PrintPreprocessedOutput.cpp revision 360660 
1//===--- PrintPreprocessedOutput.cpp - Implement the -E mode --------------===//
2//
3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4// See https://llvm.org/LICENSE.txt for license information.
5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6//
7//===----------------------------------------------------------------------===//
8//
9// This code simply runs the preprocessor on the input file and prints out the
10// result.  This is the traditional behavior of the -E option.
11//
12//===----------------------------------------------------------------------===//
13
14#include "clang/Frontend/Utils.h"
15#include "clang/Basic/CharInfo.h"
16#include "clang/Basic/Diagnostic.h"
17#include "clang/Basic/SourceManager.h"
18#include "clang/Frontend/PreprocessorOutputOptions.h"
19#include "clang/Lex/MacroInfo.h"
20#include "clang/Lex/PPCallbacks.h"
21#include "clang/Lex/Pragma.h"
22#include "clang/Lex/Preprocessor.h"
23#include "clang/Lex/TokenConcatenation.h"
24#include "llvm/ADT/STLExtras.h"
25#include "llvm/ADT/SmallString.h"
26#include "llvm/ADT/StringRef.h"
27#include "llvm/Support/ErrorHandling.h"
28#include "llvm/Support/raw_ostream.h"
29#include <cstdio>
30using namespace clang;
31
32/// PrintMacroDefinition - Print a macro definition in a form that will be
33/// properly accepted back as a definition.
34static void PrintMacroDefinition(const IdentifierInfo &II, const MacroInfo &MI,
35                                 Preprocessor &PP, raw_ostream &OS) {
36  OS << "#define " << II.getName();
37
38  if (MI.isFunctionLike()) {
39    OS << '(';
40    if (!MI.param_empty()) {
41      MacroInfo::param_iterator AI = MI.param_begin(), E = MI.param_end();
42      for (; AI+1 != E; ++AI) {
43        OS << (*AI)->getName();
44        OS << ',';
45      }
46
47      // Last argument.
48      if ((*AI)->getName() == "__VA_ARGS__")
49        OS << "...";
50      else
51        OS << (*AI)->getName();
52    }
53
54    if (MI.isGNUVarargs())
55      OS << "...";  // #define foo(x...)
56
57    OS << ')';
58  }
59
60  // GCC always emits a space, even if the macro body is empty.  However, do not
61  // want to emit two spaces if the first token has a leading space.
62  if (MI.tokens_empty() || !MI.tokens_begin()->hasLeadingSpace())
63    OS << ' ';
64
65  SmallString<128> SpellingBuffer;
66  for (const auto &T : MI.tokens()) {
67    if (T.hasLeadingSpace())
68      OS << ' ';
69
70    OS << PP.getSpelling(T, SpellingBuffer);
71  }
72}
73
74//===----------------------------------------------------------------------===//
75// Preprocessed token printer
76//===----------------------------------------------------------------------===//
77
78namespace {
79class PrintPPOutputPPCallbacks : public PPCallbacks {
80  Preprocessor &PP;
81  SourceManager &SM;
82  TokenConcatenation ConcatInfo;
83public:
84  raw_ostream &OS;
85private:
86  unsigned CurLine;
87
88  bool EmittedTokensOnThisLine;
89  bool EmittedDirectiveOnThisLine;
90  SrcMgr::CharacteristicKind FileType;
91  SmallString<512> CurFilename;
92  bool Initialized;
93  bool DisableLineMarkers;
94  bool DumpDefines;
95  bool DumpIncludeDirectives;
96  bool UseLineDirectives;
97  bool IsFirstFileEntered;
98public:
99  PrintPPOutputPPCallbacks(Preprocessor &pp, raw_ostream &os, bool lineMarkers,
100                           bool defines, bool DumpIncludeDirectives,
101                           bool UseLineDirectives)
102      : PP(pp), SM(PP.getSourceManager()), ConcatInfo(PP), OS(os),
103        DisableLineMarkers(lineMarkers), DumpDefines(defines),
104        DumpIncludeDirectives(DumpIncludeDirectives),
105        UseLineDirectives(UseLineDirectives) {
106    CurLine = 0;
107    CurFilename += "<uninit>";
108    EmittedTokensOnThisLine = false;
109    EmittedDirectiveOnThisLine = false;
110    FileType = SrcMgr::C_User;
111    Initialized = false;
112    IsFirstFileEntered = false;
113  }
114
115  void setEmittedTokensOnThisLine() { EmittedTokensOnThisLine = true; }
116  bool hasEmittedTokensOnThisLine() const { return EmittedTokensOnThisLine; }
117
118  void setEmittedDirectiveOnThisLine() { EmittedDirectiveOnThisLine = true; }
119  bool hasEmittedDirectiveOnThisLine() const {
120    return EmittedDirectiveOnThisLine;
121  }
122
123  bool startNewLineIfNeeded(bool ShouldUpdateCurrentLine = true);
124
125  void FileChanged(SourceLocation Loc, FileChangeReason Reason,
126                   SrcMgr::CharacteristicKind FileType,
127                   FileID PrevFID) override;
128  void InclusionDirective(SourceLocation HashLoc, const Token &IncludeTok,
129                          StringRef FileName, bool IsAngled,
130                          CharSourceRange FilenameRange, const FileEntry *File,
131                          StringRef SearchPath, StringRef RelativePath,
132                          const Module *Imported,
133                          SrcMgr::CharacteristicKind FileType) 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  void PragmaExecCharsetPush(SourceLocation Loc, StringRef Str) override;
147  void PragmaExecCharsetPop(SourceLocation Loc) override;
148  void PragmaAssumeNonNullBegin(SourceLocation Loc) override;
149  void PragmaAssumeNonNullEnd(SourceLocation Loc) override;
150
151  bool HandleFirstTokOnLine(Token &Tok);
152
153  /// Move to the line of the provided source location. This will
154  /// return true if the output stream required adjustment or if
155  /// the requested location is on the first line.
156  bool MoveToLine(SourceLocation Loc) {
157    PresumedLoc PLoc = SM.getPresumedLoc(Loc);
158    if (PLoc.isInvalid())
159      return false;
160    return MoveToLine(PLoc.getLine()) || (PLoc.getLine() == 1);
161  }
162  bool MoveToLine(unsigned LineNo);
163
164  bool AvoidConcat(const Token &PrevPrevTok, const Token &PrevTok,
165                   const Token &Tok) {
166    return ConcatInfo.AvoidConcat(PrevPrevTok, PrevTok, Tok);
167  }
168  void WriteLineInfo(unsigned LineNo, const char *Extra=nullptr,
169                     unsigned ExtraLen=0);
170  bool LineMarkersAreDisabled() const { return DisableLineMarkers; }
171  void HandleNewlinesInToken(const char *TokStr, unsigned Len);
172
173  /// MacroDefined - This hook is called whenever a macro definition is seen.
174  void MacroDefined(const Token &MacroNameTok,
175                    const MacroDirective *MD) override;
176
177  /// MacroUndefined - This hook is called whenever a macro #undef is seen.
178  void MacroUndefined(const Token &MacroNameTok,
179                      const MacroDefinition &MD,
180                      const MacroDirective *Undef) override;
181
182  void BeginModule(const Module *M);
183  void EndModule(const Module *M);
184};
185}  // end anonymous namespace
186
187void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
188                                             const char *Extra,
189                                             unsigned ExtraLen) {
190  startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
191
192  // Emit #line directives or GNU line markers depending on what mode we're in.
193  if (UseLineDirectives) {
194    OS << "#line" << ' ' << LineNo << ' ' << '"';
195    OS.write_escaped(CurFilename);
196    OS << '"';
197  } else {
198    OS << '#' << ' ' << LineNo << ' ' << '"';
199    OS.write_escaped(CurFilename);
200    OS << '"';
201
202    if (ExtraLen)
203      OS.write(Extra, ExtraLen);
204
205    if (FileType == SrcMgr::C_System)
206      OS.write(" 3", 2);
207    else if (FileType == SrcMgr::C_ExternCSystem)
208      OS.write(" 3 4", 4);
209  }
210  OS << '\n';
211}
212
213/// MoveToLine - Move the output to the source line specified by the location
214/// object.  We can do this by emitting some number of \n's, or be emitting a
215/// #line directive.  This returns false if already at the specified line, true
216/// if some newlines were emitted.
217bool PrintPPOutputPPCallbacks::MoveToLine(unsigned LineNo) {
218  // If this line is "close enough" to the original line, just print newlines,
219  // otherwise print a #line directive.
220  if (LineNo-CurLine <= 8) {
221    if (LineNo-CurLine == 1)
222      OS << '\n';
223    else if (LineNo == CurLine)
224      return false;    // Spelling line moved, but expansion line didn't.
225    else {
226      const char *NewLines = "\n\n\n\n\n\n\n\n";
227      OS.write(NewLines, LineNo-CurLine);
228    }
229  } else if (!DisableLineMarkers) {
230    // Emit a #line or line marker.
231    WriteLineInfo(LineNo, nullptr, 0);
232  } else {
233    // Okay, we're in -P mode, which turns off line markers.  However, we still
234    // need to emit a newline between tokens on different lines.
235    startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
236  }
237
238  CurLine = LineNo;
239  return true;
240}
241
242bool
243PrintPPOutputPPCallbacks::startNewLineIfNeeded(bool ShouldUpdateCurrentLine) {
244  if (EmittedTokensOnThisLine || EmittedDirectiveOnThisLine) {
245    OS << '\n';
246    EmittedTokensOnThisLine = false;
247    EmittedDirectiveOnThisLine = false;
248    if (ShouldUpdateCurrentLine)
249      ++CurLine;
250    return true;
251  }
252
253  return false;
254}
255
256/// FileChanged - Whenever the preprocessor enters or exits a #include file
257/// it invokes this handler.  Update our conception of the current source
258/// position.
259void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
260                                           FileChangeReason Reason,
261                                       SrcMgr::CharacteristicKind NewFileType,
262                                       FileID PrevFID) {
263  // Unless we are exiting a #include, make sure to skip ahead to the line the
264  // #include directive was at.
265  SourceManager &SourceMgr = SM;
266
267  PresumedLoc UserLoc = SourceMgr.getPresumedLoc(Loc);
268  if (UserLoc.isInvalid())
269    return;
270
271  unsigned NewLine = UserLoc.getLine();
272
273  if (Reason == PPCallbacks::EnterFile) {
274    SourceLocation IncludeLoc = UserLoc.getIncludeLoc();
275    if (IncludeLoc.isValid())
276      MoveToLine(IncludeLoc);
277  } else if (Reason == PPCallbacks::SystemHeaderPragma) {
278    // GCC emits the # directive for this directive on the line AFTER the
279    // directive and emits a bunch of spaces that aren't needed. This is because
280    // otherwise we will emit a line marker for THIS line, which requires an
281    // extra blank line after the directive to avoid making all following lines
282    // off by one. We can do better by simply incrementing NewLine here.
283    NewLine += 1;
284  }
285
286  CurLine = NewLine;
287
288  CurFilename.clear();
289  CurFilename += UserLoc.getFilename();
290  FileType = NewFileType;
291
292  if (DisableLineMarkers) {
293    startNewLineIfNeeded(/*ShouldUpdateCurrentLine=*/false);
294    return;
295  }
296
297  if (!Initialized) {
298    WriteLineInfo(CurLine);
299    Initialized = true;
300  }
301
302  // Do not emit an enter marker for the main file (which we expect is the first
303  // entered file). This matches gcc, and improves compatibility with some tools
304  // which track the # line markers as a way to determine when the preprocessed
305  // output is in the context of the main file.
306  if (Reason == PPCallbacks::EnterFile && !IsFirstFileEntered) {
307    IsFirstFileEntered = true;
308    return;
309  }
310
311  switch (Reason) {
312  case PPCallbacks::EnterFile:
313    WriteLineInfo(CurLine, " 1", 2);
314    break;
315  case PPCallbacks::ExitFile:
316    WriteLineInfo(CurLine, " 2", 2);
317    break;
318  case PPCallbacks::SystemHeaderPragma:
319  case PPCallbacks::RenameFile:
320    WriteLineInfo(CurLine);
321    break;
322  }
323}
324
325void PrintPPOutputPPCallbacks::InclusionDirective(
326    SourceLocation HashLoc,
327    const Token &IncludeTok,
328    StringRef FileName,
329    bool IsAngled,
330    CharSourceRange FilenameRange,
331    const FileEntry *File,
332    StringRef SearchPath,
333    StringRef RelativePath,
334    const Module *Imported,
335    SrcMgr::CharacteristicKind FileType) {
336  // In -dI mode, dump #include directives prior to dumping their content or
337  // interpretation.
338  if (DumpIncludeDirectives) {
339    startNewLineIfNeeded();
340    MoveToLine(HashLoc);
341    const std::string TokenText = PP.getSpelling(IncludeTok);
342    assert(!TokenText.empty());
343    OS << "#" << TokenText << " "
344       << (IsAngled ? '<' : '"') << FileName << (IsAngled ? '>' : '"')
345       << " /* clang -E -dI */";
346    setEmittedDirectiveOnThisLine();
347    startNewLineIfNeeded();
348  }
349
350  // When preprocessing, turn implicit imports into module import pragmas.
351  if (Imported) {
352    switch (IncludeTok.getIdentifierInfo()->getPPKeywordID()) {
353    case tok::pp_include:
354    case tok::pp_import:
355    case tok::pp_include_next:
356      startNewLineIfNeeded();
357      MoveToLine(HashLoc);
358      OS << "#pragma clang module import " << Imported->getFullModuleName(true)
359         << " /* clang -E: implicit import for "
360         << "#" << PP.getSpelling(IncludeTok) << " "
361         << (IsAngled ? '<' : '"') << FileName << (IsAngled ? '>' : '"')
362         << " */";
363      // Since we want a newline after the pragma, but not a #<line>, start a
364      // new line immediately.
365      EmittedTokensOnThisLine = true;
366      startNewLineIfNeeded();
367      break;
368
369    case tok::pp___include_macros:
370      // #__include_macros has no effect on a user of a preprocessed source
371      // file; the only effect is on preprocessing.
372      //
373      // FIXME: That's not *quite* true: it causes the module in question to
374      // be loaded, which can affect downstream diagnostics.
375      break;
376
377    default:
378      llvm_unreachable("unknown include directive kind");
379      break;
380    }
381  }
382}
383
384/// Handle entering the scope of a module during a module compilation.
385void PrintPPOutputPPCallbacks::BeginModule(const Module *M) {
386  startNewLineIfNeeded();
387  OS << "#pragma clang module begin " << M->getFullModuleName(true);
388  setEmittedDirectiveOnThisLine();
389}
390
391/// Handle leaving the scope of a module during a module compilation.
392void PrintPPOutputPPCallbacks::EndModule(const Module *M) {
393  startNewLineIfNeeded();
394  OS << "#pragma clang module end /*" << M->getFullModuleName(true) << "*/";
395  setEmittedDirectiveOnThisLine();
396}
397
398/// Ident - Handle #ident directives when read by the preprocessor.
399///
400void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, StringRef S) {
401  MoveToLine(Loc);
402
403  OS.write("#ident ", strlen("#ident "));
404  OS.write(S.begin(), S.size());
405  EmittedTokensOnThisLine = true;
406}
407
408/// MacroDefined - This hook is called whenever a macro definition is seen.
409void PrintPPOutputPPCallbacks::MacroDefined(const Token &MacroNameTok,
410                                            const MacroDirective *MD) {
411  const MacroInfo *MI = MD->getMacroInfo();
412  // Only print out macro definitions in -dD mode.
413  if (!DumpDefines ||
414      // Ignore __FILE__ etc.
415      MI->isBuiltinMacro()) return;
416
417  MoveToLine(MI->getDefinitionLoc());
418  PrintMacroDefinition(*MacroNameTok.getIdentifierInfo(), *MI, PP, OS);
419  setEmittedDirectiveOnThisLine();
420}
421
422void PrintPPOutputPPCallbacks::MacroUndefined(const Token &MacroNameTok,
423                                              const MacroDefinition &MD,
424                                              const MacroDirective *Undef) {
425  // Only print out macro definitions in -dD mode.
426  if (!DumpDefines) return;
427
428  MoveToLine(MacroNameTok.getLocation());
429  OS << "#undef " << MacroNameTok.getIdentifierInfo()->getName();
430  setEmittedDirectiveOnThisLine();
431}
432
433static void outputPrintable(raw_ostream &OS, StringRef Str) {
434  for (unsigned char Char : Str) {
435    if (isPrintable(Char) && Char != '\\' && Char != '"')
436      OS << (char)Char;
437    else // Output anything hard as an octal escape.
438      OS << '\\'
439         << (char)('0' + ((Char >> 6) & 7))
440         << (char)('0' + ((Char >> 3) & 7))
441         << (char)('0' + ((Char >> 0) & 7));
442  }
443}
444
445void PrintPPOutputPPCallbacks::PragmaMessage(SourceLocation Loc,
446                                             StringRef Namespace,
447                                             PragmaMessageKind Kind,
448                                             StringRef Str) {
449  startNewLineIfNeeded();
450  MoveToLine(Loc);
451  OS << "#pragma ";
452  if (!Namespace.empty())
453    OS << Namespace << ' ';
454  switch (Kind) {
455    case PMK_Message:
456      OS << "message(\"";
457      break;
458    case PMK_Warning:
459      OS << "warning \"";
460      break;
461    case PMK_Error:
462      OS << "error \"";
463      break;
464  }
465
466  outputPrintable(OS, Str);
467  OS << '"';
468  if (Kind == PMK_Message)
469    OS << ')';
470  setEmittedDirectiveOnThisLine();
471}
472
473void PrintPPOutputPPCallbacks::PragmaDebug(SourceLocation Loc,
474                                           StringRef DebugType) {
475  startNewLineIfNeeded();
476  MoveToLine(Loc);
477
478  OS << "#pragma clang __debug ";
479  OS << DebugType;
480
481  setEmittedDirectiveOnThisLine();
482}
483
484void PrintPPOutputPPCallbacks::
485PragmaDiagnosticPush(SourceLocation Loc, StringRef Namespace) {
486  startNewLineIfNeeded();
487  MoveToLine(Loc);
488  OS << "#pragma " << Namespace << " diagnostic push";
489  setEmittedDirectiveOnThisLine();
490}
491
492void PrintPPOutputPPCallbacks::
493PragmaDiagnosticPop(SourceLocation Loc, StringRef Namespace) {
494  startNewLineIfNeeded();
495  MoveToLine(Loc);
496  OS << "#pragma " << Namespace << " diagnostic pop";
497  setEmittedDirectiveOnThisLine();
498}
499
500void PrintPPOutputPPCallbacks::PragmaDiagnostic(SourceLocation Loc,
501                                                StringRef Namespace,
502                                                diag::Severity Map,
503                                                StringRef Str) {
504  startNewLineIfNeeded();
505  MoveToLine(Loc);
506  OS << "#pragma " << Namespace << " diagnostic ";
507  switch (Map) {
508  case diag::Severity::Remark:
509    OS << "remark";
510    break;
511  case diag::Severity::Warning:
512    OS << "warning";
513    break;
514  case diag::Severity::Error:
515    OS << "error";
516    break;
517  case diag::Severity::Ignored:
518    OS << "ignored";
519    break;
520  case diag::Severity::Fatal:
521    OS << "fatal";
522    break;
523  }
524  OS << " \"" << Str << '"';
525  setEmittedDirectiveOnThisLine();
526}
527
528void PrintPPOutputPPCallbacks::PragmaWarning(SourceLocation Loc,
529                                             StringRef WarningSpec,
530                                             ArrayRef<int> Ids) {
531  startNewLineIfNeeded();
532  MoveToLine(Loc);
533  OS << "#pragma warning(" << WarningSpec << ':';
534  for (ArrayRef<int>::iterator I = Ids.begin(), E = Ids.end(); I != E; ++I)
535    OS << ' ' << *I;
536  OS << ')';
537  setEmittedDirectiveOnThisLine();
538}
539
540void PrintPPOutputPPCallbacks::PragmaWarningPush(SourceLocation Loc,
541                                                 int Level) {
542  startNewLineIfNeeded();
543  MoveToLine(Loc);
544  OS << "#pragma warning(push";
545  if (Level >= 0)
546    OS << ", " << Level;
547  OS << ')';
548  setEmittedDirectiveOnThisLine();
549}
550
551void PrintPPOutputPPCallbacks::PragmaWarningPop(SourceLocation Loc) {
552  startNewLineIfNeeded();
553  MoveToLine(Loc);
554  OS << "#pragma warning(pop)";
555  setEmittedDirectiveOnThisLine();
556}
557
558void PrintPPOutputPPCallbacks::PragmaExecCharsetPush(SourceLocation Loc,
559                                                     StringRef Str) {
560  startNewLineIfNeeded();
561  MoveToLine(Loc);
562  OS << "#pragma character_execution_set(push";
563  if (!Str.empty())
564    OS << ", " << Str;
565  OS << ')';
566  setEmittedDirectiveOnThisLine();
567}
568
569void PrintPPOutputPPCallbacks::PragmaExecCharsetPop(SourceLocation Loc) {
570  startNewLineIfNeeded();
571  MoveToLine(Loc);
572  OS << "#pragma character_execution_set(pop)";
573  setEmittedDirectiveOnThisLine();
574}
575
576void PrintPPOutputPPCallbacks::
577PragmaAssumeNonNullBegin(SourceLocation Loc) {
578  startNewLineIfNeeded();
579  MoveToLine(Loc);
580  OS << "#pragma clang assume_nonnull begin";
581  setEmittedDirectiveOnThisLine();
582}
583
584void PrintPPOutputPPCallbacks::
585PragmaAssumeNonNullEnd(SourceLocation Loc) {
586  startNewLineIfNeeded();
587  MoveToLine(Loc);
588  OS << "#pragma clang assume_nonnull end";
589  setEmittedDirectiveOnThisLine();
590}
591
592/// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
593/// is called for the first token on each new line.  If this really is the start
594/// of a new logical line, handle it and return true, otherwise return false.
595/// This may not be the start of a logical line because the "start of line"
596/// marker is set for spelling lines, not expansion ones.
597bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
598  // Figure out what line we went to and insert the appropriate number of
599  // newline characters.
600  if (!MoveToLine(Tok.getLocation()))
601    return false;
602
603  // Print out space characters so that the first token on a line is
604  // indented for easy reading.
605  unsigned ColNo = SM.getExpansionColumnNumber(Tok.getLocation());
606
607  // The first token on a line can have a column number of 1, yet still expect
608  // leading white space, if a macro expansion in column 1 starts with an empty
609  // macro argument, or an empty nested macro expansion. In this case, move the
610  // token to column 2.
611  if (ColNo == 1 && Tok.hasLeadingSpace())
612    ColNo = 2;
613
614  // This hack prevents stuff like:
615  // #define HASH #
616  // HASH define foo bar
617  // From having the # character end up at column 1, which makes it so it
618  // is not handled as a #define next time through the preprocessor if in
619  // -fpreprocessed mode.
620  if (ColNo <= 1 && Tok.is(tok::hash))
621    OS << ' ';
622
623  // Otherwise, indent the appropriate number of spaces.
624  for (; ColNo > 1; --ColNo)
625    OS << ' ';
626
627  return true;
628}
629
630void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
631                                                     unsigned Len) {
632  unsigned NumNewlines = 0;
633  for (; Len; --Len, ++TokStr) {
634    if (*TokStr != '\n' &&
635        *TokStr != '\r')
636      continue;
637
638    ++NumNewlines;
639
640    // If we have \n\r or \r\n, skip both and count as one line.
641    if (Len != 1 &&
642        (TokStr[1] == '\n' || TokStr[1] == '\r') &&
643        TokStr[0] != TokStr[1]) {
644      ++TokStr;
645      --Len;
646    }
647  }
648
649  if (NumNewlines == 0) return;
650
651  CurLine += NumNewlines;
652}
653
654
655namespace {
656struct UnknownPragmaHandler : public PragmaHandler {
657  const char *Prefix;
658  PrintPPOutputPPCallbacks *Callbacks;
659
660  // Set to true if tokens should be expanded
661  bool ShouldExpandTokens;
662
663  UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks,
664                       bool RequireTokenExpansion)
665      : Prefix(prefix), Callbacks(callbacks),
666        ShouldExpandTokens(RequireTokenExpansion) {}
667  void HandlePragma(Preprocessor &PP, PragmaIntroducer Introducer,
668                    Token &PragmaTok) override {
669    // Figure out what line we went to and insert the appropriate number of
670    // newline characters.
671    Callbacks->startNewLineIfNeeded();
672    Callbacks->MoveToLine(PragmaTok.getLocation());
673    Callbacks->OS.write(Prefix, strlen(Prefix));
674
675    if (ShouldExpandTokens) {
676      // The first token does not have expanded macros. Expand them, if
677      // required.
678      auto Toks = llvm::make_unique<Token[]>(1);
679      Toks[0] = PragmaTok;
680      PP.EnterTokenStream(std::move(Toks), /*NumToks=*/1,
681                          /*DisableMacroExpansion=*/false,
682                          /*IsReinject=*/false);
683      PP.Lex(PragmaTok);
684    }
685    Token PrevToken;
686    Token PrevPrevToken;
687    PrevToken.startToken();
688    PrevPrevToken.startToken();
689
690    // Read and print all of the pragma tokens.
691    while (PragmaTok.isNot(tok::eod)) {
692      if (PragmaTok.hasLeadingSpace() ||
693          Callbacks->AvoidConcat(PrevPrevToken, PrevToken, PragmaTok))
694        Callbacks->OS << ' ';
695      std::string TokSpell = PP.getSpelling(PragmaTok);
696      Callbacks->OS.write(&TokSpell[0], TokSpell.size());
697
698      PrevPrevToken = PrevToken;
699      PrevToken = PragmaTok;
700
701      if (ShouldExpandTokens)
702        PP.Lex(PragmaTok);
703      else
704        PP.LexUnexpandedToken(PragmaTok);
705    }
706    Callbacks->setEmittedDirectiveOnThisLine();
707  }
708};
709} // end anonymous namespace
710
711
712static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
713                                    PrintPPOutputPPCallbacks *Callbacks,
714                                    raw_ostream &OS) {
715  bool DropComments = PP.getLangOpts().TraditionalCPP &&
716                      !PP.getCommentRetentionState();
717
718  char Buffer[256];
719  Token PrevPrevTok, PrevTok;
720  PrevPrevTok.startToken();
721  PrevTok.startToken();
722  while (1) {
723    if (Callbacks->hasEmittedDirectiveOnThisLine()) {
724      Callbacks->startNewLineIfNeeded();
725      Callbacks->MoveToLine(Tok.getLocation());
726    }
727
728    // If this token is at the start of a line, emit newlines if needed.
729    if (Tok.isAtStartOfLine() && Callbacks->HandleFirstTokOnLine(Tok)) {
730      // done.
731    } else if (Tok.hasLeadingSpace() ||
732               // If we haven't emitted a token on this line yet, PrevTok isn't
733               // useful to look at and no concatenation could happen anyway.
734               (Callbacks->hasEmittedTokensOnThisLine() &&
735                // Don't print "-" next to "-", it would form "--".
736                Callbacks->AvoidConcat(PrevPrevTok, PrevTok, Tok))) {
737      OS << ' ';
738    }
739
740    if (DropComments && Tok.is(tok::comment)) {
741      // Skip comments. Normally the preprocessor does not generate
742      // tok::comment nodes at all when not keeping comments, but under
743      // -traditional-cpp the lexer keeps /all/ whitespace, including comments.
744      SourceLocation StartLoc = Tok.getLocation();
745      Callbacks->MoveToLine(StartLoc.getLocWithOffset(Tok.getLength()));
746    } else if (Tok.is(tok::eod)) {
747      // Don't print end of directive tokens, since they are typically newlines
748      // that mess up our line tracking. These come from unknown pre-processor
749      // directives or hash-prefixed comments in standalone assembly files.
750      PP.Lex(Tok);
751      continue;
752    } else if (Tok.is(tok::annot_module_include)) {
753      // PrintPPOutputPPCallbacks::InclusionDirective handles producing
754      // appropriate output here. Ignore this token entirely.
755      PP.Lex(Tok);
756      continue;
757    } else if (Tok.is(tok::annot_module_begin)) {
758      // FIXME: We retrieve this token after the FileChanged callback, and
759      // retrieve the module_end token before the FileChanged callback, so
760      // we render this within the file and render the module end outside the
761      // file, but this is backwards from the token locations: the module_begin
762      // token is at the include location (outside the file) and the module_end
763      // token is at the EOF location (within the file).
764      Callbacks->BeginModule(
765          reinterpret_cast<Module *>(Tok.getAnnotationValue()));
766      PP.Lex(Tok);
767      continue;
768    } else if (Tok.is(tok::annot_module_end)) {
769      Callbacks->EndModule(
770          reinterpret_cast<Module *>(Tok.getAnnotationValue()));
771      PP.Lex(Tok);
772      continue;
773    } else if (Tok.is(tok::annot_header_unit)) {
774      // This is a header-name that has been (effectively) converted into a
775      // module-name.
776      // FIXME: The module name could contain non-identifier module name
777      // components. We don't have a good way to round-trip those.
778      Module *M = reinterpret_cast<Module *>(Tok.getAnnotationValue());
779      std::string Name = M->getFullModuleName();
780      OS.write(Name.data(), Name.size());
781      Callbacks->HandleNewlinesInToken(Name.data(), Name.size());
782    } else if (Tok.isAnnotation()) {
783      // Ignore annotation tokens created by pragmas - the pragmas themselves
784      // will be reproduced in the preprocessed output.
785      PP.Lex(Tok);
786      continue;
787    } else if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
788      OS << II->getName();
789    } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
790               Tok.getLiteralData()) {
791      OS.write(Tok.getLiteralData(), Tok.getLength());
792    } else if (Tok.getLength() < llvm::array_lengthof(Buffer)) {
793      const char *TokPtr = Buffer;
794      unsigned Len = PP.getSpelling(Tok, TokPtr);
795      OS.write(TokPtr, Len);
796
797      // Tokens that can contain embedded newlines need to adjust our current
798      // line number.
799      if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
800        Callbacks->HandleNewlinesInToken(TokPtr, Len);
801    } else {
802      std::string S = PP.getSpelling(Tok);
803      OS.write(S.data(), S.size());
804
805      // Tokens that can contain embedded newlines need to adjust our current
806      // line number.
807      if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
808        Callbacks->HandleNewlinesInToken(S.data(), S.size());
809    }
810    Callbacks->setEmittedTokensOnThisLine();
811
812    if (Tok.is(tok::eof)) break;
813
814    PrevPrevTok = PrevTok;
815    PrevTok = Tok;
816    PP.Lex(Tok);
817  }
818}
819
820typedef std::pair<const IdentifierInfo *, MacroInfo *> id_macro_pair;
821static int MacroIDCompare(const id_macro_pair *LHS, const id_macro_pair *RHS) {
822  return LHS->first->getName().compare(RHS->first->getName());
823}
824
825static void DoPrintMacros(Preprocessor &PP, raw_ostream *OS) {
826  // Ignore unknown pragmas.
827  PP.IgnorePragmas();
828
829  // -dM mode just scans and ignores all tokens in the files, then dumps out
830  // the macro table at the end.
831  PP.EnterMainSourceFile();
832
833  Token Tok;
834  do PP.Lex(Tok);
835  while (Tok.isNot(tok::eof));
836
837  SmallVector<id_macro_pair, 128> MacrosByID;
838  for (Preprocessor::macro_iterator I = PP.macro_begin(), E = PP.macro_end();
839       I != E; ++I) {
840    auto *MD = I->second.getLatest();
841    if (MD && MD->isDefined())
842      MacrosByID.push_back(id_macro_pair(I->first, MD->getMacroInfo()));
843  }
844  llvm::array_pod_sort(MacrosByID.begin(), MacrosByID.end(), MacroIDCompare);
845
846  for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
847    MacroInfo &MI = *MacrosByID[i].second;
848    // Ignore computed macros like __LINE__ and friends.
849    if (MI.isBuiltinMacro()) continue;
850
851    PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
852    *OS << '\n';
853  }
854}
855
856/// DoPrintPreprocessedInput - This implements -E mode.
857///
858void clang::DoPrintPreprocessedInput(Preprocessor &PP, raw_ostream *OS,
859                                     const PreprocessorOutputOptions &Opts) {
860  // Show macros with no output is handled specially.
861  if (!Opts.ShowCPP) {
862    assert(Opts.ShowMacros && "Not yet implemented!");
863    DoPrintMacros(PP, OS);
864    return;
865  }
866
867  // Inform the preprocessor whether we want it to retain comments or not, due
868  // to -C or -CC.
869  PP.SetCommentRetentionState(Opts.ShowComments, Opts.ShowMacroComments);
870
871  PrintPPOutputPPCallbacks *Callbacks = new PrintPPOutputPPCallbacks(
872      PP, *OS, !Opts.ShowLineMarkers, Opts.ShowMacros,
873      Opts.ShowIncludeDirectives, Opts.UseLineDirectives);
874
875  // Expand macros in pragmas with -fms-extensions.  The assumption is that
876  // the majority of pragmas in such a file will be Microsoft pragmas.
877  // Remember the handlers we will add so that we can remove them later.
878  std::unique_ptr<UnknownPragmaHandler> MicrosoftExtHandler(
879      new UnknownPragmaHandler(
880          "#pragma", Callbacks,
881          /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
882
883  std::unique_ptr<UnknownPragmaHandler> GCCHandler(new UnknownPragmaHandler(
884      "#pragma GCC", Callbacks,
885      /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
886
887  std::unique_ptr<UnknownPragmaHandler> ClangHandler(new UnknownPragmaHandler(
888      "#pragma clang", Callbacks,
889      /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
890
891  PP.AddPragmaHandler(MicrosoftExtHandler.get());
892  PP.AddPragmaHandler("GCC", GCCHandler.get());
893  PP.AddPragmaHandler("clang", ClangHandler.get());
894
895  // The tokens after pragma omp need to be expanded.
896  //
897  //  OpenMP [2.1, Directive format]
898  //  Preprocessing tokens following the #pragma omp are subject to macro
899  //  replacement.
900  std::unique_ptr<UnknownPragmaHandler> OpenMPHandler(
901      new UnknownPragmaHandler("#pragma omp", Callbacks,
902                               /*RequireTokenExpansion=*/true));
903  PP.AddPragmaHandler("omp", OpenMPHandler.get());
904
905  PP.addPPCallbacks(std::unique_ptr<PPCallbacks>(Callbacks));
906
907  // After we have configured the preprocessor, enter the main file.
908  PP.EnterMainSourceFile();
909
910  // Consume all of the tokens that come from the predefines buffer.  Those
911  // should not be emitted into the output and are guaranteed to be at the
912  // start.
913  const SourceManager &SourceMgr = PP.getSourceManager();
914  Token Tok;
915  do {
916    PP.Lex(Tok);
917    if (Tok.is(tok::eof) || !Tok.getLocation().isFileID())
918      break;
919
920    PresumedLoc PLoc = SourceMgr.getPresumedLoc(Tok.getLocation());
921    if (PLoc.isInvalid())
922      break;
923
924    if (strcmp(PLoc.getFilename(), "<built-in>"))
925      break;
926  } while (true);
927
928  // Read all the preprocessed tokens, printing them out to the stream.
929  PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
930  *OS << '\n';
931
932  // Remove the handlers we just added to leave the preprocessor in a sane state
933  // so that it can be reused (for example by a clang::Parser instance).
934  PP.RemovePragmaHandler(MicrosoftExtHandler.get());
935  PP.RemovePragmaHandler("GCC", GCCHandler.get());
936  PP.RemovePragmaHandler("clang", ClangHandler.get());
937  PP.RemovePragmaHandler("omp", OpenMPHandler.get());
938}
939