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 (MacroInfo::tokens_iterator I = MI.tokens_begin(), E = MI.tokens_end();
     68        I != E; ++I) {
     69     if (I->hasLeadingSpace())
     70       OS << ' ';
     71 
     72     OS << PP.getSpelling(*I, SpellingBuffer);
     73   }
     74 }
     75 
     76 //===----------------------------------------------------------------------===//
     77 // Preprocessed token printer
     78 //===----------------------------------------------------------------------===//
     79 
     80 namespace {
     81 class PrintPPOutputPPCallbacks : public PPCallbacks {
     82   Preprocessor &PP;
     83   SourceManager &SM;
     84   TokenConcatenation ConcatInfo;
     85 public:
     86   raw_ostream &OS;
     87 private:
     88   unsigned CurLine;
     89 
     90   bool EmittedTokensOnThisLine;
     91   bool EmittedDirectiveOnThisLine;
     92   SrcMgr::CharacteristicKind FileType;
     93   SmallString<512> CurFilename;
     94   bool Initialized;
     95   bool DisableLineMarkers;
     96   bool DumpDefines;
     97   bool UseLineDirective;
     98   bool IsFirstFileEntered;
     99 public:
    100   PrintPPOutputPPCallbacks(Preprocessor &pp, raw_ostream &os,
    101                            bool lineMarkers, bool defines)
    102      : PP(pp), SM(PP.getSourceManager()),
    103        ConcatInfo(PP), OS(os), DisableLineMarkers(lineMarkers),
    104        DumpDefines(defines) {
    105     CurLine = 0;
    106     CurFilename += "<uninit>";
    107     EmittedTokensOnThisLine = false;
    108     EmittedDirectiveOnThisLine = false;
    109     FileType = SrcMgr::C_User;
    110     Initialized = false;
    111     IsFirstFileEntered = false;
    112 
    113     // If we're in microsoft mode, use normal #line instead of line markers.
    114     UseLineDirective = PP.getLangOpts().MicrosoftExt;
    115   }
    116 
    117   void setEmittedTokensOnThisLine() { EmittedTokensOnThisLine = true; }
    118   bool hasEmittedTokensOnThisLine() const { return EmittedTokensOnThisLine; }
    119 
    120   void setEmittedDirectiveOnThisLine() { EmittedDirectiveOnThisLine = true; }
    121   bool hasEmittedDirectiveOnThisLine() const {
    122     return EmittedDirectiveOnThisLine;
    123   }
    124 
    125   bool startNewLineIfNeeded(bool ShouldUpdateCurrentLine = true);
    126 
    127   virtual void FileChanged(SourceLocation Loc, FileChangeReason Reason,
    128                            SrcMgr::CharacteristicKind FileType,
    129                            FileID PrevFID);
    130   virtual void Ident(SourceLocation Loc, const std::string &str);
    131   virtual void PragmaComment(SourceLocation Loc, const IdentifierInfo *Kind,
    132                              const std::string &Str);
    133   virtual void PragmaMessage(SourceLocation Loc, StringRef Str);
    134   virtual void PragmaDiagnosticPush(SourceLocation Loc,
    135                                     StringRef Namespace);
    136   virtual void PragmaDiagnosticPop(SourceLocation Loc,
    137                                    StringRef Namespace);
    138   virtual void PragmaDiagnostic(SourceLocation Loc, StringRef Namespace,
    139                                 diag::Mapping Map, StringRef Str);
    140 
    141   bool HandleFirstTokOnLine(Token &Tok);
    142 
    143   /// Move to the line of the provided source location. This will
    144   /// return true if the output stream required adjustment or if
    145   /// the requested location is on the first line.
    146   bool MoveToLine(SourceLocation Loc) {
    147     PresumedLoc PLoc = SM.getPresumedLoc(Loc);
    148     if (PLoc.isInvalid())
    149       return false;
    150     return MoveToLine(PLoc.getLine()) || (PLoc.getLine() == 1);
    151   }
    152   bool MoveToLine(unsigned LineNo);
    153 
    154   bool AvoidConcat(const Token &PrevPrevTok, const Token &PrevTok,
    155                    const Token &Tok) {
    156     return ConcatInfo.AvoidConcat(PrevPrevTok, PrevTok, Tok);
    157   }
    158   void WriteLineInfo(unsigned LineNo, const char *Extra=0, unsigned ExtraLen=0);
    159   bool LineMarkersAreDisabled() const { return DisableLineMarkers; }
    160   void HandleNewlinesInToken(const char *TokStr, unsigned Len);
    161 
    162   /// MacroDefined - This hook is called whenever a macro definition is seen.
    163   void MacroDefined(const Token &MacroNameTok, const MacroDirective *MD);
    164 
    165   /// MacroUndefined - This hook is called whenever a macro #undef is seen.
    166   void MacroUndefined(const Token &MacroNameTok, const MacroDirective *MD);
    167 };
    168 }  // end anonymous namespace
    169 
    170 void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
    171                                              const char *Extra,
    172                                              unsigned ExtraLen) {
    173   startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
    174 
    175   // Emit #line directives or GNU line markers depending on what mode we're in.
    176   if (UseLineDirective) {
    177     OS << "#line" << ' ' << LineNo << ' ' << '"';
    178     OS.write(CurFilename.data(), CurFilename.size());
    179     OS << '"';
    180   } else {
    181     OS << '#' << ' ' << LineNo << ' ' << '"';
    182     OS.write(CurFilename.data(), CurFilename.size());
    183     OS << '"';
    184 
    185     if (ExtraLen)
    186       OS.write(Extra, ExtraLen);
    187 
    188     if (FileType == SrcMgr::C_System)
    189       OS.write(" 3", 2);
    190     else if (FileType == SrcMgr::C_ExternCSystem)
    191       OS.write(" 3 4", 4);
    192   }
    193   OS << '\n';
    194 }
    195 
    196 /// MoveToLine - Move the output to the source line specified by the location
    197 /// object.  We can do this by emitting some number of \n's, or be emitting a
    198 /// #line directive.  This returns false if already at the specified line, true
    199 /// if some newlines were emitted.
    200 bool PrintPPOutputPPCallbacks::MoveToLine(unsigned LineNo) {
    201   // If this line is "close enough" to the original line, just print newlines,
    202   // otherwise print a #line directive.
    203   if (LineNo-CurLine <= 8) {
    204     if (LineNo-CurLine == 1)
    205       OS << '\n';
    206     else if (LineNo == CurLine)
    207       return false;    // Spelling line moved, but expansion line didn't.
    208     else {
    209       const char *NewLines = "\n\n\n\n\n\n\n\n";
    210       OS.write(NewLines, LineNo-CurLine);
    211     }
    212   } else if (!DisableLineMarkers) {
    213     // Emit a #line or line marker.
    214     WriteLineInfo(LineNo, 0, 0);
    215   } else {
    216     // Okay, we're in -P mode, which turns off line markers.  However, we still
    217     // need to emit a newline between tokens on different lines.
    218     startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
    219   }
    220 
    221   CurLine = LineNo;
    222   return true;
    223 }
    224 
    225 bool
    226 PrintPPOutputPPCallbacks::startNewLineIfNeeded(bool ShouldUpdateCurrentLine) {
    227   if (EmittedTokensOnThisLine || EmittedDirectiveOnThisLine) {
    228     OS << '\n';
    229     EmittedTokensOnThisLine = false;
    230     EmittedDirectiveOnThisLine = false;
    231     if (ShouldUpdateCurrentLine)
    232       ++CurLine;
    233     return true;
    234   }
    235 
    236   return false;
    237 }
    238 
    239 /// FileChanged - Whenever the preprocessor enters or exits a #include file
    240 /// it invokes this handler.  Update our conception of the current source
    241 /// position.
    242 void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
    243                                            FileChangeReason Reason,
    244                                        SrcMgr::CharacteristicKind NewFileType,
    245                                        FileID PrevFID) {
    246   // Unless we are exiting a #include, make sure to skip ahead to the line the
    247   // #include directive was at.
    248   SourceManager &SourceMgr = SM;
    249 
    250   PresumedLoc UserLoc = SourceMgr.getPresumedLoc(Loc);
    251   if (UserLoc.isInvalid())
    252     return;
    253 
    254   unsigned NewLine = UserLoc.getLine();
    255 
    256   if (Reason == PPCallbacks::EnterFile) {
    257     SourceLocation IncludeLoc = UserLoc.getIncludeLoc();
    258     if (IncludeLoc.isValid())
    259       MoveToLine(IncludeLoc);
    260   } else if (Reason == PPCallbacks::SystemHeaderPragma) {
    261     MoveToLine(NewLine);
    262 
    263     // TODO GCC emits the # directive for this directive on the line AFTER the
    264     // directive and emits a bunch of spaces that aren't needed.  Emulate this
    265     // strange behavior.
    266   }
    267 
    268   CurLine = NewLine;
    269 
    270   CurFilename.clear();
    271   CurFilename += UserLoc.getFilename();
    272   Lexer::Stringify(CurFilename);
    273   FileType = NewFileType;
    274 
    275   if (DisableLineMarkers) {
    276     startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
    277     return;
    278   }
    279 
    280   if (!Initialized) {
    281     WriteLineInfo(CurLine);
    282     Initialized = true;
    283   }
    284 
    285   // Do not emit an enter marker for the main file (which we expect is the first
    286   // entered file). This matches gcc, and improves compatibility with some tools
    287   // which track the # line markers as a way to determine when the preprocessed
    288   // output is in the context of the main file.
    289   if (Reason == PPCallbacks::EnterFile && !IsFirstFileEntered) {
    290     IsFirstFileEntered = true;
    291     return;
    292   }
    293 
    294   switch (Reason) {
    295   case PPCallbacks::EnterFile:
    296     WriteLineInfo(CurLine, " 1", 2);
    297     break;
    298   case PPCallbacks::ExitFile:
    299     WriteLineInfo(CurLine, " 2", 2);
    300     break;
    301   case PPCallbacks::SystemHeaderPragma:
    302   case PPCallbacks::RenameFile:
    303     WriteLineInfo(CurLine);
    304     break;
    305   }
    306 }
    307 
    308 /// Ident - Handle #ident directives when read by the preprocessor.
    309 ///
    310 void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, const std::string &S) {
    311   MoveToLine(Loc);
    312 
    313   OS.write("#ident ", strlen("#ident "));
    314   OS.write(&S[0], S.size());
    315   EmittedTokensOnThisLine = true;
    316 }
    317 
    318 /// MacroDefined - This hook is called whenever a macro definition is seen.
    319 void PrintPPOutputPPCallbacks::MacroDefined(const Token &MacroNameTok,
    320                                             const MacroDirective *MD) {
    321   const MacroInfo *MI = MD->getInfo();
    322   // Only print out macro definitions in -dD mode.
    323   if (!DumpDefines ||
    324       // Ignore __FILE__ etc.
    325       MI->isBuiltinMacro()) return;
    326 
    327   MoveToLine(MI->getDefinitionLoc());
    328   PrintMacroDefinition(*MacroNameTok.getIdentifierInfo(), *MI, PP, OS);
    329   setEmittedDirectiveOnThisLine();
    330 }
    331 
    332 void PrintPPOutputPPCallbacks::MacroUndefined(const Token &MacroNameTok,
    333                                               const MacroDirective *MD) {
    334   // Only print out macro definitions in -dD mode.
    335   if (!DumpDefines) return;
    336 
    337   MoveToLine(MacroNameTok.getLocation());
    338   OS << "#undef " << MacroNameTok.getIdentifierInfo()->getName();
    339   setEmittedDirectiveOnThisLine();
    340 }
    341 
    342 void PrintPPOutputPPCallbacks::PragmaComment(SourceLocation Loc,
    343                                              const IdentifierInfo *Kind,
    344                                              const std::string &Str) {
    345   startNewLineIfNeeded();
    346   MoveToLine(Loc);
    347   OS << "#pragma comment(" << Kind->getName();
    348 
    349   if (!Str.empty()) {
    350     OS << ", \"";
    351 
    352     for (unsigned i = 0, e = Str.size(); i != e; ++i) {
    353       unsigned char Char = Str[i];
    354       if (isPrintable(Char) && Char != '\\' && Char != '"')
    355         OS << (char)Char;
    356       else  // Output anything hard as an octal escape.
    357         OS << '\\'
    358            << (char)('0'+ ((Char >> 6) & 7))
    359            << (char)('0'+ ((Char >> 3) & 7))
    360            << (char)('0'+ ((Char >> 0) & 7));
    361     }
    362     OS << '"';
    363   }
    364 
    365   OS << ')';
    366   setEmittedDirectiveOnThisLine();
    367 }
    368 
    369 void PrintPPOutputPPCallbacks::PragmaMessage(SourceLocation Loc,
    370                                              StringRef Str) {
    371   startNewLineIfNeeded();
    372   MoveToLine(Loc);
    373   OS << "#pragma message(";
    374 
    375   OS << '"';
    376 
    377   for (unsigned i = 0, e = Str.size(); i != e; ++i) {
    378     unsigned char Char = Str[i];
    379     if (isPrintable(Char) && Char != '\\' && Char != '"')
    380       OS << (char)Char;
    381     else  // Output anything hard as an octal escape.
    382       OS << '\\'
    383          << (char)('0'+ ((Char >> 6) & 7))
    384          << (char)('0'+ ((Char >> 3) & 7))
    385          << (char)('0'+ ((Char >> 0) & 7));
    386   }
    387   OS << '"';
    388 
    389   OS << ')';
    390   setEmittedDirectiveOnThisLine();
    391 }
    392 
    393 void PrintPPOutputPPCallbacks::
    394 PragmaDiagnosticPush(SourceLocation Loc, StringRef Namespace) {
    395   startNewLineIfNeeded();
    396   MoveToLine(Loc);
    397   OS << "#pragma " << Namespace << " diagnostic push";
    398   setEmittedDirectiveOnThisLine();
    399 }
    400 
    401 void PrintPPOutputPPCallbacks::
    402 PragmaDiagnosticPop(SourceLocation Loc, StringRef Namespace) {
    403   startNewLineIfNeeded();
    404   MoveToLine(Loc);
    405   OS << "#pragma " << Namespace << " diagnostic pop";
    406   setEmittedDirectiveOnThisLine();
    407 }
    408 
    409 void PrintPPOutputPPCallbacks::
    410 PragmaDiagnostic(SourceLocation Loc, StringRef Namespace,
    411                  diag::Mapping Map, StringRef Str) {
    412   startNewLineIfNeeded();
    413   MoveToLine(Loc);
    414   OS << "#pragma " << Namespace << " diagnostic ";
    415   switch (Map) {
    416   case diag::MAP_WARNING:
    417     OS << "warning";
    418     break;
    419   case diag::MAP_ERROR:
    420     OS << "error";
    421     break;
    422   case diag::MAP_IGNORE:
    423     OS << "ignored";
    424     break;
    425   case diag::MAP_FATAL:
    426     OS << "fatal";
    427     break;
    428   }
    429   OS << " \"" << Str << '"';
    430   setEmittedDirectiveOnThisLine();
    431 }
    432 
    433 /// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
    434 /// is called for the first token on each new line.  If this really is the start
    435 /// of a new logical line, handle it and return true, otherwise return false.
    436 /// This may not be the start of a logical line because the "start of line"
    437 /// marker is set for spelling lines, not expansion ones.
    438 bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
    439   // Figure out what line we went to and insert the appropriate number of
    440   // newline characters.
    441   if (!MoveToLine(Tok.getLocation()))
    442     return false;
    443 
    444   // Print out space characters so that the first token on a line is
    445   // indented for easy reading.
    446   unsigned ColNo = SM.getExpansionColumnNumber(Tok.getLocation());
    447 
    448   // This hack prevents stuff like:
    449   // #define HASH #
    450   // HASH define foo bar
    451   // From having the # character end up at column 1, which makes it so it
    452   // is not handled as a #define next time through the preprocessor if in
    453   // -fpreprocessed mode.
    454   if (ColNo <= 1 && Tok.is(tok::hash))
    455     OS << ' ';
    456 
    457   // Otherwise, indent the appropriate number of spaces.
    458   for (; ColNo > 1; --ColNo)
    459     OS << ' ';
    460 
    461   return true;
    462 }
    463 
    464 void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
    465                                                      unsigned Len) {
    466   unsigned NumNewlines = 0;
    467   for (; Len; --Len, ++TokStr) {
    468     if (*TokStr != '\n' &&
    469         *TokStr != '\r')
    470       continue;
    471 
    472     ++NumNewlines;
    473 
    474     // If we have \n\r or \r\n, skip both and count as one line.
    475     if (Len != 1 &&
    476         (TokStr[1] == '\n' || TokStr[1] == '\r') &&
    477         TokStr[0] != TokStr[1])
    478       ++TokStr, --Len;
    479   }
    480 
    481   if (NumNewlines == 0) return;
    482 
    483   CurLine += NumNewlines;
    484 }
    485 
    486 
    487 namespace {
    488 struct UnknownPragmaHandler : public PragmaHandler {
    489   const char *Prefix;
    490   PrintPPOutputPPCallbacks *Callbacks;
    491 
    492   UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks)
    493     : Prefix(prefix), Callbacks(callbacks) {}
    494   virtual void HandlePragma(Preprocessor &PP, PragmaIntroducerKind Introducer,
    495                             Token &PragmaTok) {
    496     // Figure out what line we went to and insert the appropriate number of
    497     // newline characters.
    498     Callbacks->startNewLineIfNeeded();
    499     Callbacks->MoveToLine(PragmaTok.getLocation());
    500     Callbacks->OS.write(Prefix, strlen(Prefix));
    501     // Read and print all of the pragma tokens.
    502     while (PragmaTok.isNot(tok::eod)) {
    503       if (PragmaTok.hasLeadingSpace())
    504         Callbacks->OS << ' ';
    505       std::string TokSpell = PP.getSpelling(PragmaTok);
    506       Callbacks->OS.write(&TokSpell[0], TokSpell.size());
    507       PP.LexUnexpandedToken(PragmaTok);
    508     }
    509     Callbacks->setEmittedDirectiveOnThisLine();
    510   }
    511 };
    512 } // end anonymous namespace
    513 
    514 
    515 static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
    516                                     PrintPPOutputPPCallbacks *Callbacks,
    517                                     raw_ostream &OS) {
    518   bool DropComments = PP.getLangOpts().TraditionalCPP &&
    519                       !PP.getCommentRetentionState();
    520 
    521   char Buffer[256];
    522   Token PrevPrevTok, PrevTok;
    523   PrevPrevTok.startToken();
    524   PrevTok.startToken();
    525   while (1) {
    526     if (Callbacks->hasEmittedDirectiveOnThisLine()) {
    527       Callbacks->startNewLineIfNeeded();
    528       Callbacks->MoveToLine(Tok.getLocation());
    529     }
    530 
    531     // If this token is at the start of a line, emit newlines if needed.
    532     if (Tok.isAtStartOfLine() && Callbacks->HandleFirstTokOnLine(Tok)) {
    533       // done.
    534     } else if (Tok.hasLeadingSpace() ||
    535                // If we haven't emitted a token on this line yet, PrevTok isn't
    536                // useful to look at and no concatenation could happen anyway.
    537                (Callbacks->hasEmittedTokensOnThisLine() &&
    538                 // Don't print "-" next to "-", it would form "--".
    539                 Callbacks->AvoidConcat(PrevPrevTok, PrevTok, Tok))) {
    540       OS << ' ';
    541     }
    542 
    543     if (DropComments && Tok.is(tok::comment)) {
    544       // Skip comments. Normally the preprocessor does not generate
    545       // tok::comment nodes at all when not keeping comments, but under
    546       // -traditional-cpp the lexer keeps /all/ whitespace, including comments.
    547       SourceLocation StartLoc = Tok.getLocation();
    548       Callbacks->MoveToLine(StartLoc.getLocWithOffset(Tok.getLength()));
    549     } else if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
    550       OS << II->getName();
    551     } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
    552                Tok.getLiteralData()) {
    553       OS.write(Tok.getLiteralData(), Tok.getLength());
    554     } else if (Tok.getLength() < 256) {
    555       const char *TokPtr = Buffer;
    556       unsigned Len = PP.getSpelling(Tok, TokPtr);
    557       OS.write(TokPtr, Len);
    558 
    559       // Tokens that can contain embedded newlines need to adjust our current
    560       // line number.
    561       if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
    562         Callbacks->HandleNewlinesInToken(TokPtr, Len);
    563     } else {
    564       std::string S = PP.getSpelling(Tok);
    565       OS.write(&S[0], S.size());
    566 
    567       // Tokens that can contain embedded newlines need to adjust our current
    568       // line number.
    569       if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
    570         Callbacks->HandleNewlinesInToken(&S[0], S.size());
    571     }
    572     Callbacks->setEmittedTokensOnThisLine();
    573 
    574     if (Tok.is(tok::eof)) break;
    575 
    576     PrevPrevTok = PrevTok;
    577     PrevTok = Tok;
    578     PP.Lex(Tok);
    579   }
    580 }
    581 
    582 typedef std::pair<const IdentifierInfo *, MacroInfo *> id_macro_pair;
    583 static int MacroIDCompare(const void* a, const void* b) {
    584   const id_macro_pair *LHS = static_cast<const id_macro_pair*>(a);
    585   const id_macro_pair *RHS = static_cast<const id_macro_pair*>(b);
    586   return LHS->first->getName().compare(RHS->first->getName());
    587 }
    588 
    589 static void DoPrintMacros(Preprocessor &PP, raw_ostream *OS) {
    590   // Ignore unknown pragmas.
    591   PP.AddPragmaHandler(new EmptyPragmaHandler());
    592 
    593   // -dM mode just scans and ignores all tokens in the files, then dumps out
    594   // the macro table at the end.
    595   PP.EnterMainSourceFile();
    596 
    597   Token Tok;
    598   do PP.Lex(Tok);
    599   while (Tok.isNot(tok::eof));
    600 
    601   SmallVector<id_macro_pair, 128> MacrosByID;
    602   for (Preprocessor::macro_iterator I = PP.macro_begin(), E = PP.macro_end();
    603        I != E; ++I) {
    604     if (I->first->hasMacroDefinition())
    605       MacrosByID.push_back(id_macro_pair(I->first, I->second->getInfo()));
    606   }
    607   llvm::array_pod_sort(MacrosByID.begin(), MacrosByID.end(), MacroIDCompare);
    608 
    609   for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
    610     MacroInfo &MI = *MacrosByID[i].second;
    611     // Ignore computed macros like __LINE__ and friends.
    612     if (MI.isBuiltinMacro()) continue;
    613 
    614     PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
    615     *OS << '\n';
    616   }
    617 }
    618 
    619 /// DoPrintPreprocessedInput - This implements -E mode.
    620 ///
    621 void clang::DoPrintPreprocessedInput(Preprocessor &PP, raw_ostream *OS,
    622                                      const PreprocessorOutputOptions &Opts) {
    623   // Show macros with no output is handled specially.
    624   if (!Opts.ShowCPP) {
    625     assert(Opts.ShowMacros && "Not yet implemented!");
    626     DoPrintMacros(PP, OS);
    627     return;
    628   }
    629 
    630   // Inform the preprocessor whether we want it to retain comments or not, due
    631   // to -C or -CC.
    632   PP.SetCommentRetentionState(Opts.ShowComments, Opts.ShowMacroComments);
    633 
    634   PrintPPOutputPPCallbacks *Callbacks =
    635       new PrintPPOutputPPCallbacks(PP, *OS, !Opts.ShowLineMarkers,
    636                                    Opts.ShowMacros);
    637   PP.AddPragmaHandler(new UnknownPragmaHandler("#pragma", Callbacks));
    638   PP.AddPragmaHandler("GCC", new UnknownPragmaHandler("#pragma GCC",Callbacks));
    639   PP.AddPragmaHandler("clang",
    640                       new UnknownPragmaHandler("#pragma clang", Callbacks));
    641 
    642   PP.addPPCallbacks(Callbacks);
    643 
    644   // After we have configured the preprocessor, enter the main file.
    645   PP.EnterMainSourceFile();
    646 
    647   // Consume all of the tokens that come from the predefines buffer.  Those
    648   // should not be emitted into the output and are guaranteed to be at the
    649   // start.
    650   const SourceManager &SourceMgr = PP.getSourceManager();
    651   Token Tok;
    652   do {
    653     PP.Lex(Tok);
    654     if (Tok.is(tok::eof) || !Tok.getLocation().isFileID())
    655       break;
    656 
    657     PresumedLoc PLoc = SourceMgr.getPresumedLoc(Tok.getLocation());
    658     if (PLoc.isInvalid())
    659       break;
    660 
    661     if (strcmp(PLoc.getFilename(), "<built-in>"))
    662       break;
    663   } while (true);
    664 
    665   // Read all the preprocessed tokens, printing them out to the stream.
    666   PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
    667   *OS << '\n';
    668 }
    669