PrintPreprocessedOutput.cpp revision 29cf746aef63b1984c013448e843a290b2badf7b 
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/StringExtras.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      ;
41    else if (MI.getNumArgs() == 1)
42      OS << (*MI.arg_begin())->getName();
43    else {
44      MacroInfo::arg_iterator AI = MI.arg_begin(), E = MI.arg_end();
45      OS << (*AI++)->getName();
46      while (AI != E)
47        OS << ',' << (*AI++)->getName();
48    }
49
50    if (MI.isVariadic()) {
51      if (!MI.arg_empty())
52        OS << ',';
53      OS << "...";
54    }
55    OS << ')';
56  }
57
58  // GCC always emits a space, even if the macro body is empty.  However, do not
59  // want to emit two spaces if the first token has a leading space.
60  if (MI.tokens_empty() || !MI.tokens_begin()->hasLeadingSpace())
61    OS << ' ';
62
63  llvm::SmallVector<char, 128> SpellingBuffer;
64  for (MacroInfo::tokens_iterator I = MI.tokens_begin(), E = MI.tokens_end();
65       I != E; ++I) {
66    if (I->hasLeadingSpace())
67      OS << ' ';
68
69    // Make sure we have enough space in the spelling buffer.
70    if (I->getLength() > SpellingBuffer.size())
71      SpellingBuffer.resize(I->getLength());
72    const char *Buffer = SpellingBuffer.data();
73    unsigned SpellingLen = PP.getSpelling(*I, Buffer);
74    OS.write(Buffer, SpellingLen);
75  }
76}
77
78//===----------------------------------------------------------------------===//
79// Preprocessed token printer
80//===----------------------------------------------------------------------===//
81
82namespace {
83class PrintPPOutputPPCallbacks : public PPCallbacks {
84  Preprocessor &PP;
85  TokenConcatenation ConcatInfo;
86public:
87  llvm::raw_ostream &OS;
88private:
89  unsigned CurLine;
90  bool EmittedTokensOnThisLine;
91  bool EmittedMacroOnThisLine;
92  SrcMgr::CharacteristicKind FileType;
93  llvm::SmallString<512> CurFilename;
94  bool Initialized;
95  bool DisableLineMarkers;
96  bool DumpDefines;
97public:
98  PrintPPOutputPPCallbacks(Preprocessor &pp, llvm::raw_ostream &os,
99                           bool lineMarkers, bool defines)
100     : PP(pp), ConcatInfo(PP), OS(os), DisableLineMarkers(lineMarkers),
101       DumpDefines(defines) {
102    CurLine = 0;
103    CurFilename += "<uninit>";
104    EmittedTokensOnThisLine = false;
105    EmittedMacroOnThisLine = false;
106    FileType = SrcMgr::C_User;
107    Initialized = false;
108  }
109
110  void SetEmittedTokensOnThisLine() { EmittedTokensOnThisLine = true; }
111  bool hasEmittedTokensOnThisLine() const { return EmittedTokensOnThisLine; }
112
113  virtual void FileChanged(SourceLocation Loc, FileChangeReason Reason,
114                           SrcMgr::CharacteristicKind FileType);
115  virtual void Ident(SourceLocation Loc, const std::string &str);
116  virtual void PragmaComment(SourceLocation Loc, const IdentifierInfo *Kind,
117                             const std::string &Str);
118
119
120  bool HandleFirstTokOnLine(Token &Tok);
121  bool MoveToLine(SourceLocation Loc);
122  bool AvoidConcat(const Token &PrevTok, const Token &Tok) {
123    return ConcatInfo.AvoidConcat(PrevTok, Tok);
124  }
125  void WriteLineInfo(unsigned LineNo, const char *Extra=0, unsigned ExtraLen=0);
126
127  void HandleNewlinesInToken(const char *TokStr, unsigned Len);
128
129  /// MacroDefined - This hook is called whenever a macro definition is seen.
130  void MacroDefined(const IdentifierInfo *II, const MacroInfo *MI);
131
132};
133}  // end anonymous namespace
134
135void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
136                                             const char *Extra,
137                                             unsigned ExtraLen) {
138  if (EmittedTokensOnThisLine || EmittedMacroOnThisLine) {
139    OS << '\n';
140    EmittedTokensOnThisLine = false;
141    EmittedMacroOnThisLine = false;
142  }
143
144  OS << '#' << ' ' << LineNo << ' ' << '"';
145  OS.write(&CurFilename[0], CurFilename.size());
146  OS << '"';
147
148  if (ExtraLen)
149    OS.write(Extra, ExtraLen);
150
151  if (FileType == SrcMgr::C_System)
152    OS.write(" 3", 2);
153  else if (FileType == SrcMgr::C_ExternCSystem)
154    OS.write(" 3 4", 4);
155  OS << '\n';
156}
157
158/// MoveToLine - Move the output to the source line specified by the location
159/// object.  We can do this by emitting some number of \n's, or be emitting a
160/// #line directive.  This returns false if already at the specified line, true
161/// if some newlines were emitted.
162bool PrintPPOutputPPCallbacks::MoveToLine(SourceLocation Loc) {
163  unsigned LineNo = PP.getSourceManager().getInstantiationLineNumber(Loc);
164
165  if (DisableLineMarkers) {
166    if (LineNo == CurLine) return false;
167
168    CurLine = LineNo;
169
170    if (!EmittedTokensOnThisLine && !EmittedMacroOnThisLine)
171      return true;
172
173    OS << '\n';
174    EmittedTokensOnThisLine = false;
175    EmittedMacroOnThisLine = false;
176    return true;
177  }
178
179  // If this line is "close enough" to the original line, just print newlines,
180  // otherwise print a #line directive.
181  if (LineNo-CurLine <= 8) {
182    if (LineNo-CurLine == 1)
183      OS << '\n';
184    else if (LineNo == CurLine)
185      return false;    // Spelling line moved, but instantiation line didn't.
186    else {
187      const char *NewLines = "\n\n\n\n\n\n\n\n";
188      OS.write(NewLines, LineNo-CurLine);
189    }
190  } else {
191    WriteLineInfo(LineNo, 0, 0);
192  }
193
194  CurLine = LineNo;
195  return true;
196}
197
198
199/// FileChanged - Whenever the preprocessor enters or exits a #include file
200/// it invokes this handler.  Update our conception of the current source
201/// position.
202void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
203                                           FileChangeReason Reason,
204                                       SrcMgr::CharacteristicKind NewFileType) {
205  // Unless we are exiting a #include, make sure to skip ahead to the line the
206  // #include directive was at.
207  SourceManager &SourceMgr = PP.getSourceManager();
208  if (Reason == PPCallbacks::EnterFile) {
209    SourceLocation IncludeLoc = SourceMgr.getPresumedLoc(Loc).getIncludeLoc();
210    if (IncludeLoc.isValid())
211      MoveToLine(IncludeLoc);
212  } else if (Reason == PPCallbacks::SystemHeaderPragma) {
213    MoveToLine(Loc);
214
215    // TODO GCC emits the # directive for this directive on the line AFTER the
216    // directive and emits a bunch of spaces that aren't needed.  Emulate this
217    // strange behavior.
218  }
219
220  Loc = SourceMgr.getInstantiationLoc(Loc);
221  // FIXME: Should use presumed line #!
222  CurLine = SourceMgr.getInstantiationLineNumber(Loc);
223
224  if (DisableLineMarkers) return;
225
226  CurFilename.clear();
227  CurFilename += SourceMgr.getPresumedLoc(Loc).getFilename();
228  Lexer::Stringify(CurFilename);
229  FileType = NewFileType;
230
231  if (!Initialized) {
232    WriteLineInfo(CurLine);
233    Initialized = true;
234  }
235
236  switch (Reason) {
237  case PPCallbacks::EnterFile:
238    WriteLineInfo(CurLine, " 1", 2);
239    break;
240  case PPCallbacks::ExitFile:
241    WriteLineInfo(CurLine, " 2", 2);
242    break;
243  case PPCallbacks::SystemHeaderPragma:
244  case PPCallbacks::RenameFile:
245    WriteLineInfo(CurLine);
246    break;
247  }
248}
249
250/// Ident - Handle #ident directives when read by the preprocessor.
251///
252void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, const std::string &S) {
253  MoveToLine(Loc);
254
255  OS.write("#ident ", strlen("#ident "));
256  OS.write(&S[0], S.size());
257  EmittedTokensOnThisLine = true;
258}
259
260/// MacroDefined - This hook is called whenever a macro definition is seen.
261void PrintPPOutputPPCallbacks::MacroDefined(const IdentifierInfo *II,
262                                            const MacroInfo *MI) {
263  // Only print out macro definitions in -dD mode.
264  if (!DumpDefines ||
265      // Ignore __FILE__ etc.
266      MI->isBuiltinMacro()) return;
267
268  MoveToLine(MI->getDefinitionLoc());
269  PrintMacroDefinition(*II, *MI, PP, OS);
270  EmittedMacroOnThisLine = true;
271}
272
273
274void PrintPPOutputPPCallbacks::PragmaComment(SourceLocation Loc,
275                                             const IdentifierInfo *Kind,
276                                             const std::string &Str) {
277  MoveToLine(Loc);
278  OS << "#pragma comment(" << Kind->getName();
279
280  if (!Str.empty()) {
281    OS << ", \"";
282
283    for (unsigned i = 0, e = Str.size(); i != e; ++i) {
284      unsigned char Char = Str[i];
285      if (isprint(Char) && Char != '\\' && Char != '"')
286        OS << (char)Char;
287      else  // Output anything hard as an octal escape.
288        OS << '\\'
289           << (char)('0'+ ((Char >> 6) & 7))
290           << (char)('0'+ ((Char >> 3) & 7))
291           << (char)('0'+ ((Char >> 0) & 7));
292    }
293    OS << '"';
294  }
295
296  OS << ')';
297  EmittedTokensOnThisLine = true;
298}
299
300
301/// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
302/// is called for the first token on each new line.  If this really is the start
303/// of a new logical line, handle it and return true, otherwise return false.
304/// This may not be the start of a logical line because the "start of line"
305/// marker is set for spelling lines, not instantiation ones.
306bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
307  // Figure out what line we went to and insert the appropriate number of
308  // newline characters.
309  if (!MoveToLine(Tok.getLocation()))
310    return false;
311
312  // Print out space characters so that the first token on a line is
313  // indented for easy reading.
314  const SourceManager &SourceMgr = PP.getSourceManager();
315  unsigned ColNo = SourceMgr.getInstantiationColumnNumber(Tok.getLocation());
316
317  // This hack prevents stuff like:
318  // #define HASH #
319  // HASH define foo bar
320  // From having the # character end up at column 1, which makes it so it
321  // is not handled as a #define next time through the preprocessor if in
322  // -fpreprocessed mode.
323  if (ColNo <= 1 && Tok.is(tok::hash))
324    OS << ' ';
325
326  // Otherwise, indent the appropriate number of spaces.
327  for (; ColNo > 1; --ColNo)
328    OS << ' ';
329
330  return true;
331}
332
333void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
334                                                     unsigned Len) {
335  unsigned NumNewlines = 0;
336  for (; Len; --Len, ++TokStr) {
337    if (*TokStr != '\n' &&
338        *TokStr != '\r')
339      continue;
340
341    ++NumNewlines;
342
343    // If we have \n\r or \r\n, skip both and count as one line.
344    if (Len != 1 &&
345        (TokStr[1] == '\n' || TokStr[1] == '\r') &&
346        TokStr[0] != TokStr[1])
347      ++TokStr, --Len;
348  }
349
350  if (NumNewlines == 0) return;
351
352  CurLine += NumNewlines;
353}
354
355
356namespace {
357struct UnknownPragmaHandler : public PragmaHandler {
358  const char *Prefix;
359  PrintPPOutputPPCallbacks *Callbacks;
360
361  UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks)
362    : PragmaHandler(0), Prefix(prefix), Callbacks(callbacks) {}
363  virtual void HandlePragma(Preprocessor &PP, Token &PragmaTok) {
364    // Figure out what line we went to and insert the appropriate number of
365    // newline characters.
366    Callbacks->MoveToLine(PragmaTok.getLocation());
367    Callbacks->OS.write(Prefix, strlen(Prefix));
368
369    // Read and print all of the pragma tokens.
370    while (PragmaTok.isNot(tok::eom)) {
371      if (PragmaTok.hasLeadingSpace())
372        Callbacks->OS << ' ';
373      std::string TokSpell = PP.getSpelling(PragmaTok);
374      Callbacks->OS.write(&TokSpell[0], TokSpell.size());
375      PP.LexUnexpandedToken(PragmaTok);
376    }
377    Callbacks->OS << '\n';
378  }
379};
380} // end anonymous namespace
381
382
383static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
384                                    PrintPPOutputPPCallbacks *Callbacks,
385                                    llvm::raw_ostream &OS) {
386  char Buffer[256];
387  Token PrevTok;
388  while (1) {
389
390    // If this token is at the start of a line, emit newlines if needed.
391    if (Tok.isAtStartOfLine() && Callbacks->HandleFirstTokOnLine(Tok)) {
392      // done.
393    } else if (Tok.hasLeadingSpace() ||
394               // If we haven't emitted a token on this line yet, PrevTok isn't
395               // useful to look at and no concatenation could happen anyway.
396               (Callbacks->hasEmittedTokensOnThisLine() &&
397                // Don't print "-" next to "-", it would form "--".
398                Callbacks->AvoidConcat(PrevTok, Tok))) {
399      OS << ' ';
400    }
401
402    if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
403      OS << II->getName();
404    } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
405               Tok.getLiteralData()) {
406      OS.write(Tok.getLiteralData(), Tok.getLength());
407    } else if (Tok.getLength() < 256) {
408      const char *TokPtr = Buffer;
409      unsigned Len = PP.getSpelling(Tok, TokPtr);
410      OS.write(TokPtr, Len);
411
412      // Tokens that can contain embedded newlines need to adjust our current
413      // line number.
414      if (Tok.getKind() == tok::comment)
415        Callbacks->HandleNewlinesInToken(TokPtr, Len);
416    } else {
417      std::string S = PP.getSpelling(Tok);
418      OS.write(&S[0], S.size());
419
420      // Tokens that can contain embedded newlines need to adjust our current
421      // line number.
422      if (Tok.getKind() == tok::comment)
423        Callbacks->HandleNewlinesInToken(&S[0], S.size());
424    }
425    Callbacks->SetEmittedTokensOnThisLine();
426
427    if (Tok.is(tok::eof)) break;
428
429    PrevTok = Tok;
430    PP.Lex(Tok);
431  }
432}
433
434namespace {
435  struct SortMacrosByID {
436    typedef std::pair<IdentifierInfo*, MacroInfo*> id_macro_pair;
437    bool operator()(const id_macro_pair &LHS, const id_macro_pair &RHS) const {
438      return LHS.first->getName() < RHS.first->getName();
439    }
440  };
441}
442
443static void DoPrintMacros(Preprocessor &PP, llvm::raw_ostream *OS) {
444  // -dM mode just scans and ignores all tokens in the files, then dumps out
445  // the macro table at the end.
446  PP.EnterMainSourceFile();
447
448  Token Tok;
449  do PP.Lex(Tok);
450  while (Tok.isNot(tok::eof));
451
452  std::vector<std::pair<IdentifierInfo*, MacroInfo*> > MacrosByID;
453  for (Preprocessor::macro_iterator I = PP.macro_begin(), E = PP.macro_end();
454       I != E; ++I)
455    MacrosByID.push_back(*I);
456  std::sort(MacrosByID.begin(), MacrosByID.end(), SortMacrosByID());
457
458  for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
459    MacroInfo &MI = *MacrosByID[i].second;
460    // Ignore computed macros like __LINE__ and friends.
461    if (MI.isBuiltinMacro()) continue;
462
463    PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
464    *OS << "\n";
465  }
466}
467
468/// DoPrintPreprocessedInput - This implements -E mode.
469///
470void clang::DoPrintPreprocessedInput(Preprocessor &PP, llvm::raw_ostream *OS,
471                                     const PreprocessorOutputOptions &Opts) {
472  // Show macros with no output is handled specially.
473  if (!Opts.ShowCPP) {
474    assert(Opts.ShowMacros && "Not yet implemented!");
475    DoPrintMacros(PP, OS);
476    return;
477  }
478
479  // Inform the preprocessor whether we want it to retain comments or not, due
480  // to -C or -CC.
481  PP.SetCommentRetentionState(Opts.ShowComments, Opts.ShowMacroComments);
482
483  OS->SetBufferSize(64*1024);
484
485  PrintPPOutputPPCallbacks *Callbacks =
486      new PrintPPOutputPPCallbacks(PP, *OS, !Opts.ShowLineMarkers,
487                                   Opts.ShowMacros);
488  PP.AddPragmaHandler(0, new UnknownPragmaHandler("#pragma", Callbacks));
489  PP.AddPragmaHandler("GCC", new UnknownPragmaHandler("#pragma GCC",
490                                                      Callbacks));
491
492  PP.setPPCallbacks(Callbacks);
493
494  // After we have configured the preprocessor, enter the main file.
495  PP.EnterMainSourceFile();
496
497  // Consume all of the tokens that come from the predefines buffer.  Those
498  // should not be emitted into the output and are guaranteed to be at the
499  // start.
500  const SourceManager &SourceMgr = PP.getSourceManager();
501  Token Tok;
502  do PP.Lex(Tok);
503  while (Tok.isNot(tok::eof) && Tok.getLocation().isFileID() &&
504         !strcmp(SourceMgr.getPresumedLoc(Tok.getLocation()).getFilename(),
505                 "<built-in>"));
506
507  // Read all the preprocessed tokens, printing them out to the stream.
508  PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
509  *OS << '\n';
510}
511
512