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;
98  bool MinimizeWhitespace;
99  bool DirectivesOnly;
100
101  Token PrevTok;
102  Token PrevPrevTok;
103
104public:
105  PrintPPOutputPPCallbacks(Preprocessor &pp, raw_ostream &os, bool lineMarkers,
106                           bool defines, bool DumpIncludeDirectives,
107                           bool UseLineDirectives, bool MinimizeWhitespace,
108                           bool DirectivesOnly)
109      : PP(pp), SM(PP.getSourceManager()), ConcatInfo(PP), OS(os),
110        DisableLineMarkers(lineMarkers), DumpDefines(defines),
111        DumpIncludeDirectives(DumpIncludeDirectives),
112        UseLineDirectives(UseLineDirectives),
113        MinimizeWhitespace(MinimizeWhitespace), DirectivesOnly(DirectivesOnly) {
114    CurLine = 0;
115    CurFilename += "<uninit>";
116    EmittedTokensOnThisLine = false;
117    EmittedDirectiveOnThisLine = false;
118    FileType = SrcMgr::C_User;
119    Initialized = false;
120    IsFirstFileEntered = false;
121
122    PrevTok.startToken();
123    PrevPrevTok.startToken();
124  }
125
126  bool isMinimizeWhitespace() const { return MinimizeWhitespace; }
127
128  void setEmittedTokensOnThisLine() { EmittedTokensOnThisLine = true; }
129  bool hasEmittedTokensOnThisLine() const { return EmittedTokensOnThisLine; }
130
131  void setEmittedDirectiveOnThisLine() { EmittedDirectiveOnThisLine = true; }
132  bool hasEmittedDirectiveOnThisLine() const {
133    return EmittedDirectiveOnThisLine;
134  }
135
136  /// Ensure that the output stream position is at the beginning of a new line
137  /// and inserts one if it does not. It is intended to ensure that directives
138  /// inserted by the directives not from the input source (such as #line) are
139  /// in the first column. To insert newlines that represent the input, use
140  /// MoveToLine(/*...*/, /*RequireStartOfLine=*/true).
141  void startNewLineIfNeeded();
142
143  void FileChanged(SourceLocation Loc, FileChangeReason Reason,
144                   SrcMgr::CharacteristicKind FileType,
145                   FileID PrevFID) override;
146  void InclusionDirective(SourceLocation HashLoc, const Token &IncludeTok,
147                          StringRef FileName, bool IsAngled,
148                          CharSourceRange FilenameRange,
149                          OptionalFileEntryRef File, StringRef SearchPath,
150                          StringRef RelativePath, const Module *Imported,
151                          SrcMgr::CharacteristicKind FileType) override;
152  void Ident(SourceLocation Loc, StringRef str) override;
153  void PragmaMessage(SourceLocation Loc, StringRef Namespace,
154                     PragmaMessageKind Kind, StringRef Str) override;
155  void PragmaDebug(SourceLocation Loc, StringRef DebugType) override;
156  void PragmaDiagnosticPush(SourceLocation Loc, StringRef Namespace) override;
157  void PragmaDiagnosticPop(SourceLocation Loc, StringRef Namespace) override;
158  void PragmaDiagnostic(SourceLocation Loc, StringRef Namespace,
159                        diag::Severity Map, StringRef Str) override;
160  void PragmaWarning(SourceLocation Loc, PragmaWarningSpecifier WarningSpec,
161                     ArrayRef<int> Ids) override;
162  void PragmaWarningPush(SourceLocation Loc, int Level) override;
163  void PragmaWarningPop(SourceLocation Loc) override;
164  void PragmaExecCharsetPush(SourceLocation Loc, StringRef Str) override;
165  void PragmaExecCharsetPop(SourceLocation Loc) override;
166  void PragmaAssumeNonNullBegin(SourceLocation Loc) override;
167  void PragmaAssumeNonNullEnd(SourceLocation Loc) override;
168
169  /// Insert whitespace before emitting the next token.
170  ///
171  /// @param Tok             Next token to be emitted.
172  /// @param RequireSpace    Ensure at least one whitespace is emitted. Useful
173  ///                        if non-tokens have been emitted to the stream.
174  /// @param RequireSameLine Never emit newlines. Useful when semantics depend
175  ///                        on being on the same line, such as directives.
176  void HandleWhitespaceBeforeTok(const Token &Tok, bool RequireSpace,
177                                 bool RequireSameLine);
178
179  /// Move to the line of the provided source location. This will
180  /// return true if a newline was inserted or if
181  /// the requested location is the first token on the first line.
182  /// In these cases the next output will be the first column on the line and
183  /// make it possible to insert indention. The newline was inserted
184  /// implicitly when at the beginning of the file.
185  ///
186  /// @param Tok                 Token where to move to.
187  /// @param RequireStartOfLine  Whether the next line depends on being in the
188  ///                            first column, such as a directive.
189  ///
190  /// @return Whether column adjustments are necessary.
191  bool MoveToLine(const Token &Tok, bool RequireStartOfLine) {
192    PresumedLoc PLoc = SM.getPresumedLoc(Tok.getLocation());
193    unsigned TargetLine = PLoc.isValid() ? PLoc.getLine() : CurLine;
194    bool IsFirstInFile =
195        Tok.isAtStartOfLine() && PLoc.isValid() && PLoc.getLine() == 1;
196    return MoveToLine(TargetLine, RequireStartOfLine) || IsFirstInFile;
197  }
198
199  /// Move to the line of the provided source location. Returns true if a new
200  /// line was inserted.
201  bool MoveToLine(SourceLocation Loc, bool RequireStartOfLine) {
202    PresumedLoc PLoc = SM.getPresumedLoc(Loc);
203    unsigned TargetLine = PLoc.isValid() ? PLoc.getLine() : CurLine;
204    return MoveToLine(TargetLine, RequireStartOfLine);
205  }
206  bool MoveToLine(unsigned LineNo, bool RequireStartOfLine);
207
208  bool AvoidConcat(const Token &PrevPrevTok, const Token &PrevTok,
209                   const Token &Tok) {
210    return ConcatInfo.AvoidConcat(PrevPrevTok, PrevTok, Tok);
211  }
212  void WriteLineInfo(unsigned LineNo, const char *Extra=nullptr,
213                     unsigned ExtraLen=0);
214  bool LineMarkersAreDisabled() const { return DisableLineMarkers; }
215  void HandleNewlinesInToken(const char *TokStr, unsigned Len);
216
217  /// MacroDefined - This hook is called whenever a macro definition is seen.
218  void MacroDefined(const Token &MacroNameTok,
219                    const MacroDirective *MD) override;
220
221  /// MacroUndefined - This hook is called whenever a macro #undef is seen.
222  void MacroUndefined(const Token &MacroNameTok,
223                      const MacroDefinition &MD,
224                      const MacroDirective *Undef) override;
225
226  void BeginModule(const Module *M);
227  void EndModule(const Module *M);
228};
229}  // end anonymous namespace
230
231void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
232                                             const char *Extra,
233                                             unsigned ExtraLen) {
234  startNewLineIfNeeded();
235
236  // Emit #line directives or GNU line markers depending on what mode we're in.
237  if (UseLineDirectives) {
238    OS << "#line" << ' ' << LineNo << ' ' << '"';
239    OS.write_escaped(CurFilename);
240    OS << '"';
241  } else {
242    OS << '#' << ' ' << LineNo << ' ' << '"';
243    OS.write_escaped(CurFilename);
244    OS << '"';
245
246    if (ExtraLen)
247      OS.write(Extra, ExtraLen);
248
249    if (FileType == SrcMgr::C_System)
250      OS.write(" 3", 2);
251    else if (FileType == SrcMgr::C_ExternCSystem)
252      OS.write(" 3 4", 4);
253  }
254  OS << '\n';
255}
256
257/// MoveToLine - Move the output to the source line specified by the location
258/// object.  We can do this by emitting some number of \n's, or be emitting a
259/// #line directive.  This returns false if already at the specified line, true
260/// if some newlines were emitted.
261bool PrintPPOutputPPCallbacks::MoveToLine(unsigned LineNo,
262                                          bool RequireStartOfLine) {
263  // If it is required to start a new line or finish the current, insert
264  // vertical whitespace now and take it into account when moving to the
265  // expected line.
266  bool StartedNewLine = false;
267  if ((RequireStartOfLine && EmittedTokensOnThisLine) ||
268      EmittedDirectiveOnThisLine) {
269    OS << '\n';
270    StartedNewLine = true;
271    CurLine += 1;
272    EmittedTokensOnThisLine = false;
273    EmittedDirectiveOnThisLine = false;
274  }
275
276  // If this line is "close enough" to the original line, just print newlines,
277  // otherwise print a #line directive.
278  if (CurLine == LineNo) {
279    // Nothing to do if we are already on the correct line.
280  } else if (MinimizeWhitespace && DisableLineMarkers) {
281    // With -E -P -fminimize-whitespace, don't emit anything if not necessary.
282  } else if (!StartedNewLine && LineNo - CurLine == 1) {
283    // Printing a single line has priority over printing a #line directive, even
284    // when minimizing whitespace which otherwise would print #line directives
285    // for every single line.
286    OS << '\n';
287    StartedNewLine = true;
288  } else if (!DisableLineMarkers) {
289    if (LineNo - CurLine <= 8) {
290      const char *NewLines = "\n\n\n\n\n\n\n\n";
291      OS.write(NewLines, LineNo - CurLine);
292    } else {
293      // Emit a #line or line marker.
294      WriteLineInfo(LineNo, nullptr, 0);
295    }
296    StartedNewLine = true;
297  } else if (EmittedTokensOnThisLine) {
298    // If we are not on the correct line and don't need to be line-correct,
299    // at least ensure we start on a new line.
300    OS << '\n';
301    StartedNewLine = true;
302  }
303
304  if (StartedNewLine) {
305    EmittedTokensOnThisLine = false;
306    EmittedDirectiveOnThisLine = false;
307  }
308
309  CurLine = LineNo;
310  return StartedNewLine;
311}
312
313void PrintPPOutputPPCallbacks::startNewLineIfNeeded() {
314  if (EmittedTokensOnThisLine || EmittedDirectiveOnThisLine) {
315    OS << '\n';
316    EmittedTokensOnThisLine = false;
317    EmittedDirectiveOnThisLine = false;
318  }
319}
320
321/// FileChanged - Whenever the preprocessor enters or exits a #include file
322/// it invokes this handler.  Update our conception of the current source
323/// position.
324void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
325                                           FileChangeReason Reason,
326                                       SrcMgr::CharacteristicKind NewFileType,
327                                       FileID PrevFID) {
328  // Unless we are exiting a #include, make sure to skip ahead to the line the
329  // #include directive was at.
330  SourceManager &SourceMgr = SM;
331
332  PresumedLoc UserLoc = SourceMgr.getPresumedLoc(Loc);
333  if (UserLoc.isInvalid())
334    return;
335
336  unsigned NewLine = UserLoc.getLine();
337
338  if (Reason == PPCallbacks::EnterFile) {
339    SourceLocation IncludeLoc = UserLoc.getIncludeLoc();
340    if (IncludeLoc.isValid())
341      MoveToLine(IncludeLoc, /*RequireStartOfLine=*/false);
342  } else if (Reason == PPCallbacks::SystemHeaderPragma) {
343    // GCC emits the # directive for this directive on the line AFTER the
344    // directive and emits a bunch of spaces that aren't needed. This is because
345    // otherwise we will emit a line marker for THIS line, which requires an
346    // extra blank line after the directive to avoid making all following lines
347    // off by one. We can do better by simply incrementing NewLine here.
348    NewLine += 1;
349  }
350
351  CurLine = NewLine;
352
353  CurFilename.clear();
354  CurFilename += UserLoc.getFilename();
355  FileType = NewFileType;
356
357  if (DisableLineMarkers) {
358    if (!MinimizeWhitespace)
359      startNewLineIfNeeded();
360    return;
361  }
362
363  if (!Initialized) {
364    WriteLineInfo(CurLine);
365    Initialized = true;
366  }
367
368  // Do not emit an enter marker for the main file (which we expect is the first
369  // entered file). This matches gcc, and improves compatibility with some tools
370  // which track the # line markers as a way to determine when the preprocessed
371  // output is in the context of the main file.
372  if (Reason == PPCallbacks::EnterFile && !IsFirstFileEntered) {
373    IsFirstFileEntered = true;
374    return;
375  }
376
377  switch (Reason) {
378  case PPCallbacks::EnterFile:
379    WriteLineInfo(CurLine, " 1", 2);
380    break;
381  case PPCallbacks::ExitFile:
382    WriteLineInfo(CurLine, " 2", 2);
383    break;
384  case PPCallbacks::SystemHeaderPragma:
385  case PPCallbacks::RenameFile:
386    WriteLineInfo(CurLine);
387    break;
388  }
389}
390
391void PrintPPOutputPPCallbacks::InclusionDirective(
392    SourceLocation HashLoc, const Token &IncludeTok, StringRef FileName,
393    bool IsAngled, CharSourceRange FilenameRange, OptionalFileEntryRef File,
394    StringRef SearchPath, StringRef RelativePath, const Module *Imported,
395    SrcMgr::CharacteristicKind FileType) {
396  // In -dI mode, dump #include directives prior to dumping their content or
397  // interpretation.
398  if (DumpIncludeDirectives) {
399    MoveToLine(HashLoc, /*RequireStartOfLine=*/true);
400    const std::string TokenText = PP.getSpelling(IncludeTok);
401    assert(!TokenText.empty());
402    OS << "#" << TokenText << " "
403       << (IsAngled ? '<' : '"') << FileName << (IsAngled ? '>' : '"')
404       << " /* clang -E -dI */";
405    setEmittedDirectiveOnThisLine();
406  }
407
408  // When preprocessing, turn implicit imports into module import pragmas.
409  if (Imported) {
410    switch (IncludeTok.getIdentifierInfo()->getPPKeywordID()) {
411    case tok::pp_include:
412    case tok::pp_import:
413    case tok::pp_include_next:
414      MoveToLine(HashLoc, /*RequireStartOfLine=*/true);
415      OS << "#pragma clang module import " << Imported->getFullModuleName(true)
416         << " /* clang -E: implicit import for "
417         << "#" << PP.getSpelling(IncludeTok) << " "
418         << (IsAngled ? '<' : '"') << FileName << (IsAngled ? '>' : '"')
419         << " */";
420      setEmittedDirectiveOnThisLine();
421      break;
422
423    case tok::pp___include_macros:
424      // #__include_macros has no effect on a user of a preprocessed source
425      // file; the only effect is on preprocessing.
426      //
427      // FIXME: That's not *quite* true: it causes the module in question to
428      // be loaded, which can affect downstream diagnostics.
429      break;
430
431    default:
432      llvm_unreachable("unknown include directive kind");
433      break;
434    }
435  }
436}
437
438/// Handle entering the scope of a module during a module compilation.
439void PrintPPOutputPPCallbacks::BeginModule(const Module *M) {
440  startNewLineIfNeeded();
441  OS << "#pragma clang module begin " << M->getFullModuleName(true);
442  setEmittedDirectiveOnThisLine();
443}
444
445/// Handle leaving the scope of a module during a module compilation.
446void PrintPPOutputPPCallbacks::EndModule(const Module *M) {
447  startNewLineIfNeeded();
448  OS << "#pragma clang module end /*" << M->getFullModuleName(true) << "*/";
449  setEmittedDirectiveOnThisLine();
450}
451
452/// Ident - Handle #ident directives when read by the preprocessor.
453///
454void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, StringRef S) {
455  MoveToLine(Loc, /*RequireStartOfLine=*/true);
456
457  OS.write("#ident ", strlen("#ident "));
458  OS.write(S.begin(), S.size());
459  setEmittedTokensOnThisLine();
460}
461
462/// MacroDefined - This hook is called whenever a macro definition is seen.
463void PrintPPOutputPPCallbacks::MacroDefined(const Token &MacroNameTok,
464                                            const MacroDirective *MD) {
465  const MacroInfo *MI = MD->getMacroInfo();
466  // Print out macro definitions in -dD mode and when we have -fdirectives-only
467  // for C++20 header units.
468  if ((!DumpDefines && !DirectivesOnly) ||
469      // Ignore __FILE__ etc.
470      MI->isBuiltinMacro())
471    return;
472
473  SourceLocation DefLoc = MI->getDefinitionLoc();
474  if (DirectivesOnly && !MI->isUsed()) {
475    SourceManager &SM = PP.getSourceManager();
476    if (SM.isWrittenInBuiltinFile(DefLoc) ||
477        SM.isWrittenInCommandLineFile(DefLoc))
478      return;
479  }
480  MoveToLine(DefLoc, /*RequireStartOfLine=*/true);
481  PrintMacroDefinition(*MacroNameTok.getIdentifierInfo(), *MI, PP, OS);
482  setEmittedDirectiveOnThisLine();
483}
484
485void PrintPPOutputPPCallbacks::MacroUndefined(const Token &MacroNameTok,
486                                              const MacroDefinition &MD,
487                                              const MacroDirective *Undef) {
488  // Print out macro definitions in -dD mode and when we have -fdirectives-only
489  // for C++20 header units.
490  if (!DumpDefines && !DirectivesOnly)
491    return;
492
493  MoveToLine(MacroNameTok.getLocation(), /*RequireStartOfLine=*/true);
494  OS << "#undef " << MacroNameTok.getIdentifierInfo()->getName();
495  setEmittedDirectiveOnThisLine();
496}
497
498static void outputPrintable(raw_ostream &OS, StringRef Str) {
499  for (unsigned char Char : Str) {
500    if (isPrintable(Char) && Char != '\\' && Char != '"')
501      OS << (char)Char;
502    else // Output anything hard as an octal escape.
503      OS << '\\'
504         << (char)('0' + ((Char >> 6) & 7))
505         << (char)('0' + ((Char >> 3) & 7))
506         << (char)('0' + ((Char >> 0) & 7));
507  }
508}
509
510void PrintPPOutputPPCallbacks::PragmaMessage(SourceLocation Loc,
511                                             StringRef Namespace,
512                                             PragmaMessageKind Kind,
513                                             StringRef Str) {
514  MoveToLine(Loc, /*RequireStartOfLine=*/true);
515  OS << "#pragma ";
516  if (!Namespace.empty())
517    OS << Namespace << ' ';
518  switch (Kind) {
519    case PMK_Message:
520      OS << "message(\"";
521      break;
522    case PMK_Warning:
523      OS << "warning \"";
524      break;
525    case PMK_Error:
526      OS << "error \"";
527      break;
528  }
529
530  outputPrintable(OS, Str);
531  OS << '"';
532  if (Kind == PMK_Message)
533    OS << ')';
534  setEmittedDirectiveOnThisLine();
535}
536
537void PrintPPOutputPPCallbacks::PragmaDebug(SourceLocation Loc,
538                                           StringRef DebugType) {
539  MoveToLine(Loc, /*RequireStartOfLine=*/true);
540
541  OS << "#pragma clang __debug ";
542  OS << DebugType;
543
544  setEmittedDirectiveOnThisLine();
545}
546
547void PrintPPOutputPPCallbacks::
548PragmaDiagnosticPush(SourceLocation Loc, StringRef Namespace) {
549  MoveToLine(Loc, /*RequireStartOfLine=*/true);
550  OS << "#pragma " << Namespace << " diagnostic push";
551  setEmittedDirectiveOnThisLine();
552}
553
554void PrintPPOutputPPCallbacks::
555PragmaDiagnosticPop(SourceLocation Loc, StringRef Namespace) {
556  MoveToLine(Loc, /*RequireStartOfLine=*/true);
557  OS << "#pragma " << Namespace << " diagnostic pop";
558  setEmittedDirectiveOnThisLine();
559}
560
561void PrintPPOutputPPCallbacks::PragmaDiagnostic(SourceLocation Loc,
562                                                StringRef Namespace,
563                                                diag::Severity Map,
564                                                StringRef Str) {
565  MoveToLine(Loc, /*RequireStartOfLine=*/true);
566  OS << "#pragma " << Namespace << " diagnostic ";
567  switch (Map) {
568  case diag::Severity::Remark:
569    OS << "remark";
570    break;
571  case diag::Severity::Warning:
572    OS << "warning";
573    break;
574  case diag::Severity::Error:
575    OS << "error";
576    break;
577  case diag::Severity::Ignored:
578    OS << "ignored";
579    break;
580  case diag::Severity::Fatal:
581    OS << "fatal";
582    break;
583  }
584  OS << " \"" << Str << '"';
585  setEmittedDirectiveOnThisLine();
586}
587
588void PrintPPOutputPPCallbacks::PragmaWarning(SourceLocation Loc,
589                                             PragmaWarningSpecifier WarningSpec,
590                                             ArrayRef<int> Ids) {
591  MoveToLine(Loc, /*RequireStartOfLine=*/true);
592
593  OS << "#pragma warning(";
594  switch(WarningSpec) {
595    case PWS_Default:  OS << "default"; break;
596    case PWS_Disable:  OS << "disable"; break;
597    case PWS_Error:    OS << "error"; break;
598    case PWS_Once:     OS << "once"; break;
599    case PWS_Suppress: OS << "suppress"; break;
600    case PWS_Level1:   OS << '1'; break;
601    case PWS_Level2:   OS << '2'; break;
602    case PWS_Level3:   OS << '3'; break;
603    case PWS_Level4:   OS << '4'; break;
604  }
605  OS << ':';
606
607  for (ArrayRef<int>::iterator I = Ids.begin(), E = Ids.end(); I != E; ++I)
608    OS << ' ' << *I;
609  OS << ')';
610  setEmittedDirectiveOnThisLine();
611}
612
613void PrintPPOutputPPCallbacks::PragmaWarningPush(SourceLocation Loc,
614                                                 int Level) {
615  MoveToLine(Loc, /*RequireStartOfLine=*/true);
616  OS << "#pragma warning(push";
617  if (Level >= 0)
618    OS << ", " << Level;
619  OS << ')';
620  setEmittedDirectiveOnThisLine();
621}
622
623void PrintPPOutputPPCallbacks::PragmaWarningPop(SourceLocation Loc) {
624  MoveToLine(Loc, /*RequireStartOfLine=*/true);
625  OS << "#pragma warning(pop)";
626  setEmittedDirectiveOnThisLine();
627}
628
629void PrintPPOutputPPCallbacks::PragmaExecCharsetPush(SourceLocation Loc,
630                                                     StringRef Str) {
631  MoveToLine(Loc, /*RequireStartOfLine=*/true);
632  OS << "#pragma character_execution_set(push";
633  if (!Str.empty())
634    OS << ", " << Str;
635  OS << ')';
636  setEmittedDirectiveOnThisLine();
637}
638
639void PrintPPOutputPPCallbacks::PragmaExecCharsetPop(SourceLocation Loc) {
640  MoveToLine(Loc, /*RequireStartOfLine=*/true);
641  OS << "#pragma character_execution_set(pop)";
642  setEmittedDirectiveOnThisLine();
643}
644
645void PrintPPOutputPPCallbacks::
646PragmaAssumeNonNullBegin(SourceLocation Loc) {
647  MoveToLine(Loc, /*RequireStartOfLine=*/true);
648  OS << "#pragma clang assume_nonnull begin";
649  setEmittedDirectiveOnThisLine();
650}
651
652void PrintPPOutputPPCallbacks::
653PragmaAssumeNonNullEnd(SourceLocation Loc) {
654  MoveToLine(Loc, /*RequireStartOfLine=*/true);
655  OS << "#pragma clang assume_nonnull end";
656  setEmittedDirectiveOnThisLine();
657}
658
659void PrintPPOutputPPCallbacks::HandleWhitespaceBeforeTok(const Token &Tok,
660                                                         bool RequireSpace,
661                                                         bool RequireSameLine) {
662  // These tokens are not expanded to anything and don't need whitespace before
663  // them.
664  if (Tok.is(tok::eof) ||
665      (Tok.isAnnotation() && !Tok.is(tok::annot_header_unit) &&
666       !Tok.is(tok::annot_module_begin) && !Tok.is(tok::annot_module_end)))
667    return;
668
669  // EmittedDirectiveOnThisLine takes priority over RequireSameLine.
670  if ((!RequireSameLine || EmittedDirectiveOnThisLine) &&
671      MoveToLine(Tok, /*RequireStartOfLine=*/EmittedDirectiveOnThisLine)) {
672    if (MinimizeWhitespace) {
673      // Avoid interpreting hash as a directive under -fpreprocessed.
674      if (Tok.is(tok::hash))
675        OS << ' ';
676    } else {
677      // Print out space characters so that the first token on a line is
678      // indented for easy reading.
679      unsigned ColNo = SM.getExpansionColumnNumber(Tok.getLocation());
680
681      // The first token on a line can have a column number of 1, yet still
682      // expect leading white space, if a macro expansion in column 1 starts
683      // with an empty macro argument, or an empty nested macro expansion. In
684      // this case, move the token to column 2.
685      if (ColNo == 1 && Tok.hasLeadingSpace())
686        ColNo = 2;
687
688      // This hack prevents stuff like:
689      // #define HASH #
690      // HASH define foo bar
691      // From having the # character end up at column 1, which makes it so it
692      // is not handled as a #define next time through the preprocessor if in
693      // -fpreprocessed mode.
694      if (ColNo <= 1 && Tok.is(tok::hash))
695        OS << ' ';
696
697      // Otherwise, indent the appropriate number of spaces.
698      for (; ColNo > 1; --ColNo)
699        OS << ' ';
700    }
701  } else {
702    // Insert whitespace between the previous and next token if either
703    // - The caller requires it
704    // - The input had whitespace between them and we are not in
705    //   whitespace-minimization mode
706    // - The whitespace is necessary to keep the tokens apart and there is not
707    //   already a newline between them
708    if (RequireSpace || (!MinimizeWhitespace && Tok.hasLeadingSpace()) ||
709        ((EmittedTokensOnThisLine || EmittedDirectiveOnThisLine) &&
710         AvoidConcat(PrevPrevTok, PrevTok, Tok)))
711      OS << ' ';
712  }
713
714  PrevPrevTok = PrevTok;
715  PrevTok = Tok;
716}
717
718void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
719                                                     unsigned Len) {
720  unsigned NumNewlines = 0;
721  for (; Len; --Len, ++TokStr) {
722    if (*TokStr != '\n' &&
723        *TokStr != '\r')
724      continue;
725
726    ++NumNewlines;
727
728    // If we have \n\r or \r\n, skip both and count as one line.
729    if (Len != 1 &&
730        (TokStr[1] == '\n' || TokStr[1] == '\r') &&
731        TokStr[0] != TokStr[1]) {
732      ++TokStr;
733      --Len;
734    }
735  }
736
737  if (NumNewlines == 0) return;
738
739  CurLine += NumNewlines;
740}
741
742
743namespace {
744struct UnknownPragmaHandler : public PragmaHandler {
745  const char *Prefix;
746  PrintPPOutputPPCallbacks *Callbacks;
747
748  // Set to true if tokens should be expanded
749  bool ShouldExpandTokens;
750
751  UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks,
752                       bool RequireTokenExpansion)
753      : Prefix(prefix), Callbacks(callbacks),
754        ShouldExpandTokens(RequireTokenExpansion) {}
755  void HandlePragma(Preprocessor &PP, PragmaIntroducer Introducer,
756                    Token &PragmaTok) override {
757    // Figure out what line we went to and insert the appropriate number of
758    // newline characters.
759    Callbacks->MoveToLine(PragmaTok.getLocation(), /*RequireStartOfLine=*/true);
760    Callbacks->OS.write(Prefix, strlen(Prefix));
761    Callbacks->setEmittedTokensOnThisLine();
762
763    if (ShouldExpandTokens) {
764      // The first token does not have expanded macros. Expand them, if
765      // required.
766      auto Toks = std::make_unique<Token[]>(1);
767      Toks[0] = PragmaTok;
768      PP.EnterTokenStream(std::move(Toks), /*NumToks=*/1,
769                          /*DisableMacroExpansion=*/false,
770                          /*IsReinject=*/false);
771      PP.Lex(PragmaTok);
772    }
773
774    // Read and print all of the pragma tokens.
775    bool IsFirst = true;
776    while (PragmaTok.isNot(tok::eod)) {
777      Callbacks->HandleWhitespaceBeforeTok(PragmaTok, /*RequireSpace=*/IsFirst,
778                                           /*RequireSameLine=*/true);
779      IsFirst = false;
780      std::string TokSpell = PP.getSpelling(PragmaTok);
781      Callbacks->OS.write(&TokSpell[0], TokSpell.size());
782      Callbacks->setEmittedTokensOnThisLine();
783
784      if (ShouldExpandTokens)
785        PP.Lex(PragmaTok);
786      else
787        PP.LexUnexpandedToken(PragmaTok);
788    }
789    Callbacks->setEmittedDirectiveOnThisLine();
790  }
791};
792} // end anonymous namespace
793
794
795static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
796                                    PrintPPOutputPPCallbacks *Callbacks,
797                                    raw_ostream &OS) {
798  bool DropComments = PP.getLangOpts().TraditionalCPP &&
799                      !PP.getCommentRetentionState();
800
801  bool IsStartOfLine = false;
802  char Buffer[256];
803  while (true) {
804    // Two lines joined with line continuation ('\' as last character on the
805    // line) must be emitted as one line even though Tok.getLine() returns two
806    // different values. In this situation Tok.isAtStartOfLine() is false even
807    // though it may be the first token on the lexical line. When
808    // dropping/skipping a token that is at the start of a line, propagate the
809    // start-of-line-ness to the next token to not append it to the previous
810    // line.
811    IsStartOfLine = IsStartOfLine || Tok.isAtStartOfLine();
812
813    Callbacks->HandleWhitespaceBeforeTok(Tok, /*RequireSpace=*/false,
814                                         /*RequireSameLine=*/!IsStartOfLine);
815
816    if (DropComments && Tok.is(tok::comment)) {
817      // Skip comments. Normally the preprocessor does not generate
818      // tok::comment nodes at all when not keeping comments, but under
819      // -traditional-cpp the lexer keeps /all/ whitespace, including comments.
820      PP.Lex(Tok);
821      continue;
822    } else if (Tok.is(tok::eod)) {
823      // Don't print end of directive tokens, since they are typically newlines
824      // that mess up our line tracking. These come from unknown pre-processor
825      // directives or hash-prefixed comments in standalone assembly files.
826      PP.Lex(Tok);
827      // FIXME: The token on the next line after #include should have
828      // Tok.isAtStartOfLine() set.
829      IsStartOfLine = true;
830      continue;
831    } else if (Tok.is(tok::annot_module_include)) {
832      // PrintPPOutputPPCallbacks::InclusionDirective handles producing
833      // appropriate output here. Ignore this token entirely.
834      PP.Lex(Tok);
835      IsStartOfLine = true;
836      continue;
837    } else if (Tok.is(tok::annot_module_begin)) {
838      // FIXME: We retrieve this token after the FileChanged callback, and
839      // retrieve the module_end token before the FileChanged callback, so
840      // we render this within the file and render the module end outside the
841      // file, but this is backwards from the token locations: the module_begin
842      // token is at the include location (outside the file) and the module_end
843      // token is at the EOF location (within the file).
844      Callbacks->BeginModule(
845          reinterpret_cast<Module *>(Tok.getAnnotationValue()));
846      PP.Lex(Tok);
847      IsStartOfLine = true;
848      continue;
849    } else if (Tok.is(tok::annot_module_end)) {
850      Callbacks->EndModule(
851          reinterpret_cast<Module *>(Tok.getAnnotationValue()));
852      PP.Lex(Tok);
853      IsStartOfLine = true;
854      continue;
855    } else if (Tok.is(tok::annot_header_unit)) {
856      // This is a header-name that has been (effectively) converted into a
857      // module-name.
858      // FIXME: The module name could contain non-identifier module name
859      // components. We don't have a good way to round-trip those.
860      Module *M = reinterpret_cast<Module *>(Tok.getAnnotationValue());
861      std::string Name = M->getFullModuleName();
862      OS.write(Name.data(), Name.size());
863      Callbacks->HandleNewlinesInToken(Name.data(), Name.size());
864    } else if (Tok.isAnnotation()) {
865      // Ignore annotation tokens created by pragmas - the pragmas themselves
866      // will be reproduced in the preprocessed output.
867      PP.Lex(Tok);
868      continue;
869    } else if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
870      OS << II->getName();
871    } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
872               Tok.getLiteralData()) {
873      OS.write(Tok.getLiteralData(), Tok.getLength());
874    } else if (Tok.getLength() < std::size(Buffer)) {
875      const char *TokPtr = Buffer;
876      unsigned Len = PP.getSpelling(Tok, TokPtr);
877      OS.write(TokPtr, Len);
878
879      // Tokens that can contain embedded newlines need to adjust our current
880      // line number.
881      // FIXME: The token may end with a newline in which case
882      // setEmittedDirectiveOnThisLine/setEmittedTokensOnThisLine afterwards is
883      // wrong.
884      if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
885        Callbacks->HandleNewlinesInToken(TokPtr, Len);
886      if (Tok.is(tok::comment) && Len >= 2 && TokPtr[0] == '/' &&
887          TokPtr[1] == '/') {
888        // It's a line comment;
889        // Ensure that we don't concatenate anything behind it.
890        Callbacks->setEmittedDirectiveOnThisLine();
891      }
892    } else {
893      std::string S = PP.getSpelling(Tok);
894      OS.write(S.data(), S.size());
895
896      // Tokens that can contain embedded newlines need to adjust our current
897      // line number.
898      if (Tok.getKind() == tok::comment || Tok.getKind() == tok::unknown)
899        Callbacks->HandleNewlinesInToken(S.data(), S.size());
900      if (Tok.is(tok::comment) && S.size() >= 2 && S[0] == '/' && S[1] == '/') {
901        // It's a line comment;
902        // Ensure that we don't concatenate anything behind it.
903        Callbacks->setEmittedDirectiveOnThisLine();
904      }
905    }
906    Callbacks->setEmittedTokensOnThisLine();
907    IsStartOfLine = false;
908
909    if (Tok.is(tok::eof)) break;
910
911    PP.Lex(Tok);
912  }
913}
914
915typedef std::pair<const IdentifierInfo *, MacroInfo *> id_macro_pair;
916static int MacroIDCompare(const id_macro_pair *LHS, const id_macro_pair *RHS) {
917  return LHS->first->getName().compare(RHS->first->getName());
918}
919
920static void DoPrintMacros(Preprocessor &PP, raw_ostream *OS) {
921  // Ignore unknown pragmas.
922  PP.IgnorePragmas();
923
924  // -dM mode just scans and ignores all tokens in the files, then dumps out
925  // the macro table at the end.
926  PP.EnterMainSourceFile();
927
928  Token Tok;
929  do PP.Lex(Tok);
930  while (Tok.isNot(tok::eof));
931
932  SmallVector<id_macro_pair, 128> MacrosByID;
933  for (Preprocessor::macro_iterator I = PP.macro_begin(), E = PP.macro_end();
934       I != E; ++I) {
935    auto *MD = I->second.getLatest();
936    if (MD && MD->isDefined())
937      MacrosByID.push_back(id_macro_pair(I->first, MD->getMacroInfo()));
938  }
939  llvm::array_pod_sort(MacrosByID.begin(), MacrosByID.end(), MacroIDCompare);
940
941  for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
942    MacroInfo &MI = *MacrosByID[i].second;
943    // Ignore computed macros like __LINE__ and friends.
944    if (MI.isBuiltinMacro()) continue;
945
946    PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
947    *OS << '\n';
948  }
949}
950
951/// DoPrintPreprocessedInput - This implements -E mode.
952///
953void clang::DoPrintPreprocessedInput(Preprocessor &PP, raw_ostream *OS,
954                                     const PreprocessorOutputOptions &Opts) {
955  // Show macros with no output is handled specially.
956  if (!Opts.ShowCPP) {
957    assert(Opts.ShowMacros && "Not yet implemented!");
958    DoPrintMacros(PP, OS);
959    return;
960  }
961
962  // Inform the preprocessor whether we want it to retain comments or not, due
963  // to -C or -CC.
964  PP.SetCommentRetentionState(Opts.ShowComments, Opts.ShowMacroComments);
965
966  PrintPPOutputPPCallbacks *Callbacks = new PrintPPOutputPPCallbacks(
967      PP, *OS, !Opts.ShowLineMarkers, Opts.ShowMacros,
968      Opts.ShowIncludeDirectives, Opts.UseLineDirectives,
969      Opts.MinimizeWhitespace, Opts.DirectivesOnly);
970
971  // Expand macros in pragmas with -fms-extensions.  The assumption is that
972  // the majority of pragmas in such a file will be Microsoft pragmas.
973  // Remember the handlers we will add so that we can remove them later.
974  std::unique_ptr<UnknownPragmaHandler> MicrosoftExtHandler(
975      new UnknownPragmaHandler(
976          "#pragma", Callbacks,
977          /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
978
979  std::unique_ptr<UnknownPragmaHandler> GCCHandler(new UnknownPragmaHandler(
980      "#pragma GCC", Callbacks,
981      /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
982
983  std::unique_ptr<UnknownPragmaHandler> ClangHandler(new UnknownPragmaHandler(
984      "#pragma clang", Callbacks,
985      /*RequireTokenExpansion=*/PP.getLangOpts().MicrosoftExt));
986
987  PP.AddPragmaHandler(MicrosoftExtHandler.get());
988  PP.AddPragmaHandler("GCC", GCCHandler.get());
989  PP.AddPragmaHandler("clang", ClangHandler.get());
990
991  // The tokens after pragma omp need to be expanded.
992  //
993  //  OpenMP [2.1, Directive format]
994  //  Preprocessing tokens following the #pragma omp are subject to macro
995  //  replacement.
996  std::unique_ptr<UnknownPragmaHandler> OpenMPHandler(
997      new UnknownPragmaHandler("#pragma omp", Callbacks,
998                               /*RequireTokenExpansion=*/true));
999  PP.AddPragmaHandler("omp", OpenMPHandler.get());
1000
1001  PP.addPPCallbacks(std::unique_ptr<PPCallbacks>(Callbacks));
1002
1003  // After we have configured the preprocessor, enter the main file.
1004  PP.EnterMainSourceFile();
1005  if (Opts.DirectivesOnly)
1006    PP.SetMacroExpansionOnlyInDirectives();
1007
1008  // Consume all of the tokens that come from the predefines buffer.  Those
1009  // should not be emitted into the output and are guaranteed to be at the
1010  // start.
1011  const SourceManager &SourceMgr = PP.getSourceManager();
1012  Token Tok;
1013  do {
1014    PP.Lex(Tok);
1015    if (Tok.is(tok::eof) || !Tok.getLocation().isFileID())
1016      break;
1017
1018    PresumedLoc PLoc = SourceMgr.getPresumedLoc(Tok.getLocation());
1019    if (PLoc.isInvalid())
1020      break;
1021
1022    if (strcmp(PLoc.getFilename(), "<built-in>"))
1023      break;
1024  } while (true);
1025
1026  // Read all the preprocessed tokens, printing them out to the stream.
1027  PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
1028  *OS << '\n';
1029
1030  // Remove the handlers we just added to leave the preprocessor in a sane state
1031  // so that it can be reused (for example by a clang::Parser instance).
1032  PP.RemovePragmaHandler(MicrosoftExtHandler.get());
1033  PP.RemovePragmaHandler("GCC", GCCHandler.get());
1034  PP.RemovePragmaHandler("clang", ClangHandler.get());
1035  PP.RemovePragmaHandler("omp", OpenMPHandler.get());
1036}
1037