PrintPreprocessedOutput.cpp revision 076eea20b80024fc63bbd71fb019375983680ea6 
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>
31using namespace clang;
32
33/// PrintMacroDefinition - Print a macro definition in a form that will be
34/// properly accepted back as a definition.
35static 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
80namespace {
81class PrintPPOutputPPCallbacks : public PPCallbacks {
82  Preprocessor &PP;
83  SourceManager &SM;
84  TokenConcatenation ConcatInfo;
85public:
86  raw_ostream &OS;
87private:
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;
99public:
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 InclusionDirective(SourceLocation HashLoc,
131                                  const Token &IncludeTok,
132                                  StringRef FileName,
133                                  bool IsAngled,
134                                  CharSourceRange FilenameRange,
135                                  const FileEntry *File,
136                                  StringRef SearchPath,
137                                  StringRef RelativePath,
138                                  const Module *Imported);
139  virtual void Ident(SourceLocation Loc, const std::string &str);
140  virtual void PragmaCaptured(SourceLocation Loc, StringRef Str);
141  virtual void PragmaComment(SourceLocation Loc, const IdentifierInfo *Kind,
142                             const std::string &Str);
143  virtual void PragmaMessage(SourceLocation Loc, StringRef Namespace,
144                             PragmaMessageKind Kind, StringRef Str);
145  virtual void PragmaDebug(SourceLocation Loc, StringRef DebugType);
146  virtual void PragmaDiagnosticPush(SourceLocation Loc,
147                                    StringRef Namespace);
148  virtual void PragmaDiagnosticPop(SourceLocation Loc,
149                                   StringRef Namespace);
150  virtual void PragmaDiagnostic(SourceLocation Loc, StringRef Namespace,
151                                diag::Mapping Map, StringRef Str);
152
153  bool HandleFirstTokOnLine(Token &Tok);
154
155  /// Move to the line of the provided source location. This will
156  /// return true if the output stream required adjustment or if
157  /// the requested location is on the first line.
158  bool MoveToLine(SourceLocation Loc) {
159    PresumedLoc PLoc = SM.getPresumedLoc(Loc);
160    if (PLoc.isInvalid())
161      return false;
162    return MoveToLine(PLoc.getLine()) || (PLoc.getLine() == 1);
163  }
164  bool MoveToLine(unsigned LineNo);
165
166  bool AvoidConcat(const Token &PrevPrevTok, const Token &PrevTok,
167                   const Token &Tok) {
168    return ConcatInfo.AvoidConcat(PrevPrevTok, PrevTok, Tok);
169  }
170  void WriteLineInfo(unsigned LineNo, const char *Extra=0, unsigned ExtraLen=0);
171  bool LineMarkersAreDisabled() const { return DisableLineMarkers; }
172  void HandleNewlinesInToken(const char *TokStr, unsigned Len);
173
174  /// MacroDefined - This hook is called whenever a macro definition is seen.
175  void MacroDefined(const Token &MacroNameTok, const MacroDirective *MD);
176
177  /// MacroUndefined - This hook is called whenever a macro #undef is seen.
178  void MacroUndefined(const Token &MacroNameTok, const MacroDirective *MD);
179};
180}  // end anonymous namespace
181
182void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
183                                             const char *Extra,
184                                             unsigned ExtraLen) {
185  startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
186
187  // Emit #line directives or GNU line markers depending on what mode we're in.
188  if (UseLineDirective) {
189    OS << "#line" << ' ' << LineNo << ' ' << '"';
190    OS.write(CurFilename.data(), CurFilename.size());
191    OS << '"';
192  } else {
193    OS << '#' << ' ' << LineNo << ' ' << '"';
194    OS.write(CurFilename.data(), CurFilename.size());
195    OS << '"';
196
197    if (ExtraLen)
198      OS.write(Extra, ExtraLen);
199
200    if (FileType == SrcMgr::C_System)
201      OS.write(" 3", 2);
202    else if (FileType == SrcMgr::C_ExternCSystem)
203      OS.write(" 3 4", 4);
204  }
205  OS << '\n';
206}
207
208/// MoveToLine - Move the output to the source line specified by the location
209/// object.  We can do this by emitting some number of \n's, or be emitting a
210/// #line directive.  This returns false if already at the specified line, true
211/// if some newlines were emitted.
212bool PrintPPOutputPPCallbacks::MoveToLine(unsigned LineNo) {
213  // If this line is "close enough" to the original line, just print newlines,
214  // otherwise print a #line directive.
215  if (LineNo-CurLine <= 8) {
216    if (LineNo-CurLine == 1)
217      OS << '\n';
218    else if (LineNo == CurLine)
219      return false;    // Spelling line moved, but expansion line didn't.
220    else {
221      const char *NewLines = "\n\n\n\n\n\n\n\n";
222      OS.write(NewLines, LineNo-CurLine);
223    }
224  } else if (!DisableLineMarkers) {
225    // Emit a #line or line marker.
226    WriteLineInfo(LineNo, 0, 0);
227  } else {
228    // Okay, we're in -P mode, which turns off line markers.  However, we still
229    // need to emit a newline between tokens on different lines.
230    startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
231  }
232
233  CurLine = LineNo;
234  return true;
235}
236
237bool
238PrintPPOutputPPCallbacks::startNewLineIfNeeded(bool ShouldUpdateCurrentLine) {
239  if (EmittedTokensOnThisLine || EmittedDirectiveOnThisLine) {
240    OS << '\n';
241    EmittedTokensOnThisLine = false;
242    EmittedDirectiveOnThisLine = false;
243    if (ShouldUpdateCurrentLine)
244      ++CurLine;
245    return true;
246  }
247
248  return false;
249}
250
251/// FileChanged - Whenever the preprocessor enters or exits a #include file
252/// it invokes this handler.  Update our conception of the current source
253/// position.
254void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
255                                           FileChangeReason Reason,
256                                       SrcMgr::CharacteristicKind NewFileType,
257                                       FileID PrevFID) {
258  // Unless we are exiting a #include, make sure to skip ahead to the line the
259  // #include directive was at.
260  SourceManager &SourceMgr = SM;
261
262  PresumedLoc UserLoc = SourceMgr.getPresumedLoc(Loc);
263  if (UserLoc.isInvalid())
264    return;
265
266  unsigned NewLine = UserLoc.getLine();
267
268  if (Reason == PPCallbacks::EnterFile) {
269    SourceLocation IncludeLoc = UserLoc.getIncludeLoc();
270    if (IncludeLoc.isValid())
271      MoveToLine(IncludeLoc);
272  } else if (Reason == PPCallbacks::SystemHeaderPragma) {
273    MoveToLine(NewLine);
274
275    // TODO GCC emits the # directive for this directive on the line AFTER the
276    // directive and emits a bunch of spaces that aren't needed.  Emulate this
277    // strange behavior.
278  }
279
280  CurLine = NewLine;
281
282  CurFilename.clear();
283  CurFilename += UserLoc.getFilename();
284  Lexer::Stringify(CurFilename);
285  FileType = NewFileType;
286
287  if (DisableLineMarkers) {
288    startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
289    return;
290  }
291
292  if (!Initialized) {
293    WriteLineInfo(CurLine);
294    Initialized = true;
295  }
296
297  // Do not emit an enter marker for the main file (which we expect is the first
298  // entered file). This matches gcc, and improves compatibility with some tools
299  // which track the # line markers as a way to determine when the preprocessed
300  // output is in the context of the main file.
301  if (Reason == PPCallbacks::EnterFile && !IsFirstFileEntered) {
302    IsFirstFileEntered = true;
303    return;
304  }
305
306  switch (Reason) {
307  case PPCallbacks::EnterFile:
308    WriteLineInfo(CurLine, " 1", 2);
309    break;
310  case PPCallbacks::ExitFile:
311    WriteLineInfo(CurLine, " 2", 2);
312    break;
313  case PPCallbacks::SystemHeaderPragma:
314  case PPCallbacks::RenameFile:
315    WriteLineInfo(CurLine);
316    break;
317  }
318}
319
320void PrintPPOutputPPCallbacks::InclusionDirective(SourceLocation HashLoc,
321                                                  const Token &IncludeTok,
322                                                  StringRef FileName,
323                                                  bool IsAngled,
324                                                  CharSourceRange FilenameRange,
325                                                  const FileEntry *File,
326                                                  StringRef SearchPath,
327                                                  StringRef RelativePath,
328                                                  const Module *Imported) {
329  // When preprocessing, turn implicit imports into @imports.
330  // FIXME: This is a stop-gap until a more comprehensive "preprocessing with
331  // modules" solution is introduced.
332  if (Imported) {
333    startNewLineIfNeeded();
334    MoveToLine(HashLoc);
335    OS << "@import " << Imported->getFullModuleName() << ";"
336       << " /* clang -E: implicit import for \"" << File->getName() << "\" */";
337  }
338}
339
340/// Ident - Handle #ident directives when read by the preprocessor.
341///
342void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, const std::string &S) {
343  MoveToLine(Loc);
344
345  OS.write("#ident ", strlen("#ident "));
346  OS.write(&S[0], S.size());
347  EmittedTokensOnThisLine = true;
348}
349
350void PrintPPOutputPPCallbacks::PragmaCaptured(SourceLocation Loc,
351                                              StringRef Str) {
352  startNewLineIfNeeded();
353  MoveToLine(Loc);
354  OS << "#pragma captured";
355
356  setEmittedDirectiveOnThisLine();
357}
358
359/// MacroDefined - This hook is called whenever a macro definition is seen.
360void PrintPPOutputPPCallbacks::MacroDefined(const Token &MacroNameTok,
361                                            const MacroDirective *MD) {
362  const MacroInfo *MI = MD->getMacroInfo();
363  // Only print out macro definitions in -dD mode.
364  if (!DumpDefines ||
365      // Ignore __FILE__ etc.
366      MI->isBuiltinMacro()) return;
367
368  MoveToLine(MI->getDefinitionLoc());
369  PrintMacroDefinition(*MacroNameTok.getIdentifierInfo(), *MI, PP, OS);
370  setEmittedDirectiveOnThisLine();
371}
372
373void PrintPPOutputPPCallbacks::MacroUndefined(const Token &MacroNameTok,
374                                              const MacroDirective *MD) {
375  // Only print out macro definitions in -dD mode.
376  if (!DumpDefines) return;
377
378  MoveToLine(MacroNameTok.getLocation());
379  OS << "#undef " << MacroNameTok.getIdentifierInfo()->getName();
380  setEmittedDirectiveOnThisLine();
381}
382
383void PrintPPOutputPPCallbacks::PragmaComment(SourceLocation Loc,
384                                             const IdentifierInfo *Kind,
385                                             const std::string &Str) {
386  startNewLineIfNeeded();
387  MoveToLine(Loc);
388  OS << "#pragma comment(" << Kind->getName();
389
390  if (!Str.empty()) {
391    OS << ", \"";
392
393    for (unsigned i = 0, e = Str.size(); i != e; ++i) {
394      unsigned char Char = Str[i];
395      if (isPrintable(Char) && Char != '\\' && Char != '"')
396        OS << (char)Char;
397      else  // Output anything hard as an octal escape.
398        OS << '\\'
399           << (char)('0'+ ((Char >> 6) & 7))
400           << (char)('0'+ ((Char >> 3) & 7))
401           << (char)('0'+ ((Char >> 0) & 7));
402    }
403    OS << '"';
404  }
405
406  OS << ')';
407  setEmittedDirectiveOnThisLine();
408}
409
410void PrintPPOutputPPCallbacks::PragmaMessage(SourceLocation Loc,
411                                             StringRef Namespace,
412                                             PragmaMessageKind Kind,
413                                             StringRef Str) {
414  startNewLineIfNeeded();
415  MoveToLine(Loc);
416  OS << "#pragma ";
417  if (!Namespace.empty())
418    OS << Namespace << ' ';
419  switch (Kind) {
420    case PMK_Message:
421      OS << "message(\"";
422      break;
423    case PMK_Warning:
424      OS << "warning(\"";
425      break;
426    case PMK_Error:
427      OS << "error(\"";
428      break;
429  }
430
431  for (unsigned i = 0, e = Str.size(); i != e; ++i) {
432    unsigned char Char = Str[i];
433    if (isPrintable(Char) && Char != '\\' && Char != '"')
434      OS << (char)Char;
435    else  // Output anything hard as an octal escape.
436      OS << '\\'
437         << (char)('0'+ ((Char >> 6) & 7))
438         << (char)('0'+ ((Char >> 3) & 7))
439         << (char)('0'+ ((Char >> 0) & 7));
440  }
441  OS << '"';
442
443  OS << ')';
444  setEmittedDirectiveOnThisLine();
445}
446
447void PrintPPOutputPPCallbacks::PragmaDebug(SourceLocation Loc,
448                                           StringRef DebugType) {
449  startNewLineIfNeeded();
450  MoveToLine(Loc);
451
452  OS << "#pragma clang __debug ";
453  OS << DebugType;
454
455  setEmittedDirectiveOnThisLine();
456}
457
458void PrintPPOutputPPCallbacks::
459PragmaDiagnosticPush(SourceLocation Loc, StringRef Namespace) {
460  startNewLineIfNeeded();
461  MoveToLine(Loc);
462  OS << "#pragma " << Namespace << " diagnostic push";
463  setEmittedDirectiveOnThisLine();
464}
465
466void PrintPPOutputPPCallbacks::
467PragmaDiagnosticPop(SourceLocation Loc, StringRef Namespace) {
468  startNewLineIfNeeded();
469  MoveToLine(Loc);
470  OS << "#pragma " << Namespace << " diagnostic pop";
471  setEmittedDirectiveOnThisLine();
472}
473
474void PrintPPOutputPPCallbacks::
475PragmaDiagnostic(SourceLocation Loc, StringRef Namespace,
476                 diag::Mapping Map, StringRef Str) {
477  startNewLineIfNeeded();
478  MoveToLine(Loc);
479  OS << "#pragma " << Namespace << " diagnostic ";
480  switch (Map) {
481  case diag::MAP_WARNING:
482    OS << "warning";
483    break;
484  case diag::MAP_ERROR:
485    OS << "error";
486    break;
487  case diag::MAP_IGNORE:
488    OS << "ignored";
489    break;
490  case diag::MAP_FATAL:
491    OS << "fatal";
492    break;
493  }
494  OS << " \"" << Str << '"';
495  setEmittedDirectiveOnThisLine();
496}
497
498/// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
499/// is called for the first token on each new line.  If this really is the start
500/// of a new logical line, handle it and return true, otherwise return false.
501/// This may not be the start of a logical line because the "start of line"
502/// marker is set for spelling lines, not expansion ones.
503bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
504  // Figure out what line we went to and insert the appropriate number of
505  // newline characters.
506  if (!MoveToLine(Tok.getLocation()))
507    return false;
508
509  // Print out space characters so that the first token on a line is
510  // indented for easy reading.
511  unsigned ColNo = SM.getExpansionColumnNumber(Tok.getLocation());
512
513  // This hack prevents stuff like:
514  // #define HASH #
515  // HASH define foo bar
516  // From having the # character end up at column 1, which makes it so it
517  // is not handled as a #define next time through the preprocessor if in
518  // -fpreprocessed mode.
519  if (ColNo <= 1 && Tok.is(tok::hash))
520    OS << ' ';
521
522  // Otherwise, indent the appropriate number of spaces.
523  for (; ColNo > 1; --ColNo)
524    OS << ' ';
525
526  return true;
527}
528
529void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
530                                                     unsigned Len) {
531  unsigned NumNewlines = 0;
532  for (; Len; --Len, ++TokStr) {
533    if (*TokStr != '\n' &&
534        *TokStr != '\r')
535      continue;
536
537    ++NumNewlines;
538
539    // If we have \n\r or \r\n, skip both and count as one line.
540    if (Len != 1 &&
541        (TokStr[1] == '\n' || TokStr[1] == '\r') &&
542        TokStr[0] != TokStr[1])
543      ++TokStr, --Len;
544  }
545
546  if (NumNewlines == 0) return;
547
548  CurLine += NumNewlines;
549}
550
551
552namespace {
553struct UnknownPragmaHandler : public PragmaHandler {
554  const char *Prefix;
555  PrintPPOutputPPCallbacks *Callbacks;
556
557  UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks)
558    : Prefix(prefix), Callbacks(callbacks) {}
559  virtual void HandlePragma(Preprocessor &PP, PragmaIntroducerKind Introducer,
560                            Token &PragmaTok) {
561    // Figure out what line we went to and insert the appropriate number of
562    // newline characters.
563    Callbacks->startNewLineIfNeeded();
564    Callbacks->MoveToLine(PragmaTok.getLocation());
565    Callbacks->OS.write(Prefix, strlen(Prefix));
566    // Read and print all of the pragma tokens.
567    while (PragmaTok.isNot(tok::eod)) {
568      if (PragmaTok.hasLeadingSpace())
569        Callbacks->OS << ' ';
570      std::string TokSpell = PP.getSpelling(PragmaTok);
571      Callbacks->OS.write(&TokSpell[0], TokSpell.size());
572      PP.LexUnexpandedToken(PragmaTok);
573    }
574    Callbacks->setEmittedDirectiveOnThisLine();
575  }
576};
577} // end anonymous namespace
578
579
580static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
581                                    PrintPPOutputPPCallbacks *Callbacks,
582                                    raw_ostream &OS) {
583  bool DropComments = PP.getLangOpts().TraditionalCPP &&
584                      !PP.getCommentRetentionState();
585
586  char Buffer[256];
587  Token PrevPrevTok, PrevTok;
588  PrevPrevTok.startToken();
589  PrevTok.startToken();
590  while (1) {
591    if (Callbacks->hasEmittedDirectiveOnThisLine()) {
592      Callbacks->startNewLineIfNeeded();
593      Callbacks->MoveToLine(Tok.getLocation());
594    }
595
596    // If this token is at the start of a line, emit newlines if needed.
597    if (Tok.isAtStartOfLine() && Callbacks->HandleFirstTokOnLine(Tok)) {
598      // done.
599    } else if (Tok.hasLeadingSpace() ||
600               // If we haven't emitted a token on this line yet, PrevTok isn't
601               // useful to look at and no concatenation could happen anyway.
602               (Callbacks->hasEmittedTokensOnThisLine() &&
603                // Don't print "-" next to "-", it would form "--".
604                Callbacks->AvoidConcat(PrevPrevTok, PrevTok, Tok))) {
605      OS << ' ';
606    }
607
608    if (DropComments && Tok.is(tok::comment)) {
609      // Skip comments. Normally the preprocessor does not generate
610      // tok::comment nodes at all when not keeping comments, but under
611      // -traditional-cpp the lexer keeps /all/ whitespace, including comments.
612      SourceLocation StartLoc = Tok.getLocation();
613      Callbacks->MoveToLine(StartLoc.getLocWithOffset(Tok.getLength()));
614    } else if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
615      OS << II->getName();
616    } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
617               Tok.getLiteralData()) {
618      OS.write(Tok.getLiteralData(), Tok.getLength());
619    } else if (Tok.getLength() < 256) {
620      const char *TokPtr = Buffer;
621      unsigned Len = PP.getSpelling(Tok, TokPtr);
622      OS.write(TokPtr, Len);
623
624      // Tokens that can contain embedded newlines need to adjust our current
625      // line number.
626      if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
627        Callbacks->HandleNewlinesInToken(TokPtr, Len);
628    } else {
629      std::string S = PP.getSpelling(Tok);
630      OS.write(&S[0], S.size());
631
632      // Tokens that can contain embedded newlines need to adjust our current
633      // line number.
634      if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
635        Callbacks->HandleNewlinesInToken(&S[0], S.size());
636    }
637    Callbacks->setEmittedTokensOnThisLine();
638
639    if (Tok.is(tok::eof)) break;
640
641    PrevPrevTok = PrevTok;
642    PrevTok = Tok;
643    PP.Lex(Tok);
644  }
645}
646
647typedef std::pair<const IdentifierInfo *, MacroInfo *> id_macro_pair;
648static int MacroIDCompare(const void* a, const void* b) {
649  const id_macro_pair *LHS = static_cast<const id_macro_pair*>(a);
650  const id_macro_pair *RHS = static_cast<const id_macro_pair*>(b);
651  return LHS->first->getName().compare(RHS->first->getName());
652}
653
654static void DoPrintMacros(Preprocessor &PP, raw_ostream *OS) {
655  // Ignore unknown pragmas.
656  PP.AddPragmaHandler(new EmptyPragmaHandler());
657
658  // -dM mode just scans and ignores all tokens in the files, then dumps out
659  // the macro table at the end.
660  PP.EnterMainSourceFile();
661
662  Token Tok;
663  do PP.Lex(Tok);
664  while (Tok.isNot(tok::eof));
665
666  SmallVector<id_macro_pair, 128> MacrosByID;
667  for (Preprocessor::macro_iterator I = PP.macro_begin(), E = PP.macro_end();
668       I != E; ++I) {
669    if (I->first->hasMacroDefinition())
670      MacrosByID.push_back(id_macro_pair(I->first, I->second->getMacroInfo()));
671  }
672  llvm::array_pod_sort(MacrosByID.begin(), MacrosByID.end(), MacroIDCompare);
673
674  for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
675    MacroInfo &MI = *MacrosByID[i].second;
676    // Ignore computed macros like __LINE__ and friends.
677    if (MI.isBuiltinMacro()) continue;
678
679    PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
680    *OS << '\n';
681  }
682}
683
684/// DoPrintPreprocessedInput - This implements -E mode.
685///
686void clang::DoPrintPreprocessedInput(Preprocessor &PP, raw_ostream *OS,
687                                     const PreprocessorOutputOptions &Opts) {
688  // Show macros with no output is handled specially.
689  if (!Opts.ShowCPP) {
690    assert(Opts.ShowMacros && "Not yet implemented!");
691    DoPrintMacros(PP, OS);
692    return;
693  }
694
695  // Inform the preprocessor whether we want it to retain comments or not, due
696  // to -C or -CC.
697  PP.SetCommentRetentionState(Opts.ShowComments, Opts.ShowMacroComments);
698
699  PrintPPOutputPPCallbacks *Callbacks =
700      new PrintPPOutputPPCallbacks(PP, *OS, !Opts.ShowLineMarkers,
701                                   Opts.ShowMacros);
702  PP.AddPragmaHandler(new UnknownPragmaHandler("#pragma", Callbacks));
703  PP.AddPragmaHandler("GCC", new UnknownPragmaHandler("#pragma GCC",Callbacks));
704  PP.AddPragmaHandler("clang",
705                      new UnknownPragmaHandler("#pragma clang", Callbacks));
706
707  PP.addPPCallbacks(Callbacks);
708
709  // After we have configured the preprocessor, enter the main file.
710  PP.EnterMainSourceFile();
711
712  // Consume all of the tokens that come from the predefines buffer.  Those
713  // should not be emitted into the output and are guaranteed to be at the
714  // start.
715  const SourceManager &SourceMgr = PP.getSourceManager();
716  Token Tok;
717  do {
718    PP.Lex(Tok);
719    if (Tok.is(tok::eof) || !Tok.getLocation().isFileID())
720      break;
721
722    PresumedLoc PLoc = SourceMgr.getPresumedLoc(Tok.getLocation());
723    if (PLoc.isInvalid())
724      break;
725
726    if (strcmp(PLoc.getFilename(), "<built-in>"))
727      break;
728  } while (true);
729
730  // Read all the preprocessed tokens, printing them out to the stream.
731  PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
732  *OS << '\n';
733}
734