PrintPreprocessedOutput.cpp revision ec0df5f4b74dd3fb37fdb574defc36f5397594b9 
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  /// MacroDefined - This hook is called whenever a macro definition is seen.
127  void MacroDefined(const IdentifierInfo *II, const MacroInfo *MI);
128
129};
130}  // end anonymous namespace
131
132void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
133                                             const char *Extra,
134                                             unsigned ExtraLen) {
135  if (EmittedTokensOnThisLine || EmittedMacroOnThisLine) {
136    OS << '\n';
137    EmittedTokensOnThisLine = false;
138    EmittedMacroOnThisLine = false;
139  }
140
141  OS << '#' << ' ' << LineNo << ' ' << '"';
142  OS.write(&CurFilename[0], CurFilename.size());
143  OS << '"';
144
145  if (ExtraLen)
146    OS.write(Extra, ExtraLen);
147
148  if (FileType == SrcMgr::C_System)
149    OS.write(" 3", 2);
150  else if (FileType == SrcMgr::C_ExternCSystem)
151    OS.write(" 3 4", 4);
152  OS << '\n';
153}
154
155/// MoveToLine - Move the output to the source line specified by the location
156/// object.  We can do this by emitting some number of \n's, or be emitting a
157/// #line directive.  This returns false if already at the specified line, true
158/// if some newlines were emitted.
159bool PrintPPOutputPPCallbacks::MoveToLine(SourceLocation Loc) {
160  unsigned LineNo = PP.getSourceManager().getInstantiationLineNumber(Loc);
161
162  if (DisableLineMarkers) {
163    if (LineNo == CurLine) return false;
164
165    CurLine = LineNo;
166
167    if (!EmittedTokensOnThisLine && !EmittedMacroOnThisLine)
168      return true;
169
170    OS << '\n';
171    EmittedTokensOnThisLine = false;
172    EmittedMacroOnThisLine = false;
173    return true;
174  }
175
176  // If this line is "close enough" to the original line, just print newlines,
177  // otherwise print a #line directive.
178  if (LineNo-CurLine <= 8) {
179    if (LineNo-CurLine == 1)
180      OS << '\n';
181    else if (LineNo == CurLine)
182      return false;    // Spelling line moved, but instantiation line didn't.
183    else {
184      const char *NewLines = "\n\n\n\n\n\n\n\n";
185      OS.write(NewLines, LineNo-CurLine);
186    }
187  } else {
188    WriteLineInfo(LineNo, 0, 0);
189  }
190
191  CurLine = LineNo;
192  return true;
193}
194
195
196/// FileChanged - Whenever the preprocessor enters or exits a #include file
197/// it invokes this handler.  Update our conception of the current source
198/// position.
199void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
200                                           FileChangeReason Reason,
201                                       SrcMgr::CharacteristicKind NewFileType) {
202  // Unless we are exiting a #include, make sure to skip ahead to the line the
203  // #include directive was at.
204  SourceManager &SourceMgr = PP.getSourceManager();
205  if (Reason == PPCallbacks::EnterFile) {
206    SourceLocation IncludeLoc = SourceMgr.getPresumedLoc(Loc).getIncludeLoc();
207    if (IncludeLoc.isValid())
208      MoveToLine(IncludeLoc);
209  } else if (Reason == PPCallbacks::SystemHeaderPragma) {
210    MoveToLine(Loc);
211
212    // TODO GCC emits the # directive for this directive on the line AFTER the
213    // directive and emits a bunch of spaces that aren't needed.  Emulate this
214    // strange behavior.
215  }
216
217  Loc = SourceMgr.getInstantiationLoc(Loc);
218  // FIXME: Should use presumed line #!
219  CurLine = SourceMgr.getInstantiationLineNumber(Loc);
220
221  if (DisableLineMarkers) return;
222
223  CurFilename.clear();
224  CurFilename += SourceMgr.getPresumedLoc(Loc).getFilename();
225  Lexer::Stringify(CurFilename);
226  FileType = NewFileType;
227
228  if (!Initialized) {
229    WriteLineInfo(CurLine);
230    Initialized = true;
231  }
232
233  switch (Reason) {
234  case PPCallbacks::EnterFile:
235    WriteLineInfo(CurLine, " 1", 2);
236    break;
237  case PPCallbacks::ExitFile:
238    WriteLineInfo(CurLine, " 2", 2);
239    break;
240  case PPCallbacks::SystemHeaderPragma:
241  case PPCallbacks::RenameFile:
242    WriteLineInfo(CurLine);
243    break;
244  }
245}
246
247/// Ident - Handle #ident directives when read by the preprocessor.
248///
249void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, const std::string &S) {
250  MoveToLine(Loc);
251
252  OS.write("#ident ", strlen("#ident "));
253  OS.write(&S[0], S.size());
254  EmittedTokensOnThisLine = true;
255}
256
257/// MacroDefined - This hook is called whenever a macro definition is seen.
258void PrintPPOutputPPCallbacks::MacroDefined(const IdentifierInfo *II,
259                                            const MacroInfo *MI) {
260  // Only print out macro definitions in -dD mode.
261  if (!DumpDefines ||
262      // Ignore __FILE__ etc.
263      MI->isBuiltinMacro()) return;
264
265  MoveToLine(MI->getDefinitionLoc());
266  PrintMacroDefinition(*II, *MI, PP, OS);
267  EmittedMacroOnThisLine = true;
268}
269
270
271void PrintPPOutputPPCallbacks::PragmaComment(SourceLocation Loc,
272                                             const IdentifierInfo *Kind,
273                                             const std::string &Str) {
274  MoveToLine(Loc);
275  OS << "#pragma comment(" << Kind->getName();
276
277  if (!Str.empty()) {
278    OS << ", \"";
279
280    for (unsigned i = 0, e = Str.size(); i != e; ++i) {
281      unsigned char Char = Str[i];
282      if (isprint(Char) && Char != '\\' && Char != '"')
283        OS << (char)Char;
284      else  // Output anything hard as an octal escape.
285        OS << '\\'
286           << (char)('0'+ ((Char >> 6) & 7))
287           << (char)('0'+ ((Char >> 3) & 7))
288           << (char)('0'+ ((Char >> 0) & 7));
289    }
290    OS << '"';
291  }
292
293  OS << ')';
294  EmittedTokensOnThisLine = true;
295}
296
297
298/// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
299/// is called for the first token on each new line.  If this really is the start
300/// of a new logical line, handle it and return true, otherwise return false.
301/// This may not be the start of a logical line because the "start of line"
302/// marker is set for spelling lines, not instantiation ones.
303bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
304  // Figure out what line we went to and insert the appropriate number of
305  // newline characters.
306  if (!MoveToLine(Tok.getLocation()))
307    return false;
308
309  // Print out space characters so that the first token on a line is
310  // indented for easy reading.
311  const SourceManager &SourceMgr = PP.getSourceManager();
312  unsigned ColNo = SourceMgr.getInstantiationColumnNumber(Tok.getLocation());
313
314  // This hack prevents stuff like:
315  // #define HASH #
316  // HASH define foo bar
317  // From having the # character end up at column 1, which makes it so it
318  // is not handled as a #define next time through the preprocessor if in
319  // -fpreprocessed mode.
320  if (ColNo <= 1 && Tok.is(tok::hash))
321    OS << ' ';
322
323  // Otherwise, indent the appropriate number of spaces.
324  for (; ColNo > 1; --ColNo)
325    OS << ' ';
326
327  return true;
328}
329
330namespace {
331struct UnknownPragmaHandler : public PragmaHandler {
332  const char *Prefix;
333  PrintPPOutputPPCallbacks *Callbacks;
334
335  UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks)
336    : PragmaHandler(0), Prefix(prefix), Callbacks(callbacks) {}
337  virtual void HandlePragma(Preprocessor &PP, Token &PragmaTok) {
338    // Figure out what line we went to and insert the appropriate number of
339    // newline characters.
340    Callbacks->MoveToLine(PragmaTok.getLocation());
341    Callbacks->OS.write(Prefix, strlen(Prefix));
342
343    // Read and print all of the pragma tokens.
344    while (PragmaTok.isNot(tok::eom)) {
345      if (PragmaTok.hasLeadingSpace())
346        Callbacks->OS << ' ';
347      std::string TokSpell = PP.getSpelling(PragmaTok);
348      Callbacks->OS.write(&TokSpell[0], TokSpell.size());
349      PP.LexUnexpandedToken(PragmaTok);
350    }
351    Callbacks->OS << '\n';
352  }
353};
354} // end anonymous namespace
355
356
357static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
358                                    PrintPPOutputPPCallbacks *Callbacks,
359                                    llvm::raw_ostream &OS) {
360  char Buffer[256];
361  Token PrevTok;
362  while (1) {
363
364    // If this token is at the start of a line, emit newlines if needed.
365    if (Tok.isAtStartOfLine() && Callbacks->HandleFirstTokOnLine(Tok)) {
366      // done.
367    } else if (Tok.hasLeadingSpace() ||
368               // If we haven't emitted a token on this line yet, PrevTok isn't
369               // useful to look at and no concatenation could happen anyway.
370               (Callbacks->hasEmittedTokensOnThisLine() &&
371                // Don't print "-" next to "-", it would form "--".
372                Callbacks->AvoidConcat(PrevTok, Tok))) {
373      OS << ' ';
374    }
375
376    if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
377      OS.write(II->getName(), II->getLength());
378    } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
379               Tok.getLiteralData()) {
380      OS.write(Tok.getLiteralData(), Tok.getLength());
381    } else if (Tok.getLength() < 256) {
382      const char *TokPtr = Buffer;
383      unsigned Len = PP.getSpelling(Tok, TokPtr);
384      OS.write(TokPtr, Len);
385    } else {
386      std::string S = PP.getSpelling(Tok);
387      OS.write(&S[0], S.size());
388    }
389    Callbacks->SetEmittedTokensOnThisLine();
390
391    if (Tok.is(tok::eof)) break;
392
393    PrevTok = Tok;
394    PP.Lex(Tok);
395  }
396}
397
398namespace {
399  struct SortMacrosByID {
400    typedef std::pair<IdentifierInfo*, MacroInfo*> id_macro_pair;
401    bool operator()(const id_macro_pair &LHS, const id_macro_pair &RHS) const {
402      return strcmp(LHS.first->getName(), RHS.first->getName()) < 0;
403    }
404  };
405}
406
407void clang::DoPrintMacros(Preprocessor &PP, llvm::raw_ostream *OS) {
408  // -dM mode just scans and ignores all tokens in the files, then dumps out
409  // the macro table at the end.
410  PP.EnterMainSourceFile();
411
412  Token Tok;
413  do PP.Lex(Tok);
414  while (Tok.isNot(tok::eof));
415
416  std::vector<std::pair<IdentifierInfo*, MacroInfo*> > MacrosByID;
417  for (Preprocessor::macro_iterator I = PP.macro_begin(), E = PP.macro_end();
418       I != E; ++I)
419    MacrosByID.push_back(*I);
420  std::sort(MacrosByID.begin(), MacrosByID.end(), SortMacrosByID());
421
422  for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
423    MacroInfo &MI = *MacrosByID[i].second;
424    // Ignore computed macros like __LINE__ and friends.
425    if (MI.isBuiltinMacro()) continue;
426
427    PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
428    *OS << "\n";
429  }
430}
431
432/// DoPrintPreprocessedInput - This implements -E mode.
433///
434void clang::DoPrintPreprocessedInput(Preprocessor &PP, llvm::raw_ostream *OS,
435                                     bool EnableCommentOutput,
436                                     bool EnableMacroCommentOutput,
437                                     bool DisableLineMarkers,
438                                     bool DumpDefines) {
439  // Inform the preprocessor whether we want it to retain comments or not, due
440  // to -C or -CC.
441  PP.SetCommentRetentionState(EnableCommentOutput, EnableMacroCommentOutput);
442
443  OS->SetBufferSize(64*1024);
444
445  PrintPPOutputPPCallbacks *Callbacks =
446      new PrintPPOutputPPCallbacks(PP, *OS, DisableLineMarkers, DumpDefines);
447  PP.AddPragmaHandler(0, new UnknownPragmaHandler("#pragma", Callbacks));
448  PP.AddPragmaHandler("GCC", new UnknownPragmaHandler("#pragma GCC",
449                                                      Callbacks));
450
451  PP.setPPCallbacks(Callbacks);
452
453  // After we have configured the preprocessor, enter the main file.
454  PP.EnterMainSourceFile();
455
456  // Consume all of the tokens that come from the predefines buffer.  Those
457  // should not be emitted into the output and are guaranteed to be at the
458  // start.
459  const SourceManager &SourceMgr = PP.getSourceManager();
460  Token Tok;
461  do PP.Lex(Tok);
462  while (Tok.isNot(tok::eof) && Tok.getLocation().isFileID() &&
463         !strcmp(SourceMgr.getPresumedLoc(Tok.getLocation()).getFilename(),
464                 "<built-in>"));
465
466  // Read all the preprocessed tokens, printing them out to the stream.
467  PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
468  *OS << '\n';
469}
470
471