PrintPreprocessedOutput.cpp revision 314564 
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 (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
79namespace {
80class PrintPPOutputPPCallbacks : public PPCallbacks {
81  Preprocessor &PP;
82  SourceManager &SM;
83  TokenConcatenation ConcatInfo;
84public:
85  raw_ostream &OS;
86private:
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 DumpIncludeDirectives;
97  bool UseLineDirectives;
98  bool IsFirstFileEntered;
99public:
100  PrintPPOutputPPCallbacks(Preprocessor &pp, raw_ostream &os, bool lineMarkers,
101                           bool defines, bool DumpIncludeDirectives,
102                           bool UseLineDirectives)
103      : PP(pp), SM(PP.getSourceManager()), ConcatInfo(PP), OS(os),
104        DisableLineMarkers(lineMarkers), DumpDefines(defines),
105        DumpIncludeDirectives(DumpIncludeDirectives),
106        UseLineDirectives(UseLineDirectives) {
107    CurLine = 0;
108    CurFilename += "<uninit>";
109    EmittedTokensOnThisLine = false;
110    EmittedDirectiveOnThisLine = false;
111    FileType = SrcMgr::C_User;
112    Initialized = false;
113    IsFirstFileEntered = false;
114  }
115
116  void setEmittedTokensOnThisLine() { EmittedTokensOnThisLine = true; }
117  bool hasEmittedTokensOnThisLine() const { return EmittedTokensOnThisLine; }
118
119  void setEmittedDirectiveOnThisLine() { EmittedDirectiveOnThisLine = true; }
120  bool hasEmittedDirectiveOnThisLine() const {
121    return EmittedDirectiveOnThisLine;
122  }
123
124  bool startNewLineIfNeeded(bool ShouldUpdateCurrentLine = true);
125
126  void FileChanged(SourceLocation Loc, FileChangeReason Reason,
127                   SrcMgr::CharacteristicKind FileType,
128                   FileID PrevFID) override;
129  void InclusionDirective(SourceLocation HashLoc, const Token &IncludeTok,
130                          StringRef FileName, bool IsAngled,
131                          CharSourceRange FilenameRange, const FileEntry *File,
132                          StringRef SearchPath, StringRef RelativePath,
133                          const Module *Imported) override;
134  void Ident(SourceLocation Loc, StringRef str) override;
135  void PragmaMessage(SourceLocation Loc, StringRef Namespace,
136                     PragmaMessageKind Kind, StringRef Str) override;
137  void PragmaDebug(SourceLocation Loc, StringRef DebugType) override;
138  void PragmaDiagnosticPush(SourceLocation Loc, StringRef Namespace) override;
139  void PragmaDiagnosticPop(SourceLocation Loc, StringRef Namespace) override;
140  void PragmaDiagnostic(SourceLocation Loc, StringRef Namespace,
141                        diag::Severity Map, StringRef Str) override;
142  void PragmaWarning(SourceLocation Loc, StringRef WarningSpec,
143                     ArrayRef<int> Ids) override;
144  void PragmaWarningPush(SourceLocation Loc, int Level) override;
145  void PragmaWarningPop(SourceLocation Loc) override;
146
147  bool HandleFirstTokOnLine(Token &Tok);
148
149  /// Move to the line of the provided source location. This will
150  /// return true if the output stream required adjustment or if
151  /// the requested location is on the first line.
152  bool MoveToLine(SourceLocation Loc) {
153    PresumedLoc PLoc = SM.getPresumedLoc(Loc);
154    if (PLoc.isInvalid())
155      return false;
156    return MoveToLine(PLoc.getLine()) || (PLoc.getLine() == 1);
157  }
158  bool MoveToLine(unsigned LineNo);
159
160  bool AvoidConcat(const Token &PrevPrevTok, const Token &PrevTok,
161                   const Token &Tok) {
162    return ConcatInfo.AvoidConcat(PrevPrevTok, PrevTok, Tok);
163  }
164  void WriteLineInfo(unsigned LineNo, const char *Extra=nullptr,
165                     unsigned ExtraLen=0);
166  bool LineMarkersAreDisabled() const { return DisableLineMarkers; }
167  void HandleNewlinesInToken(const char *TokStr, unsigned Len);
168
169  /// MacroDefined - This hook is called whenever a macro definition is seen.
170  void MacroDefined(const Token &MacroNameTok,
171                    const MacroDirective *MD) override;
172
173  /// MacroUndefined - This hook is called whenever a macro #undef is seen.
174  void MacroUndefined(const Token &MacroNameTok,
175                      const MacroDefinition &MD) override;
176};
177}  // end anonymous namespace
178
179void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
180                                             const char *Extra,
181                                             unsigned ExtraLen) {
182  startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
183
184  // Emit #line directives or GNU line markers depending on what mode we're in.
185  if (UseLineDirectives) {
186    OS << "#line" << ' ' << LineNo << ' ' << '"';
187    OS.write_escaped(CurFilename);
188    OS << '"';
189  } else {
190    OS << '#' << ' ' << LineNo << ' ' << '"';
191    OS.write_escaped(CurFilename);
192    OS << '"';
193
194    if (ExtraLen)
195      OS.write(Extra, ExtraLen);
196
197    if (FileType == SrcMgr::C_System)
198      OS.write(" 3", 2);
199    else if (FileType == SrcMgr::C_ExternCSystem)
200      OS.write(" 3 4", 4);
201  }
202  OS << '\n';
203}
204
205/// MoveToLine - Move the output to the source line specified by the location
206/// object.  We can do this by emitting some number of \n's, or be emitting a
207/// #line directive.  This returns false if already at the specified line, true
208/// if some newlines were emitted.
209bool PrintPPOutputPPCallbacks::MoveToLine(unsigned LineNo) {
210  // If this line is "close enough" to the original line, just print newlines,
211  // otherwise print a #line directive.
212  if (LineNo-CurLine <= 8) {
213    if (LineNo-CurLine == 1)
214      OS << '\n';
215    else if (LineNo == CurLine)
216      return false;    // Spelling line moved, but expansion line didn't.
217    else {
218      const char *NewLines = "\n\n\n\n\n\n\n\n";
219      OS.write(NewLines, LineNo-CurLine);
220    }
221  } else if (!DisableLineMarkers) {
222    // Emit a #line or line marker.
223    WriteLineInfo(LineNo, nullptr, 0);
224  } else {
225    // Okay, we're in -P mode, which turns off line markers.  However, we still
226    // need to emit a newline between tokens on different lines.
227    startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
228  }
229
230  CurLine = LineNo;
231  return true;
232}
233
234bool
235PrintPPOutputPPCallbacks::startNewLineIfNeeded(bool ShouldUpdateCurrentLine) {
236  if (EmittedTokensOnThisLine || EmittedDirectiveOnThisLine) {
237    OS << '\n';
238    EmittedTokensOnThisLine = false;
239    EmittedDirectiveOnThisLine = false;
240    if (ShouldUpdateCurrentLine)
241      ++CurLine;
242    return true;
243  }
244
245  return false;
246}
247
248/// FileChanged - Whenever the preprocessor enters or exits a #include file
249/// it invokes this handler.  Update our conception of the current source
250/// position.
251void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
252                                           FileChangeReason Reason,
253                                       SrcMgr::CharacteristicKind NewFileType,
254                                       FileID PrevFID) {
255  // Unless we are exiting a #include, make sure to skip ahead to the line the
256  // #include directive was at.
257  SourceManager &SourceMgr = SM;
258
259  PresumedLoc UserLoc = SourceMgr.getPresumedLoc(Loc);
260  if (UserLoc.isInvalid())
261    return;
262
263  unsigned NewLine = UserLoc.getLine();
264
265  if (Reason == PPCallbacks::EnterFile) {
266    SourceLocation IncludeLoc = UserLoc.getIncludeLoc();
267    if (IncludeLoc.isValid())
268      MoveToLine(IncludeLoc);
269  } else if (Reason == PPCallbacks::SystemHeaderPragma) {
270    // GCC emits the # directive for this directive on the line AFTER the
271    // directive and emits a bunch of spaces that aren't needed. This is because
272    // otherwise we will emit a line marker for THIS line, which requires an
273    // extra blank line after the directive to avoid making all following lines
274    // off by one. We can do better by simply incrementing NewLine here.
275    NewLine += 1;
276  }
277
278  CurLine = NewLine;
279
280  CurFilename.clear();
281  CurFilename += UserLoc.getFilename();
282  FileType = NewFileType;
283
284  if (DisableLineMarkers) {
285    startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
286    return;
287  }
288
289  if (!Initialized) {
290    WriteLineInfo(CurLine);
291    Initialized = true;
292  }
293
294  // Do not emit an enter marker for the main file (which we expect is the first
295  // entered file). This matches gcc, and improves compatibility with some tools
296  // which track the # line markers as a way to determine when the preprocessed
297  // output is in the context of the main file.
298  if (Reason == PPCallbacks::EnterFile && !IsFirstFileEntered) {
299    IsFirstFileEntered = true;
300    return;
301  }
302
303  switch (Reason) {
304  case PPCallbacks::EnterFile:
305    WriteLineInfo(CurLine, " 1", 2);
306    break;
307  case PPCallbacks::ExitFile:
308    WriteLineInfo(CurLine, " 2", 2);
309    break;
310  case PPCallbacks::SystemHeaderPragma:
311  case PPCallbacks::RenameFile:
312    WriteLineInfo(CurLine);
313    break;
314  }
315}
316
317void PrintPPOutputPPCallbacks::InclusionDirective(SourceLocation HashLoc,
318                                                  const Token &IncludeTok,
319                                                  StringRef FileName,
320                                                  bool IsAngled,
321                                                  CharSourceRange FilenameRange,
322                                                  const FileEntry *File,
323                                                  StringRef SearchPath,
324                                                  StringRef RelativePath,
325                                                  const Module *Imported) {
326  if (Imported) {
327    // When preprocessing, turn implicit imports into @imports.
328    // FIXME: This is a stop-gap until a more comprehensive "preprocessing with
329    // modules" solution is introduced.
330    startNewLineIfNeeded();
331    MoveToLine(HashLoc);
332    if (PP.getLangOpts().ObjC2) {
333      OS << "@import " << Imported->getFullModuleName() << ";"
334         << " /* clang -E: implicit import for \"" << File->getName()
335         << "\" */";
336    } else {
337      const std::string TokenText = PP.getSpelling(IncludeTok);
338      assert(!TokenText.empty());
339      OS << "#" << TokenText << " "
340         << (IsAngled ? '<' : '"')
341         << FileName
342         << (IsAngled ? '>' : '"')
343         << " /* clang -E: implicit import for module "
344         << Imported->getFullModuleName() << " */";
345    }
346    // Since we want a newline after the @import, but not a #<line>, start a new
347    // line immediately.
348    EmittedTokensOnThisLine = true;
349    startNewLineIfNeeded();
350  } else {
351    // Not a module import; it's a more vanilla inclusion of some file using one
352    // of: #include, #import, #include_next, #include_macros.
353    if (DumpIncludeDirectives) {
354      startNewLineIfNeeded();
355      MoveToLine(HashLoc);
356      const std::string TokenText = PP.getSpelling(IncludeTok);
357      assert(!TokenText.empty());
358      OS << "#" << TokenText << " "
359         << (IsAngled ? '<' : '"') << FileName << (IsAngled ? '>' : '"')
360         << " /* clang -E -dI */";
361      setEmittedDirectiveOnThisLine();
362      startNewLineIfNeeded();
363    }
364  }
365}
366
367/// Ident - Handle #ident directives when read by the preprocessor.
368///
369void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, StringRef S) {
370  MoveToLine(Loc);
371
372  OS.write("#ident ", strlen("#ident "));
373  OS.write(S.begin(), S.size());
374  EmittedTokensOnThisLine = true;
375}
376
377/// MacroDefined - This hook is called whenever a macro definition is seen.
378void PrintPPOutputPPCallbacks::MacroDefined(const Token &MacroNameTok,
379                                            const MacroDirective *MD) {
380  const MacroInfo *MI = MD->getMacroInfo();
381  // Only print out macro definitions in -dD mode.
382  if (!DumpDefines ||
383      // Ignore __FILE__ etc.
384      MI->isBuiltinMacro()) return;
385
386  MoveToLine(MI->getDefinitionLoc());
387  PrintMacroDefinition(*MacroNameTok.getIdentifierInfo(), *MI, PP, OS);
388  setEmittedDirectiveOnThisLine();
389}
390
391void PrintPPOutputPPCallbacks::MacroUndefined(const Token &MacroNameTok,
392                                              const MacroDefinition &MD) {
393  // Only print out macro definitions in -dD mode.
394  if (!DumpDefines) return;
395
396  MoveToLine(MacroNameTok.getLocation());
397  OS << "#undef " << MacroNameTok.getIdentifierInfo()->getName();
398  setEmittedDirectiveOnThisLine();
399}
400
401static void outputPrintable(raw_ostream &OS, StringRef Str) {
402  for (unsigned char Char : Str) {
403    if (isPrintable(Char) && Char != '\\' && Char != '"')
404      OS << (char)Char;
405    else // Output anything hard as an octal escape.
406      OS << '\\'
407         << (char)('0' + ((Char >> 6) & 7))
408         << (char)('0' + ((Char >> 3) & 7))
409         << (char)('0' + ((Char >> 0) & 7));
410  }
411}
412
413void PrintPPOutputPPCallbacks::PragmaMessage(SourceLocation Loc,
414                                             StringRef Namespace,
415                                             PragmaMessageKind Kind,
416                                             StringRef Str) {
417  startNewLineIfNeeded();
418  MoveToLine(Loc);
419  OS << "#pragma ";
420  if (!Namespace.empty())
421    OS << Namespace << ' ';
422  switch (Kind) {
423    case PMK_Message:
424      OS << "message(\"";
425      break;
426    case PMK_Warning:
427      OS << "warning \"";
428      break;
429    case PMK_Error:
430      OS << "error \"";
431      break;
432  }
433
434  outputPrintable(OS, Str);
435  OS << '"';
436  if (Kind == PMK_Message)
437    OS << ')';
438  setEmittedDirectiveOnThisLine();
439}
440
441void PrintPPOutputPPCallbacks::PragmaDebug(SourceLocation Loc,
442                                           StringRef DebugType) {
443  startNewLineIfNeeded();
444  MoveToLine(Loc);
445
446  OS << "#pragma clang __debug ";
447  OS << DebugType;
448
449  setEmittedDirectiveOnThisLine();
450}
451
452void PrintPPOutputPPCallbacks::
453PragmaDiagnosticPush(SourceLocation Loc, StringRef Namespace) {
454  startNewLineIfNeeded();
455  MoveToLine(Loc);
456  OS << "#pragma " << Namespace << " diagnostic push";
457  setEmittedDirectiveOnThisLine();
458}
459
460void PrintPPOutputPPCallbacks::
461PragmaDiagnosticPop(SourceLocation Loc, StringRef Namespace) {
462  startNewLineIfNeeded();
463  MoveToLine(Loc);
464  OS << "#pragma " << Namespace << " diagnostic pop";
465  setEmittedDirectiveOnThisLine();
466}
467
468void PrintPPOutputPPCallbacks::PragmaDiagnostic(SourceLocation Loc,
469                                                StringRef Namespace,
470                                                diag::Severity Map,
471                                                StringRef Str) {
472  startNewLineIfNeeded();
473  MoveToLine(Loc);
474  OS << "#pragma " << Namespace << " diagnostic ";
475  switch (Map) {
476  case diag::Severity::Remark:
477    OS << "remark";
478    break;
479  case diag::Severity::Warning:
480    OS << "warning";
481    break;
482  case diag::Severity::Error:
483    OS << "error";
484    break;
485  case diag::Severity::Ignored:
486    OS << "ignored";
487    break;
488  case diag::Severity::Fatal:
489    OS << "fatal";
490    break;
491  }
492  OS << " \"" << Str << '"';
493  setEmittedDirectiveOnThisLine();
494}
495
496void PrintPPOutputPPCallbacks::PragmaWarning(SourceLocation Loc,
497                                             StringRef WarningSpec,
498                                             ArrayRef<int> Ids) {
499  startNewLineIfNeeded();
500  MoveToLine(Loc);
501  OS << "#pragma warning(" << WarningSpec << ':';
502  for (ArrayRef<int>::iterator I = Ids.begin(), E = Ids.end(); I != E; ++I)
503    OS << ' ' << *I;
504  OS << ')';
505  setEmittedDirectiveOnThisLine();
506}
507
508void PrintPPOutputPPCallbacks::PragmaWarningPush(SourceLocation Loc,
509                                                 int Level) {
510  startNewLineIfNeeded();
511  MoveToLine(Loc);
512  OS << "#pragma warning(push";
513  if (Level >= 0)
514    OS << ", " << Level;
515  OS << ')';
516  setEmittedDirectiveOnThisLine();
517}
518
519void PrintPPOutputPPCallbacks::PragmaWarningPop(SourceLocation Loc) {
520  startNewLineIfNeeded();
521  MoveToLine(Loc);
522  OS << "#pragma warning(pop)";
523  setEmittedDirectiveOnThisLine();
524}
525
526/// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
527/// is called for the first token on each new line.  If this really is the start
528/// of a new logical line, handle it and return true, otherwise return false.
529/// This may not be the start of a logical line because the "start of line"
530/// marker is set for spelling lines, not expansion ones.
531bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
532  // Figure out what line we went to and insert the appropriate number of
533  // newline characters.
534  if (!MoveToLine(Tok.getLocation()))
535    return false;
536
537  // Print out space characters so that the first token on a line is
538  // indented for easy reading.
539  unsigned ColNo = SM.getExpansionColumnNumber(Tok.getLocation());
540
541  // The first token on a line can have a column number of 1, yet still expect
542  // leading white space, if a macro expansion in column 1 starts with an empty
543  // macro argument, or an empty nested macro expansion. In this case, move the
544  // token to column 2.
545  if (ColNo == 1 && Tok.hasLeadingSpace())
546    ColNo = 2;
547
548  // This hack prevents stuff like:
549  // #define HASH #
550  // HASH define foo bar
551  // From having the # character end up at column 1, which makes it so it
552  // is not handled as a #define next time through the preprocessor if in
553  // -fpreprocessed mode.
554  if (ColNo <= 1 && Tok.is(tok::hash))
555    OS << ' ';
556
557  // Otherwise, indent the appropriate number of spaces.
558  for (; ColNo > 1; --ColNo)
559    OS << ' ';
560
561  return true;
562}
563
564void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
565                                                     unsigned Len) {
566  unsigned NumNewlines = 0;
567  for (; Len; --Len, ++TokStr) {
568    if (*TokStr != '\n' &&
569        *TokStr != '\r')
570      continue;
571
572    ++NumNewlines;
573
574    // If we have \n\r or \r\n, skip both and count as one line.
575    if (Len != 1 &&
576        (TokStr[1] == '\n' || TokStr[1] == '\r') &&
577        TokStr[0] != TokStr[1]) {
578      ++TokStr;
579      --Len;
580    }
581  }
582
583  if (NumNewlines == 0) return;
584
585  CurLine += NumNewlines;
586}
587
588
589namespace {
590struct UnknownPragmaHandler : public PragmaHandler {
591  const char *Prefix;
592  PrintPPOutputPPCallbacks *Callbacks;
593
594  // Set to true if tokens should be expanded
595  bool ShouldExpandTokens;
596
597  UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks,
598                       bool RequireTokenExpansion)
599      : Prefix(prefix), Callbacks(callbacks),
600        ShouldExpandTokens(RequireTokenExpansion) {}
601  void HandlePragma(Preprocessor &PP, PragmaIntroducerKind Introducer,
602                    Token &PragmaTok) override {
603    // Figure out what line we went to and insert the appropriate number of
604    // newline characters.
605    Callbacks->startNewLineIfNeeded();
606    Callbacks->MoveToLine(PragmaTok.getLocation());
607    Callbacks->OS.write(Prefix, strlen(Prefix));
608
609    if (ShouldExpandTokens) {
610      // The first token does not have expanded macros. Expand them, if
611      // required.
612      auto Toks = llvm::make_unique<Token[]>(1);
613      Toks[0] = PragmaTok;
614      PP.EnterTokenStream(std::move(Toks), /*NumToks=*/1,
615                          /*DisableMacroExpansion=*/false);
616      PP.Lex(PragmaTok);
617    }
618    Token PrevToken;
619    Token PrevPrevToken;
620    PrevToken.startToken();
621    PrevPrevToken.startToken();
622
623    // Read and print all of the pragma tokens.
624    while (PragmaTok.isNot(tok::eod)) {
625      if (PragmaTok.hasLeadingSpace() ||
626          Callbacks->AvoidConcat(PrevPrevToken, PrevToken, PragmaTok))
627        Callbacks->OS << ' ';
628      std::string TokSpell = PP.getSpelling(PragmaTok);
629      Callbacks->OS.write(&TokSpell[0], TokSpell.size());
630
631      PrevPrevToken = PrevToken;
632      PrevToken = PragmaTok;
633
634      if (ShouldExpandTokens)
635        PP.Lex(PragmaTok);
636      else
637        PP.LexUnexpandedToken(PragmaTok);
638    }
639    Callbacks->setEmittedDirectiveOnThisLine();
640  }
641};
642} // end anonymous namespace
643
644
645static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
646                                    PrintPPOutputPPCallbacks *Callbacks,
647                                    raw_ostream &OS) {
648  bool DropComments = PP.getLangOpts().TraditionalCPP &&
649                      !PP.getCommentRetentionState();
650
651  char Buffer[256];
652  Token PrevPrevTok, PrevTok;
653  PrevPrevTok.startToken();
654  PrevTok.startToken();
655  while (1) {
656    if (Callbacks->hasEmittedDirectiveOnThisLine()) {
657      Callbacks->startNewLineIfNeeded();
658      Callbacks->MoveToLine(Tok.getLocation());
659    }
660
661    // If this token is at the start of a line, emit newlines if needed.
662    if (Tok.isAtStartOfLine() && Callbacks->HandleFirstTokOnLine(Tok)) {
663      // done.
664    } else if (Tok.hasLeadingSpace() ||
665               // If we haven't emitted a token on this line yet, PrevTok isn't
666               // useful to look at and no concatenation could happen anyway.
667               (Callbacks->hasEmittedTokensOnThisLine() &&
668                // Don't print "-" next to "-", it would form "--".
669                Callbacks->AvoidConcat(PrevPrevTok, PrevTok, Tok))) {
670      OS << ' ';
671    }
672
673    if (DropComments && Tok.is(tok::comment)) {
674      // Skip comments. Normally the preprocessor does not generate
675      // tok::comment nodes at all when not keeping comments, but under
676      // -traditional-cpp the lexer keeps /all/ whitespace, including comments.
677      SourceLocation StartLoc = Tok.getLocation();
678      Callbacks->MoveToLine(StartLoc.getLocWithOffset(Tok.getLength()));
679    } else if (Tok.is(tok::annot_module_include) ||
680               Tok.is(tok::annot_module_begin) ||
681               Tok.is(tok::annot_module_end)) {
682      // PrintPPOutputPPCallbacks::InclusionDirective handles producing
683      // appropriate output here. Ignore this token entirely.
684      PP.Lex(Tok);
685      continue;
686    } else if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
687      OS << II->getName();
688    } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
689               Tok.getLiteralData()) {
690      OS.write(Tok.getLiteralData(), Tok.getLength());
691    } else if (Tok.getLength() < 256) {
692      const char *TokPtr = Buffer;
693      unsigned Len = PP.getSpelling(Tok, TokPtr);
694      OS.write(TokPtr, Len);
695
696      // Tokens that can contain embedded newlines need to adjust our current
697      // line number.
698      if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
699        Callbacks->HandleNewlinesInToken(TokPtr, Len);
700    } else {
701      std::string S = PP.getSpelling(Tok);
702      OS.write(&S[0], S.size());
703
704      // Tokens that can contain embedded newlines need to adjust our current
705      // line number.
706      if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
707        Callbacks->HandleNewlinesInToken(&S[0], S.size());
708    }
709    Callbacks->setEmittedTokensOnThisLine();
710
711    if (Tok.is(tok::eof)) break;
712
713    PrevPrevTok = PrevTok;
714    PrevTok = Tok;
715    PP.Lex(Tok);
716  }
717}
718
719typedef std::pair<const IdentifierInfo *, MacroInfo *> id_macro_pair;
720static int MacroIDCompare(const id_macro_pair *LHS, const id_macro_pair *RHS) {
721  return LHS->first->getName().compare(RHS->first->getName());
722}
723
724static void DoPrintMacros(Preprocessor &PP, raw_ostream *OS) {
725  // Ignore unknown pragmas.
726  PP.IgnorePragmas();
727
728  // -dM mode just scans and ignores all tokens in the files, then dumps out
729  // the macro table at the end.
730  PP.EnterMainSourceFile();
731
732  Token Tok;
733  do PP.Lex(Tok);
734  while (Tok.isNot(tok::eof));
735
736  SmallVector<id_macro_pair, 128> MacrosByID;
737  for (Preprocessor::macro_iterator I = PP.macro_begin(), E = PP.macro_end();
738       I != E; ++I) {
739    auto *MD = I->second.getLatest();
740    if (MD && MD->isDefined())
741      MacrosByID.push_back(id_macro_pair(I->first, MD->getMacroInfo()));
742  }
743  llvm::array_pod_sort(MacrosByID.begin(), MacrosByID.end(), MacroIDCompare);
744
745  for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
746    MacroInfo &MI = *MacrosByID[i].second;
747    // Ignore computed macros like __LINE__ and friends.
748    if (MI.isBuiltinMacro()) continue;
749
750    PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
751    *OS << '\n';
752  }
753}
754
755/// DoPrintPreprocessedInput - This implements -E mode.
756///
757void clang::DoPrintPreprocessedInput(Preprocessor &PP, raw_ostream *OS,
758                                     const PreprocessorOutputOptions &Opts) {
759  // Show macros with no output is handled specially.
760  if (!Opts.ShowCPP) {
761    assert(Opts.ShowMacros && "Not yet implemented!");
762    DoPrintMacros(PP, OS);
763    return;
764  }
765
766  // Inform the preprocessor whether we want it to retain comments or not, due
767  // to -C or -CC.
768  PP.SetCommentRetentionState(Opts.ShowComments, Opts.ShowMacroComments);
769
770  PrintPPOutputPPCallbacks *Callbacks = new PrintPPOutputPPCallbacks(
771      PP, *OS, !Opts.ShowLineMarkers, Opts.ShowMacros,
772      Opts.ShowIncludeDirectives, Opts.UseLineDirectives);
773
774  // Expand macros in pragmas with -fms-extensions.  The assumption is that
775  // the majority of pragmas in such a file will be Microsoft pragmas.
776  PP.AddPragmaHandler(new UnknownPragmaHandler(
777      "#pragma", Callbacks,
778      /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
779  PP.AddPragmaHandler(
780      "GCC", new UnknownPragmaHandler(
781                 "#pragma GCC", Callbacks,
782                 /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
783  PP.AddPragmaHandler(
784      "clang", new UnknownPragmaHandler(
785                   "#pragma clang", Callbacks,
786                   /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
787
788  // The tokens after pragma omp need to be expanded.
789  //
790  //  OpenMP [2.1, Directive format]
791  //  Preprocessing tokens following the #pragma omp are subject to macro
792  //  replacement.
793  PP.AddPragmaHandler("omp",
794                      new UnknownPragmaHandler("#pragma omp", Callbacks,
795                                               /*RequireTokenExpansion=*/true));
796
797  PP.addPPCallbacks(std::unique_ptr<PPCallbacks>(Callbacks));
798
799  // After we have configured the preprocessor, enter the main file.
800  PP.EnterMainSourceFile();
801
802  // Consume all of the tokens that come from the predefines buffer.  Those
803  // should not be emitted into the output and are guaranteed to be at the
804  // start.
805  const SourceManager &SourceMgr = PP.getSourceManager();
806  Token Tok;
807  do {
808    PP.Lex(Tok);
809    if (Tok.is(tok::eof) || !Tok.getLocation().isFileID())
810      break;
811
812    PresumedLoc PLoc = SourceMgr.getPresumedLoc(Tok.getLocation());
813    if (PLoc.isInvalid())
814      break;
815
816    if (strcmp(PLoc.getFilename(), "<built-in>"))
817      break;
818  } while (true);
819
820  // Read all the preprocessed tokens, printing them out to the stream.
821  PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
822  *OS << '\n';
823}
824