PrintPreprocessedOutput.cpp revision d76fbda5da19c34752aa582e690a4dbd1cae39cd 
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 << '#';
145  if (PP.getLangOptions().Microsoft)
146    OS << "line";
147  OS << ' ' << LineNo << ' ' << '"';
148
149  OS.write(&CurFilename[0], CurFilename.size());
150  OS << '"';
151
152  if (!PP.getLangOptions().Microsoft) {
153    if (ExtraLen)
154      OS.write(Extra, ExtraLen);
155
156    if (FileType == SrcMgr::C_System)
157      OS.write(" 3", 2);
158    else if (FileType == SrcMgr::C_ExternCSystem)
159      OS.write(" 3 4", 4);
160  }
161  OS << '\n';
162}
163
164/// MoveToLine - Move the output to the source line specified by the location
165/// object.  We can do this by emitting some number of \n's, or be emitting a
166/// #line directive.  This returns false if already at the specified line, true
167/// if some newlines were emitted.
168bool PrintPPOutputPPCallbacks::MoveToLine(SourceLocation Loc) {
169  unsigned LineNo = PP.getSourceManager().getInstantiationLineNumber(Loc);
170
171  if (DisableLineMarkers) {
172    if (LineNo == CurLine) return false;
173
174    CurLine = LineNo;
175
176    if (!EmittedTokensOnThisLine && !EmittedMacroOnThisLine)
177      return true;
178
179    OS << '\n';
180    EmittedTokensOnThisLine = false;
181    EmittedMacroOnThisLine = false;
182    return true;
183  }
184
185  // If this line is "close enough" to the original line, just print newlines,
186  // otherwise print a #line directive.
187  if (LineNo-CurLine <= 8) {
188    if (LineNo-CurLine == 1)
189      OS << '\n';
190    else if (LineNo == CurLine)
191      return false;    // Spelling line moved, but instantiation line didn't.
192    else {
193      const char *NewLines = "\n\n\n\n\n\n\n\n";
194      OS.write(NewLines, LineNo-CurLine);
195    }
196  } else {
197    WriteLineInfo(LineNo, 0, 0);
198  }
199
200  CurLine = LineNo;
201  return true;
202}
203
204
205/// FileChanged - Whenever the preprocessor enters or exits a #include file
206/// it invokes this handler.  Update our conception of the current source
207/// position.
208void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
209                                           FileChangeReason Reason,
210                                       SrcMgr::CharacteristicKind NewFileType) {
211  // Unless we are exiting a #include, make sure to skip ahead to the line the
212  // #include directive was at.
213  SourceManager &SourceMgr = PP.getSourceManager();
214  if (Reason == PPCallbacks::EnterFile) {
215    SourceLocation IncludeLoc = SourceMgr.getPresumedLoc(Loc).getIncludeLoc();
216    if (IncludeLoc.isValid())
217      MoveToLine(IncludeLoc);
218  } else if (Reason == PPCallbacks::SystemHeaderPragma) {
219    MoveToLine(Loc);
220
221    // TODO GCC emits the # directive for this directive on the line AFTER the
222    // directive and emits a bunch of spaces that aren't needed.  Emulate this
223    // strange behavior.
224  }
225
226  Loc = SourceMgr.getInstantiationLoc(Loc);
227  // FIXME: Should use presumed line #!
228  CurLine = SourceMgr.getInstantiationLineNumber(Loc);
229
230  if (DisableLineMarkers) return;
231
232  CurFilename.clear();
233  CurFilename += SourceMgr.getPresumedLoc(Loc).getFilename();
234  Lexer::Stringify(CurFilename);
235  FileType = NewFileType;
236
237  if (!Initialized) {
238    WriteLineInfo(CurLine);
239    Initialized = true;
240  }
241
242  switch (Reason) {
243  case PPCallbacks::EnterFile:
244    WriteLineInfo(CurLine, " 1", 2);
245    break;
246  case PPCallbacks::ExitFile:
247    WriteLineInfo(CurLine, " 2", 2);
248    break;
249  case PPCallbacks::SystemHeaderPragma:
250  case PPCallbacks::RenameFile:
251    WriteLineInfo(CurLine);
252    break;
253  }
254}
255
256/// Ident - Handle #ident directives when read by the preprocessor.
257///
258void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, const std::string &S) {
259  MoveToLine(Loc);
260
261  OS.write("#ident ", strlen("#ident "));
262  OS.write(&S[0], S.size());
263  EmittedTokensOnThisLine = true;
264}
265
266/// MacroDefined - This hook is called whenever a macro definition is seen.
267void PrintPPOutputPPCallbacks::MacroDefined(const IdentifierInfo *II,
268                                            const MacroInfo *MI) {
269  // Only print out macro definitions in -dD mode.
270  if (!DumpDefines ||
271      // Ignore __FILE__ etc.
272      MI->isBuiltinMacro()) return;
273
274  MoveToLine(MI->getDefinitionLoc());
275  PrintMacroDefinition(*II, *MI, PP, OS);
276  EmittedMacroOnThisLine = true;
277}
278
279
280void PrintPPOutputPPCallbacks::PragmaComment(SourceLocation Loc,
281                                             const IdentifierInfo *Kind,
282                                             const std::string &Str) {
283  MoveToLine(Loc);
284  OS << "#pragma comment(" << Kind->getName();
285
286  if (!Str.empty()) {
287    OS << ", \"";
288
289    for (unsigned i = 0, e = Str.size(); i != e; ++i) {
290      unsigned char Char = Str[i];
291      if (isprint(Char) && Char != '\\' && Char != '"')
292        OS << (char)Char;
293      else  // Output anything hard as an octal escape.
294        OS << '\\'
295           << (char)('0'+ ((Char >> 6) & 7))
296           << (char)('0'+ ((Char >> 3) & 7))
297           << (char)('0'+ ((Char >> 0) & 7));
298    }
299    OS << '"';
300  }
301
302  OS << ')';
303  EmittedTokensOnThisLine = true;
304}
305
306
307/// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
308/// is called for the first token on each new line.  If this really is the start
309/// of a new logical line, handle it and return true, otherwise return false.
310/// This may not be the start of a logical line because the "start of line"
311/// marker is set for spelling lines, not instantiation ones.
312bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
313  // Figure out what line we went to and insert the appropriate number of
314  // newline characters.
315  if (!MoveToLine(Tok.getLocation()))
316    return false;
317
318  // Print out space characters so that the first token on a line is
319  // indented for easy reading.
320  const SourceManager &SourceMgr = PP.getSourceManager();
321  unsigned ColNo = SourceMgr.getInstantiationColumnNumber(Tok.getLocation());
322
323  // This hack prevents stuff like:
324  // #define HASH #
325  // HASH define foo bar
326  // From having the # character end up at column 1, which makes it so it
327  // is not handled as a #define next time through the preprocessor if in
328  // -fpreprocessed mode.
329  if (ColNo <= 1 && Tok.is(tok::hash))
330    OS << ' ';
331
332  // Otherwise, indent the appropriate number of spaces.
333  for (; ColNo > 1; --ColNo)
334    OS << ' ';
335
336  return true;
337}
338
339void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
340                                                     unsigned Len) {
341  unsigned NumNewlines = 0;
342  for (; Len; --Len, ++TokStr) {
343    if (*TokStr != '\n' &&
344        *TokStr != '\r')
345      continue;
346
347    ++NumNewlines;
348
349    // If we have \n\r or \r\n, skip both and count as one line.
350    if (Len != 1 &&
351        (TokStr[1] == '\n' || TokStr[1] == '\r') &&
352        TokStr[0] != TokStr[1])
353      ++TokStr, --Len;
354  }
355
356  if (NumNewlines == 0) return;
357
358  CurLine += NumNewlines;
359}
360
361
362namespace {
363struct UnknownPragmaHandler : public PragmaHandler {
364  const char *Prefix;
365  PrintPPOutputPPCallbacks *Callbacks;
366
367  UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks)
368    : PragmaHandler(0), Prefix(prefix), Callbacks(callbacks) {}
369  virtual void HandlePragma(Preprocessor &PP, Token &PragmaTok) {
370    // Figure out what line we went to and insert the appropriate number of
371    // newline characters.
372    Callbacks->MoveToLine(PragmaTok.getLocation());
373    Callbacks->OS.write(Prefix, strlen(Prefix));
374
375    // Read and print all of the pragma tokens.
376    while (PragmaTok.isNot(tok::eom)) {
377      if (PragmaTok.hasLeadingSpace())
378        Callbacks->OS << ' ';
379      std::string TokSpell = PP.getSpelling(PragmaTok);
380      Callbacks->OS.write(&TokSpell[0], TokSpell.size());
381      PP.LexUnexpandedToken(PragmaTok);
382    }
383    Callbacks->OS << '\n';
384  }
385};
386} // end anonymous namespace
387
388
389static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
390                                    PrintPPOutputPPCallbacks *Callbacks,
391                                    llvm::raw_ostream &OS) {
392  char Buffer[256];
393  Token PrevTok;
394  while (1) {
395
396    // If this token is at the start of a line, emit newlines if needed.
397    if (Tok.isAtStartOfLine() && Callbacks->HandleFirstTokOnLine(Tok)) {
398      // done.
399    } else if (Tok.hasLeadingSpace() ||
400               // If we haven't emitted a token on this line yet, PrevTok isn't
401               // useful to look at and no concatenation could happen anyway.
402               (Callbacks->hasEmittedTokensOnThisLine() &&
403                // Don't print "-" next to "-", it would form "--".
404                Callbacks->AvoidConcat(PrevTok, Tok))) {
405      OS << ' ';
406    }
407
408    if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
409      OS << II->getName();
410    } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
411               Tok.getLiteralData()) {
412      OS.write(Tok.getLiteralData(), Tok.getLength());
413    } else if (Tok.getLength() < 256) {
414      const char *TokPtr = Buffer;
415      unsigned Len = PP.getSpelling(Tok, TokPtr);
416      OS.write(TokPtr, Len);
417
418      // Tokens that can contain embedded newlines need to adjust our current
419      // line number.
420      if (Tok.getKind() == tok::comment)
421        Callbacks->HandleNewlinesInToken(TokPtr, Len);
422    } else {
423      std::string S = PP.getSpelling(Tok);
424      OS.write(&S[0], S.size());
425
426      // Tokens that can contain embedded newlines need to adjust our current
427      // line number.
428      if (Tok.getKind() == tok::comment)
429        Callbacks->HandleNewlinesInToken(&S[0], S.size());
430    }
431    Callbacks->SetEmittedTokensOnThisLine();
432
433    if (Tok.is(tok::eof)) break;
434
435    PrevTok = Tok;
436    PP.Lex(Tok);
437  }
438}
439
440namespace {
441  struct SortMacrosByID {
442    typedef std::pair<IdentifierInfo*, MacroInfo*> id_macro_pair;
443    bool operator()(const id_macro_pair &LHS, const id_macro_pair &RHS) const {
444      return LHS.first->getName() < RHS.first->getName();
445    }
446  };
447}
448
449static void DoPrintMacros(Preprocessor &PP, llvm::raw_ostream *OS) {
450  // -dM mode just scans and ignores all tokens in the files, then dumps out
451  // the macro table at the end.
452  PP.EnterMainSourceFile();
453
454  Token Tok;
455  do PP.Lex(Tok);
456  while (Tok.isNot(tok::eof));
457
458  std::vector<std::pair<IdentifierInfo*, MacroInfo*> > MacrosByID;
459  for (Preprocessor::macro_iterator I = PP.macro_begin(), E = PP.macro_end();
460       I != E; ++I)
461    MacrosByID.push_back(*I);
462  std::sort(MacrosByID.begin(), MacrosByID.end(), SortMacrosByID());
463
464  for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
465    MacroInfo &MI = *MacrosByID[i].second;
466    // Ignore computed macros like __LINE__ and friends.
467    if (MI.isBuiltinMacro()) continue;
468
469    PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
470    *OS << "\n";
471  }
472}
473
474/// DoPrintPreprocessedInput - This implements -E mode.
475///
476void clang::DoPrintPreprocessedInput(Preprocessor &PP, llvm::raw_ostream *OS,
477                                     const PreprocessorOutputOptions &Opts) {
478  // Show macros with no output is handled specially.
479  if (!Opts.ShowCPP) {
480    assert(Opts.ShowMacros && "Not yet implemented!");
481    DoPrintMacros(PP, OS);
482    return;
483  }
484
485  // Inform the preprocessor whether we want it to retain comments or not, due
486  // to -C or -CC.
487  PP.SetCommentRetentionState(Opts.ShowComments, Opts.ShowMacroComments);
488
489  OS->SetBufferSize(64*1024);
490
491  PrintPPOutputPPCallbacks *Callbacks =
492      new PrintPPOutputPPCallbacks(PP, *OS, !Opts.ShowLineMarkers,
493                                   Opts.ShowMacros);
494  PP.AddPragmaHandler(0, new UnknownPragmaHandler("#pragma", Callbacks));
495  PP.AddPragmaHandler("GCC", new UnknownPragmaHandler("#pragma GCC",
496                                                      Callbacks));
497
498  PP.setPPCallbacks(Callbacks);
499
500  // After we have configured the preprocessor, enter the main file.
501  PP.EnterMainSourceFile();
502
503  // Consume all of the tokens that come from the predefines buffer.  Those
504  // should not be emitted into the output and are guaranteed to be at the
505  // start.
506  const SourceManager &SourceMgr = PP.getSourceManager();
507  Token Tok;
508  do PP.Lex(Tok);
509  while (Tok.isNot(tok::eof) && Tok.getLocation().isFileID() &&
510         !strcmp(SourceMgr.getPresumedLoc(Tok.getLocation()).getFilename(),
511                 "<built-in>"));
512
513  // Read all the preprocessed tokens, printing them out to the stream.
514  PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
515  *OS << '\n';
516}
517
518