PrintPreprocessedOutput.cpp revision df721408f54f87a790fb986dfd64984472a1f892 
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/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/SmallString.h"
25#include "llvm/ADT/STLExtras.h"
26#include "llvm/Config/config.h"
27#include "llvm/Support/raw_ostream.h"
28#include <cstdio>
29using namespace clang;
30
31/// PrintMacroDefinition - Print a macro definition in a form that will be
32/// properly accepted back as a definition.
33static void PrintMacroDefinition(const IdentifierInfo &II, const MacroInfo &MI,
34                                 Preprocessor &PP, llvm::raw_ostream &OS) {
35  OS << "#define " << II.getName();
36
37  if (MI.isFunctionLike()) {
38    OS << '(';
39    if (!MI.arg_empty()) {
40      MacroInfo::arg_iterator AI = MI.arg_begin(), E = MI.arg_end();
41      for (; AI+1 != E; ++AI) {
42        OS << (*AI)->getName();
43        OS << ',';
44      }
45
46      // Last argument.
47      if ((*AI)->getName() == "__VA_ARGS__")
48        OS << "...";
49      else
50        OS << (*AI)->getName();
51    }
52
53    if (MI.isGNUVarargs())
54      OS << "...";  // #define foo(x...)
55
56    OS << ')';
57  }
58
59  // GCC always emits a space, even if the macro body is empty.  However, do not
60  // want to emit two spaces if the first token has a leading space.
61  if (MI.tokens_empty() || !MI.tokens_begin()->hasLeadingSpace())
62    OS << ' ';
63
64  llvm::SmallString<128> SpellingBuffer;
65  for (MacroInfo::tokens_iterator I = MI.tokens_begin(), E = MI.tokens_end();
66       I != E; ++I) {
67    if (I->hasLeadingSpace())
68      OS << ' ';
69
70    OS << PP.getSpelling(*I, 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  llvm::raw_ostream &OS;
85private:
86  unsigned CurLine;
87  bool EmittedTokensOnThisLine;
88  bool EmittedMacroOnThisLine;
89  SrcMgr::CharacteristicKind FileType;
90  llvm::SmallString<512> CurFilename;
91  bool Initialized;
92  bool DisableLineMarkers;
93  bool DumpDefines;
94  bool UseLineDirective;
95public:
96  PrintPPOutputPPCallbacks(Preprocessor &pp, llvm::raw_ostream &os,
97                           bool lineMarkers, bool defines)
98     : PP(pp), SM(PP.getSourceManager()),
99       ConcatInfo(PP), OS(os), DisableLineMarkers(lineMarkers),
100       DumpDefines(defines) {
101    CurLine = 0;
102    CurFilename += "<uninit>";
103    EmittedTokensOnThisLine = false;
104    EmittedMacroOnThisLine = false;
105    FileType = SrcMgr::C_User;
106    Initialized = false;
107
108    // If we're in microsoft mode, use normal #line instead of line markers.
109    UseLineDirective = PP.getLangOptions().Microsoft;
110  }
111
112  void SetEmittedTokensOnThisLine() { EmittedTokensOnThisLine = true; }
113  bool hasEmittedTokensOnThisLine() const { return EmittedTokensOnThisLine; }
114
115  virtual void FileChanged(SourceLocation Loc, FileChangeReason Reason,
116                           SrcMgr::CharacteristicKind FileType);
117  virtual void Ident(SourceLocation Loc, const std::string &str);
118  virtual void PragmaComment(SourceLocation Loc, const IdentifierInfo *Kind,
119                             const std::string &Str);
120
121
122  bool HandleFirstTokOnLine(Token &Tok);
123  bool MoveToLine(SourceLocation Loc) {
124    return MoveToLine(SM.getPresumedLoc(Loc).getLine());
125  }
126  bool MoveToLine(unsigned LineNo);
127
128  bool AvoidConcat(const Token &PrevTok, const Token &Tok) {
129    return ConcatInfo.AvoidConcat(PrevTok, Tok);
130  }
131  void WriteLineInfo(unsigned LineNo, const char *Extra=0, unsigned ExtraLen=0);
132
133  void HandleNewlinesInToken(const char *TokStr, unsigned Len);
134
135  /// MacroDefined - This hook is called whenever a macro definition is seen.
136  void MacroDefined(const IdentifierInfo *II, const MacroInfo *MI);
137
138};
139}  // end anonymous namespace
140
141void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
142                                             const char *Extra,
143                                             unsigned ExtraLen) {
144  if (EmittedTokensOnThisLine || EmittedMacroOnThisLine) {
145    OS << '\n';
146    EmittedTokensOnThisLine = false;
147    EmittedMacroOnThisLine = false;
148  }
149
150  // Emit #line directives or GNU line markers depending on what mode we're in.
151  if (UseLineDirective) {
152    OS << "#line" << ' ' << LineNo << ' ' << '"';
153    OS.write(&CurFilename[0], CurFilename.size());
154    OS << '"';
155  } else {
156    OS << '#' << ' ' << LineNo << ' ' << '"';
157    OS.write(&CurFilename[0], CurFilename.size());
158    OS << '"';
159
160    if (ExtraLen)
161      OS.write(Extra, ExtraLen);
162
163    if (FileType == SrcMgr::C_System)
164      OS.write(" 3", 2);
165    else if (FileType == SrcMgr::C_ExternCSystem)
166      OS.write(" 3 4", 4);
167  }
168  OS << '\n';
169}
170
171/// MoveToLine - Move the output to the source line specified by the location
172/// object.  We can do this by emitting some number of \n's, or be emitting a
173/// #line directive.  This returns false if already at the specified line, true
174/// if some newlines were emitted.
175bool PrintPPOutputPPCallbacks::MoveToLine(unsigned LineNo) {
176  if (DisableLineMarkers) {
177    if (LineNo == CurLine) return false;
178
179    CurLine = LineNo;
180
181    if (!EmittedTokensOnThisLine && !EmittedMacroOnThisLine)
182      return true;
183
184    OS << '\n';
185    EmittedTokensOnThisLine = false;
186    EmittedMacroOnThisLine = false;
187    return true;
188  }
189
190  // If this line is "close enough" to the original line, just print newlines,
191  // otherwise print a #line directive.
192  if (LineNo-CurLine <= 8) {
193    if (LineNo-CurLine == 1)
194      OS << '\n';
195    else if (LineNo == CurLine)
196      return false;    // Spelling line moved, but instantiation line didn't.
197    else {
198      const char *NewLines = "\n\n\n\n\n\n\n\n";
199      OS.write(NewLines, LineNo-CurLine);
200    }
201  } else {
202    WriteLineInfo(LineNo, 0, 0);
203  }
204
205  CurLine = LineNo;
206  return true;
207}
208
209
210/// FileChanged - Whenever the preprocessor enters or exits a #include file
211/// it invokes this handler.  Update our conception of the current source
212/// position.
213void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
214                                           FileChangeReason Reason,
215                                       SrcMgr::CharacteristicKind NewFileType) {
216  // Unless we are exiting a #include, make sure to skip ahead to the line the
217  // #include directive was at.
218  SourceManager &SourceMgr = SM;
219
220  PresumedLoc UserLoc = SourceMgr.getPresumedLoc(Loc);
221  unsigned NewLine = UserLoc.getLine()+1;
222
223  if (Reason == PPCallbacks::EnterFile) {
224    SourceLocation IncludeLoc = SourceMgr.getPresumedLoc(Loc).getIncludeLoc();
225    if (IncludeLoc.isValid())
226      MoveToLine(IncludeLoc);
227  } else if (Reason == PPCallbacks::SystemHeaderPragma) {
228    MoveToLine(NewLine);
229
230    // TODO GCC emits the # directive for this directive on the line AFTER the
231    // directive and emits a bunch of spaces that aren't needed.  Emulate this
232    // strange behavior.
233  }
234
235  CurLine = NewLine;
236
237  if (DisableLineMarkers) return;
238
239  CurFilename.clear();
240  CurFilename += UserLoc.getFilename();
241  Lexer::Stringify(CurFilename);
242  FileType = NewFileType;
243
244  if (!Initialized) {
245    WriteLineInfo(CurLine);
246    Initialized = true;
247  }
248
249  switch (Reason) {
250  case PPCallbacks::EnterFile:
251    WriteLineInfo(CurLine, " 1", 2);
252    break;
253  case PPCallbacks::ExitFile:
254    WriteLineInfo(CurLine, " 2", 2);
255    break;
256  case PPCallbacks::SystemHeaderPragma:
257  case PPCallbacks::RenameFile:
258    WriteLineInfo(CurLine);
259    break;
260  }
261}
262
263/// Ident - Handle #ident directives when read by the preprocessor.
264///
265void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, const std::string &S) {
266  MoveToLine(Loc);
267
268  OS.write("#ident ", strlen("#ident "));
269  OS.write(&S[0], S.size());
270  EmittedTokensOnThisLine = true;
271}
272
273/// MacroDefined - This hook is called whenever a macro definition is seen.
274void PrintPPOutputPPCallbacks::MacroDefined(const IdentifierInfo *II,
275                                            const MacroInfo *MI) {
276  // Only print out macro definitions in -dD mode.
277  if (!DumpDefines ||
278      // Ignore __FILE__ etc.
279      MI->isBuiltinMacro()) return;
280
281  MoveToLine(MI->getDefinitionLoc());
282  PrintMacroDefinition(*II, *MI, PP, OS);
283  EmittedMacroOnThisLine = true;
284}
285
286
287void PrintPPOutputPPCallbacks::PragmaComment(SourceLocation Loc,
288                                             const IdentifierInfo *Kind,
289                                             const std::string &Str) {
290  MoveToLine(Loc);
291  OS << "#pragma comment(" << Kind->getName();
292
293  if (!Str.empty()) {
294    OS << ", \"";
295
296    for (unsigned i = 0, e = Str.size(); i != e; ++i) {
297      unsigned char Char = Str[i];
298      if (isprint(Char) && Char != '\\' && Char != '"')
299        OS << (char)Char;
300      else  // Output anything hard as an octal escape.
301        OS << '\\'
302           << (char)('0'+ ((Char >> 6) & 7))
303           << (char)('0'+ ((Char >> 3) & 7))
304           << (char)('0'+ ((Char >> 0) & 7));
305    }
306    OS << '"';
307  }
308
309  OS << ')';
310  EmittedTokensOnThisLine = true;
311}
312
313
314/// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
315/// is called for the first token on each new line.  If this really is the start
316/// of a new logical line, handle it and return true, otherwise return false.
317/// This may not be the start of a logical line because the "start of line"
318/// marker is set for spelling lines, not instantiation ones.
319bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
320  // Figure out what line we went to and insert the appropriate number of
321  // newline characters.
322  if (!MoveToLine(Tok.getLocation()))
323    return false;
324
325  // Print out space characters so that the first token on a line is
326  // indented for easy reading.
327  unsigned ColNo = SM.getInstantiationColumnNumber(Tok.getLocation());
328
329  // This hack prevents stuff like:
330  // #define HASH #
331  // HASH define foo bar
332  // From having the # character end up at column 1, which makes it so it
333  // is not handled as a #define next time through the preprocessor if in
334  // -fpreprocessed mode.
335  if (ColNo <= 1 && Tok.is(tok::hash))
336    OS << ' ';
337
338  // Otherwise, indent the appropriate number of spaces.
339  for (; ColNo > 1; --ColNo)
340    OS << ' ';
341
342  return true;
343}
344
345void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
346                                                     unsigned Len) {
347  unsigned NumNewlines = 0;
348  for (; Len; --Len, ++TokStr) {
349    if (*TokStr != '\n' &&
350        *TokStr != '\r')
351      continue;
352
353    ++NumNewlines;
354
355    // If we have \n\r or \r\n, skip both and count as one line.
356    if (Len != 1 &&
357        (TokStr[1] == '\n' || TokStr[1] == '\r') &&
358        TokStr[0] != TokStr[1])
359      ++TokStr, --Len;
360  }
361
362  if (NumNewlines == 0) return;
363
364  CurLine += NumNewlines;
365}
366
367
368namespace {
369struct UnknownPragmaHandler : public PragmaHandler {
370  const char *Prefix;
371  PrintPPOutputPPCallbacks *Callbacks;
372
373  UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks)
374    : PragmaHandler(0), Prefix(prefix), Callbacks(callbacks) {}
375  virtual void HandlePragma(Preprocessor &PP, Token &PragmaTok) {
376    // Figure out what line we went to and insert the appropriate number of
377    // newline characters.
378    Callbacks->MoveToLine(PragmaTok.getLocation());
379    Callbacks->OS.write(Prefix, strlen(Prefix));
380
381    // Read and print all of the pragma tokens.
382    while (PragmaTok.isNot(tok::eom)) {
383      if (PragmaTok.hasLeadingSpace())
384        Callbacks->OS << ' ';
385      std::string TokSpell = PP.getSpelling(PragmaTok);
386      Callbacks->OS.write(&TokSpell[0], TokSpell.size());
387      PP.LexUnexpandedToken(PragmaTok);
388    }
389    Callbacks->OS << '\n';
390  }
391};
392} // end anonymous namespace
393
394
395static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
396                                    PrintPPOutputPPCallbacks *Callbacks,
397                                    llvm::raw_ostream &OS) {
398  char Buffer[256];
399  Token PrevTok;
400  while (1) {
401
402    // If this token is at the start of a line, emit newlines if needed.
403    if (Tok.isAtStartOfLine() && Callbacks->HandleFirstTokOnLine(Tok)) {
404      // done.
405    } else if (Tok.hasLeadingSpace() ||
406               // If we haven't emitted a token on this line yet, PrevTok isn't
407               // useful to look at and no concatenation could happen anyway.
408               (Callbacks->hasEmittedTokensOnThisLine() &&
409                // Don't print "-" next to "-", it would form "--".
410                Callbacks->AvoidConcat(PrevTok, Tok))) {
411      OS << ' ';
412    }
413
414    if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
415      OS << II->getName();
416    } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
417               Tok.getLiteralData()) {
418      OS.write(Tok.getLiteralData(), Tok.getLength());
419    } else if (Tok.getLength() < 256) {
420      const char *TokPtr = Buffer;
421      unsigned Len = PP.getSpelling(Tok, TokPtr);
422      OS.write(TokPtr, Len);
423
424      // Tokens that can contain embedded newlines need to adjust our current
425      // line number.
426      if (Tok.getKind() == tok::comment)
427        Callbacks->HandleNewlinesInToken(TokPtr, Len);
428    } else {
429      std::string S = PP.getSpelling(Tok);
430      OS.write(&S[0], S.size());
431
432      // Tokens that can contain embedded newlines need to adjust our current
433      // line number.
434      if (Tok.getKind() == tok::comment)
435        Callbacks->HandleNewlinesInToken(&S[0], S.size());
436    }
437    Callbacks->SetEmittedTokensOnThisLine();
438
439    if (Tok.is(tok::eof)) break;
440
441    PrevTok = Tok;
442    PP.Lex(Tok);
443  }
444}
445
446typedef std::pair<IdentifierInfo*, MacroInfo*> id_macro_pair;
447static int MacroIDCompare(const void* a, const void* b) {
448  const id_macro_pair *LHS = static_cast<const id_macro_pair*>(a);
449  const id_macro_pair *RHS = static_cast<const id_macro_pair*>(b);
450  return LHS->first->getName().compare(RHS->first->getName());
451}
452
453static void DoPrintMacros(Preprocessor &PP, llvm::raw_ostream *OS) {
454  // -dM mode just scans and ignores all tokens in the files, then dumps out
455  // the macro table at the end.
456  if (PP.EnterMainSourceFile())
457    return;
458
459  Token Tok;
460  do PP.Lex(Tok);
461  while (Tok.isNot(tok::eof));
462
463  llvm::SmallVector<id_macro_pair, 128>
464    MacrosByID(PP.macro_begin(), PP.macro_end());
465  llvm::array_pod_sort(MacrosByID.begin(), MacrosByID.end(), MacroIDCompare);
466
467  for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
468    MacroInfo &MI = *MacrosByID[i].second;
469    // Ignore computed macros like __LINE__ and friends.
470    if (MI.isBuiltinMacro()) continue;
471
472    PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
473    *OS << '\n';
474  }
475}
476
477/// DoPrintPreprocessedInput - This implements -E mode.
478///
479void clang::DoPrintPreprocessedInput(Preprocessor &PP, llvm::raw_ostream *OS,
480                                     const PreprocessorOutputOptions &Opts) {
481  // Show macros with no output is handled specially.
482  if (!Opts.ShowCPP) {
483    assert(Opts.ShowMacros && "Not yet implemented!");
484    DoPrintMacros(PP, OS);
485    return;
486  }
487
488  // Inform the preprocessor whether we want it to retain comments or not, due
489  // to -C or -CC.
490  PP.SetCommentRetentionState(Opts.ShowComments, Opts.ShowMacroComments);
491
492  OS->SetBufferSize(64*1024);
493
494  PrintPPOutputPPCallbacks *Callbacks =
495      new PrintPPOutputPPCallbacks(PP, *OS, !Opts.ShowLineMarkers,
496                                   Opts.ShowMacros);
497  PP.AddPragmaHandler(0, new UnknownPragmaHandler("#pragma", Callbacks));
498  PP.AddPragmaHandler("GCC", new UnknownPragmaHandler("#pragma GCC",
499                                                      Callbacks));
500
501  PP.addPPCallbacks(Callbacks);
502
503  // After we have configured the preprocessor, enter the main file.
504  if (PP.EnterMainSourceFile())
505    return;
506
507  // Consume all of the tokens that come from the predefines buffer.  Those
508  // should not be emitted into the output and are guaranteed to be at the
509  // start.
510  const SourceManager &SourceMgr = PP.getSourceManager();
511  Token Tok;
512  do PP.Lex(Tok);
513  while (Tok.isNot(tok::eof) && Tok.getLocation().isFileID() &&
514         !strcmp(SourceMgr.getPresumedLoc(Tok.getLocation()).getFilename(),
515                 "<built-in>"));
516
517  // Read all the preprocessed tokens, printing them out to the stream.
518  PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
519  *OS << '\n';
520}
521
522