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 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 PragmaMessage(SourceLocation Loc, StringRef Namespace,
142                              PragmaMessageKind Kind, StringRef Str);
143   virtual void PragmaDebug(SourceLocation Loc, StringRef DebugType);
144   virtual void PragmaDiagnosticPush(SourceLocation Loc,
145                                     StringRef Namespace);
146   virtual void PragmaDiagnosticPop(SourceLocation Loc,
147                                    StringRef Namespace);
148   virtual void PragmaDiagnostic(SourceLocation Loc, StringRef Namespace,
149                                 diag::Mapping Map, StringRef Str);
150   virtual void PragmaWarning(SourceLocation Loc, StringRef WarningSpec,
151                              ArrayRef<int> Ids);
152   virtual void PragmaWarningPush(SourceLocation Loc, int Level);
153   virtual void PragmaWarningPop(SourceLocation Loc);
154 
155   bool HandleFirstTokOnLine(Token &Tok);
156 
157   /// Move to the line of the provided source location. This will
158   /// return true if the output stream required adjustment or if
159   /// the requested location is on the first line.
160   bool MoveToLine(SourceLocation Loc) {
161     PresumedLoc PLoc = SM.getPresumedLoc(Loc);
162     if (PLoc.isInvalid())
163       return false;
164     return MoveToLine(PLoc.getLine()) || (PLoc.getLine() == 1);
165   }
166   bool MoveToLine(unsigned LineNo);
167 
168   bool AvoidConcat(const Token &PrevPrevTok, const Token &PrevTok,
169                    const Token &Tok) {
170     return ConcatInfo.AvoidConcat(PrevPrevTok, PrevTok, Tok);
171   }
172   void WriteLineInfo(unsigned LineNo, const char *Extra=0, unsigned ExtraLen=0);
173   bool LineMarkersAreDisabled() const { return DisableLineMarkers; }
174   void HandleNewlinesInToken(const char *TokStr, unsigned Len);
175 
176   /// MacroDefined - This hook is called whenever a macro definition is seen.
177   void MacroDefined(const Token &MacroNameTok, const MacroDirective *MD);
178 
179   /// MacroUndefined - This hook is called whenever a macro #undef is seen.
180   void MacroUndefined(const Token &MacroNameTok, const MacroDirective *MD);
181 };
182 }  // end anonymous namespace
183 
184 void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
185                                              const char *Extra,
186                                              unsigned ExtraLen) {
187   startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
188 
189   // Emit #line directives or GNU line markers depending on what mode we're in.
190   if (UseLineDirective) {
191     OS << "#line" << ' ' << LineNo << ' ' << '"';
192     OS.write_escaped(CurFilename);
193     OS << '"';
194   } else {
195     OS << '#' << ' ' << LineNo << ' ' << '"';
196     OS.write_escaped(CurFilename);
197     OS << '"';
198 
199     if (ExtraLen)
200       OS.write(Extra, ExtraLen);
201 
202     if (FileType == SrcMgr::C_System)
203       OS.write(" 3", 2);
204     else if (FileType == SrcMgr::C_ExternCSystem)
205       OS.write(" 3 4", 4);
206   }
207   OS << '\n';
208 }
209 
210 /// MoveToLine - Move the output to the source line specified by the location
211 /// object.  We can do this by emitting some number of \n's, or be emitting a
212 /// #line directive.  This returns false if already at the specified line, true
213 /// if some newlines were emitted.
214 bool PrintPPOutputPPCallbacks::MoveToLine(unsigned LineNo) {
215   // If this line is "close enough" to the original line, just print newlines,
216   // otherwise print a #line directive.
217   if (LineNo-CurLine <= 8) {
218     if (LineNo-CurLine == 1)
219       OS << '\n';
220     else if (LineNo == CurLine)
221       return false;    // Spelling line moved, but expansion line didn't.
222     else {
223       const char *NewLines = "\n\n\n\n\n\n\n\n";
224       OS.write(NewLines, LineNo-CurLine);
225     }
226   } else if (!DisableLineMarkers) {
227     // Emit a #line or line marker.
228     WriteLineInfo(LineNo, 0, 0);
229   } else {
230     // Okay, we're in -P mode, which turns off line markers.  However, we still
231     // need to emit a newline between tokens on different lines.
232     startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
233   }
234 
235   CurLine = LineNo;
236   return true;
237 }
238 
239 bool
240 PrintPPOutputPPCallbacks::startNewLineIfNeeded(bool ShouldUpdateCurrentLine) {
241   if (EmittedTokensOnThisLine || EmittedDirectiveOnThisLine) {
242     OS << '\n';
243     EmittedTokensOnThisLine = false;
244     EmittedDirectiveOnThisLine = false;
245     if (ShouldUpdateCurrentLine)
246       ++CurLine;
247     return true;
248   }
249 
250   return false;
251 }
252 
253 /// FileChanged - Whenever the preprocessor enters or exits a #include file
254 /// it invokes this handler.  Update our conception of the current source
255 /// position.
256 void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
257                                            FileChangeReason Reason,
258                                        SrcMgr::CharacteristicKind NewFileType,
259                                        FileID PrevFID) {
260   // Unless we are exiting a #include, make sure to skip ahead to the line the
261   // #include directive was at.
262   SourceManager &SourceMgr = SM;
263 
264   PresumedLoc UserLoc = SourceMgr.getPresumedLoc(Loc);
265   if (UserLoc.isInvalid())
266     return;
267 
268   unsigned NewLine = UserLoc.getLine();
269 
270   if (Reason == PPCallbacks::EnterFile) {
271     SourceLocation IncludeLoc = UserLoc.getIncludeLoc();
272     if (IncludeLoc.isValid())
273       MoveToLine(IncludeLoc);
274   } else if (Reason == PPCallbacks::SystemHeaderPragma) {
275     // GCC emits the # directive for this directive on the line AFTER the
276     // directive and emits a bunch of spaces that aren't needed. This is because
277     // otherwise we will emit a line marker for THIS line, which requires an
278     // extra blank line after the directive to avoid making all following lines
279     // off by one. We can do better by simply incrementing NewLine here.
280     NewLine += 1;
281   }
282 
283   CurLine = NewLine;
284 
285   CurFilename.clear();
286   CurFilename += UserLoc.getFilename();
287   FileType = NewFileType;
288 
289   if (DisableLineMarkers) {
290     startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
291     return;
292   }
293 
294   if (!Initialized) {
295     WriteLineInfo(CurLine);
296     Initialized = true;
297   }
298 
299   // Do not emit an enter marker for the main file (which we expect is the first
300   // entered file). This matches gcc, and improves compatibility with some tools
301   // which track the # line markers as a way to determine when the preprocessed
302   // output is in the context of the main file.
303   if (Reason == PPCallbacks::EnterFile && !IsFirstFileEntered) {
304     IsFirstFileEntered = true;
305     return;
306   }
307 
308   switch (Reason) {
309   case PPCallbacks::EnterFile:
310     WriteLineInfo(CurLine, " 1", 2);
311     break;
312   case PPCallbacks::ExitFile:
313     WriteLineInfo(CurLine, " 2", 2);
314     break;
315   case PPCallbacks::SystemHeaderPragma:
316   case PPCallbacks::RenameFile:
317     WriteLineInfo(CurLine);
318     break;
319   }
320 }
321 
322 void PrintPPOutputPPCallbacks::InclusionDirective(SourceLocation HashLoc,
323                                                   const Token &IncludeTok,
324                                                   StringRef FileName,
325                                                   bool IsAngled,
326                                                   CharSourceRange FilenameRange,
327                                                   const FileEntry *File,
328                                                   StringRef SearchPath,
329                                                   StringRef RelativePath,
330                                                   const Module *Imported) {
331   // When preprocessing, turn implicit imports into @imports.
332   // FIXME: This is a stop-gap until a more comprehensive "preprocessing with
333   // modules" solution is introduced.
334   if (Imported) {
335     startNewLineIfNeeded();
336     MoveToLine(HashLoc);
337     OS << "@import " << Imported->getFullModuleName() << ";"
338        << " /* clang -E: implicit import for \"" << File->getName() << "\" */";
339     EmittedTokensOnThisLine = true;
340   }
341 }
342 
343 /// Ident - Handle #ident directives when read by the preprocessor.
344 ///
345 void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, const std::string &S) {
346   MoveToLine(Loc);
347 
348   OS.write("#ident ", strlen("#ident "));
349   OS.write(&S[0], S.size());
350   EmittedTokensOnThisLine = true;
351 }
352 
353 void PrintPPOutputPPCallbacks::PragmaCaptured(SourceLocation Loc,
354                                               StringRef Str) {
355   startNewLineIfNeeded();
356   MoveToLine(Loc);
357   OS << "#pragma captured";
358 
359   setEmittedDirectiveOnThisLine();
360 }
361 
362 /// MacroDefined - This hook is called whenever a macro definition is seen.
363 void PrintPPOutputPPCallbacks::MacroDefined(const Token &MacroNameTok,
364                                             const MacroDirective *MD) {
365   const MacroInfo *MI = MD->getMacroInfo();
366   // Only print out macro definitions in -dD mode.
367   if (!DumpDefines ||
368       // Ignore __FILE__ etc.
369       MI->isBuiltinMacro()) return;
370 
371   MoveToLine(MI->getDefinitionLoc());
372   PrintMacroDefinition(*MacroNameTok.getIdentifierInfo(), *MI, PP, OS);
373   setEmittedDirectiveOnThisLine();
374 }
375 
376 void PrintPPOutputPPCallbacks::MacroUndefined(const Token &MacroNameTok,
377                                               const MacroDirective *MD) {
378   // Only print out macro definitions in -dD mode.
379   if (!DumpDefines) return;
380 
381   MoveToLine(MacroNameTok.getLocation());
382   OS << "#undef " << MacroNameTok.getIdentifierInfo()->getName();
383   setEmittedDirectiveOnThisLine();
384 }
385 
386 static void outputPrintable(llvm::raw_ostream& OS,
387                                              const std::string &Str) {
388     for (unsigned i = 0, e = Str.size(); i != e; ++i) {
389       unsigned char Char = Str[i];
390       if (isPrintable(Char) && Char != '\\' && Char != '"')
391         OS << (char)Char;
392       else  // Output anything hard as an octal escape.
393         OS << '\\'
394            << (char)('0'+ ((Char >> 6) & 7))
395            << (char)('0'+ ((Char >> 3) & 7))
396            << (char)('0'+ ((Char >> 0) & 7));
397     }
398 }
399 
400 void PrintPPOutputPPCallbacks::PragmaMessage(SourceLocation Loc,
401                                              StringRef Namespace,
402                                              PragmaMessageKind Kind,
403                                              StringRef Str) {
404   startNewLineIfNeeded();
405   MoveToLine(Loc);
406   OS << "#pragma ";
407   if (!Namespace.empty())
408     OS << Namespace << ' ';
409   switch (Kind) {
410     case PMK_Message:
411       OS << "message(\"";
412       break;
413     case PMK_Warning:
414       OS << "warning \"";
415       break;
416     case PMK_Error:
417       OS << "error \"";
418       break;
419   }
420 
421   outputPrintable(OS, Str);
422   OS << '"';
423   if (Kind == PMK_Message)
424     OS << ')';
425   setEmittedDirectiveOnThisLine();
426 }
427 
428 void PrintPPOutputPPCallbacks::PragmaDebug(SourceLocation Loc,
429                                            StringRef DebugType) {
430   startNewLineIfNeeded();
431   MoveToLine(Loc);
432 
433   OS << "#pragma clang __debug ";
434   OS << DebugType;
435 
436   setEmittedDirectiveOnThisLine();
437 }
438 
439 void PrintPPOutputPPCallbacks::
440 PragmaDiagnosticPush(SourceLocation Loc, StringRef Namespace) {
441   startNewLineIfNeeded();
442   MoveToLine(Loc);
443   OS << "#pragma " << Namespace << " diagnostic push";
444   setEmittedDirectiveOnThisLine();
445 }
446 
447 void PrintPPOutputPPCallbacks::
448 PragmaDiagnosticPop(SourceLocation Loc, StringRef Namespace) {
449   startNewLineIfNeeded();
450   MoveToLine(Loc);
451   OS << "#pragma " << Namespace << " diagnostic pop";
452   setEmittedDirectiveOnThisLine();
453 }
454 
455 void PrintPPOutputPPCallbacks::
456 PragmaDiagnostic(SourceLocation Loc, StringRef Namespace,
457                  diag::Mapping Map, StringRef Str) {
458   startNewLineIfNeeded();
459   MoveToLine(Loc);
460   OS << "#pragma " << Namespace << " diagnostic ";
461   switch (Map) {
462   case diag::MAP_REMARK:
463     OS << "remark";
464     break;
465   case diag::MAP_WARNING:
466     OS << "warning";
467     break;
468   case diag::MAP_ERROR:
469     OS << "error";
470     break;
471   case diag::MAP_IGNORE:
472     OS << "ignored";
473     break;
474   case diag::MAP_FATAL:
475     OS << "fatal";
476     break;
477   }
478   OS << " \"" << Str << '"';
479   setEmittedDirectiveOnThisLine();
480 }
481 
482 void PrintPPOutputPPCallbacks::PragmaWarning(SourceLocation Loc,
483                                              StringRef WarningSpec,
484                                              ArrayRef<int> Ids) {
485   startNewLineIfNeeded();
486   MoveToLine(Loc);
487   OS << "#pragma warning(" << WarningSpec << ':';
488   for (ArrayRef<int>::iterator I = Ids.begin(), E = Ids.end(); I != E; ++I)
489     OS << ' ' << *I;
490   OS << ')';
491   setEmittedDirectiveOnThisLine();
492 }
493 
494 void PrintPPOutputPPCallbacks::PragmaWarningPush(SourceLocation Loc,
495                                                  int Level) {
496   startNewLineIfNeeded();
497   MoveToLine(Loc);
498   OS << "#pragma warning(push";
499   if (Level >= 0)
500     OS << ", " << Level;
501   OS << ')';
502   setEmittedDirectiveOnThisLine();
503 }
504 
505 void PrintPPOutputPPCallbacks::PragmaWarningPop(SourceLocation Loc) {
506   startNewLineIfNeeded();
507   MoveToLine(Loc);
508   OS << "#pragma warning(pop)";
509   setEmittedDirectiveOnThisLine();
510 }
511 
512 /// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
513 /// is called for the first token on each new line.  If this really is the start
514 /// of a new logical line, handle it and return true, otherwise return false.
515 /// This may not be the start of a logical line because the "start of line"
516 /// marker is set for spelling lines, not expansion ones.
517 bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
518   // Figure out what line we went to and insert the appropriate number of
519   // newline characters.
520   if (!MoveToLine(Tok.getLocation()))
521     return false;
522 
523   // Print out space characters so that the first token on a line is
524   // indented for easy reading.
525   unsigned ColNo = SM.getExpansionColumnNumber(Tok.getLocation());
526 
527   // The first token on a line can have a column number of 1, yet still expect
528   // leading white space, if a macro expansion in column 1 starts with an empty
529   // macro argument, or an empty nested macro expansion. In this case, move the
530   // token to column 2.
531   if (ColNo == 1 && Tok.hasLeadingSpace())
532     ColNo = 2;
533 
534   // This hack prevents stuff like:
535   // #define HASH #
536   // HASH define foo bar
537   // From having the # character end up at column 1, which makes it so it
538   // is not handled as a #define next time through the preprocessor if in
539   // -fpreprocessed mode.
540   if (ColNo <= 1 && Tok.is(tok::hash))
541     OS << ' ';
542 
543   // Otherwise, indent the appropriate number of spaces.
544   for (; ColNo > 1; --ColNo)
545     OS << ' ';
546 
547   return true;
548 }
549 
550 void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
551                                                      unsigned Len) {
552   unsigned NumNewlines = 0;
553   for (; Len; --Len, ++TokStr) {
554     if (*TokStr != '\n' &&
555         *TokStr != '\r')
556       continue;
557 
558     ++NumNewlines;
559 
560     // If we have \n\r or \r\n, skip both and count as one line.
561     if (Len != 1 &&
562         (TokStr[1] == '\n' || TokStr[1] == '\r') &&
563         TokStr[0] != TokStr[1])
564       ++TokStr, --Len;
565   }
566 
567   if (NumNewlines == 0) return;
568 
569   CurLine += NumNewlines;
570 }
571 
572 
573 namespace {
574 struct UnknownPragmaHandler : public PragmaHandler {
575   const char *Prefix;
576   PrintPPOutputPPCallbacks *Callbacks;
577 
578   UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks)
579     : Prefix(prefix), Callbacks(callbacks) {}
580   virtual void HandlePragma(Preprocessor &PP, PragmaIntroducerKind Introducer,
581                             Token &PragmaTok) {
582     // Figure out what line we went to and insert the appropriate number of
583     // newline characters.
584     Callbacks->startNewLineIfNeeded();
585     Callbacks->MoveToLine(PragmaTok.getLocation());
586     Callbacks->OS.write(Prefix, strlen(Prefix));
587     // Read and print all of the pragma tokens.
588     while (PragmaTok.isNot(tok::eod)) {
589       if (PragmaTok.hasLeadingSpace())
590         Callbacks->OS << ' ';
591       std::string TokSpell = PP.getSpelling(PragmaTok);
592       Callbacks->OS.write(&TokSpell[0], TokSpell.size());
593 
594       // Expand macros in pragmas with -fms-extensions.  The assumption is that
595       // the majority of pragmas in such a file will be Microsoft pragmas.
596       if (PP.getLangOpts().MicrosoftExt)
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.AddPragmaHandler(new EmptyPragmaHandler());
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     if (I->first->hasMacroDefinition())
702       MacrosByID.push_back(id_macro_pair(I->first, I->second->getMacroInfo()));
703   }
704   llvm::array_pod_sort(MacrosByID.begin(), MacrosByID.end(), MacroIDCompare);
705 
706   for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
707     MacroInfo &MI = *MacrosByID[i].second;
708     // Ignore computed macros like __LINE__ and friends.
709     if (MI.isBuiltinMacro()) continue;
710 
711     PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
712     *OS << '\n';
713   }
714 }
715 
716 /// DoPrintPreprocessedInput - This implements -E mode.
717 ///
718 void clang::DoPrintPreprocessedInput(Preprocessor &PP, raw_ostream *OS,
719                                      const PreprocessorOutputOptions &Opts) {
720   // Show macros with no output is handled specially.
721   if (!Opts.ShowCPP) {
722     assert(Opts.ShowMacros && "Not yet implemented!");
723     DoPrintMacros(PP, OS);
724     return;
725   }
726 
727   // Inform the preprocessor whether we want it to retain comments or not, due
728   // to -C or -CC.
729   PP.SetCommentRetentionState(Opts.ShowComments, Opts.ShowMacroComments);
730 
731   PrintPPOutputPPCallbacks *Callbacks =
732       new PrintPPOutputPPCallbacks(PP, *OS, !Opts.ShowLineMarkers,
733                                    Opts.ShowMacros);
734   PP.AddPragmaHandler(new UnknownPragmaHandler("#pragma", Callbacks));
735   PP.AddPragmaHandler("GCC", new UnknownPragmaHandler("#pragma GCC",Callbacks));
736   PP.AddPragmaHandler("clang",
737                       new UnknownPragmaHandler("#pragma clang", Callbacks));
738 
739   PP.addPPCallbacks(Callbacks);
740 
741   // After we have configured the preprocessor, enter the main file.
742   PP.EnterMainSourceFile();
743 
744   // Consume all of the tokens that come from the predefines buffer.  Those
745   // should not be emitted into the output and are guaranteed to be at the
746   // start.
747   const SourceManager &SourceMgr = PP.getSourceManager();
748   Token Tok;
749   do {
750     PP.Lex(Tok);
751     if (Tok.is(tok::eof) || !Tok.getLocation().isFileID())
752       break;
753 
754     PresumedLoc PLoc = SourceMgr.getPresumedLoc(Tok.getLocation());
755     if (PLoc.isInvalid())
756       break;
757 
758     if (strcmp(PLoc.getFilename(), "<built-in>"))
759       break;
760   } while (true);
761 
762   // Read all the preprocessed tokens, printing them out to the stream.
763   PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
764   *OS << '\n';
765 }
766