1 /* 2 * globals.c -- File containing all variables/tables visible to all files. 3 * 4 * SOFTWARE RIGHTS 5 * 6 * We reserve no LEGAL rights to the Purdue Compiler Construction Tool 7 * Set (PCCTS) -- PCCTS is in the public domain. An individual or 8 * company may do whatever they wish with source code distributed with 9 * PCCTS or the code generated by PCCTS, including the incorporation of 10 * PCCTS, or its output, into commerical software. 11 * 12 * We encourage users to develop software with PCCTS. However, we do ask 13 * that credit is given to us for developing PCCTS. By "credit", 14 * we mean that if you incorporate our source code into one of your 15 * programs (commercial product, research project, or otherwise) that you 16 * acknowledge this fact somewhere in the documentation, research report, 17 * etc... If you like PCCTS and have developed a nice tool with the 18 * output, please mention that you developed it using PCCTS. In 19 * addition, we ask that this header remain intact in our source code. 20 * As long as these guidelines are kept, we expect to continue enhancing 21 * this system and expect to make other tools available as they are 22 * completed. 23 * 24 * ANTLR 1.33 25 * Terence Parr 26 * Parr Research Corporation 27 * with Purdue University and AHPCRC, University of Minnesota 28 * 1989-2001 29 */ 30 31 #include <stdio.h> 32 33 #include "pcctscfg.h" 34 35 #include "set.h" 36 #include "syn.h" 37 #include "hash.h" 38 #include "generic.h" 39 40 char Version[] = "1.33MR33" ; /* PCCTS version number */ /* MRXXX */ 41 char VersionDef[] = "13333"; /* same (except int equiv for preproc symbol) */ /* MRXXX */ 42 43 char LexStartSymbol[] = "START";/* Name of starting lexical class/automaton */ 44 char *RemapFileName = "remap.h"; 45 char *DlgFileName = "parser.dlg"; 46 char *DefFileName = "tokens.h"; 47 char *ErrFileName = "err.c"; 48 char *ModeFileName = "mode.h"; 49 char *StdMsgName = NULL; 50 51 char *ParserName = DefaultParserName; 52 53 /* list of PCCTS supplied support symbols; these are renamed when more than 54 * one ANTLR-generated parsers are linked together to avoid name conflicts. 55 * Can't use '##' ANSIC preprocessor concat operator with K&R and: 56 * #define zzskip zzparser ## skip 57 * will not work for ANSI/C++ as 'zzparserskip' is created w/o zzparser 58 * being substituted--ack!!! 59 */ 60 char *StandardSymbols[] = { 61 /* ANTLR stuff */ 62 "zzStackOvfMsg", 63 "zzasp", 64 "zzaStack", 65 "inf_tokens", 66 "inf_text", 67 "inf_text_buffer", 68 "inf_text_buffer_ptr", 69 "inf_text_buffer_size", 70 "inf_labase", 71 "inf_last", 72 "inf_lap", 73 "zztokenLA", 74 "zztextLA", 75 "zzlap", 76 "zzlabase", 77 "zztoktext", 78 "zztoken", 79 "zzdirty", 80 "zzguessing", 81 "zzguess_start", 82 "zzresynch", 83 "zzinf_tokens", 84 "zzinf_text", 85 "zzinf_text_buffer", 86 "zzinf_labase", 87 "zzinf_last", 88 "zzfill_inf_look", 89 "zzFAIL", 90 "zzsave_antlr_state", 91 "zzrestore_antlr_state", 92 "zzsyn", 93 "zzset_el", 94 "zzset_deg", 95 "zzedecode", 96 "_zzsetmatch", 97 "_zzmatch", 98 "_inf_zzgettok", 99 "zzconsumeUntil", 100 "zzconsumeUntilToken", 101 "_zzmatch_wsig", 102 "_zzsetmatch_wsig", 103 "_zzmatch_wdfltsig", 104 "_zzsetmatch_wdfltsig", 105 "zzdflthandlers", 106 /* DLG stuff */ 107 "zzreal_line", 108 "zzcharfull", 109 "zzerr", 110 "zzlextext", 111 "zzbegexpr", 112 "zzendexpr", 113 "zzbufsize", 114 "zzbegcol", 115 "zzendcol", 116 "zzline", 117 "zzchar", 118 "zzbufovf", 119 "zzrdstream", 120 "zzrdfunc", 121 "zzrdstr", 122 "zzclose_stream", 123 "zzsave_dlg_state", 124 "zzrestore_dlg_state", 125 "zzmode", 126 "zzskip", 127 "zzmore", 128 "zzreplchar", 129 "zzreplstr", 130 "zzgettok", 131 "zzadvance", 132 "zzerrstd", 133 "zzerr_in", 134 "zzconstr_attr", 135 "zzempty_attr", 136 "zzerraction", 137 "zztokens", /* list of token regular expressions */ 138 "dfa", 139 "accepts", 140 "actions", 141 "zzTraceOptionValue", /* MR10 */ 142 "zzTraceGuessOptionValue", /* MR10 */ 143 "zzTraceCurrentRuleName", /* MR10 */ 144 "zzTraceDepth", /* MR10 */ 145 "zzGuessSeq", /* MR10 */ 146 "zzSyntaxErrCount", /* MR11 */ 147 "zzLexErrCount", /* MR11 */ 148 "zzTraceGuessDone", /* MR13 - BJS */ 149 "zzTraceGuessFail", /* MR13 - BJS */ 150 "zzTraceGuessOption", /* MR13 - BJS */ 151 "zzTraceIn", /* MR13 - BJS */ 152 "zzTraceOption", /* MR13 - BJS */ 153 "zzTraceOut", /* MR13 - BJS */ 154 "zzTraceReset", /* MR13 - BJS */ 155 NULL /* must be present */ 156 }; 157 158 /* list of PCCTS supplied support functions; these are renamed when more than 159 * one ANTLR-generated parsers are linked together to avoid name conflicts. 160 */ 161 char *ASTSymbols[] = { 162 "AST", 163 "zzast_sp", 164 "zzastStack", 165 "zzlink", 166 "zzastnew", 167 "zzsubchild", 168 "zzsubroot", 169 "zzpre_ast", 170 "zzfree_ast", 171 "zztmake", 172 "zzdup_ast", 173 "zztfree", 174 "zzdouble_link", 175 NULL /* must be present */ 176 }; 177 178 /* Current ambiguity examination information */ 179 int CurAmbigAlt1, CurAmbigAlt2, CurAmbigline, CurAmbigfile; 180 char *CurAmbigbtype; 181 182 183 /* M e t h o d T a b l e s */ 184 /* 185 * The following tables are used to fill syntax diagram nodes with the correct 186 * function pointers for computing FIRST sets and printing themselves. 187 */ 188 189 /* fpTraverse[node type] == pointer to function that calculates trees 190 * representing the FIRST sets for that node (maintains spatial info). 191 * We use 'struct _tree' not 'tree' due to a g++ 2.4.3 bug. 192 */ 193 #ifdef __cplusplus 194 struct _tree *(*fpTraverse[NumNodeTypes+1])(... /* Node *, int, set * */) = { 195 NULL, 196 (struct _tree *(*)(...)) tJunc, 197 (struct _tree *(*)(...)) tRuleRef, 198 (struct _tree *(*)(...)) tToken, 199 (struct _tree *(*)(...)) tAction 200 }; 201 #else 202 Tree *(*fpTraverse[NumNodeTypes+1])() = { 203 NULL, 204 tJunc, 205 tRuleRef, 206 tToken, 207 tAction 208 }; 209 #endif 210 211 /* fpReach[node type] == pointer to function that calculates FIRST set for 212 * that node. (r stands for reach). We use 'struct _set' not 'set' 213 * due to a g++ 2.4.3 bug. 214 */ 215 #ifdef __cplusplus 216 struct _set (*fpReach[NumNodeTypes+1])(... /* Node *, int, set * */) = { 217 NULL, 218 (struct _set (*)(...)) rJunc, 219 (struct _set (*)(...)) rRuleRef, 220 (struct _set (*)(...)) rToken, 221 (struct _set (*)(...)) rAction 222 }; 223 #else 224 set (*fpReach[NumNodeTypes+1])() = { 225 NULL, 226 rJunc, 227 rRuleRef, 228 rToken, 229 rAction 230 }; 231 #endif 232 233 /* fpPrint[node type] == pointer to function that knows how to print that node. */ 234 #ifdef __cplusplus 235 void (*fpPrint[NumNodeTypes+1])(... /* Node * */) = { 236 NULL, 237 (void (*)(...)) pJunc, 238 (void (*)(...)) pRuleRef, 239 (void (*)(...)) pToken, 240 (void (*)(...)) pAction 241 }; 242 #else 243 void (*fpPrint[NumNodeTypes+1])() = { 244 NULL, 245 pJunc, 246 pRuleRef, 247 pToken, 248 pAction 249 }; 250 #endif 251 252 char *decodeJType[] = { 253 "invalid", 254 "aSubBlk", 255 "aOptBlk", 256 "aLoopBlk", 257 "EndBlk", 258 "RuleBlk", 259 "Generic", 260 "EndRule", 261 "aPlusBlk", 262 "aLoopBegin" 263 }; 264 265 266 /* H a s h T a b l e s */ 267 268 Entry **Tname, /* Table of all token names (maps name to tok num)*/ 269 **Texpr, /* Table of all token expressions 270 (maps expr to tok num) */ 271 **Rname, /* Table of all Rules (has ptr to start of rule) */ 272 **Fcache, /* Cache of First/Follow Computations */ 273 **Tcache; /* Tree cache; First/Follow for permute trees */ 274 Entry **Elabel; /* Table of all element label names */ 275 Entry **Sname; /* Signal names */ 276 Entry **Pname; /* symbolic predicate names MR11 */ 277 278 279 /* V a r i a b l e s */ 280 281 int Save_argc; /* MR10 */ 282 char **Save_argv; /* MR10 */ 283 int EpToken=0; /* Imaginary Epsilon token number */ 284 int WildCardToken=0; 285 int CurFile= -1; /* Index into FileStr table */ 286 char *CurPredName=NULL; /* MR11 */ 287 char *CurRule=NULL; /* Pointer to current rule name */ 288 int CurRuleDebug=0; /* MR13 debug flag */ 289 RuleEntry *CurRuleNode=NULL;/* Pointer to current rule node in syntax tree */ 290 char *CurRetDef=NULL; /* Pointer to current return type definition */ 291 char *CurParmDef=NULL; /* Pointer to current parameter definition */ 292 Junction *CurRuleBlk=NULL; /* Pointer to current block node for enclosing block */ 293 ListNode *CurExGroups=NULL; /* Current list of exception groups for rule/alts */ 294 ListNode *CurElementLabels=NULL; 295 ListNode *CurAstLabelsInActions=NULL; /* MR27 */ 296 297 /* MR10 used by <<>>? to set "label_used_in_semantic_pred" */ 298 /* MR10 this will force LT(i) assignment even in guess mode */ 299 300 ListNode *CurActionLabels=NULL; /* MR10 Element Labels appearing in last action */ 301 int numericActionLabel=0 ; /* MR10 << ... $1 ... >> or << ... $1 ... >>? */ 302 ListNode *NumericPredLabels=NULL; /* MR10 << ... $1 ... >>? ONLY */ 303 ListNode *ContextGuardPredicateList=NULL; /* MR13 for re-evaluating predicates 304 after meta tokens are defined */ 305 306 int CurBlockID=0; /* Unique int for each block */ 307 int CurAltNum=0; 308 Junction *CurAltStart = NULL; /* Junction node that starts the alt */ 309 Junction *OuterAltStart = NULL; /* For chaining exception groups MR7 */ 310 int NumRules=0; /* Rules are from 1 to n */ 311 FILE *output=NULL; /* current parser output file */ 312 FILE *input=NULL; /* current grammar input file */ 313 char *FileStr[MaxNumFiles];/* Ptr to array of file names on command-line */ 314 int NumFiles=0; /* current grammar file number */ 315 #ifdef __cplusplus 316 void (**fpTrans)(...), /* array of ptrs to funcs that translate nodes */ 317 (**fpJTrans)(...); /* ... that translate junctions */ 318 #else 319 void (**fpTrans)(), /* array of ptrs to funcs that translate nodes */ 320 (**fpJTrans)(); /* ... that translate junctions */ 321 #endif 322 int **FoStack; /* Array of LL_k ptrs to stacks of rule numbers */ 323 int **FoTOS; /* FOLLOW stack top-of-stack pointers */ 324 Junction *SynDiag = NULL; /* Pointer to start of syntax diagram */ 325 int BlkLevel=1; /* Current block level. Set by antlr.g, used by 326 * scanner to translate $i.j attributes */ 327 set reserved_positions; /* set of token positions reserved by '#token T=i' cmds */ 328 set all_tokens; /* set of all token types */ 329 set imag_tokens; /* set of all imaginary token types (EpToken, errclasses...) */ 330 set tokclasses; /* set of all token class token types */ 331 ListNode *ForcedTokens = 0; /* list of token_id/token_num pairs to remap */ 332 ListNode *MetaTokenNodes=NULL; /* list of meta token refs such as token classes etc... */ 333 int *TokenInd=NULL; /* an indirection level between token num and position 334 * of that token def in TokenStr and ExprStr */ 335 int LastTokenCounted=0; /* ==TokenNum if no token renumbering (same as old TokenNum) */ 336 int TokenNum=TokenStart; 337 char **TokenStr=NULL; /* map token # to token name */ 338 char **ExprStr=NULL; /* map token # to expr */ 339 Junction **RulePtr=NULL; /* map rule # to RuleBlk node of rule */ 340 ListNode *ExprOrder=NULL; /* list of exprs as they are found in grammar */ 341 ListNode *BeforeActions=NULL;/* list of grammar actions before rules */ 342 ListNode *AfterActions=NULL;/* list of grammar actions after rules */ 343 ListNode *LexActions=NULL; /* list of lexical actions */ 344 345 /* MR1 */ 346 /* MR1 11-Apr-97 Provide mechanism for inserting code into DLG class */ 347 /* MR1 via #lexmember <<....>> */ 348 /* MR1 via #lexprefix <<....>> */ 349 /* MR1 */ 350 351 ListNode *LexMemberActions=NULL;/* list of lexical header member decl MR1 */ 352 ListNode *LexPrefixActions=NULL;/* list of lexical header #include decl MR1 */ 353 ListNode **Cycles=NULL; /* list of cycles (for each k) found when 354 doing FOLLOWs */ 355 ListNode *eclasses=NULL; /* list of error classes */ 356 ListNode *tclasses=NULL; /* list of token classes */ 357 LClass lclass[MaxLexClasses]; /* array of lex class definitions */ 358 int CurrentLexClass; /* index into lclass */ 359 int NumLexClasses=0; /* in range 1..MaxLexClasses (init 0) */ 360 361 char *HdrAction=NULL; /* action defined with #header */ 362 char *FirstAction=NULL; /* action defined with #first MR11 */ 363 FILE *ErrFile; /* sets and error recovery stuff */ 364 FILE *DefFile=NULL; /* list of tokens, return value structs, setwd defs */ 365 FILE *MRinfoFile=NULL; /* MR10 information file */ 366 int MRinfo=0; /* MR10 */ 367 int MRinfoSeq=0; /* MR10 */ 368 int InfoP=0; /* MR10 predicates */ 369 int InfoT=0; /* MR10 tnodes */ 370 int InfoF=0; /* MR10 first/follow sets */ 371 int InfoM=0; /* MR10 monitor progress */ 372 int InfoO=0; /* MR12 orphan rules */ 373 int TnodesInUse=0; /* MR10 */ 374 int TnodesPeak=0; /* MR10 */ 375 int TnodesAllocated=0; /* MR10 */ 376 int TnodesReportThreshold=0; /* MR11 */ 377 int PotentialSuppression=0; /* MR10 */ 378 int PotentialDummy=0; /* MR10 */ 379 int CannotContinue=FALSE; 380 int OutputLL_k = 1; /* LL_k for parsing must be power of 2 */ 381 int action_file; /* used to track start of action */ 382 int action_line; 383 int FoundGuessBlk=0; /* there is a (...)? block somewhere in grammar */ 384 int FoundException=0; /* there is an exception somewhere in grammar */ 385 /* MR6 Distinguish between @ operator and real exception */ 386 /* MR6 by keeping separate flags for @ operator and real exceptions */ 387 int FoundAtOperator=0; /* MR6 */ 388 int FoundExceptionGroup=0; /* MR6 */ 389 int pLevel=0; /* print Level */ 390 int pAlt1,pAlt2; /* print "==>" in front of these alts */ 391 392 /* C++ output stuff */ 393 FILE *Parser_h, /* where subclass of ANTLRParser goes */ 394 *Parser_c; /* where code for subclass of ANTLRParser goes */ 395 char Parser_h_Name[MaxFileName+1] = ""; 396 char Parser_c_Name[MaxFileName+1] = ""; 397 char MRinfoFile_Name[MaxFileName+1] = ""; /* MR10 */ 398 char *ClassDeclStuff=NULL; /* MR10 */ 399 char *BaseClassName=NULL; /* MR22 */ 400 /* list of actions inside the #class {...} defs */ 401 ListNode *class_before_actions=NULL; 402 ListNode *class_after_actions=NULL; 403 404 char CurrentClassName[MaxRuleName]=""; 405 int no_classes_found=1; 406 char *UserTokenDefsFile; 407 int UserDefdTokens=0; /* found #tokdefs? */ 408 char *OutputDirectory=TopDirectory; 409 ExceptionGroup *DefaultExGroup = NULL; 410 int NumSignals = NumPredefinedSignals; 411 int ContextGuardTRAV=0; 412 413 char *MR_AmbAidRule=NULL; /* MR11 */ 414 int MR_AmbAidLine=0; /* MR11 */ 415 int MR_AmbAidDepth=0; /* MR11 */ 416 int MR_AmbAidMultiple=0; /* MR11 */ 417 int MR_skipped_e3_report=0; /* MR11 */ 418 int MR_usingPredNames=0; /* MR11 */ 419 int MR_BadExprSets=0; /* MR13 */ 420 int MR_Inhibit_Tokens_h_Gen=0; /* MR13 */ 421 int NewAST=0; /* MR13 */ 422 int tmakeInParser=0; /* MR23 */ 423 int AlphaBetaTrace=0; /* MR14 */ 424 int MR_BlkErr=0; /* MR21 */ 425 int MR_AlphaBetaMessageCount=0; /* MR14 */ 426 int MR_AlphaBetaWarning=0; /* MR14 */ 427 int MR_ErrorSetComputationActive=0; /* MR14 */ 428 int MR_MaintainBackTrace=0; /* MR14 */ 429 set MR_CompromisedRules; /* MR14 */ 430 431 Junction *MR_RuleBlkWithHalt; /* MR10 */ 432 433 /* C m d - L i n e O p t i o n s */ 434 435 int LL_k=1; /* how many tokens of full lookahead */ 436 int CLL_k= -1; /* how many tokens of compressed lookahead */ 437 int PrintOut = FALSE; /* print out the grammar */ 438 int PrintAnnotate = FALSE;/* annotate printout with FIRST sets */ 439 int CodeGen=TRUE; /* Generate output code? */ 440 int LexGen=TRUE; /* Generate lexical files? (tokens.h, parser.dlg) */ 441 int GenAST=FALSE; /* Generate AST's? */ 442 int GenANSI=FALSE; /* Generate ANSI code where necessary */ 443 int GenExprSetsOpt=TRUE;/* use sets not (LA(1)==tok) expression lists */ 444 int GenCR=FALSE; /* Generate cross reference? */ 445 int GenLineInfo=FALSE; /* Generate # line "file" stuff? */ 446 int GenLineInfoMS=FALSE;/* Like -gl but replace "\" with "/" for MS C/C++ systems */ 447 int TraceGen=FALSE; /* Generate code to trace rule invocation */ 448 int elevel=1; /* error level for ambiguity messages */ 449 int GenEClasseForRules=0;/* don't generate eclass for each rule */ 450 int TreeResourceLimit= -1;/* don't limit tree resource */ 451 int DemandLookahead = 0;/* demand/delayed lookahead or not */ 452 char *RulePrefix = ""; /* prefix each generated rule with this */ 453 char *stdpccts = "stdpccts.h";/* where to generate std pccts include file */ 454 int GenStdPccts = 0; /* don't gen stdpccts.h? */ 455 int ParseWithPredicates = 1; 456 int WarningLevel = 1; 457 int UseStdout = 0; /* MR6 */ 458 int TabWidth = 2; /* MR6 */ /* MR27 */ 459 int HoistPredicateContext = 0; 460 int MRhoisting = 0; /* MR9 */ 461 int MRhoistingk = 0; /* MR13 */ 462 int MR_debugGenRule=0; /* MR11 */ 463 464 int GenCC = 0; /* Generate C++ output */ 465 466 PointerStack MR_BackTraceStack={0,0,NULL}; /* MR10 */ 467 PointerStack MR_PredRuleRefStack={0,0,NULL}; /* MR10 */ 468 PointerStack MR_RuleBlkWithHaltStack={0,0,NULL}; /* MR10 */ 469 470 /* DontCopyTokens and Pragma_DupLabeledTokens were a bad idea. I've just 471 turned them off rather than backpatching the code. Who knows? We 472 may need them in the future. 473 */ 474 int DontCopyTokens = 1; /* in C++, don't copy ANTLRToken passed to ANTLR */ 475 476 /* Remember if LT(i), LA(i), or LATEXT(i) used in an action which is not 477 a predicate. If so, give a warning for novice users. 478 */ 479 480 int LTinTokenAction = 0; /* MR23 */ 481 int PURIFY = 1; /* MR23 */ 482 483 int CurBlockID_array[MAX_BLK_LEVEL]; /* MR23 */ 484 int CurAltNum_array[MAX_BLK_LEVEL]; /* MR23 */ 485
