ragel/main.cpp

   1 /*
   2  *  Copyright 2001-2007 Adrian Thurston <thurston@cs.queensu.ca>
   3  */
   4
   5 /*  This file is part of Ragel.
   6  *
   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.
  11  *
  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.
  16  *
  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
  20  */
  21
  22 #include <stdlib.h>
  23 #include <string.h>
  24 #include <stdio.h>
  25 #include <iostream>
  26 #include <fstream>
  27 #include <unistd.h>
  28 #include <sstream>
  29 #include <unistd.h>
  30 #include <sys/types.h>
  31 #include <sys/stat.h>
  32 #include <fcntl.h>
  33 #include <errno.h>
  34
  35 #ifdef _WIN32
  36 #include <windows.h>
  37 #include <psapi.h>
  38 #include <time.h>
  39 #include <io.h>
  40 #include <process.h>
  41
  42 #if _MSC_VER
  43 #define S_IRUSR _S_IREAD
  44 #define S_IWUSR _S_IWRITE
  45 #endif
  46 #endif
  47
  48 /* Parsing. */
  49 #include "ragel.h"
  50 #include "rlscan.h"
  51
  52 /* Parameters and output. */
  53 #include "pcheck.h"
  54 #include "vector.h"
  55 #include "version.h"
  56 #include "common.h"
  57 #include "xmlparse.h"
  58
  59 using std::istream;
  60 using std::ostream;
  61 using std::ifstream;
  62 using std::ofstream;
  63 using std::cin;
  64 using std::cout;
  65 using std::cerr;
  66 using std::endl;
  67 using std::ios;
  68 using std::streamsize;
  69
  70 /* Controls minimization. */
  71 MinimizeLevel minimizeLevel = MinimizePartition2;
  72 MinimizeOpt minimizeOpt = MinimizeMostOps;
  73
  74 /* Graphviz dot file generation. */
  75 const char *machineSpec = 0, *machineName = 0;
  76 bool machineSpecFound = false;
  77 bool wantDupsRemoved = true;
  78
  79 bool printStatistics = false;
  80 bool frontendOnly = false;
  81 bool generateDot = false;
  82
  83 /* Target language and output style. */
  84 CodeStyleEnum codeStyle = GenTables;
  85
  86 int numSplitPartitions = 0;
  87 bool noLineDirectives = false;
  88
  89 bool displayPrintables = false;
  90 bool graphvizDone = false;
  91
  92 /* Target ruby impl */
  93 RubyImplEnum rubyImpl = MRI;
  94
  95 ArgsVector includePaths;
  96
  97 istream *inStream = 0;
  98 ostream *outStream = 0;
  99 output_filter *outFilter = 0;
 100 const char *outputFileName = 0;
 101
 102 /* Print a summary of the options. */
 103 void usage()
 104 {
 105         cout <<
 106 "usage: ragel [options] file\n"
 107 "general:\n"
 108 "   -h, -H, -?, --help   Print this usage and exit\n"
 109 "   -v, --version        Print version information and exit\n"
 110 "   -o <file>            Write output to <file>\n"
 111 "   -s                   Print some statistics on stderr\n"
 112 "   -d                   Do not remove duplicates from action lists\n"
 113 "   -I <dir>             Add <dir> to the list of directories to search\n"
 114 "                        for included an imported files\n"
 115 "error reporting format:\n"
 116 "   --error-format=gnu   file:line:column: message (default)\n"
 117 "   --error-format=msvc  file(line,column): message\n"
 118 "fsm minimization:\n"
 119 "   -n                   Do not perform minimization\n"
 120 "   -m                   Minimize at the end of the compilation\n"
 121 "   -l                   Minimize after most operations (default)\n"
 122 "   -e                   Minimize after every operation\n"
 123 "visualization:\n"
 124 "   -x                   Run the frontend only: emit XML intermediate format\n"
 125 "   -V                   Generate a dot file for Graphviz\n"
 126 "   -p                   Display printable characters on labels\n"
 127 "   -S <spec>            FSM specification to output (for rlgen-dot)\n"
 128 "   -M <machine>         Machine definition/instantiation to output (for rlgen-dot)\n"
 129 "host language:\n"
 130 "   -C                   The host language is C, C++, Obj-C or Obj-C++ (default)\n"
 131 "   -D                   The host language is D\n"
 132 "   -J                   The host language is Java\n"
 133 "   -R                   The host language is Ruby\n"
 134 "   -A                   The host language is C#\n"
 135 "line direcives: (C/D/C# only)\n"
 136 "   -L                   Inhibit writing of #line directives\n"
 137 "code style: (C/Ruby/C# only)\n"
 138 "   -T0                  Table driven FSM (default)\n"
 139 "   -T1                  Faster table driven FSM\n"
 140 "   -F0                  Flat table driven FSM\n"
 141 "   -F1                  Faster flat table-driven FSM\n"
 142 "code style: (C/C# only)\n"
 143 "   -G0                  Goto-driven FSM\n"
 144 "   -G1                  Faster goto-driven FSM\n"
 145 "code style: (C only)\n"
 146 "   -G2                  Really fast goto-driven FSM\n"
 147 "   -P<N>                N-Way Split really fast goto-driven FSM\n"
 148         ;
 149
 150         exit(0);
 151 }
 152
 153 /* Print version information and exit. */
 154 void version()
 155 {
 156         cout << "Ragel State Machine Compiler version " VERSION << " " PUBDATE << endl <<
 157                         "Copyright (c) 2001-2007 by Adrian Thurston" << endl;
 158         exit(0);
 159 }
 160
 161 /* Error reporting format. */
 162 ErrorFormat errorFormat = ErrorFormatGNU;
 163
 164 InputLoc makeInputLoc( const char *fileName, int line, int col)
 165 {
 166         InputLoc loc = { fileName, line, col };
 167         return loc;
 168 }
 169
 170 ostream &operator<<( ostream &out, const InputLoc &loc )
 171 {
 172         assert( loc.fileName != 0 );
 173         switch ( errorFormat ) {
 174         case ErrorFormatMSVC:
 175                 out << loc.fileName << "(" << loc.line;
 176                 if ( loc.col )
 177                         out << "," << loc.col;
 178                 out << ")";
 179                 break;
 180
 181         default:
 182                 out << loc.fileName << ":" << loc.line;
 183                 if ( loc.col )
 184                         out << ":" << loc.col;
 185                 break;
 186         }
 187         return out;
 188 }
 189
 190 /* Total error count. */
 191 int gblErrorCount = 0;
 192
 193 /* Print the opening to a warning in the input, then return the error ostream. */
 194 ostream &warning( const InputLoc &loc )
 195 {
 196         cerr << loc << ": warning: ";
 197         return cerr;
 198 }
 199
 200 /* Print the opening to a program error, then return the error stream. */
 201 ostream &error()
 202 {
 203         gblErrorCount += 1;
 204         cerr << PROGNAME ": ";
 205         return cerr;
 206 }
 207
 208 ostream &error( const InputLoc &loc )
 209 {
 210         gblErrorCount += 1;
 211         cerr << loc << ": ";
 212         return cerr;
 213 }
 214
 215 void escapeLineDirectivePath( std::ostream &out, char *path )
 216 {
 217         for ( char *pc = path; *pc != 0; pc++ ) {
 218                 if ( *pc == '\\' )
 219                         out << "\\\\";
 220                 else
 221                         out << *pc;
 222         }
 223 }
 224
 225 void processArgs( int argc, const char **argv, const char *&inputFileName )
 226 {
 227         ParamCheck pc("xo:dnmleabjkS:M:I:CDJRAvHh?-:sT:F:G:P:LpV", argc, argv);
 228
 229         /* FIXME: Need to check code styles VS langauge. */
 230
 231         while ( pc.check() ) {
 232                 switch ( pc.state ) {
 233                 case ParamCheck::match:
 234                         switch ( pc.parameter ) {
 235                         case 'V':
 236                                 generateDot = true;
 237                                 break;
 238
 239                         case 'x':
 240                                 frontendOnly = true;
 241                                 break;
 242
 243                         /* Output. */
 244                         case 'o':
 245                                 if ( *pc.paramArg == 0 )
 246                                         error() << "a zero length output file name was given" << endl;
 247                                 else if ( outputFileName != 0 )
 248                                         error() << "more than one output file name was given" << endl;
 249                                 else {
 250                                         /* Ok, remember the output file name. */
 251                                         outputFileName = pc.paramArg;
 252                                 }
 253                                 break;
 254
 255                         /* Flag for turning off duplicate action removal. */
 256                         case 'd':
 257                                 wantDupsRemoved = false;
 258                                 break;
 259
 260                         /* Minimization, mostly hidden options. */
 261                         case 'n':
 262                                 minimizeOpt = MinimizeNone;
 263                                 break;
 264                         case 'm':
 265                                 minimizeOpt = MinimizeEnd;
 266                                 break;
 267                         case 'l':
 268                                 minimizeOpt = MinimizeMostOps;
 269                                 break;
 270                         case 'e':
 271                                 minimizeOpt = MinimizeEveryOp;
 272                                 break;
 273                         case 'a':
 274                                 minimizeLevel = MinimizeApprox;
 275                                 break;
 276                         case 'b':
 277                                 minimizeLevel = MinimizeStable;
 278                                 break;
 279                         case 'j':
 280                                 minimizeLevel = MinimizePartition1;
 281                                 break;
 282                         case 'k':
 283                                 minimizeLevel = MinimizePartition2;
 284                                 break;
 285
 286                         /* Machine spec. */
 287                         case 'S':
 288                                 if ( *pc.paramArg == 0 )
 289                                         error() << "please specify an argument to -S" << endl;
 290                                 else if ( machineSpec != 0 )
 291                                         error() << "more than one -S argument was given" << endl;
 292                                 else {
 293                                         /* Ok, remember the path to the machine to generate. */
 294                                         machineSpec = pc.paramArg;
 295                                 }
 296                                 break;
 297
 298                         /* Machine path. */
 299                         case 'M':
 300                                 if ( *pc.paramArg == 0 )
 301                                         error() << "please specify an argument to -M" << endl;
 302                                 else if ( machineName != 0 )
 303                                         error() << "more than one -M argument was given" << endl;
 304                                 else {
 305                                         /* Ok, remember the machine name to generate. */
 306                                         machineName = pc.paramArg;
 307                                 }
 308                                 break;
 309
 310                         case 'I':
 311                                 if ( *pc.paramArg == 0 )
 312                                         error() << "please specify an argument to -I" << endl;
 313                                 else {
 314                                         includePaths.append( pc.paramArg );
 315                                 }
 316                                 break;
 317
 318                         /* Host language types. */
 319                         case 'C':
 320                                 hostLang = &hostLangC;
 321                                 break;
 322                         case 'D':
 323                                 hostLang = &hostLangD;
 324                                 break;
 325                         case 'J':
 326                                 hostLang = &hostLangJava;
 327                                 break;
 328                         case 'R':
 329                                 hostLang = &hostLangRuby;
 330                                 break;
 331                         case 'A':
 332                                 hostLang = &hostLangCSharp;
 333                                 break;
 334
 335                         /* Version and help. */
 336                         case 'v':
 337                                 version();
 338                                 break;
 339                         case 'H': case 'h': case '?':
 340                                 usage();
 341                                 break;
 342                         case 's':
 343                                 printStatistics = true;
 344                                 break;
 345                         case '-': {
 346                                 char *eq = strchr( pc.paramArg, '=' );
 347
 348                                 if ( eq != 0 )
 349                                         *eq++ = 0;
 350
 351                                 if ( strcmp( pc.paramArg, "help" ) == 0 )
 352                                         usage();
 353                                 else if ( strcmp( pc.paramArg, "version" ) == 0 )
 354                                         version();
 355                                 else if ( strcmp( pc.paramArg, "error-format" ) == 0 ) {
 356                                         if ( eq == 0 )
 357                                                 error() << "expecting '=value' for error-format" << endl;
 358                                         else if ( strcmp( eq, "gnu" ) == 0 )
 359                                                 errorFormat = ErrorFormatGNU;
 360                                         else if ( strcmp( eq, "msvc" ) == 0 )
 361                                                 errorFormat = ErrorFormatMSVC;
 362                                         else
 363                                                 error() << "invalid value for error-format" << endl;
 364                                 }
 365                                 else if ( strcmp( pc.paramArg, "rbx" ) == 0 )
 366                                         rubyImpl = Rubinius;
 367                                 else {
 368                                         error() << "--" << pc.paramArg <<
 369                                                         " is an invalid argument" << endl;
 370                                 }
 371                                 break;
 372                         }
 373
 374                         /* Passthrough args. */
 375                         case 'T':
 376                                 if ( pc.paramArg[0] == '0' )
 377                                         codeStyle = GenTables;
 378                                 else if ( pc.paramArg[0] == '1' )
 379                                         codeStyle = GenFTables;
 380                                 else {
 381                                         error() << "-T" << pc.paramArg[0] <<
 382                                                         " is an invalid argument" << endl;
 383                                         exit(1);
 384                                 }
 385                                 break;
 386                         case 'F':
 387                                 if ( pc.paramArg[0] == '0' )
 388                                         codeStyle = GenFlat;
 389                                 else if ( pc.paramArg[0] == '1' )
 390                                         codeStyle = GenFFlat;
 391                                 else {
 392                                         error() << "-F" << pc.paramArg[0] <<
 393                                                         " is an invalid argument" << endl;
 394                                         exit(1);
 395                                 }
 396                                 break;
 397                         case 'G':
 398                                 if ( pc.paramArg[0] == '0' )
 399                                         codeStyle = GenGoto;
 400                                 else if ( pc.paramArg[0] == '1' )
 401                                         codeStyle = GenFGoto;
 402                                 else if ( pc.paramArg[0] == '2' )
 403                                         codeStyle = GenIpGoto;
 404                                 else {
 405                                         error() << "-G" << pc.paramArg[0] <<
 406                                                         " is an invalid argument" << endl;
 407                                         exit(1);
 408                                 }
 409                                 break;
 410                         case 'P':
 411                                 codeStyle = GenSplit;
 412                                 numSplitPartitions = atoi( pc.paramArg );
 413                                 break;
 414
 415                         case 'p':
 416                                 displayPrintables = true;
 417                                 break;
 418
 419                         case 'L':
 420                                 noLineDirectives = true;
 421                                 break;
 422                         }
 423                         break;
 424
 425                 case ParamCheck::invalid:
 426                         error() << "-" << pc.parameter << " is an invalid argument" << endl;
 427                         break;
 428
 429                 case ParamCheck::noparam:
 430                         /* It is interpreted as an input file. */
 431                         if ( *pc.curArg == 0 )
 432                                 error() << "a zero length input file name was given" << endl;
 433                         else if ( inputFileName != 0 )
 434                                 error() << "more than one input file name was given" << endl;
 435                         else {
 436                                 /* OK, Remember the filename. */
 437                                 inputFileName = pc.curArg;
 438                         }
 439                         break;
 440                 }
 441         }
 442 }
 443
 444 void process( const char *inputFileName, const char *intermed )
 445 {
 446         const char *xmlFileName = intermed;
 447         bool wantComplete = true;
 448         bool outputActive = true;
 449
 450         /* Open the input file for reading. */
 451         assert( inputFileName != 0 );
 452         ifstream *inFile = new ifstream( inputFileName );
 453         if ( ! inFile->is_open() )
 454                 error() << "could not open " << inputFileName << " for reading" << endp;
 455
 456         /* Used for just a few things. */
 457         std::ostringstream hostData;
 458
 459         if ( machineSpec == 0 && machineName == 0 )
 460                 hostData << "<host line=\"1\" col=\"1\">";
 461
 462         Scanner scanner( inputFileName, *inFile, hostData, 0, 0, 0, false );
 463         scanner.do_scan();
 464
 465         /* Finished, final check for errors.. */
 466         if ( gblErrorCount > 0 )
 467                 exit(1);
 468
 469         /* Now send EOF to all parsers. */
 470         terminateAllParsers();
 471
 472         /* Finished, final check for errors.. */
 473         if ( gblErrorCount > 0 )
 474                 exit(1);
 475
 476         if ( machineSpec == 0 && machineName == 0 )
 477                 hostData << "</host>\n";
 478
 479         if ( gblErrorCount > 0 )
 480                 exit(1);
 481
 482         /* Open the XML file for writing. */
 483         ostream *xmlOutFile = new ofstream( xmlFileName );
 484
 485         /* Open the XML file for reading. */
 486         ifstream *xmlInFile = new ifstream( xmlFileName );
 487         if ( ! xmlInFile->is_open() )
 488                 error() << "could not open " << xmlFileName << " for reading" << endl;
 489
 490         /* Bail on above error. */
 491         if ( gblErrorCount > 0 )
 492                 exit(1);
 493
 494         /* Locate the backend program */
 495         if ( generateDot ) {
 496                 wantComplete = false;
 497                 outputActive = false;
 498         }
 499
 500         XmlScanner xmlScanner( xmlFileName, *xmlInFile );
 501         XmlParser xmlParser( xmlFileName, outputActive, wantComplete );
 502         xmlParser.init();
 503
 504         /* Write the machines, then the surrounding code. */
 505         writeMachines( *xmlOutFile, hostData.str(), inputFileName, xmlParser );
 506
 507         /* Close the input and the intermediate file. */
 508         delete xmlOutFile;
 509         delete inFile;
 510
 511         /* Bail on above error. */
 512         if ( gblErrorCount > 0 )
 513                 exit(1);
 514
 515         xml_parse( *xmlInFile, xmlFileName,
 516                 outputActive, wantComplete,
 517                 xmlScanner, xmlParser );
 518
 519         /* If writing to a file, delete the ostream, causing it to flush.
 520          * Standard out is flushed automatically. */
 521         if ( outputFileName != 0 ) {
 522                 delete outStream;
 523                 delete outFilter;
 524         }
 525
 526         /* Finished, final check for errors.. */
 527         if ( gblErrorCount > 0 ) {
 528                 /* If we opened an output file, remove it. */
 529                 if ( outputFileName != 0 )
 530                         unlink( outputFileName );
 531                 exit(1);
 532         }
 533 }
 534
 535 char *makeIntermedTemplate( const char *baseFileName )
 536 {
 537         char *result = 0;
 538         const char *templ = "ragel-XXXXXX.xml";
 539         char *lastSlash = strrchr( baseFileName, '/' );
 540         if ( lastSlash == 0 ) {
 541                 result = new char[strlen(templ)+1];
 542                 strcpy( result, templ );
 543         }
 544         else {
 545                 int baseLen = lastSlash - baseFileName + 1;
 546                 result = new char[baseLen + strlen(templ) + 1];
 547                 memcpy( result, baseFileName, baseLen );
 548                 strcpy( result+baseLen, templ );
 549         }
 550         return result;
 551 };
 552
 553 const char *openIntermed( const char *inputFileName, const char *outputFileName )
 554 {
 555         srand(time(0));
 556         const char *result = 0;
 557
 558         /* Which filename do we use as the base? */
 559         const char *baseFileName = outputFileName != 0 ? outputFileName : inputFileName;
 560
 561         /* The template for the intermediate file name. */
 562         const char *intermedFileName = makeIntermedTemplate( baseFileName );
 563
 564         /* Randomize the name and try to open. */
 565         char fnChars[] = "abcdefghijklmnopqrstuvwxyz0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ";
 566         char *firstX = strrchr( intermedFileName, 'X' ) - 5;
 567         for ( int tries = 0; tries < 20; tries++ ) {
 568                 /* Choose a random name. */
 569                 for ( int x = 0; x < 6; x++ )
 570                         firstX[x] = fnChars[rand() % 52];
 571
 572                 /* Try to open the file. */
 573                 int fd = ::open( intermedFileName, O_WRONLY|O_EXCL|O_CREAT, S_IRUSR|S_IWUSR );
 574
 575                 if ( fd > 0 ) {
 576                         /* Success. Close the file immediately and return the name for use
 577                          * by the child processes. */
 578                         ::close( fd );
 579                         result = intermedFileName;
 580                         break;
 581                 }
 582
 583                 if ( errno == EACCES ) {
 584                         error() << "failed to open temp file " << intermedFileName <<
 585                                         ", access denied" << endp;
 586                 }
 587         }
 588
 589         if ( result == 0 )
 590                 error() << "abnormal error: cannot find unique name for temp file" << endp;
 591
 592         return result;
 593 }
 594
 595
 596 void cleanExit( const char *intermed, int status )
 597 {
 598         unlink( intermed );
 599         exit( status );
 600 }
 601
 602 /* Main, process args and call yyparse to start scanning input. */
 603 int main( int argc, const char **argv )
 604 {
 605         const char *inputFileName = 0;
 606         processArgs( argc, argv, inputFileName );
 607
 608         /* If -M or -S are given and we're not generating a dot file then invoke
 609          * the frontend. These options are not useful with code generators. */
 610         if ( machineName != 0 || machineSpec != 0 ) {
 611                 if ( !generateDot )
 612                         frontendOnly = true;
 613         }
 614
 615         /* Require an input file. If we use standard in then we won't have a file
 616          * name on which to base the output. */
 617         if ( inputFileName == 0 )
 618                 error() << "no input file given" << endl;
 619
 620         /* Bail on argument processing errors. */
 621         if ( gblErrorCount > 0 )
 622                 exit(1);
 623
 624         /* Make sure we are not writing to the same file as the input file. */
 625         if ( inputFileName != 0 && outputFileName != 0 &&
 626                         strcmp( inputFileName, outputFileName  ) == 0 )
 627         {
 628                 error() << "output file \"" << outputFileName  <<
 629                                 "\" is the same as the input file" << endp;
 630         }
 631
 632         const char *intermed = openIntermed( inputFileName, outputFileName );
 633         process( inputFileName, intermed );
 634
 635         /* Clean up the intermediate. */
 636         cleanExit( intermed, 0 );
 637
 638         return 0;
 639 }