2 * Copyright 2001-2007 Adrian Thurston <thurston@cs.queensu.ca>
5 /* This file is part of Ragel.
7 * Ragel is free software; you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation; either version 2 of the License, or
10 * (at your option) any later version.
12 * Ragel is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 * GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with Ragel; if not, write to the Free Software
19 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
30 #include <sys/types.h>
45 #define S_IRUSR _S_IREAD
46 #define S_IWUSR _S_IWRITE
54 /* Parameters and output. */
69 using std::streamsize;
71 /* Controls minimization. */
72 MinimizeLevel minimizeLevel = MinimizePartition2;
73 MinimizeOpt minimizeOpt = MinimizeMostOps;
75 /* Graphviz dot file generation. */
76 char *machineSpec = 0, *machineName = 0;
77 bool machineSpecFound = false;
78 bool wantDupsRemoved = true;
80 bool printStatistics = false;
81 bool frontendOnly = false;
82 bool generateDot = false;
84 ArgsVector frontendArgs;
85 ArgsVector backendArgs;
86 ArgsVector includePaths;
88 /* Print a summary of the options. */
92 "usage: ragel [options] file\n"
94 " -h, -H, -?, --help Print this usage and exit\n"
95 " -v, --version Print version information and exit\n"
96 " -o <file> Write output to <file>\n"
97 " -s Print some statistics on stderr\n"
98 " -d Do not remove duplicates from action lists\n"
99 " -I <dir> Add <dir> to the list of directories to search\n"
100 " for included an imported files\n"
101 "error reporting format:\n"
102 " --error-format=gnu file:line:column: message (default)\n"
103 " --error-format=msvc file(line,column): message\n"
104 "fsm minimization:\n"
105 " -n Do not perform minimization\n"
106 " -m Minimize at the end of the compilation\n"
107 " -l Minimize after most operations (default)\n"
108 " -e Minimize after every operation\n"
110 " -x Run the frontend only: emit XML intermediate format\n"
111 " -V Generate a dot file for Graphviz\n"
112 " -p Display printable characters on labels\n"
113 " -S <spec> FSM specification to output (for rlgen-dot)\n"
114 " -M <machine> Machine definition/instantiation to output (for rlgen-dot)\n"
116 " -C The host language is C, C++, Obj-C or Obj-C++ (default)\n"
117 " -D The host language is D\n"
118 " -J The host language is Java\n"
119 " -R The host language is Ruby\n"
120 " -A The host language is C#\n"
121 "line direcives: (C/D/C# only)\n"
122 " -L Inhibit writing of #line directives\n"
123 "code style: (C/Ruby/C# only)\n"
124 " -T0 Table driven FSM (default)\n"
125 " -T1 Faster table driven FSM\n"
126 " -F0 Flat table driven FSM\n"
127 " -F1 Faster flat table-driven FSM\n"
128 "code style: (C/C# only)\n"
129 " -G0 Goto-driven FSM\n"
130 " -G1 Faster goto-driven FSM\n"
131 "code style: (C only)\n"
132 " -G2 Really fast goto-driven FSM\n"
133 " -P<N> N-Way Split really fast goto-driven FSM\n"
139 /* Print version information and exit. */
142 cout << "Ragel State Machine Compiler version " VERSION << " " PUBDATE << endl <<
143 "Copyright (c) 2001-2007 by Adrian Thurston" << endl;
147 /* Error reporting format. */
148 ErrorFormat errorFormat = ErrorFormatGNU;
150 InputLoc makeInputLoc( const char *fileName, int line, int col)
152 InputLoc loc = { fileName, line, col };
156 ostream &operator<<( ostream &out, const InputLoc &loc )
158 assert( loc.fileName != 0 );
159 switch ( errorFormat ) {
160 case ErrorFormatMSVC:
161 out << loc.fileName << "(" << loc.line;
163 out << "," << loc.col;
168 out << loc.fileName << ":" << loc.line;
170 out << ":" << loc.col;
176 /* Total error count. */
177 int gblErrorCount = 0;
179 /* Print the opening to a warning in the input, then return the error ostream. */
180 ostream &warning( const InputLoc &loc )
182 cerr << loc << ": warning: ";
186 /* Print the opening to a program error, then return the error stream. */
190 cerr << PROGNAME ": ";
194 ostream &error( const InputLoc &loc )
201 void escapeLineDirectivePath( std::ostream &out, char *path )
203 for ( char *pc = path; *pc != 0; pc++ ) {
211 void processArgs( int argc, char **argv, char *&inputFileName, char *&outputFileName )
213 ParamCheck pc("xo:dnmleabjkS:M:I:CDJRAvHh?-:sT:F:G:P:LpV", argc, argv);
215 while ( pc.check() ) {
216 switch ( pc.state ) {
217 case ParamCheck::match:
218 switch ( pc.parameter ) {
229 if ( *pc.paramArg == 0 )
230 error() << "a zero length output file name was given" << endl;
231 else if ( outputFileName != 0 )
232 error() << "more than one output file name was given" << endl;
234 /* Ok, remember the output file name. */
235 outputFileName = pc.paramArg;
239 /* Minimization, mostly hidden options. */
241 wantDupsRemoved = false;
242 frontendArgs.append( "-d" );
245 /* Minimization, mostly hidden options. */
247 minimizeOpt = MinimizeNone;
248 frontendArgs.append( "-n" );
251 minimizeOpt = MinimizeEnd;
252 frontendArgs.append( "-m" );
255 minimizeOpt = MinimizeMostOps;
256 frontendArgs.append( "-l" );
259 minimizeOpt = MinimizeEveryOp;
260 frontendArgs.append( "-e" );
263 minimizeLevel = MinimizeApprox;
264 frontendArgs.append( "-a" );
267 minimizeLevel = MinimizeStable;
268 frontendArgs.append( "-b" );
271 minimizeLevel = MinimizePartition1;
272 frontendArgs.append( "-j" );
275 minimizeLevel = MinimizePartition2;
276 frontendArgs.append( "-k" );
281 if ( *pc.paramArg == 0 )
282 error() << "please specify an argument to -S" << endl;
283 else if ( machineSpec != 0 )
284 error() << "more than one -S argument was given" << endl;
286 /* Ok, remember the path to the machine to generate. */
287 machineSpec = pc.paramArg;
288 frontendArgs.append( "-S" );
289 frontendArgs.append( pc.paramArg );
295 if ( *pc.paramArg == 0 )
296 error() << "please specify an argument to -M" << endl;
297 else if ( machineName != 0 )
298 error() << "more than one -M argument was given" << endl;
300 /* Ok, remember the machine name to generate. */
301 machineName = pc.paramArg;
302 frontendArgs.append( "-M" );
303 frontendArgs.append( pc.paramArg );
308 if ( *pc.paramArg == 0 )
309 error() << "please specify an argument to -I" << endl;
311 includePaths.append( pc.paramArg );
312 frontendArgs.append( "-I" );
313 frontendArgs.append( pc.paramArg );
317 /* Host language types. */
319 hostLang = &hostLangC;
320 frontendArgs.append( "-C" );
323 hostLang = &hostLangD;
324 frontendArgs.append( "-D" );
327 hostLang = &hostLangJava;
328 frontendArgs.append( "-J" );
331 hostLang = &hostLangRuby;
332 frontendArgs.append( "-R" );
335 hostLang = &hostLangCSharp;
336 frontendArgs.append( "-A" );
339 /* Version and help. */
343 case 'H': case 'h': case '?':
347 printStatistics = true;
348 frontendArgs.append( "-s" );
351 char *eq = strchr( pc.paramArg, '=' );
356 if ( strcmp( pc.paramArg, "help" ) == 0 )
358 else if ( strcmp( pc.paramArg, "version" ) == 0 )
360 else if ( strcmp( pc.paramArg, "error-format" ) == 0 ) {
362 error() << "expecting '=value' for error-format" << endl;
363 else if ( strcmp( eq, "gnu" ) == 0 ) {
364 errorFormat = ErrorFormatGNU;
365 frontendArgs.append( "--error-format=gnu" );
367 else if ( strcmp( eq, "msvc" ) == 0 ) {
368 errorFormat = ErrorFormatMSVC;
369 frontendArgs.append( "--error-format=msvc" );
372 error() << "invalid value for error-format" << endl;
376 error() << "--" << pc.paramArg <<
377 " is an invalid argument" << endl;
382 /* Passthrough args. */
384 backendArgs.append( "-T" );
385 backendArgs.append( pc.paramArg );
388 backendArgs.append( "-F" );
389 backendArgs.append( pc.paramArg );
392 backendArgs.append( "-G" );
393 backendArgs.append( pc.paramArg );
396 backendArgs.append( "-P" );
397 backendArgs.append( pc.paramArg );
400 backendArgs.append( "-p" );
403 backendArgs.append( "-L" );
408 case ParamCheck::invalid:
409 error() << "-" << pc.parameter << " is an invalid argument" << endl;
412 case ParamCheck::noparam:
413 /* It is interpreted as an input file. */
414 if ( *pc.curArg == 0 )
415 error() << "a zero length input file name was given" << endl;
416 else if ( inputFileName != 0 )
417 error() << "more than one input file name was given" << endl;
419 /* OK, Remember the filename. */
420 inputFileName = pc.curArg;
427 int frontend( char *inputFileName, char *outputFileName )
429 /* Open the input file for reading. */
430 assert( inputFileName != 0 );
431 ifstream *inFile = new ifstream( inputFileName );
432 istream *inStream = inFile;
433 if ( ! inFile->is_open() )
434 error() << "could not open " << inputFileName << " for reading" << endp;
436 /* Used for just a few things. */
437 std::ostringstream hostData;
439 if ( machineSpec == 0 && machineName == 0 )
440 hostData << "<host line=\"1\" col=\"1\">";
442 Scanner scanner( inputFileName, *inStream, hostData, 0, 0, 0, false );
445 /* Finished, final check for errors.. */
446 if ( gblErrorCount > 0 )
449 /* Now send EOF to all parsers. */
450 terminateAllParsers();
452 /* Finished, final check for errors.. */
453 if ( gblErrorCount > 0 )
456 if ( machineSpec == 0 && machineName == 0 )
457 hostData << "</host>\n";
459 if ( gblErrorCount > 0 )
462 ostream *outputFile = 0;
463 if ( outputFileName != 0 )
464 outputFile = new ofstream( outputFileName );
468 /* Write the machines, then the surrounding code. */
469 writeMachines( *outputFile, hostData.str(), inputFileName );
471 /* Close the intermediate file. */
472 if ( outputFileName != 0 )
475 return gblErrorCount > 0;
478 char *makeIntermedTemplate( char *baseFileName )
481 const char *templ = "ragel-XXXXXX.xml";
482 char *lastSlash = strrchr( baseFileName, '/' );
483 if ( lastSlash == 0 ) {
484 result = new char[strlen(templ)+1];
485 strcpy( result, templ );
488 int baseLen = lastSlash - baseFileName + 1;
489 result = new char[baseLen + strlen(templ) + 1];
490 memcpy( result, baseFileName, baseLen );
491 strcpy( result+baseLen, templ );
496 char *openIntermed( char *inputFileName, char *outputFileName )
501 /* Which filename do we use as the base? */
502 char *baseFileName = outputFileName != 0 ? outputFileName : inputFileName;
504 /* The template for the intermediate file name. */
505 char *intermedFileName = makeIntermedTemplate( baseFileName );
507 /* Randomize the name and try to open. */
508 char fnChars[] = "abcdefghijklmnopqrstuvwxyz0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ";
509 char *firstX = strrchr( intermedFileName, 'X' ) - 5;
510 for ( int tries = 0; tries < 20; tries++ ) {
511 /* Choose a random name. */
512 for ( int x = 0; x < 6; x++ )
513 firstX[x] = fnChars[rand() % 52];
515 /* Try to open the file. */
516 int fd = ::open( intermedFileName, O_WRONLY|O_EXCL|O_CREAT, S_IRUSR|S_IWUSR );
519 /* Success. Close the file immediately and return the name for use
520 * by the child processes. */
522 result = intermedFileName;
526 if ( errno == EACCES ) {
527 error() << "failed to open temp file " << intermedFileName <<
528 ", access denied" << endp;
533 error() << "abnormal error: cannot find unique name for temp file" << endp;
539 void cleanExit( char *intermed, int status )
547 /* If any forward slash is found in argv0 then it is assumed that the path is
548 * explicit and the path to the backend executable should be derived from
549 * that. Whe check that location and also go up one then inside a directory of
550 * the same name in case we are executing from the source tree. If no forward
551 * slash is found it is assumed the file is being run from the installed
552 * location. The PREFIX supplied during configuration is used. */
553 char **makePathChecksUnix( const char *argv0, const char *progName )
555 char **result = new char*[3];
556 const char *lastSlash = strrchr( argv0, '/' );
559 if ( lastSlash != 0 ) {
560 char *path = strdup( argv0 );
561 int givenPathLen = (lastSlash - argv0) + 1;
562 path[givenPathLen] = 0;
564 int progNameLen = strlen(progName);
565 int length = givenPathLen + progNameLen + 1;
566 char *check = new char[length];
567 sprintf( check, "%s%s", path, progName );
568 result[numChecks++] = check;
570 length = givenPathLen + 3 + progNameLen + 1 + progNameLen + 1;
571 check = new char[length];
572 sprintf( check, "%s../%s/%s", path, progName, progName );
573 result[numChecks++] = check;
576 int prefixLen = strlen(PREFIX);
577 int progNameLen = strlen(progName);
578 int length = prefixLen + 5 + progNameLen + 1;
579 char *check = new char[length];
581 sprintf( check, PREFIX "/bin/%s", progName );
582 result[numChecks++] = check;
585 result[numChecks] = 0;
590 void forkAndExec( const char *progName, char **pathChecks,
591 ArgsVector &args, char *intermed )
595 /* Error, no child created. */
596 error() << "failed to fork for " << progName << endl;
597 cleanExit( intermed, 1 );
599 else if ( pid == 0 ) {
601 while ( *pathChecks != 0 ) {
602 /* Execv does not modify argv, it just uses the const form that is
603 * compatible with the most code. Ours not included. */
604 execv( *pathChecks, (char *const*) args.data );
607 error() << "failed to exec " << progName << endl;
608 cleanExit( intermed, 1 );
611 /* Parent process, wait for the child. */
615 /* What happened with the child. */
616 if ( ! WIFEXITED( status ) ) {
617 error() << progName << " did not exit normally" << endl;
618 cleanExit( intermed, 1 );
621 if ( WEXITSTATUS(status) != 0 )
622 cleanExit( intermed, WEXITSTATUS(status) );
627 /* GetModuleFileNameEx is used to find out where the the current process's
628 * binary is. That location is searched first. If that fails then we go up one
629 * directory and look for the executable inside a directory of the same name
630 * in case we are executing from the source tree.
632 char **makePathChecksWin( const char *progName )
635 char *imageFileName = new char[len];
636 HANDLE h = GetCurrentProcess();
637 len = GetModuleFileNameEx( h, NULL, imageFileName, len );
638 imageFileName[len] = 0;
640 char **result = new char*[3];
641 const char *lastSlash = strrchr( imageFileName, '\\' );
644 assert( lastSlash != 0 );
645 char *path = strdup( imageFileName );
646 int givenPathLen = (lastSlash - imageFileName) + 1;
647 path[givenPathLen] = 0;
649 int progNameLen = strlen(progName);
650 int length = givenPathLen + progNameLen + 1;
651 char *check = new char[length];
652 sprintf( check, "%s%s", path, progName );
653 result[numChecks++] = check;
655 length = givenPathLen + 3 + progNameLen + 1 + progNameLen + 1;
656 check = new char[length];
657 sprintf( check, "%s..\\%s\\%s", path, progName, progName );
658 result[numChecks++] = check;
660 result[numChecks] = 0;
664 void spawn( const char *progName, char **pathChecks,
665 ArgsVector &args, char *intermed )
668 while ( *pathChecks != 0 ) {
669 //cerr << "trying to execute " << *pathChecks << endl;
670 result = _spawnv( _P_WAIT, *pathChecks, args.data );
671 if ( result >= 0 || errno != ENOENT )
677 error() << "failed to spawn " << progName << endl;
678 cleanExit( intermed, 1 );
682 cleanExit( intermed, 1 );
687 void execFrontend( const char *argv0, char *inputFileName, char *intermed )
689 /* The frontend program name. */
690 const char *progName = "ragel";
692 frontendArgs.insert( 0, progName );
693 frontendArgs.insert( 1, "-x" );
694 frontendArgs.append( "-o" );
695 frontendArgs.append( intermed );
696 frontendArgs.append( inputFileName );
697 frontendArgs.append( 0 );
700 char **pathChecks = makePathChecksUnix( argv0, progName );
701 forkAndExec( progName, pathChecks, frontendArgs, intermed );
703 char **pathChecks = makePathChecksWin( progName );
704 spawn( progName, pathChecks, frontendArgs, intermed );
708 void execBackend( const char *argv0, char *intermed, char *outputFileName )
710 /* Locate the backend program */
711 const char *progName = 0;
713 progName = "rlgen-dot";
715 switch ( hostLang->lang ) {
718 progName = "rlgen-cd";
721 progName = "rlgen-java";
724 progName = "rlgen-ruby";
726 case HostLang::CSharp:
727 progName = "rlgen-csharp";
731 backendArgs.insert( 0, progName );
732 if ( outputFileName != 0 ) {
733 backendArgs.append( "-o" );
734 backendArgs.append( outputFileName );
736 backendArgs.append( intermed );
737 backendArgs.append( 0 );
740 char **pathChecks = makePathChecksUnix( argv0, progName );
741 forkAndExec( progName, pathChecks, backendArgs, intermed );
743 char **pathChecks = makePathChecksWin( progName );
744 spawn( progName, pathChecks, backendArgs, intermed );
748 /* Main, process args and call yyparse to start scanning input. */
749 int main(int argc, char **argv)
751 char *inputFileName = 0;
752 char *outputFileName = 0;
754 processArgs( argc, argv, inputFileName, outputFileName );
756 /* If -M or -S are given and we're not generating a dot file then invoke
757 * the frontend. These options are not useful with code generators. */
758 if ( machineName != 0 || machineSpec != 0 ) {
763 /* Require an input file. If we use standard in then we won't have a file
764 * name on which to base the output. */
765 if ( inputFileName == 0 )
766 error() << "no input file given" << endl;
768 /* Bail on argument processing errors. */
769 if ( gblErrorCount > 0 )
772 /* Make sure we are not writing to the same file as the input file. */
773 if ( inputFileName != 0 && outputFileName != 0 &&
774 strcmp( inputFileName, outputFileName ) == 0 )
776 error() << "output file \"" << outputFileName <<
777 "\" is the same as the input file" << endp;
781 return frontend( inputFileName, outputFileName );
783 char *intermed = openIntermed( inputFileName, outputFileName );
785 /* From here on in the cleanExit function should be used to exit. */
787 /* Run the frontend, then the backend processes. */
788 execFrontend( argv[0], inputFileName, intermed );
789 execBackend( argv[0], intermed, outputFileName );
791 /* Clean up the intermediate. */
792 cleanExit( intermed, 0 );