PrintPreprocessedOutput.cpp revision 198398 
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/Lex/MacroInfo.h"
17#include "clang/Lex/PPCallbacks.h"
18#include "clang/Lex/Preprocessor.h"
19#include "clang/Lex/Pragma.h"
20#include "clang/Lex/TokenConcatenation.h"
21#include "clang/Basic/SourceManager.h"
22#include "clang/Basic/Diagnostic.h"
23#include "llvm/ADT/SmallString.h"
24#include "llvm/ADT/StringExtras.h"
25#include "llvm/Config/config.h"
26#include "llvm/Support/raw_ostream.h"
27#include <cstdio>
28using namespace clang;
29
30/// PrintMacroDefinition - Print a macro definition in a form that will be
31/// properly accepted back as a definition.
32static void PrintMacroDefinition(const IdentifierInfo &II, const MacroInfo &MI,
33                                 Preprocessor &PP, llvm::raw_ostream &OS) {
34  OS << "#define " << II.getName();
35
36  if (MI.isFunctionLike()) {
37    OS << '(';
38    if (MI.arg_empty())
39      ;
40    else if (MI.getNumArgs() == 1)
41      OS << (*MI.arg_begin())->getName();
42    else {
43      MacroInfo::arg_iterator AI = MI.arg_begin(), E = MI.arg_end();
44      OS << (*AI++)->getName();
45      while (AI != E)
46        OS << ',' << (*AI++)->getName();
47    }
48
49    if (MI.isVariadic()) {
50      if (!MI.arg_empty())
51        OS << ',';
52      OS << "...";
53    }
54    OS << ')';
55  }
56
57  // GCC always emits a space, even if the macro body is empty.  However, do not
58  // want to emit two spaces if the first token has a leading space.
59  if (MI.tokens_empty() || !MI.tokens_begin()->hasLeadingSpace())
60    OS << ' ';
61
62  llvm::SmallVector<char, 128> SpellingBuffer;
63  for (MacroInfo::tokens_iterator I = MI.tokens_begin(), E = MI.tokens_end();
64       I != E; ++I) {
65    if (I->hasLeadingSpace())
66      OS << ' ';
67
68    // Make sure we have enough space in the spelling buffer.
69    if (I->getLength() < SpellingBuffer.size())
70      SpellingBuffer.resize(I->getLength());
71    const char *Buffer = SpellingBuffer.data();
72    unsigned SpellingLen = PP.getSpelling(*I, Buffer);
73    OS.write(Buffer, SpellingLen);
74  }
75}
76
77//===----------------------------------------------------------------------===//
78// Preprocessed token printer
79//===----------------------------------------------------------------------===//
80
81namespace {
82class PrintPPOutputPPCallbacks : public PPCallbacks {
83  Preprocessor &PP;
84  TokenConcatenation ConcatInfo;
85public:
86  llvm::raw_ostream &OS;
87private:
88  unsigned CurLine;
89  bool EmittedTokensOnThisLine;
90  bool EmittedMacroOnThisLine;
91  SrcMgr::CharacteristicKind FileType;
92  llvm::SmallString<512> CurFilename;
93  bool Initialized;
94  bool DisableLineMarkers;
95  bool DumpDefines;
96public:
97  PrintPPOutputPPCallbacks(Preprocessor &pp, llvm::raw_ostream &os,
98                           bool lineMarkers, bool defines)
99     : PP(pp), 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
109  void SetEmittedTokensOnThisLine() { EmittedTokensOnThisLine = true; }
110  bool hasEmittedTokensOnThisLine() const { return EmittedTokensOnThisLine; }
111
112  virtual void FileChanged(SourceLocation Loc, FileChangeReason Reason,
113                           SrcMgr::CharacteristicKind FileType);
114  virtual void Ident(SourceLocation Loc, const std::string &str);
115  virtual void PragmaComment(SourceLocation Loc, const IdentifierInfo *Kind,
116                             const std::string &Str);
117
118
119  bool HandleFirstTokOnLine(Token &Tok);
120  bool MoveToLine(SourceLocation Loc);
121  bool AvoidConcat(const Token &PrevTok, const Token &Tok) {
122    return ConcatInfo.AvoidConcat(PrevTok, Tok);
123  }
124  void WriteLineInfo(unsigned LineNo, const char *Extra=0, unsigned ExtraLen=0);
125
126  void HandleNewlinesInToken(const char *TokStr, unsigned Len);
127
128  /// MacroDefined - This hook is called whenever a macro definition is seen.
129  void MacroDefined(const IdentifierInfo *II, const MacroInfo *MI);
130
131};
132}  // end anonymous namespace
133
134void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
135                                             const char *Extra,
136                                             unsigned ExtraLen) {
137  if (EmittedTokensOnThisLine || EmittedMacroOnThisLine) {
138    OS << '\n';
139    EmittedTokensOnThisLine = false;
140    EmittedMacroOnThisLine = false;
141  }
142
143  OS << '#' << ' ' << LineNo << ' ' << '"';
144  OS.write(&CurFilename[0], CurFilename.size());
145  OS << '"';
146
147  if (ExtraLen)
148    OS.write(Extra, ExtraLen);
149
150  if (FileType == SrcMgr::C_System)
151    OS.write(" 3", 2);
152  else if (FileType == SrcMgr::C_ExternCSystem)
153    OS.write(" 3 4", 4);
154  OS << '\n';
155}
156
157/// MoveToLine - Move the output to the source line specified by the location
158/// object.  We can do this by emitting some number of \n's, or be emitting a
159/// #line directive.  This returns false if already at the specified line, true
160/// if some newlines were emitted.
161bool PrintPPOutputPPCallbacks::MoveToLine(SourceLocation Loc) {
162  unsigned LineNo = PP.getSourceManager().getInstantiationLineNumber(Loc);
163
164  if (DisableLineMarkers) {
165    if (LineNo == CurLine) return false;
166
167    CurLine = LineNo;
168
169    if (!EmittedTokensOnThisLine && !EmittedMacroOnThisLine)
170      return true;
171
172    OS << '\n';
173    EmittedTokensOnThisLine = false;
174    EmittedMacroOnThisLine = false;
175    return true;
176  }
177
178  // If this line is "close enough" to the original line, just print newlines,
179  // otherwise print a #line directive.
180  if (LineNo-CurLine <= 8) {
181    if (LineNo-CurLine == 1)
182      OS << '\n';
183    else if (LineNo == CurLine)
184      return false;    // Spelling line moved, but instantiation line didn't.
185    else {
186      const char *NewLines = "\n\n\n\n\n\n\n\n";
187      OS.write(NewLines, LineNo-CurLine);
188    }
189  } else {
190    WriteLineInfo(LineNo, 0, 0);
191  }
192
193  CurLine = LineNo;
194  return true;
195}
196
197
198/// FileChanged - Whenever the preprocessor enters or exits a #include file
199/// it invokes this handler.  Update our conception of the current source
200/// position.
201void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
202                                           FileChangeReason Reason,
203                                       SrcMgr::CharacteristicKind NewFileType) {
204  // Unless we are exiting a #include, make sure to skip ahead to the line the
205  // #include directive was at.
206  SourceManager &SourceMgr = PP.getSourceManager();
207  if (Reason == PPCallbacks::EnterFile) {
208    SourceLocation IncludeLoc = SourceMgr.getPresumedLoc(Loc).getIncludeLoc();
209    if (IncludeLoc.isValid())
210      MoveToLine(IncludeLoc);
211  } else if (Reason == PPCallbacks::SystemHeaderPragma) {
212    MoveToLine(Loc);
213
214    // TODO GCC emits the # directive for this directive on the line AFTER the
215    // directive and emits a bunch of spaces that aren't needed.  Emulate this
216    // strange behavior.
217  }
218
219  Loc = SourceMgr.getInstantiationLoc(Loc);
220  // FIXME: Should use presumed line #!
221  CurLine = SourceMgr.getInstantiationLineNumber(Loc);
222
223  if (DisableLineMarkers) return;
224
225  CurFilename.clear();
226  CurFilename += SourceMgr.getPresumedLoc(Loc).getFilename();
227  Lexer::Stringify(CurFilename);
228  FileType = NewFileType;
229
230  if (!Initialized) {
231    WriteLineInfo(CurLine);
232    Initialized = true;
233  }
234
235  switch (Reason) {
236  case PPCallbacks::EnterFile:
237    WriteLineInfo(CurLine, " 1", 2);
238    break;
239  case PPCallbacks::ExitFile:
240    WriteLineInfo(CurLine, " 2", 2);
241    break;
242  case PPCallbacks::SystemHeaderPragma:
243  case PPCallbacks::RenameFile:
244    WriteLineInfo(CurLine);
245    break;
246  }
247}
248
249/// Ident - Handle #ident directives when read by the preprocessor.
250///
251void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, const std::string &S) {
252  MoveToLine(Loc);
253
254  OS.write("#ident ", strlen("#ident "));
255  OS.write(&S[0], S.size());
256  EmittedTokensOnThisLine = true;
257}
258
259/// MacroDefined - This hook is called whenever a macro definition is seen.
260void PrintPPOutputPPCallbacks::MacroDefined(const IdentifierInfo *II,
261                                            const MacroInfo *MI) {
262  // Only print out macro definitions in -dD mode.
263  if (!DumpDefines ||
264      // Ignore __FILE__ etc.
265      MI->isBuiltinMacro()) return;
266
267  MoveToLine(MI->getDefinitionLoc());
268  PrintMacroDefinition(*II, *MI, PP, OS);
269  EmittedMacroOnThisLine = true;
270}
271
272
273void PrintPPOutputPPCallbacks::PragmaComment(SourceLocation Loc,
274                                             const IdentifierInfo *Kind,
275                                             const std::string &Str) {
276  MoveToLine(Loc);
277  OS << "#pragma comment(" << Kind->getName();
278
279  if (!Str.empty()) {
280    OS << ", \"";
281
282    for (unsigned i = 0, e = Str.size(); i != e; ++i) {
283      unsigned char Char = Str[i];
284      if (isprint(Char) && Char != '\\' && Char != '"')
285        OS << (char)Char;
286      else  // Output anything hard as an octal escape.
287        OS << '\\'
288           << (char)('0'+ ((Char >> 6) & 7))
289           << (char)('0'+ ((Char >> 3) & 7))
290           << (char)('0'+ ((Char >> 0) & 7));
291    }
292    OS << '"';
293  }
294
295  OS << ')';
296  EmittedTokensOnThisLine = true;
297}
298
299
300/// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
301/// is called for the first token on each new line.  If this really is the start
302/// of a new logical line, handle it and return true, otherwise return false.
303/// This may not be the start of a logical line because the "start of line"
304/// marker is set for spelling lines, not instantiation ones.
305bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
306  // Figure out what line we went to and insert the appropriate number of
307  // newline characters.
308  if (!MoveToLine(Tok.getLocation()))
309    return false;
310
311  // Print out space characters so that the first token on a line is
312  // indented for easy reading.
313  const SourceManager &SourceMgr = PP.getSourceManager();
314  unsigned ColNo = SourceMgr.getInstantiationColumnNumber(Tok.getLocation());
315
316  // This hack prevents stuff like:
317  // #define HASH #
318  // HASH define foo bar
319  // From having the # character end up at column 1, which makes it so it
320  // is not handled as a #define next time through the preprocessor if in
321  // -fpreprocessed mode.
322  if (ColNo <= 1 && Tok.is(tok::hash))
323    OS << ' ';
324
325  // Otherwise, indent the appropriate number of spaces.
326  for (; ColNo > 1; --ColNo)
327    OS << ' ';
328
329  return true;
330}
331
332void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
333                                                     unsigned Len) {
334  unsigned NumNewlines = 0;
335  for (; Len; --Len, ++TokStr) {
336    if (*TokStr != '\n' &&
337        *TokStr != '\r')
338      continue;
339
340    ++NumNewlines;
341
342    // If we have \n\r or \r\n, skip both and count as one line.
343    if (Len != 1 &&
344        (TokStr[1] == '\n' || TokStr[1] == '\r') &&
345        TokStr[0] != TokStr[1])
346      ++TokStr, --Len;
347  }
348
349  if (NumNewlines == 0) return;
350
351  CurLine += NumNewlines;
352}
353
354
355namespace {
356struct UnknownPragmaHandler : public PragmaHandler {
357  const char *Prefix;
358  PrintPPOutputPPCallbacks *Callbacks;
359
360  UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks)
361    : PragmaHandler(0), Prefix(prefix), Callbacks(callbacks) {}
362  virtual void HandlePragma(Preprocessor &PP, Token &PragmaTok) {
363    // Figure out what line we went to and insert the appropriate number of
364    // newline characters.
365    Callbacks->MoveToLine(PragmaTok.getLocation());
366    Callbacks->OS.write(Prefix, strlen(Prefix));
367
368    // Read and print all of the pragma tokens.
369    while (PragmaTok.isNot(tok::eom)) {
370      if (PragmaTok.hasLeadingSpace())
371        Callbacks->OS << ' ';
372      std::string TokSpell = PP.getSpelling(PragmaTok);
373      Callbacks->OS.write(&TokSpell[0], TokSpell.size());
374      PP.LexUnexpandedToken(PragmaTok);
375    }
376    Callbacks->OS << '\n';
377  }
378};
379} // end anonymous namespace
380
381
382static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
383                                    PrintPPOutputPPCallbacks *Callbacks,
384                                    llvm::raw_ostream &OS) {
385  char Buffer[256];
386  Token PrevTok;
387  while (1) {
388
389    // If this token is at the start of a line, emit newlines if needed.
390    if (Tok.isAtStartOfLine() && Callbacks->HandleFirstTokOnLine(Tok)) {
391      // done.
392    } else if (Tok.hasLeadingSpace() ||
393               // If we haven't emitted a token on this line yet, PrevTok isn't
394               // useful to look at and no concatenation could happen anyway.
395               (Callbacks->hasEmittedTokensOnThisLine() &&
396                // Don't print "-" next to "-", it would form "--".
397                Callbacks->AvoidConcat(PrevTok, Tok))) {
398      OS << ' ';
399    }
400
401    if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
402      OS << II->getName();
403    } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
404               Tok.getLiteralData()) {
405      OS.write(Tok.getLiteralData(), Tok.getLength());
406    } else if (Tok.getLength() < 256) {
407      const char *TokPtr = Buffer;
408      unsigned Len = PP.getSpelling(Tok, TokPtr);
409      OS.write(TokPtr, Len);
410
411      // Tokens that can contain embedded newlines need to adjust our current
412      // line number.
413      if (Tok.getKind() == tok::comment)
414        Callbacks->HandleNewlinesInToken(TokPtr, Len);
415    } else {
416      std::string S = PP.getSpelling(Tok);
417      OS.write(&S[0], S.size());
418
419      // Tokens that can contain embedded newlines need to adjust our current
420      // line number.
421      if (Tok.getKind() == tok::comment)
422        Callbacks->HandleNewlinesInToken(&S[0], S.size());
423    }
424    Callbacks->SetEmittedTokensOnThisLine();
425
426    if (Tok.is(tok::eof)) break;
427
428    PrevTok = Tok;
429    PP.Lex(Tok);
430  }
431}
432
433namespace {
434  struct SortMacrosByID {
435    typedef std::pair<IdentifierInfo*, MacroInfo*> id_macro_pair;
436    bool operator()(const id_macro_pair &LHS, const id_macro_pair &RHS) const {
437      return LHS.first->getName() < RHS.first->getName();
438    }
439  };
440}
441
442void clang::DoPrintMacros(Preprocessor &PP, llvm::raw_ostream *OS) {
443  // -dM mode just scans and ignores all tokens in the files, then dumps out
444  // the macro table at the end.
445  PP.EnterMainSourceFile();
446
447  Token Tok;
448  do PP.Lex(Tok);
449  while (Tok.isNot(tok::eof));
450
451  std::vector<std::pair<IdentifierInfo*, MacroInfo*> > MacrosByID;
452  for (Preprocessor::macro_iterator I = PP.macro_begin(), E = PP.macro_end();
453       I != E; ++I)
454    MacrosByID.push_back(*I);
455  std::sort(MacrosByID.begin(), MacrosByID.end(), SortMacrosByID());
456
457  for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
458    MacroInfo &MI = *MacrosByID[i].second;
459    // Ignore computed macros like __LINE__ and friends.
460    if (MI.isBuiltinMacro()) continue;
461
462    PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
463    *OS << "\n";
464  }
465}
466
467/// DoPrintPreprocessedInput - This implements -E mode.
468///
469void clang::DoPrintPreprocessedInput(Preprocessor &PP, llvm::raw_ostream *OS,
470                                     bool EnableCommentOutput,
471                                     bool EnableMacroCommentOutput,
472                                     bool DisableLineMarkers,
473                                     bool DumpDefines) {
474  // Inform the preprocessor whether we want it to retain comments or not, due
475  // to -C or -CC.
476  PP.SetCommentRetentionState(EnableCommentOutput, EnableMacroCommentOutput);
477
478  OS->SetBufferSize(64*1024);
479
480  PrintPPOutputPPCallbacks *Callbacks =
481      new PrintPPOutputPPCallbacks(PP, *OS, DisableLineMarkers, DumpDefines);
482  PP.AddPragmaHandler(0, new UnknownPragmaHandler("#pragma", Callbacks));
483  PP.AddPragmaHandler("GCC", new UnknownPragmaHandler("#pragma GCC",
484                                                      Callbacks));
485
486  PP.setPPCallbacks(Callbacks);
487
488  // After we have configured the preprocessor, enter the main file.
489  PP.EnterMainSourceFile();
490
491  // Consume all of the tokens that come from the predefines buffer.  Those
492  // should not be emitted into the output and are guaranteed to be at the
493  // start.
494  const SourceManager &SourceMgr = PP.getSourceManager();
495  Token Tok;
496  do PP.Lex(Tok);
497  while (Tok.isNot(tok::eof) && Tok.getLocation().isFileID() &&
498         !strcmp(SourceMgr.getPresumedLoc(Tok.getLocation()).getFilename(),
499                 "<built-in>"));
500
501  // Read all the preprocessed tokens, printing them out to the stream.
502  PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
503  *OS << '\n';
504}
505
506