
    377         OS << (char)Char;
    378       else  // Output anything hard as an octal escape.
    379         OS << '\\'
    380            << (char)('0'+ ((Char >> 6) & 7))
    381            << (char)('0'+ ((Char >> 3) & 7))
    382            << (char)('0'+ ((Char >> 0) & 7));
    383     }
    384 }
    385 
    386 void PrintPPOutputPPCallbacks::PragmaMessage(SourceLocation Loc,
    387                                              StringRef Namespace,
    388                                              PragmaMessageKind Kind,
    389                                              StringRef Str) {
    390   startNewLineIfNeeded();
    391   MoveToLine(Loc);
    392   OS << "#pragma ";
    393   if (!Namespace.empty())
    394     OS << Namespace << ' ';
    395   switch (Kind) {
    396     case PMK_Message:
    397       OS << "message(\"";
    398       break;
    399     case PMK_Warning:
    400       OS << "warning \"";
    401       break;
    402     case PMK_Error:
    403       OS << "error \"";
    404       break;
    405   }
    406 
    407   outputPrintable(OS, Str);
    408   OS << '"';
    409   if (Kind == PMK_Message)
    410     OS << ')';
    411   setEmittedDirectiveOnThisLine();
    412 }
    413 
    414 void PrintPPOutputPPCallbacks::PragmaDebug(SourceLocation Loc,
    415                                            StringRef DebugType) {
    416   startNewLineIfNeeded();
    417   MoveToLine(Loc);
    418 
    419   OS << "#pragma clang __debug ";
    420   OS << DebugType;
    421 
    422   setEmittedDirectiveOnThisLine();
    423 }
    424 
    425 void PrintPPOutputPPCallbacks::
    426 PragmaDiagnosticPush(SourceLocation Loc, StringRef Namespace) {
    427   startNewLineIfNeeded();
    428   MoveToLine(Loc);
    429   OS << "#pragma " << Namespace << " diagnostic push";
    430   setEmittedDirectiveOnThisLine();
    431 }
    432 
    433 void PrintPPOutputPPCallbacks::
    434 PragmaDiagnosticPop(SourceLocation Loc, StringRef Namespace) {
    435   startNewLineIfNeeded();
    436   MoveToLine(Loc);
    437   OS << "#pragma " << Namespace << " diagnostic pop";
    438   setEmittedDirectiveOnThisLine();
    439 }
    440 
    441 void PrintPPOutputPPCallbacks::PragmaDiagnostic(SourceLocation Loc,
    442                                                 StringRef Namespace,
    443                                                 diag::Severity Map,
    444                                                 StringRef Str) {
    445   startNewLineIfNeeded();
    446   MoveToLine(Loc);
    447   OS << "#pragma " << Namespace << " diagnostic ";
    448   switch (Map) {
    449   case diag::Severity::Remark:
    450     OS << "remark";
    451     break;
    452   case diag::Severity::Warning:
    453     OS << "warning";
    454     break;
    455   case diag::Severity::Error:
    456     OS << "error";
    457     break;
    458   case diag::Severity::Ignored:
    459     OS << "ignored";
    460     break;
    461   case diag::Severity::Fatal:
    462     OS << "fatal";
    463     break;
    464   }
    465   OS << " \"" << Str << '"';
    466   setEmittedDirectiveOnThisLine();
    467 }
    468 
    469 void PrintPPOutputPPCallbacks::PragmaWarning(SourceLocation Loc,
    470                                              StringRef WarningSpec,
    471                                              ArrayRef<int> Ids) {
    472   startNewLineIfNeeded();
    473   MoveToLine(Loc);
    474   OS << "#pragma warning(" << WarningSpec << ':';
    475   for (ArrayRef<int>::iterator I = Ids.begin(), E = Ids.end(); I != E; ++I)
    476     OS << ' ' << *I;
    477   OS << ')';
    478   setEmittedDirectiveOnThisLine();
    479 }
    480 
    481 void PrintPPOutputPPCallbacks::PragmaWarningPush(SourceLocation Loc,
    482                                                  int Level) {
    483   startNewLineIfNeeded();
    484   MoveToLine(Loc);
    485   OS << "#pragma warning(push";
    486   if (Level >= 0)
    487     OS << ", " << Level;
    488   OS << ')';
    489   setEmittedDirectiveOnThisLine();
    490 }
    491 
    492 void PrintPPOutputPPCallbacks::PragmaWarningPop(SourceLocation Loc) {
    493   startNewLineIfNeeded();
    494   MoveToLine(Loc);
    495   OS << "#pragma warning(pop)";
    496   setEmittedDirectiveOnThisLine();
    497 }
    498 
    499 /// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
    500 /// is called for the first token on each new line.  If this really is the start
    501 /// of a new logical line, handle it and return true, otherwise return false.
    502 /// This may not be the start of a logical line because the "start of line"
    503 /// marker is set for spelling lines, not expansion ones.
    504 bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
    505   // Figure out what line we went to and insert the appropriate number of
    506   // newline characters.
    507   if (!MoveToLine(Tok.getLocation()))
    508     return false;
    509 
    510   // Print out space characters so that the first token on a line is
    511   // indented for easy reading.
    512   unsigned ColNo = SM.getExpansionColumnNumber(Tok.getLocation());
    513 
    514   // The first token on a line can have a column number of 1, yet still expect
    515   // leading white space, if a macro expansion in column 1 starts with an empty
    516   // macro argument, or an empty nested macro expansion. In this case, move the
    517   // token to column 2.
    518   if (ColNo == 1 && Tok.hasLeadingSpace())
    519     ColNo = 2;
    520 
    521   // This hack prevents stuff like:
    522   // #define HASH #
    523   // HASH define foo bar
    524   // From having the # character end up at column 1, which makes it so it
    525   // is not handled as a #define next time through the preprocessor if in
    526   // -fpreprocessed mode.
    527   if (ColNo <= 1 && Tok.is(tok::hash))
    528     OS << ' ';
    529 
    530   // Otherwise, indent the appropriate number of spaces.
    531   for (; ColNo > 1; --ColNo)
    532     OS << ' ';
    533 
    534   return true;
    535 }
    536 
    537 void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
    538                                                      unsigned Len) {
    539   unsigned NumNewlines = 0;
    540   for (; Len; --Len, ++TokStr) {
    541     if (*TokStr != '\n' &&
    542         *TokStr != '\r')
    543       continue;
    544 
    545     ++NumNewlines;
    546 
    547     // If we have \n\r or \r\n, skip both and count as one line.
    548     if (Len != 1 &&
    549         (TokStr[1] == '\n' || TokStr[1] == '\r') &&
    550         TokStr[0] != TokStr[1])
    551       ++TokStr, --Len;
    552   }
    553 
    554   if (NumNewlines == 0) return;
    555 
    556   CurLine += NumNewlines;
    557 }
    558 
    559 
    560 namespace {
    561 struct UnknownPragmaHandler : public PragmaHandler {
    562   const char *Prefix;
    563   PrintPPOutputPPCallbacks *Callbacks;
    564 
    565   // Set to true if tokens should be expanded
    566   bool ShouldExpandTokens;
    567 
    568   UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks,
    569                        bool RequireTokenExpansion)
    570       : Prefix(prefix), Callbacks(callbacks),
    571         ShouldExpandTokens(RequireTokenExpansion) {}
    572   void HandlePragma(Preprocessor &PP, PragmaIntroducerKind Introducer,
    573                     Token &PragmaTok) override {
    574     // Figure out what line we went to and insert the appropriate number of
    575     // newline characters.
    576     Callbacks->startNewLineIfNeeded();
    577     Callbacks->MoveToLine(PragmaTok.getLocation());
    578     Callbacks->OS.write(Prefix, strlen(Prefix));
    579 
    580     Token PrevToken;
    581     Token PrevPrevToken;
    582     PrevToken.startToken();
    583     PrevPrevToken.startToken();
    584 
    585     // Read and print all of the pragma tokens.
    586     while (PragmaTok.isNot(tok::eod)) {
    587       if (PragmaTok.hasLeadingSpace() ||
    588           Callbacks->AvoidConcat(PrevPrevToken, PrevToken, PragmaTok))
    589         Callbacks->OS << ' ';
    590       std::string TokSpell = PP.getSpelling(PragmaTok);
    591       Callbacks->OS.write(&TokSpell[0], TokSpell.size());
    592 
    593       PrevPrevToken = PrevToken;
    594       PrevToken = PragmaTok;
    595 
    596       if (ShouldExpandTokens)
    597         PP.Lex(PragmaTok);
    598       else
    599         PP.LexUnexpandedToken(PragmaTok);
    600     }
    601     Callbacks->setEmittedDirectiveOnThisLine();
    602   }
    603 };
    604 } // end anonymous namespace
    605 
    606 
    607 static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
    608                                     PrintPPOutputPPCallbacks *Callbacks,
    609                                     raw_ostream &OS) {
    610   bool DropComments = PP.getLangOpts().TraditionalCPP &&
    611                       !PP.getCommentRetentionState();
    612 
    613   char Buffer[256];
    614   Token PrevPrevTok, PrevTok;
    615   PrevPrevTok.startToken();
    616   PrevTok.startToken();
    617   while (1) {
    618     if (Callbacks->hasEmittedDirectiveOnThisLine()) {
    619       Callbacks->startNewLineIfNeeded();
    620       Callbacks->MoveToLine(Tok.getLocation());
    621     }
    622 
    623     // If this token is at the start of a line, emit newlines if needed.
    624     if (Tok.isAtStartOfLine() && Callbacks->HandleFirstTokOnLine(Tok)) {
    625       // done.
    626     } else if (Tok.hasLeadingSpace() ||
    627                // If we haven't emitted a token on this line yet, PrevTok isn't
    628                // useful to look at and no concatenation could happen anyway.
    629                (Callbacks->hasEmittedTokensOnThisLine() &&
    630                 // Don't print "-" next to "-", it would form "--".
    631                 Callbacks->AvoidConcat(PrevPrevTok, PrevTok, Tok))) {
    632       OS << ' ';
    633     }
    634 
    635     if (DropComments && Tok.is(tok::comment)) {
    636       // Skip comments. Normally the preprocessor does not generate
    637       // tok::comment nodes at all when not keeping comments, but under
    638       // -traditional-cpp the lexer keeps /all/ whitespace, including comments.
    639       SourceLocation StartLoc = Tok.getLocation();
    640       Callbacks->MoveToLine(StartLoc.getLocWithOffset(Tok.getLength()));
    641     } else if (Tok.is(tok::annot_module_include) ||
    642                Tok.is(tok::annot_module_begin) ||
    643                Tok.is(tok::annot_module_end)) {
    644       // PrintPPOutputPPCallbacks::InclusionDirective handles producing
    645       // appropriate output here. Ignore this token entirely.
    646       PP.Lex(Tok);
    647       continue;
    648     } else if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
    649       OS << II->getName();
    650     } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
    651                Tok.getLiteralData()) {
    652       OS.write(Tok.getLiteralData(), Tok.getLength());
    653     } else if (Tok.getLength() < 256) {
    654       const char *TokPtr = Buffer;
    655       unsigned Len = PP.getSpelling(Tok, TokPtr);
    656       OS.write(TokPtr, Len);
    657 
    658       // Tokens that can contain embedded newlines need to adjust our current
    659       // line number.
    660       if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
    661         Callbacks->HandleNewlinesInToken(TokPtr, Len);
    662     } else {
    663       std::string S = PP.getSpelling(Tok);
    664       OS.write(&S[0], S.size());
    665 
    666       // Tokens that can contain embedded newlines need to adjust our current
    667       // line number.
    668       if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
    669         Callbacks->HandleNewlinesInToken(&S[0], S.size());
    670     }
    671     Callbacks->setEmittedTokensOnThisLine();
    672 
    673     if (Tok.is(tok::eof)) break;
    674 
    675     PrevPrevTok = PrevTok;
    676     PrevTok = Tok;
    677     PP.Lex(Tok);
    678   }
    679 }
    680 
    681 typedef std::pair<const IdentifierInfo *, MacroInfo *> id_macro_pair;
    682 static int MacroIDCompare(const id_macro_pair *LHS, const id_macro_pair *RHS) {
    683   return LHS->first->getName().compare(RHS->first->getName());
    684 }
    685 
    686 static void DoPrintMacros(Preprocessor &PP, raw_ostream *OS) {
    687   // Ignore unknown pragmas.
    688   PP.IgnorePragmas();
    689 
    690   // -dM mode just scans and ignores all tokens in the files, then dumps out
    691   // the macro table at the end.
    692   PP.EnterMainSourceFile();
    693 
    694   Token Tok;
    695   do PP.Lex(Tok);
    696   while (Tok.isNot(tok::eof));
    697 
    698   SmallVector<id_macro_pair, 128> MacrosByID;
    699   for (Preprocessor::macro_iterator I = PP.macro_begin(), E = PP.macro_end();
    700        I != E; ++I) {
    701     auto *MD = I->second.getLatest();
    702     if (MD && MD->isDefined())
    703       MacrosByID.push_back(id_macro_pair(I->first, MD->getMacroInfo()));
    704   }
    705   llvm::array_pod_sort(MacrosByID.begin(), MacrosByID.end(), MacroIDCompare);
    706 
    707   for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
    708     MacroInfo &MI = *MacrosByID[i].second;
    709     // Ignore computed macros like __LINE__ and friends.
    710     if (MI.isBuiltinMacro()) continue;
    711 
    712     PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
    713     *OS << '\n';
    714   }
    715 }
    716 
    717 /// DoPrintPreprocessedInput - This implements -E mode.
    718 ///
    719 void clang::DoPrintPreprocessedInput(Preprocessor &PP, raw_ostream *OS,
    720                                      const PreprocessorOutputOptions &Opts) {
    721   // Show macros with no output is handled specially.
    722   if (!Opts.ShowCPP) {
    723     assert(Opts.ShowMacros && "Not yet implemented!");
    724     DoPrintMacros(PP, OS);
    725     return;
    726   }
    727 
    728   // Inform the preprocessor whether we want it to retain comments or not, due
    729   // to -C or -CC.
    730   PP.SetCommentRetentionState(Opts.ShowComments, Opts.ShowMacroComments);
    731 
    732   PrintPPOutputPPCallbacks *Callbacks = new PrintPPOutputPPCallbacks(
    733       PP, *OS, !Opts.ShowLineMarkers, Opts.ShowMacros, Opts.UseLineDirectives);
    734 
    735   // Expand macros in pragmas with -fms-extensions.  The assumption is that
    736   // the majority of pragmas in such a file will be Microsoft pragmas.
    737   PP.AddPragmaHandler(new UnknownPragmaHandler(
    738       "#pragma", Callbacks,
    739       /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
    740   PP.AddPragmaHandler(
    741       "GCC", new UnknownPragmaHandler(
    742                  "#pragma GCC", Callbacks,
    743                  /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
    744   PP.AddPragmaHandler(
    745       "clang", new UnknownPragmaHandler(
    746                    "#pragma clang", Callbacks,
    747                    /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
    748 
    749   // The tokens after pragma omp need to be expanded.
    750   //
    751   //  OpenMP [2.1, Directive format]
    752   //  Preprocessing tokens following the #pragma omp are subject to macro
    753   //  replacement.
    754   PP.AddPragmaHandler("omp",
    755                       new UnknownPragmaHandler("#pragma omp", Callbacks,
    756                                                /*RequireTokenExpansion=*/true));
    757 
    758   PP.addPPCallbacks(std::unique_ptr<PPCallbacks>(Callbacks));
    759 
    760   // After we have configured the preprocessor, enter the main file.
    761   PP.EnterMainSourceFile();
    762 
    763   // Consume all of the tokens that come from the predefines buffer.  Those
    764   // should not be emitted into the output and are guaranteed to be at the
    765   // start.
    766   const SourceManager &SourceMgr = PP.getSourceManager();
    767   Token Tok;
    768   do {
    769     PP.Lex(Tok);
    770     if (Tok.is(tok::eof) || !Tok.getLocation().isFileID())
    771       break;
    772 
    773     PresumedLoc PLoc = SourceMgr.getPresumedLoc(Tok.getLocation());
    774     if (PLoc.isInvalid())
    775       break;
    776 
    777     if (strcmp(PLoc.getFilename(), "<built-in>"))
    778       break;
    779   } while (true);
    780 
    781   // Read all the preprocessed tokens, printing them out to the stream.
    782   PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
    783   *OS << '\n';
    784 }
    785