1: /*
2: ** This file contains all sources (including headers) to the LEMON
3: ** LALR(1) parser generator. The sources have been combined into a
4: ** single file to make it easy to include LEMON in the source tree
5: ** and Makefile of another program.
6: **
7: ** The author of this program disclaims copyright.
8: */
9: #include <stdio.h>
10: #include <stdarg.h>
11: #include <string.h>
12: #include <ctype.h>
13: #include <stdlib.h>
14:
15: #ifdef HAVE_CONFIG_H
16: #include "config.h"
17: #endif
18:
19: #ifdef HAVE_STDINT_H
20: # include <stdint.h>
21: #endif
22: #ifdef HAVE_INTTYPES_H
23: # include <inttypes.h>
24: #endif
25:
26: #define UNUSED(x) ( (void)(x) )
27:
28: extern void qsort();
29: extern double strtod();
30: extern long strtol();
31: extern void free();
32: extern int access();
33: extern int atoi();
34: extern char *getenv();
35:
36: #ifndef __WIN32__
37: # if defined(_WIN32) || defined(WIN32)
38: # define __WIN32__
39: # endif
40: #endif
41:
42: #if __GNUC__ > 2
43: #define NORETURN __attribute__ ((__noreturn__))
44: #else
45: #define NORETURN
46: #endif
47:
48: /* #define PRIVATE static */
49: #define PRIVATE static
50:
51: #ifdef TEST
52: #define MAXRHS 5 /* Set low to exercise exception code */
53: #else
54: #define MAXRHS 1000
55: #endif
56:
57: char *msort();
58: extern void *malloc();
59:
60: extern void memory_error() NORETURN;
61:
62: /******** From the file "action.h" *************************************/
63: struct action *Action_new();
64: struct action *Action_sort();
65: void Action_add();
66:
67: /********* From the file "assert.h" ************************************/
68: void myassert() NORETURN;
69: #ifndef NDEBUG
70: # define assert(X) if(!(X))myassert(__FILE__,__LINE__)
71: #else
72: # define assert(X)
73: #endif
74:
75: /********** From the file "build.h" ************************************/
76: void FindRulePrecedences();
77: void FindFirstSets();
78: void FindStates();
79: void FindLinks();
80: void FindFollowSets();
81: void FindActions();
82:
83: /********* From the file "configlist.h" *********************************/
84: void Configlist_init(/* void */);
85: struct config *Configlist_add(/* struct rule *, int */);
86: struct config *Configlist_addbasis(/* struct rule *, int */);
87: void Configlist_closure(/* void */);
88: void Configlist_sort(/* void */);
89: void Configlist_sortbasis(/* void */);
90: struct config *Configlist_return(/* void */);
91: struct config *Configlist_basis(/* void */);
92: void Configlist_eat(/* struct config * */);
93: void Configlist_reset(/* void */);
94:
95: /********* From the file "error.h" ***************************************/
96: void ErrorMsg(const char *, int,const char *, ...);
97:
98: /****** From the file "option.h" ******************************************/
99: struct s_options {
100: enum { OPT_FLAG=1, OPT_INT, OPT_DBL, OPT_STR,
101: OPT_FFLAG, OPT_FINT, OPT_FDBL, OPT_FSTR} type;
102: char *label;
103: char *arg;
104: char *message;
105: };
106: int OptInit(/* char**,struct s_options*,FILE* */);
107: int OptNArgs(/* void */);
108: char *OptArg(/* int */);
109: void OptErr(/* int */);
110: void OptPrint(/* void */);
111:
112: /******** From the file "parse.h" *****************************************/
113: void Parse(/* struct lemon *lemp */);
114:
115: /********* From the file "plink.h" ***************************************/
116: struct plink *Plink_new(/* void */);
117: void Plink_add(/* struct plink **, struct config * */);
118: void Plink_copy(/* struct plink **, struct plink * */);
119: void Plink_delete(/* struct plink * */);
120:
121: /********** From the file "report.h" *************************************/
122: void Reprint(/* struct lemon * */);
123: void ReportOutput(/* struct lemon * */);
124: void ReportTable(/* struct lemon * */);
125: void ReportHeader(/* struct lemon * */);
126: void CompressTables(/* struct lemon * */);
127:
128: /********** From the file "set.h" ****************************************/
129: void SetSize(/* int N */); /* All sets will be of size N */
130: char *SetNew(/* void */); /* A new set for element 0..N */
131: void SetFree(/* char* */); /* Deallocate a set */
132:
133: int SetAdd(/* char*,int */); /* Add element to a set */
134: int SetUnion(/* char *A,char *B */); /* A <- A U B, thru element N */
135:
136: #define SetFind(X,Y) (X[Y]) /* True if Y is in set X */
137:
138: /********** From the file "struct.h" *************************************/
139: /*
140: ** Principal data structures for the LEMON parser generator.
141: */
142:
143: typedef enum {Bo_FALSE=0, Bo_TRUE} Boolean;
144:
145: /* Symbols (terminals and nonterminals) of the grammar are stored
146: ** in the following: */
147: struct symbol {
148: char *name; /* Name of the symbol */
149: int index; /* Index number for this symbol */
150: enum {
151: TERMINAL,
152: NONTERMINAL
153: } type; /* Symbols are all either TERMINALS or NTs */
154: struct rule *rule; /* Linked list of rules of this (if an NT) */
155: struct symbol *fallback; /* fallback token in case this token doesn't parse */
156: int prec; /* Precedence if defined (-1 otherwise) */
157: enum e_assoc {
158: LEFT,
159: RIGHT,
160: NONE,
161: UNK
162: } assoc; /* Associativity if predecence is defined */
163: char *firstset; /* First-set for all rules of this symbol */
164: Boolean lambda; /* True if NT and can generate an empty string */
165: char *destructor; /* Code which executes whenever this symbol is
166: ** popped from the stack during error processing */
167: int destructorln; /* Line number of destructor code */
168: char *datatype; /* The data type of information held by this
169: ** object. Only used if type==NONTERMINAL */
170: int dtnum; /* The data type number. In the parser, the value
171: ** stack is a union. The .yy%d element of this
172: ** union is the correct data type for this object */
173: };
174:
175: /* Each production rule in the grammar is stored in the following
176: ** structure. */
177: struct rule {
178: struct symbol *lhs; /* Left-hand side of the rule */
179: char *lhsalias; /* Alias for the LHS (NULL if none) */
180: int ruleline; /* Line number for the rule */
181: int nrhs; /* Number of RHS symbols */
182: struct symbol **rhs; /* The RHS symbols */
183: char **rhsalias; /* An alias for each RHS symbol (NULL if none) */
184: int line; /* Line number at which code begins */
185: char *code; /* The code executed when this rule is reduced */
186: struct symbol *precsym; /* Precedence symbol for this rule */
187: int index; /* An index number for this rule */
188: Boolean canReduce; /* True if this rule is ever reduced */
189: struct rule *nextlhs; /* Next rule with the same LHS */
190: struct rule *next; /* Next rule in the global list */
191: };
192:
193: /* A configuration is a production rule of the grammar together with
194: ** a mark (dot) showing how much of that rule has been processed so far.
195: ** Configurations also contain a follow-set which is a list of terminal
196: ** symbols which are allowed to immediately follow the end of the rule.
197: ** Every configuration is recorded as an instance of the following: */
198: struct config {
199: struct rule *rp; /* The rule upon which the configuration is based */
200: int dot; /* The parse point */
201: char *fws; /* Follow-set for this configuration only */
202: struct plink *fplp; /* Follow-set forward propagation links */
203: struct plink *bplp; /* Follow-set backwards propagation links */
204: struct state *stp; /* Pointer to state which contains this */
205: enum {
206: COMPLETE, /* The status is used during followset and */
207: INCOMPLETE /* shift computations */
208: } status;
209: struct config *next; /* Next configuration in the state */
210: struct config *bp; /* The next basis configuration */
211: };
212:
213: /* Every shift or reduce operation is stored as one of the following */
214: struct action {
215: struct symbol *sp; /* The look-ahead symbol */
216: enum e_action {
217: SHIFT,
218: ACCEPT,
219: REDUCE,
220: ERROR,
221: CONFLICT, /* Was a reduce, but part of a conflict */
222: SH_RESOLVED, /* Was a shift. Precedence resolved conflict */
223: RD_RESOLVED, /* Was reduce. Precedence resolved conflict */
224: NOT_USED /* Deleted by compression */
225: } type;
226: union {
227: struct state *stp; /* The new state, if a shift */
228: struct rule *rp; /* The rule, if a reduce */
229: } x;
230: struct action *next; /* Next action for this state */
231: struct action *collide; /* Next action with the same hash */
232: };
233:
234: /* Each state of the generated parser's finite state machine
235: ** is encoded as an instance of the following structure. */
236: struct state {
237: struct config *bp; /* The basis configurations for this state */
238: struct config *cfp; /* All configurations in this set */
239: int index; /* Sequencial number for this state */
240: struct action *ap; /* Array of actions for this state */
241: int nTknAct, nNtAct; /* Number of actions on terminals and nonterminals */
242: int iTknOfst, iNtOfst; /* yy_action[] offset for terminals and nonterms */
243: int iDflt; /* Default action */
244: };
245: #define NO_OFFSET (-2147483647)
246:
247: /* A followset propagation link indicates that the contents of one
248: ** configuration followset should be propagated to another whenever
249: ** the first changes. */
250: struct plink {
251: struct config *cfp; /* The configuration to which linked */
252: struct plink *next; /* The next propagate link */
253: };
254:
255: /* The state vector for the entire parser generator is recorded as
256: ** follows. (LEMON uses no global variables and makes little use of
257: ** static variables. Fields in the following structure can be thought
258: ** of as begin global variables in the program.) */
259: struct lemon {
260: struct state **sorted; /* Table of states sorted by state number */
261: struct rule *rule; /* List of all rules */
262: int nstate; /* Number of states */
263: int nrule; /* Number of rules */
264: int nsymbol; /* Number of terminal and nonterminal symbols */
265: int nterminal; /* Number of terminal symbols */
266: struct symbol **symbols; /* Sorted array of pointers to symbols */
267: int errorcnt; /* Number of errors */
268: struct symbol *errsym; /* The error symbol */
269: char *name; /* Name of the generated parser */
270: char *arg; /* Declaration of the 3th argument to parser */
271: char *tokentype; /* Type of terminal symbols in the parser stack */
272: char *vartype; /* The default type of non-terminal symbols */
273: char *start; /* Name of the start symbol for the grammar */
274: char *stacksize; /* Size of the parser stack */
275: char *include; /* Code to put at the start of the C file */
276: int includeln; /* Line number for start of include code */
277: char *error; /* Code to execute when an error is seen */
278: int errorln; /* Line number for start of error code */
279: char *overflow; /* Code to execute on a stack overflow */
280: int overflowln; /* Line number for start of overflow code */
281: char *failure; /* Code to execute on parser failure */
282: int failureln; /* Line number for start of failure code */
283: char *accept; /* Code to execute when the parser excepts */
284: int acceptln; /* Line number for the start of accept code */
285: char *extracode; /* Code appended to the generated file */
286: int extracodeln; /* Line number for the start of the extra code */
287: char *tokendest; /* Code to execute to destroy token data */
288: int tokendestln; /* Line number for token destroyer code */
289: char *vardest; /* Code for the default non-terminal destructor */
290: int vardestln; /* Line number for default non-term destructor code*/
291: char *filename; /* Name of the input file */
292: char *tmplname; /* Name of the template file */
293: char *outname; /* Name of the current output file */
294: char *tokenprefix; /* A prefix added to token names in the .h file */
295: int nconflict; /* Number of parsing conflicts */
296: int tablesize; /* Size of the parse tables */
297: int basisflag; /* Print only basis configurations */
298: int has_fallback; /* True if any %fallback is seen in the grammer */
299: char *argv0; /* Name of the program */
300: };
301:
302: #define MemoryCheck(X) if((X)==0){ \
303: memory_error(); \
304: }
305:
306: /**************** From the file "table.h" *********************************/
307: /*
308: ** All code in this file has been automatically generated
309: ** from a specification in the file
310: ** "table.q"
311: ** by the associative array code building program "aagen".
312: ** Do not edit this file! Instead, edit the specification
313: ** file, then rerun aagen.
314: */
315: /*
316: ** Code for processing tables in the LEMON parser generator.
317: */
318:
319: /* Routines for handling a strings */
320:
321: char *Strsafe();
322:
323: void Strsafe_init(/* void */);
324: int Strsafe_insert(/* char * */);
325: char *Strsafe_find(/* char * */);
326:
327: /* Routines for handling symbols of the grammar */
328:
329: struct symbol *Symbol_new();
330: int Symbolcmpp(/* struct symbol **, struct symbol ** */);
331: void Symbol_init(/* void */);
332: int Symbol_insert(/* struct symbol *, char * */);
333: struct symbol *Symbol_find(/* char * */);
334: struct symbol *Symbol_Nth(/* int */);
335: int Symbol_count(/* */);
336: struct symbol **Symbol_arrayof(/* */);
337:
338: /* Routines to manage the state table */
339:
340: int Configcmp(/* struct config *, struct config * */);
341: struct state *State_new();
342: void State_init(/* void */);
343: int State_insert(/* struct state *, struct config * */);
344: struct state *State_find(/* struct config * */);
345: struct state **State_arrayof(/* */);
346:
347: /* Routines used for efficiency in Configlist_add */
348:
349: void Configtable_init(/* void */);
350: int Configtable_insert(/* struct config * */);
351: struct config *Configtable_find(/* struct config * */);
352: void Configtable_clear(/* int(*)(struct config *) */);
353: /****************** From the file "action.c" *******************************/
354: /*
355: ** Routines processing parser actions in the LEMON parser generator.
356: */
357:
358: /* Allocate a new parser action */
359: struct action *Action_new(){
360: static struct action *freelist = NULL;
361: struct action *new;
362:
363: if( freelist==NULL ){
364: int i;
365: int amt = 100;
366: freelist = (struct action *)malloc( sizeof(struct action)*amt );
367: if( freelist==0 ){
368: fprintf(stderr,"Unable to allocate memory for a new parser action.");
369: exit(1);
370: }
371: for(i=0; i<amt-1; i++) freelist[i].next = &freelist[i+1];
372: freelist[amt-1].next = 0;
373: }
374: new = freelist;
375: freelist = freelist->next;
376: return new;
377: }
378:
379: /* Compare two actions */
380: static int actioncmp(ap1,ap2)
381: struct action *ap1;
382: struct action *ap2;
383: {
384: int rc;
385: rc = ap1->sp->index - ap2->sp->index;
386: if( rc==0 ) rc = (int)ap1->type - (int)ap2->type;
387: if( rc==0 ){
388: assert( ap1->type==REDUCE || ap1->type==RD_RESOLVED || ap1->type==CONFLICT);
389: assert( ap2->type==REDUCE || ap2->type==RD_RESOLVED || ap2->type==CONFLICT);
390: rc = ap1->x.rp->index - ap2->x.rp->index;
391: }
392: return rc;
393: }
394:
395: /* Sort parser actions */
396: struct action *Action_sort(ap)
397: struct action *ap;
398: {
399: ap = (struct action *)msort(ap,&ap->next,actioncmp);
400: return ap;
401: }
402:
403: void Action_add(app,type,sp,arg)
404: struct action **app;
405: enum e_action type;
406: struct symbol *sp;
407: void *arg;
408: {
409: struct action *new;
410: new = Action_new();
411: new->next = *app;
412: *app = new;
413: new->type = type;
414: new->sp = sp;
415: if( type==SHIFT ){
416: new->x.stp = (struct state *)arg;
417: }else{
418: new->x.rp = (struct rule *)arg;
419: }
420: }
421: /********************** New code to implement the "acttab" module ***********/
422: /*
423: ** This module implements routines use to construct the yy_action[] table.
424: */
425:
426: /*
427: ** The state of the yy_action table under construction is an instance of
428: ** the following structure
429: */
430: typedef struct acttab acttab;
431: struct acttab {
432: int nAction; /* Number of used slots in aAction[] */
433: int nActionAlloc; /* Slots allocated for aAction[] */
434: struct {
435: int lookahead; /* Value of the lookahead token */
436: int action; /* Action to take on the given lookahead */
437: } *aAction, /* The yy_action[] table under construction */
438: *aLookahead; /* A single new transaction set */
439: int mnLookahead; /* Minimum aLookahead[].lookahead */
440: int mnAction; /* Action associated with mnLookahead */
441: int mxLookahead; /* Maximum aLookahead[].lookahead */
442: int nLookahead; /* Used slots in aLookahead[] */
443: int nLookaheadAlloc; /* Slots allocated in aLookahead[] */
444: };
445:
446: /* Return the number of entries in the yy_action table */
447: #define acttab_size(X) ((X)->nAction)
448:
449: /* The value for the N-th entry in yy_action */
450: #define acttab_yyaction(X,N) ((X)->aAction[N].action)
451:
452: /* The value for the N-th entry in yy_lookahead */
453: #define acttab_yylookahead(X,N) ((X)->aAction[N].lookahead)
454:
455: /* Free all memory associated with the given acttab */
456: /*
457: PRIVATE void acttab_free(acttab *p){
458: free( p->aAction );
459: free( p->aLookahead );
460: free( p );
461: }
462: */
463:
464: /* Allocate a new acttab structure */
465: PRIVATE acttab *acttab_alloc(void){
466: acttab *p = malloc( sizeof(*p) );
467: if( p==0 ){
468: fprintf(stderr,"Unable to allocate memory for a new acttab.");
469: exit(1);
470: }
471: memset(p, 0, sizeof(*p));
472: return p;
473: }
474:
475: /* Add a new action to the current transaction set
476: */
477: PRIVATE void acttab_action(acttab *p, int lookahead, int action){
478: if( p->nLookahead>=p->nLookaheadAlloc ){
479: p->nLookaheadAlloc += 25;
480: p->aLookahead = realloc( p->aLookahead,
481: sizeof(p->aLookahead[0])*p->nLookaheadAlloc );
482: if( p->aLookahead==0 ){
483: fprintf(stderr,"malloc failed\n");
484: exit(1);
485: }
486: }
487: if( p->nLookahead==0 ){
488: p->mxLookahead = lookahead;
489: p->mnLookahead = lookahead;
490: p->mnAction = action;
491: }else{
492: if( p->mxLookahead<lookahead ) p->mxLookahead = lookahead;
493: if( p->mnLookahead>lookahead ){
494: p->mnLookahead = lookahead;
495: p->mnAction = action;
496: }
497: }
498: p->aLookahead[p->nLookahead].lookahead = lookahead;
499: p->aLookahead[p->nLookahead].action = action;
500: p->nLookahead++;
501: }
502:
503: /*
504: ** Add the transaction set built up with prior calls to acttab_action()
505: ** into the current action table. Then reset the transaction set back
506: ** to an empty set in preparation for a new round of acttab_action() calls.
507: **
508: ** Return the offset into the action table of the new transaction.
509: */
510: PRIVATE int acttab_insert(acttab *p){
511: int i, j, k, n;
512: assert( p->nLookahead>0 );
513:
514: /* Make sure we have enough space to hold the expanded action table
515: ** in the worst case. The worst case occurs if the transaction set
516: ** must be appended to the current action table
517: */
518: n = p->mxLookahead + 1;
519: if( p->nAction + n >= p->nActionAlloc ){
520: int oldAlloc = p->nActionAlloc;
521: p->nActionAlloc = p->nAction + n + p->nActionAlloc + 20;
522: p->aAction = realloc( p->aAction,
523: sizeof(p->aAction[0])*p->nActionAlloc);
524: if( p->aAction==0 ){
525: fprintf(stderr,"malloc failed\n");
526: exit(1);
527: }
528: for(i=oldAlloc; i<p->nActionAlloc; i++){
529: p->aAction[i].lookahead = -1;
530: p->aAction[i].action = -1;
531: }
532: }
533:
534: /* Scan the existing action table looking for an offset where we can
535: ** insert the current transaction set. Fall out of the loop when that
536: ** offset is found. In the worst case, we fall out of the loop when
537: ** i reaches p->nAction, which means we append the new transaction set.
538: **
539: ** i is the index in p->aAction[] where p->mnLookahead is inserted.
540: */
541: for(i=0; i<p->nAction+p->mnLookahead; i++){
542: if( p->aAction[i].lookahead<0 ){
543: for(j=0; j<p->nLookahead; j++){
544: k = p->aLookahead[j].lookahead - p->mnLookahead + i;
545: if( k<0 ) break;
546: if( p->aAction[k].lookahead>=0 ) break;
547: }
548: if( j<p->nLookahead ) continue;
549: for(j=0; j<p->nAction; j++){
550: if( p->aAction[j].lookahead==j+p->mnLookahead-i ) break;
551: }
552: if( j==p->nAction ){
553: break; /* Fits in empty slots */
554: }
555: }else if( p->aAction[i].lookahead==p->mnLookahead ){
556: if( p->aAction[i].action!=p->mnAction ) continue;
557: for(j=0; j<p->nLookahead; j++){
558: k = p->aLookahead[j].lookahead - p->mnLookahead + i;
559: if( k<0 || k>=p->nAction ) break;
560: if( p->aLookahead[j].lookahead!=p->aAction[k].lookahead ) break;
561: if( p->aLookahead[j].action!=p->aAction[k].action ) break;
562: }
563: if( j<p->nLookahead ) continue;
564: n = 0;
565: for(j=0; j<p->nAction; j++){
566: if( p->aAction[j].lookahead<0 ) continue;
567: if( p->aAction[j].lookahead==j+p->mnLookahead-i ) n++;
568: }
569: if( n==p->nLookahead ){
570: break; /* Same as a prior transaction set */
571: }
572: }
573: }
574: /* Insert transaction set at index i. */
575: for(j=0; j<p->nLookahead; j++){
576: k = p->aLookahead[j].lookahead - p->mnLookahead + i;
577: p->aAction[k] = p->aLookahead[j];
578: if( k>=p->nAction ) p->nAction = k+1;
579: }
580: p->nLookahead = 0;
581:
582: /* Return the offset that is added to the lookahead in order to get the
583: ** index into yy_action of the action */
584: return i - p->mnLookahead;
585: }
586:
587: /********************** From the file "assert.c" ****************************/
588: /*
589: ** A more efficient way of handling assertions.
590: */
591: void myassert(file,line)
592: char *file;
593: int line;
594: {
595: fprintf(stderr,"Assertion failed on line %d of file \"%s\"\n",line,file);
596: exit(1);
597: }
598: /********************** From the file "build.c" *****************************/
599: /*
600: ** Routines to construction the finite state machine for the LEMON
601: ** parser generator.
602: */
603:
604: /* Find a precedence symbol of every rule in the grammar.
605: **
606: ** Those rules which have a precedence symbol coded in the input
607: ** grammar using the "[symbol]" construct will already have the
608: ** rp->precsym field filled. Other rules take as their precedence
609: ** symbol the first RHS symbol with a defined precedence. If there
610: ** are not RHS symbols with a defined precedence, the precedence
611: ** symbol field is left blank.
612: */
613: void FindRulePrecedences(xp)
614: struct lemon *xp;
615: {
616: struct rule *rp;
617: for(rp=xp->rule; rp; rp=rp->next){
618: if( rp->precsym==0 ){
619: int i;
620: for(i=0; i<rp->nrhs; i++){
621: if( rp->rhs[i]->prec>=0 ){
622: rp->precsym = rp->rhs[i];
623: break;
624: }
625: }
626: }
627: }
628: return;
629: }
630:
631: /* Find all nonterminals which will generate the empty string.
632: ** Then go back and compute the first sets of every nonterminal.
633: ** The first set is the set of all terminal symbols which can begin
634: ** a string generated by that nonterminal.
635: */
636: void FindFirstSets(lemp)
637: struct lemon *lemp;
638: {
639: int i;
640: struct rule *rp;
641: int progress;
642:
643: for(i=0; i<lemp->nsymbol; i++){
644: lemp->symbols[i]->lambda = Bo_FALSE;
645: }
646: for(i=lemp->nterminal; i<lemp->nsymbol; i++){
647: lemp->symbols[i]->firstset = SetNew();
648: }
649:
650: /* First compute all lambdas */
651: do{
652: progress = 0;
653: for(rp=lemp->rule; rp; rp=rp->next){
654: if( rp->lhs->lambda ) continue;
655: for(i=0; i<rp->nrhs; i++){
656: if( rp->rhs[i]->lambda==Bo_FALSE ) break;
657: }
658: if( i==rp->nrhs ){
659: rp->lhs->lambda = Bo_TRUE;
660: progress = 1;
661: }
662: }
663: }while( progress );
664:
665: /* Now compute all first sets */
666: do{
667: struct symbol *s1, *s2;
668: progress = 0;
669: for(rp=lemp->rule; rp; rp=rp->next){
670: s1 = rp->lhs;
671: for(i=0; i<rp->nrhs; i++){
672: s2 = rp->rhs[i];
673: if( s2->type==TERMINAL ){
674: progress += SetAdd(s1->firstset,s2->index);
675: break;
676: }else if( s1==s2 ){
677: if( s1->lambda==Bo_FALSE ) break;
678: }else{
679: progress += SetUnion(s1->firstset,s2->firstset);
680: if( s2->lambda==Bo_FALSE ) break;
681: }
682: }
683: }
684: }while( progress );
685: return;
686: }
687:
688: /* Compute all LR(0) states for the grammar. Links
689: ** are added to between some states so that the LR(1) follow sets
690: ** can be computed later.
691: */
692: PRIVATE struct state *getstate(/* struct lemon * */); /* forward reference */
693: void FindStates(lemp)
694: struct lemon *lemp;
695: {
696: struct symbol *sp;
697: struct rule *rp;
698:
699: Configlist_init();
700:
701: /* Find the start symbol */
702: if( lemp->start ){
703: sp = Symbol_find(lemp->start);
704: if( sp==0 ){
705: ErrorMsg(lemp->filename,0,
706: "The specified start symbol \"%s\" is not \
707: in a nonterminal of the grammar. \"%s\" will be used as the start \
708: symbol instead.",lemp->start,lemp->rule->lhs->name);
709: lemp->errorcnt++;
710: sp = lemp->rule->lhs;
711: }
712: }else{
713: sp = lemp->rule->lhs;
714: }
715:
716: /* Make sure the start symbol doesn't occur on the right-hand side of
717: ** any rule. Report an error if it does. (YACC would generate a new
718: ** start symbol in this case.) */
719: for(rp=lemp->rule; rp; rp=rp->next){
720: int i;
721: for(i=0; i<rp->nrhs; i++){
722: if( rp->rhs[i]==sp ){
723: ErrorMsg(lemp->filename,0,
724: "The start symbol \"%s\" occurs on the \
725: right-hand side of a rule. This will result in a parser which \
726: does not work properly.",sp->name);
727: lemp->errorcnt++;
728: }
729: }
730: }
731:
732: /* The basis configuration set for the first state
733: ** is all rules which have the start symbol as their
734: ** left-hand side */
735: for(rp=sp->rule; rp; rp=rp->nextlhs){
736: struct config *newcfp;
737: newcfp = Configlist_addbasis(rp,0);
738: SetAdd(newcfp->fws,0);
739: }
740:
741: /* Compute the first state. All other states will be
742: ** computed automatically during the computation of the first one.
743: ** The returned pointer to the first state is not used. */
744: (void)getstate(lemp);
745: return;
746: }
747:
748: /* Return a pointer to a state which is described by the configuration
749: ** list which has been built from calls to Configlist_add.
750: */
751: PRIVATE void buildshifts(/* struct lemon *, struct state * */); /* Forwd ref */
752: PRIVATE struct state *getstate(lemp)
753: struct lemon *lemp;
754: {
755: struct config *cfp, *bp;
756: struct state *stp;
757:
758: /* Extract the sorted basis of the new state. The basis was constructed
759: ** by prior calls to "Configlist_addbasis()". */
760: Configlist_sortbasis();
761: bp = Configlist_basis();
762:
763: /* Get a state with the same basis */
764: stp = State_find(bp);
765: if( stp ){
766: /* A state with the same basis already exists! Copy all the follow-set
767: ** propagation links from the state under construction into the
768: ** preexisting state, then return a pointer to the preexisting state */
769: struct config *x, *y;
770: for(x=bp, y=stp->bp; x && y; x=x->bp, y=y->bp){
771: Plink_copy(&y->bplp,x->bplp);
772: Plink_delete(x->fplp);
773: x->fplp = x->bplp = 0;
774: }
775: cfp = Configlist_return();
776: Configlist_eat(cfp);
777: }else{
778: /* This really is a new state. Construct all the details */
779: Configlist_closure(lemp); /* Compute the configuration closure */
780: Configlist_sort(); /* Sort the configuration closure */
781: cfp = Configlist_return(); /* Get a pointer to the config list */
782: stp = State_new(); /* A new state structure */
783: MemoryCheck(stp);
784: stp->bp = bp; /* Remember the configuration basis */
785: stp->cfp = cfp; /* Remember the configuration closure */
786: stp->index = lemp->nstate++; /* Every state gets a sequence number */
787: stp->ap = 0; /* No actions, yet. */
788: State_insert(stp,stp->bp); /* Add to the state table */
789: buildshifts(lemp,stp); /* Recursively compute successor states */
790: }
791: return stp;
792: }
793:
794: /* Construct all successor states to the given state. A "successor"
795: ** state is any state which can be reached by a shift action.
796: */
797: PRIVATE void buildshifts(lemp,stp)
798: struct lemon *lemp;
799: struct state *stp; /* The state from which successors are computed */
800: {
801: struct config *cfp; /* For looping thru the config closure of "stp" */
802: struct config *bcfp; /* For the inner loop on config closure of "stp" */
803: struct config *new; /* */
804: struct symbol *sp; /* Symbol following the dot in configuration "cfp" */
805: struct symbol *bsp; /* Symbol following the dot in configuration "bcfp" */
806: struct state *newstp; /* A pointer to a successor state */
807:
808: /* Each configuration becomes complete after it contibutes to a successor
809: ** state. Initially, all configurations are incomplete */
810: for(cfp=stp->cfp; cfp; cfp=cfp->next) cfp->status = INCOMPLETE;
811:
812: /* Loop through all configurations of the state "stp" */
813: for(cfp=stp->cfp; cfp; cfp=cfp->next){
814: if( cfp->status==COMPLETE ) continue; /* Already used by inner loop */
815: if( cfp->dot>=cfp->rp->nrhs ) continue; /* Can't shift this config */
816: Configlist_reset(); /* Reset the new config set */
817: sp = cfp->rp->rhs[cfp->dot]; /* Symbol after the dot */
818:
819: /* For every configuration in the state "stp" which has the symbol "sp"
820: ** following its dot, add the same configuration to the basis set under
821: ** construction but with the dot shifted one symbol to the right. */
822: for(bcfp=cfp; bcfp; bcfp=bcfp->next){
823: if( bcfp->status==COMPLETE ) continue; /* Already used */
824: if( bcfp->dot>=bcfp->rp->nrhs ) continue; /* Can't shift this one */
825: bsp = bcfp->rp->rhs[bcfp->dot]; /* Get symbol after dot */
826: if( bsp!=sp ) continue; /* Must be same as for "cfp" */
827: bcfp->status = COMPLETE; /* Mark this config as used */
828: new = Configlist_addbasis(bcfp->rp,bcfp->dot+1);
829: Plink_add(&new->bplp,bcfp);
830: }
831:
832: /* Get a pointer to the state described by the basis configuration set
833: ** constructed in the preceding loop */
834: newstp = getstate(lemp);
835:
836: /* The state "newstp" is reached from the state "stp" by a shift action
837: ** on the symbol "sp" */
838: Action_add(&stp->ap,SHIFT,sp,newstp);
839: }
840: }
841:
842: /*
843: ** Construct the propagation links
844: */
845: void FindLinks(lemp)
846: struct lemon *lemp;
847: {
848: int i;
849: struct config *cfp, *other;
850: struct state *stp;
851: struct plink *plp;
852:
853: /* Housekeeping detail:
854: ** Add to every propagate link a pointer back to the state to
855: ** which the link is attached. */
856: for(i=0; i<lemp->nstate; i++){
857: stp = lemp->sorted[i];
858: for(cfp=stp->cfp; cfp; cfp=cfp->next){
859: cfp->stp = stp;
860: }
861: }
862:
863: /* Convert all backlinks into forward links. Only the forward
864: ** links are used in the follow-set computation. */
865: for(i=0; i<lemp->nstate; i++){
866: stp = lemp->sorted[i];
867: for(cfp=stp->cfp; cfp; cfp=cfp->next){
868: for(plp=cfp->bplp; plp; plp=plp->next){
869: other = plp->cfp;
870: Plink_add(&other->fplp,cfp);
871: }
872: }
873: }
874: }
875:
876: /* Compute all followsets.
877: **
878: ** A followset is the set of all symbols which can come immediately
879: ** after a configuration.
880: */
881: void FindFollowSets(lemp)
882: struct lemon *lemp;
883: {
884: int i;
885: struct config *cfp;
886: struct plink *plp;
887: int progress;
888: int change;
889:
890: for(i=0; i<lemp->nstate; i++){
891: for(cfp=lemp->sorted[i]->cfp; cfp; cfp=cfp->next){
892: cfp->status = INCOMPLETE;
893: }
894: }
895:
896: do{
897: progress = 0;
898: for(i=0; i<lemp->nstate; i++){
899: for(cfp=lemp->sorted[i]->cfp; cfp; cfp=cfp->next){
900: if( cfp->status==COMPLETE ) continue;
901: for(plp=cfp->fplp; plp; plp=plp->next){
902: change = SetUnion(plp->cfp->fws,cfp->fws);
903: if( change ){
904: plp->cfp->status = INCOMPLETE;
905: progress = 1;
906: }
907: }
908: cfp->status = COMPLETE;
909: }
910: }
911: }while( progress );
912: }
913:
914: static int resolve_conflict();
915:
916: /* Compute the reduce actions, and resolve conflicts.
917: */
918: void FindActions(lemp)
919: struct lemon *lemp;
920: {
921: int i,j;
922: struct config *cfp;
923: struct symbol *sp;
924: struct rule *rp;
925:
926: /* Add all of the reduce actions
927: ** A reduce action is added for each element of the followset of
928: ** a configuration which has its dot at the extreme right.
929: */
930: for(i=0; i<lemp->nstate; i++){ /* Loop over all states */
931: struct state *stp;
932: stp = lemp->sorted[i];
933: for(cfp=stp->cfp; cfp; cfp=cfp->next){ /* Loop over all configurations */
934: if( cfp->rp->nrhs==cfp->dot ){ /* Is dot at extreme right? */
935: for(j=0; j<lemp->nterminal; j++){
936: if( SetFind(cfp->fws,j) ){
937: /* Add a reduce action to the state "stp" which will reduce by the
938: ** rule "cfp->rp" if the lookahead symbol is "lemp->symbols[j]" */
939: Action_add(&stp->ap,REDUCE,lemp->symbols[j],cfp->rp);
940: }
941: }
942: }
943: }
944: }
945:
946: /* Add the accepting token */
947: if( lemp->start ){
948: sp = Symbol_find(lemp->start);
949: if( sp==0 ) sp = lemp->rule->lhs;
950: }else{
951: sp = lemp->rule->lhs;
952: }
953: /* Add to the first state (which is always the starting state of the
954: ** finite state machine) an action to ACCEPT if the lookahead is the
955: ** start nonterminal. */
956: Action_add(&lemp->sorted[0]->ap,ACCEPT,sp,0);
957:
958: /* Resolve conflicts */
959: for(i=0; i<lemp->nstate; i++){
960: struct action *ap, *nap;
961: struct state *stp;
962: stp = lemp->sorted[i];
963: assert( stp->ap );
964: stp->ap = Action_sort(stp->ap);
965: for(ap=stp->ap; ap && ap->next; ap=ap->next){
966: for(nap=ap->next; nap && nap->sp==ap->sp; nap=nap->next){
967: /* The two actions "ap" and "nap" have the same lookahead.
968: ** Figure out which one should be used */
969: lemp->nconflict += resolve_conflict(ap,nap,lemp->errsym);
970: }
971: }
972: }
973:
974: /* Report an error for each rule that can never be reduced. */
975: for(rp=lemp->rule; rp; rp=rp->next) rp->canReduce = Bo_FALSE;
976: for(i=0; i<lemp->nstate; i++){
977: struct action *ap;
978: for(ap=lemp->sorted[i]->ap; ap; ap=ap->next){
979: if( ap->type==REDUCE ) ap->x.rp->canReduce = Bo_TRUE;
980: }
981: }
982: for(rp=lemp->rule; rp; rp=rp->next){
983: if( rp->canReduce ) continue;
984: ErrorMsg(lemp->filename,rp->ruleline,"This rule can not be reduced.\n");
985: lemp->errorcnt++;
986: }
987: }
988:
989: /* Resolve a conflict between the two given actions. If the
990: ** conflict can't be resolve, return non-zero.
991: **
992: ** NO LONGER TRUE:
993: ** To resolve a conflict, first look to see if either action
994: ** is on an error rule. In that case, take the action which
995: ** is not associated with the error rule. If neither or both
996: ** actions are associated with an error rule, then try to
997: ** use precedence to resolve the conflict.
998: **
999: ** If either action is a SHIFT, then it must be apx. This
1000: ** function won't work if apx->type==REDUCE and apy->type==SHIFT.
1001: */
1002: static int resolve_conflict(apx,apy,errsym)
1003: struct action *apx;
1004: struct action *apy;
1005: struct symbol *errsym; /* The error symbol (if defined. NULL otherwise) */
1006: {
1007: struct symbol *spx, *spy;
1008: int errcnt = 0;
1009: UNUSED(errsym);
1010: assert( apx->sp==apy->sp ); /* Otherwise there would be no conflict */
1011: if( apx->type==SHIFT && apy->type==REDUCE ){
1012: spx = apx->sp;
1013: spy = apy->x.rp->precsym;
1014: if( spy==0 || spx->prec<0 || spy->prec<0 ){
1015: /* Not enough precedence information. */
1016: apy->type = CONFLICT;
1017: errcnt++;
1018: }else if( spx->prec>spy->prec ){ /* Lower precedence wins */
1019: apy->type = RD_RESOLVED;
1020: }else if( spx->prec<spy->prec ){
1021: apx->type = SH_RESOLVED;
1022: }else if( spx->prec==spy->prec && spx->assoc==RIGHT ){ /* Use operator */
1023: apy->type = RD_RESOLVED; /* associativity */
1024: }else if( spx->prec==spy->prec && spx->assoc==LEFT ){ /* to break tie */
1025: apx->type = SH_RESOLVED;
1026: }else{
1027: assert( spx->prec==spy->prec && spx->assoc==NONE );
1028: apy->type = CONFLICT;
1029: errcnt++;
1030: }
1031: }else if( apx->type==REDUCE && apy->type==REDUCE ){
1032: spx = apx->x.rp->precsym;
1033: spy = apy->x.rp->precsym;
1034: if( spx==0 || spy==0 || spx->prec<0 ||
1035: spy->prec<0 || spx->prec==spy->prec ){
1036: apy->type = CONFLICT;
1037: errcnt++;
1038: }else if( spx->prec>spy->prec ){
1039: apy->type = RD_RESOLVED;
1040: }else if( spx->prec<spy->prec ){
1041: apx->type = RD_RESOLVED;
1042: }
1043: }else{
1044: assert(
1045: apx->type==SH_RESOLVED ||
1046: apx->type==RD_RESOLVED ||
1047: apx->type==CONFLICT ||
1048: apy->type==SH_RESOLVED ||
1049: apy->type==RD_RESOLVED ||
1050: apy->type==CONFLICT
1051: );
1052: /* The REDUCE/SHIFT case cannot happen because SHIFTs come before
1053: ** REDUCEs on the list. If we reach this point it must be because
1054: ** the parser conflict had already been resolved. */
1055: }
1056: return errcnt;
1057: }
1058: /********************* From the file "configlist.c" *************************/
1059: /*
1060: ** Routines to processing a configuration list and building a state
1061: ** in the LEMON parser generator.
1062: */
1063:
1064: static struct config *freelist = 0; /* List of free configurations */
1065: static struct config *current = 0; /* Top of list of configurations */
1066: static struct config **currentend = 0; /* Last on list of configs */
1067: static struct config *basis = 0; /* Top of list of basis configs */
1068: static struct config **basisend = 0; /* End of list of basis configs */
1069:
1070: /* Return a pointer to a new configuration */
1071: PRIVATE struct config *newconfig(){
1072: struct config *new;
1073: if( freelist==0 ){
1074: int i;
1075: int amt = 3;
1076: freelist = (struct config *)malloc( sizeof(struct config)*amt );
1077: if( freelist==0 ){
1078: fprintf(stderr,"Unable to allocate memory for a new configuration.");
1079: exit(1);
1080: }
1081: for(i=0; i<amt-1; i++) freelist[i].next = &freelist[i+1];
1082: freelist[amt-1].next = 0;
1083: }
1084: new = freelist;
1085: freelist = freelist->next;
1086: return new;
1087: }
1088:
1089: /* The configuration "old" is no longer used */
1090: PRIVATE void deleteconfig(old)
1091: struct config *old;
1092: {
1093: old->next = freelist;
1094: freelist = old;
1095: }
1096:
1097: /* Initialized the configuration list builder */
1098: void Configlist_init(){
1099: current = 0;
1100: currentend = ¤t;
1101: basis = 0;
1102: basisend = &basis;
1103: Configtable_init();
1104: return;
1105: }
1106:
1107: /* Initialized the configuration list builder */
1108: void Configlist_reset(){
1109: current = 0;
1110: currentend = ¤t;
1111: basis = 0;
1112: basisend = &basis;
1113: Configtable_clear(0);
1114: return;
1115: }
1116:
1117: /* Add another configuration to the configuration list */
1118: struct config *Configlist_add(rp,dot)
1119: struct rule *rp; /* The rule */
1120: int dot; /* Index into the RHS of the rule where the dot goes */
1121: {
1122: struct config *cfp, model;
1123:
1124: assert( currentend!=0 );
1125: model.rp = rp;
1126: model.dot = dot;
1127: cfp = Configtable_find(&model);
1128: if( cfp==0 ){
1129: cfp = newconfig();
1130: cfp->rp = rp;
1131: cfp->dot = dot;
1132: cfp->fws = SetNew();
1133: cfp->stp = 0;
1134: cfp->fplp = cfp->bplp = 0;
1135: cfp->next = 0;
1136: cfp->bp = 0;
1137: *currentend = cfp;
1138: currentend = &cfp->next;
1139: Configtable_insert(cfp);
1140: }
1141: return cfp;
1142: }
1143:
1144: /* Add a basis configuration to the configuration list */
1145: struct config *Configlist_addbasis(rp,dot)
1146: struct rule *rp;
1147: int dot;
1148: {
1149: struct config *cfp, model;
1150:
1151: assert( basisend!=0 );
1152: assert( currentend!=0 );
1153: model.rp = rp;
1154: model.dot = dot;
1155: cfp = Configtable_find(&model);
1156: if( cfp==0 ){
1157: cfp = newconfig();
1158: cfp->rp = rp;
1159: cfp->dot = dot;
1160: cfp->fws = SetNew();
1161: cfp->stp = 0;
1162: cfp->fplp = cfp->bplp = 0;
1163: cfp->next = 0;
1164: cfp->bp = 0;
1165: *currentend = cfp;
1166: currentend = &cfp->next;
1167: *basisend = cfp;
1168: basisend = &cfp->bp;
1169: Configtable_insert(cfp);
1170: }
1171: return cfp;
1172: }
1173:
1174: /* Compute the closure of the configuration list */
1175: void Configlist_closure(lemp)
1176: struct lemon *lemp;
1177: {
1178: struct config *cfp, *newcfp;
1179: struct rule *rp, *newrp;
1180: struct symbol *sp, *xsp;
1181: int i, dot;
1182:
1183: assert( currentend!=0 );
1184: for(cfp=current; cfp; cfp=cfp->next){
1185: rp = cfp->rp;
1186: dot = cfp->dot;
1187: if( dot>=rp->nrhs ) continue;
1188: sp = rp->rhs[dot];
1189: if( sp->type==NONTERMINAL ){
1190: if( sp->rule==0 && sp!=lemp->errsym ){
1191: ErrorMsg(lemp->filename,rp->line,"Nonterminal \"%s\" has no rules.",
1192: sp->name);
1193: lemp->errorcnt++;
1194: }
1195: for(newrp=sp->rule; newrp; newrp=newrp->nextlhs){
1196: newcfp = Configlist_add(newrp,0);
1197: for(i=dot+1; i<rp->nrhs; i++){
1198: xsp = rp->rhs[i];
1199: if( xsp->type==TERMINAL ){
1200: SetAdd(newcfp->fws,xsp->index);
1201: break;
1202: }else{
1203: SetUnion(newcfp->fws,xsp->firstset);
1204: if( xsp->lambda==Bo_FALSE ) break;
1205: }
1206: }
1207: if( i==rp->nrhs ) Plink_add(&cfp->fplp,newcfp);
1208: }
1209: }
1210: }
1211: return;
1212: }
1213:
1214: /* Sort the configuration list */
1215: void Configlist_sort(){
1216: current = (struct config *)msort(current,&(current->next),Configcmp);
1217: currentend = 0;
1218: return;
1219: }
1220:
1221: /* Sort the basis configuration list */
1222: void Configlist_sortbasis(){
1223: basis = (struct config *)msort(current,&(current->bp),Configcmp);
1224: basisend = 0;
1225: return;
1226: }
1227:
1228: /* Return a pointer to the head of the configuration list and
1229: ** reset the list */
1230: struct config *Configlist_return(){
1231: struct config *old;
1232: old = current;
1233: current = 0;
1234: currentend = 0;
1235: return old;
1236: }
1237:
1238: /* Return a pointer to the head of the configuration list and
1239: ** reset the list */
1240: struct config *Configlist_basis(){
1241: struct config *old;
1242: old = basis;
1243: basis = 0;
1244: basisend = 0;
1245: return old;
1246: }
1247:
1248: /* Free all elements of the given configuration list */
1249: void Configlist_eat(cfp)
1250: struct config *cfp;
1251: {
1252: struct config *nextcfp;
1253: for(; cfp; cfp=nextcfp){
1254: nextcfp = cfp->next;
1255: assert( cfp->fplp==0 );
1256: assert( cfp->bplp==0 );
1257: if( cfp->fws ) SetFree(cfp->fws);
1258: deleteconfig(cfp);
1259: }
1260: return;
1261: }
1262: /***************** From the file "error.c" *********************************/
1263: /*
1264: ** Code for printing error message.
1265: */
1266:
1267: /* Find a good place to break "msg" so that its length is at least "min"
1268: ** but no more than "max". Make the point as close to max as possible.
1269: */
1270: static int findbreak(msg,min,max)
1271: char *msg;
1272: int min;
1273: int max;
1274: {
1275: int i,spot;
1276: char c;
1277: for(i=spot=min; i<=max; i++){
1278: c = msg[i];
1279: if( c=='\t' ) msg[i] = ' ';
1280: if( c=='\n' ){ msg[i] = ' '; spot = i; break; }
1281: if( c==0 ){ spot = i; break; }
1282: if( c=='-' && i<max-1 ) spot = i+1;
1283: if( c==' ' ) spot = i;
1284: }
1285: return spot;
1286: }
1287:
1288: /*
1289: ** The error message is split across multiple lines if necessary. The
1290: ** splits occur at a space, if there is a space available near the end
1291: ** of the line.
1292: */
1293: #define ERRMSGSIZE 10000 /* Hope this is big enough. No way to error check */
1294: #define LINEWIDTH 79 /* Max width of any output line */
1295: #define PREFIXLIMIT 30 /* Max width of the prefix on each line */
1296: void ErrorMsg(const char *filename, int lineno, const char *format, ...){
1297: char errmsg[ERRMSGSIZE];
1298: char prefix[PREFIXLIMIT+10];
1299: int errmsgsize;
1300: int prefixsize;
1301: int availablewidth;
1302: va_list ap;
1303: int end, restart, base;
1304:
1305: va_start(ap, format);
1306: /* Prepare a prefix to be prepended to every output line */
1307: if( lineno>0 ){
1308: sprintf(prefix,"%.*s:%d: ",PREFIXLIMIT-10,filename,lineno);
1309: }else{
1310: sprintf(prefix,"%.*s: ",PREFIXLIMIT-10,filename);
1311: }
1312: prefixsize = strlen(prefix);
1313: availablewidth = LINEWIDTH - prefixsize;
1314:
1315: /* Generate the error message */
1316: vsprintf(errmsg,format,ap);
1317: va_end(ap);
1318: errmsgsize = strlen(errmsg);
1319: /* Remove trailing '\n's from the error message. */
1320: while( errmsgsize>0 && errmsg[errmsgsize-1]=='\n' ){
1321: errmsg[--errmsgsize] = 0;
1322: }
1323:
1324: /* Print the error message */
1325: base = 0;
1326: while( errmsg[base]!=0 ){
1327: end = restart = findbreak(&errmsg[base],0,availablewidth);
1328: restart += base;
1329: while( errmsg[restart]==' ' ) restart++;
1330: fprintf(stdout,"%s%.*s\n",prefix,end,&errmsg[base]);
1331: base = restart;
1332: }
1333: }
1334: /**************** From the file "main.c" ************************************/
1335: /*
1336: ** Main program file for the LEMON parser generator.
1337: */
1338:
1339: /* Report an out-of-memory condition and abort. This function
1340: ** is used mostly by the "MemoryCheck" macro in struct.h
1341: */
1342: void memory_error() {
1343: fprintf(stderr,"Out of memory. Aborting...\n");
1344: exit(1);
1345: }
1346:
1347:
1348: /* The main program. Parse the command line and do it... */
1349: int main(argc,argv)
1350: int argc;
1351: char **argv;
1352: {
1353: static int version = 0;
1354: static int rpflag = 0;
1355: static int basisflag = 0;
1356: static int compress = 0;
1357: static int quiet = 0;
1358: static int statistics = 0;
1359: static int mhflag = 0;
1360: static struct s_options options[] = {
1361: {OPT_FLAG, "b", (char*)&basisflag, "Print only the basis in report."},
1362: {OPT_FLAG, "c", (char*)&compress, "Don't compress the action table."},
1363: {OPT_FLAG, "g", (char*)&rpflag, "Print grammar without actions."},
1364: {OPT_FLAG, "m", (char*)&mhflag, "Output a makeheaders compatible file"},
1365: {OPT_FLAG, "q", (char*)&quiet, "(Quiet) Don't print the report file."},
1366: {OPT_FLAG, "s", (char*)&statistics, "Print parser stats to standard output."},
1367: {OPT_FLAG, "x", (char*)&version, "Print the version number."},
1368: {OPT_FLAG,0,0,0}
1369: };
1370: int i;
1371: struct lemon lem;
1372: char *def_tmpl_name = "lempar.c";
1373:
1374: UNUSED(argc);
1375: OptInit(argv,options,stderr);
1376: if( version ){
1377: printf("Lemon version 1.0\n");
1378: exit(0);
1379: }
1380: if( OptNArgs() < 1 ){
1381: fprintf(stderr,"Exactly one filename argument is required.\n");
1382: exit(1);
1383: }
1384: lem.errorcnt = 0;
1385:
1386: /* Initialize the machine */
1387: Strsafe_init();
1388: Symbol_init();
1389: State_init();
1390: lem.argv0 = argv[0];
1391: lem.filename = OptArg(0);
1392: lem.tmplname = (OptNArgs() == 2) ? OptArg(1) : def_tmpl_name;
1393: lem.basisflag = basisflag;
1394: lem.has_fallback = 0;
1395: lem.nconflict = 0;
1396: lem.name = lem.include = lem.arg = lem.tokentype = lem.start = 0;
1397: lem.vartype = 0;
1398: lem.stacksize = 0;
1399: lem.error = lem.overflow = lem.failure = lem.accept = lem.tokendest =
1400: lem.tokenprefix = lem.outname = lem.extracode = 0;
1401: lem.vardest = 0;
1402: lem.tablesize = 0;
1403: Symbol_new("$");
1404: lem.errsym = Symbol_new("error");
1405:
1406: /* Parse the input file */
1407: Parse(&lem);
1408: if( lem.errorcnt ) exit(lem.errorcnt);
1409: if( lem.rule==0 ){
1410: fprintf(stderr,"Empty grammar.\n");
1411: exit(1);
1412: }
1413:
1414: /* Count and index the symbols of the grammar */
1415: lem.nsymbol = Symbol_count();
1416: Symbol_new("{default}");
1417: lem.symbols = Symbol_arrayof();
1418: for(i=0; i<=lem.nsymbol; i++) lem.symbols[i]->index = i;
1419: qsort(lem.symbols,lem.nsymbol+1,sizeof(struct symbol*),
1420: (int(*)())Symbolcmpp);
1421: for(i=0; i<=lem.nsymbol; i++) lem.symbols[i]->index = i;
1422: for(i=1; isupper(lem.symbols[i]->name[0]); i++);
1423: lem.nterminal = i;
1424:
1425: /* Generate a reprint of the grammar, if requested on the command line */
1426: if( rpflag ){
1427: Reprint(&lem);
1428: }else{
1429: /* Initialize the size for all follow and first sets */
1430: SetSize(lem.nterminal);
1431:
1432: /* Find the precedence for every production rule (that has one) */
1433: FindRulePrecedences(&lem);
1434:
1435: /* Compute the lambda-nonterminals and the first-sets for every
1436: ** nonterminal */
1437: FindFirstSets(&lem);
1438:
1439: /* Compute all LR(0) states. Also record follow-set propagation
1440: ** links so that the follow-set can be computed later */
1441: lem.nstate = 0;
1442: FindStates(&lem);
1443: lem.sorted = State_arrayof();
1444:
1445: /* Tie up loose ends on the propagation links */
1446: FindLinks(&lem);
1447:
1448: /* Compute the follow set of every reducible configuration */
1449: FindFollowSets(&lem);
1450:
1451: /* Compute the action tables */
1452: FindActions(&lem);
1453:
1454: /* Compress the action tables */
1455: if( compress==0 ) CompressTables(&lem);
1456:
1457: /* Generate a report of the parser generated. (the "y.output" file) */
1458: if( !quiet ) ReportOutput(&lem);
1459:
1460: /* Generate the source code for the parser */
1461: ReportTable(&lem, mhflag);
1462:
1463: /* Produce a header file for use by the scanner. (This step is
1464: ** omitted if the "-m" option is used because makeheaders will
1465: ** generate the file for us.) */
1466: if( !mhflag ) ReportHeader(&lem);
1467: }
1468: if( statistics ){
1469: printf("Parser statistics: %d terminals, %d nonterminals, %d rules\n",
1470: lem.nterminal, lem.nsymbol - lem.nterminal, lem.nrule);
1471: printf(" %d states, %d parser table entries, %d conflicts\n",
1472: lem.nstate, lem.tablesize, lem.nconflict);
1473: }
1474: if( lem.nconflict ){
1475: fprintf(stderr,"%d parsing conflicts.\n",lem.nconflict);
1476: }
1477: exit(lem.errorcnt + lem.nconflict);
1478: }
1479: /******************** From the file "msort.c" *******************************/
1480: /*
1481: ** A generic merge-sort program.
1482: **
1483: ** USAGE:
1484: ** Let "ptr" be a pointer to some structure which is at the head of
1485: ** a null-terminated list. Then to sort the list call:
1486: **
1487: ** ptr = msort(ptr,&(ptr->next),cmpfnc);
1488: **
1489: ** In the above, "cmpfnc" is a pointer to a function which compares
1490: ** two instances of the structure and returns an integer, as in
1491: ** strcmp. The second argument is a pointer to the pointer to the
1492: ** second element of the linked list. This address is used to compute
1493: ** the offset to the "next" field within the structure. The offset to
1494: ** the "next" field must be constant for all structures in the list.
1495: **
1496: ** The function returns a new pointer which is the head of the list
1497: ** after sorting.
1498: **
1499: ** ALGORITHM:
1500: ** Merge-sort.
1501: */
1502:
1503: /*
1504: ** Return a pointer to the next structure in the linked list.
1505: */
1506: #define NEXT(A) (*(char**)(((unsigned long)A)+offset))
1507:
1508: /*
1509: ** Inputs:
1510: ** a: A sorted, null-terminated linked list. (May be null).
1511: ** b: A sorted, null-terminated linked list. (May be null).
1512: ** cmp: A pointer to the comparison function.
1513: ** offset: Offset in the structure to the "next" field.
1514: **
1515: ** Return Value:
1516: ** A pointer to the head of a sorted list containing the elements
1517: ** of both a and b.
1518: **
1519: ** Side effects:
1520: ** The "next" pointers for elements in the lists a and b are
1521: ** changed.
1522: */
1523: static char *merge(a,b,cmp,offset)
1524: char *a;
1525: char *b;
1526: int (*cmp)();
1527: int offset;
1528: {
1529: char *ptr, *head;
1530:
1531: if( a==0 ){
1532: head = b;
1533: }else if( b==0 ){
1534: head = a;
1535: }else{
1536: if( (*cmp)(a,b)<0 ){
1537: ptr = a;
1538: a = NEXT(a);
1539: }else{
1540: ptr = b;
1541: b = NEXT(b);
1542: }
1543: head = ptr;
1544: while( a && b ){
1545: if( (*cmp)(a,b)<0 ){
1546: NEXT(ptr) = a;
1547: ptr = a;
1548: a = NEXT(a);
1549: }else{
1550: NEXT(ptr) = b;
1551: ptr = b;
1552: b = NEXT(b);
1553: }
1554: }
1555: if( a ) NEXT(ptr) = a;
1556: else NEXT(ptr) = b;
1557: }
1558: return head;
1559: }
1560:
1561: /*
1562: ** Inputs:
1563: ** list: Pointer to a singly-linked list of structures.
1564: ** next: Pointer to pointer to the second element of the list.
1565: ** cmp: A comparison function.
1566: **
1567: ** Return Value:
1568: ** A pointer to the head of a sorted list containing the elements
1569: ** orginally in list.
1570: **
1571: ** Side effects:
1572: ** The "next" pointers for elements in list are changed.
1573: */
1574: #define LISTSIZE 30
1575: char *msort(list,next,cmp)
1576: char *list;
1577: char **next;
1578: int (*cmp)();
1579: {
1580: unsigned long offset;
1581: char *ep;
1582: char *set[LISTSIZE];
1583: int i;
1584: offset = (unsigned long)next - (unsigned long)list;
1585: for(i=0; i<LISTSIZE; i++) set[i] = 0;
1586: while( list ){
1587: ep = list;
1588: list = NEXT(list);
1589: NEXT(ep) = 0;
1590: for(i=0; i<LISTSIZE-1 && set[i]!=0; i++){
1591: ep = merge(ep,set[i],cmp,offset);
1592: set[i] = 0;
1593: }
1594: set[i] = ep;
1595: }
1596: ep = 0;
1597: for(i=0; i<LISTSIZE; i++) if( set[i] ) ep = merge(ep,set[i],cmp,offset);
1598: return ep;
1599: }
1600: /************************ From the file "option.c" **************************/
1601: static char **argv;
1602: static struct s_options *op;
1603: static FILE *errstream;
1604:
1605: #define ISOPT(X) ((X)[0]=='-'||(X)[0]=='+'||strchr((X),'=')!=0)
1606:
1607: /*
1608: ** Print the command line with a carrot pointing to the k-th character
1609: ** of the n-th field.
1610: */
1611: static void errline(n,k,err)
1612: int n;
1613: int k;
1614: FILE *err;
1615: {
1616: int spcnt = 0, i;
1617: if( argv[0] ) {
1618: fprintf(err,"%s",argv[0]);
1619: spcnt += strlen(argv[0]) + 1;
1620: }
1621: for(i=1; i<n && argv[i]; i++){
1622: fprintf(err," %s",argv[i]);
1623: spcnt += strlen(argv[i]) + 1;
1624: }
1625: spcnt += k;
1626: for(; argv[i]; i++) fprintf(err," %s",argv[i]);
1627: if( spcnt<20 ){
1628: fprintf(err,"\n%*s^-- here\n",spcnt,"");
1629: }else{
1630: fprintf(err,"\n%*shere --^\n",spcnt-7,"");
1631: }
1632: }
1633:
1634: /*
1635: ** Return the index of the N-th non-switch argument. Return -1
1636: ** if N is out of range.
1637: */
1638: static int argindex(n)
1639: int n;
1640: {
1641: int i;
1642: int dashdash = 0;
1643: if( argv!=0 && *argv!=0 ){
1644: for(i=1; argv[i]; i++){
1645: if( dashdash || !ISOPT(argv[i]) ){
1646: if( n==0 ) return i;
1647: n--;
1648: }
1649: if( strcmp(argv[i],"--")==0 ) dashdash = 1;
1650: }
1651: }
1652: return -1;
1653: }
1654:
1655: static char emsg[] = "Command line syntax error: ";
1656:
1657: /*
1658: ** Process a flag command line argument.
1659: */
1660: static int handleflags(i,err)
1661: int i;
1662: FILE *err;
1663: {
1664: int v;
1665: int errcnt = 0;
1666: int j;
1667: for(j=0; op[j].label; j++){
1668: if( strcmp(&argv[i][1],op[j].label)==0 ) break;
1669: }
1670: v = argv[i][0]=='-' ? 1 : 0;
1671: if( op[j].label==0 ){
1672: if( err ){
1673: fprintf(err,"%sundefined option.\n",emsg);
1674: errline(i,1,err);
1675: }
1676: errcnt++;
1677: }else if( op[j].type==OPT_FLAG ){
1678: *((int*)op[j].arg) = v;
1679: }else if( op[j].type==OPT_FFLAG ){
1680: (*(void(*)())(intptr_t)(op[j].arg))(v);
1681: }else{
1682: if( err ){
1683: fprintf(err,"%smissing argument on switch.\n",emsg);
1684: errline(i,1,err);
1685: }
1686: errcnt++;
1687: }
1688: return errcnt;
1689: }
1690:
1691: /*
1692: ** Process a command line switch which has an argument.
1693: */
1694: static int handleswitch(i,err)
1695: int i;
1696: FILE *err;
1697: {
1698: int lv = 0;
1699: double dv = 0.0;
1700: char *sv = 0, *end;
1701: char *cp;
1702: int j;
1703: int errcnt = 0;
1704: cp = strchr(argv[i],'=');
1705: *cp = 0;
1706: for(j=0; op[j].label; j++){
1707: if( strcmp(argv[i],op[j].label)==0 ) break;
1708: }
1709: *cp = '=';
1710: if( op[j].label==0 ){
1711: if( err ){
1712: fprintf(err,"%sundefined option.\n",emsg);
1713: errline(i,0,err);
1714: }
1715: errcnt++;
1716: }else{
1717: cp++;
1718: switch( op[j].type ){
1719: case OPT_FLAG:
1720: case OPT_FFLAG:
1721: if( err ){
1722: fprintf(err,"%soption requires an argument.\n",emsg);
1723: errline(i,0,err);
1724: }
1725: errcnt++;
1726: break;
1727: case OPT_DBL:
1728: case OPT_FDBL:
1729: dv = strtod(cp,&end);
1730: if( *end ){
1731: if( err ){
1732: fprintf(err,"%sillegal character in floating-point argument.\n",emsg);
1733: errline(i,((unsigned long)end)-(unsigned long)argv[i],err);
1734: }
1735: errcnt++;
1736: }
1737: break;
1738: case OPT_INT:
1739: case OPT_FINT:
1740: lv = strtol(cp,&end,0);
1741: if( *end ){
1742: if( err ){
1743: fprintf(err,"%sillegal character in integer argument.\n",emsg);
1744: errline(i,((unsigned long)end)-(unsigned long)argv[i],err);
1745: }
1746: errcnt++;
1747: }
1748: break;
1749: case OPT_STR:
1750: case OPT_FSTR:
1751: sv = cp;
1752: break;
1753: }
1754: switch( op[j].type ){
1755: case OPT_FLAG:
1756: case OPT_FFLAG:
1757: break;
1758: case OPT_DBL:
1759: *(double*)(op[j].arg) = dv;
1760: break;
1761: case OPT_FDBL:
1762: (*(void(*)())(intptr_t)(op[j].arg))(dv);
1763: break;
1764: case OPT_INT:
1765: *(int*)(op[j].arg) = lv;
1766: break;
1767: case OPT_FINT:
1768: (*(void(*)())(intptr_t)(op[j].arg))((int)lv);
1769: break;
1770: case OPT_STR:
1771: *(char**)(op[j].arg) = sv;
1772: break;
1773: case OPT_FSTR:
1774: (*(void(*)())(intptr_t)(op[j].arg))(sv);
1775: break;
1776: }
1777: }
1778: return errcnt;
1779: }
1780:
1781: int OptInit(a,o,err)
1782: char **a;
1783: struct s_options *o;
1784: FILE *err;
1785: {
1786: int errcnt = 0;
1787: argv = a;
1788: op = o;
1789: errstream = err;
1790: if( argv && *argv && op ){
1791: int i;
1792: for(i=1; argv[i]; i++){
1793: if( argv[i][0]=='+' || argv[i][0]=='-' ){
1794: errcnt += handleflags(i,err);
1795: }else if( strchr(argv[i],'=') ){
1796: errcnt += handleswitch(i,err);
1797: }
1798: }
1799: }
1800: if( errcnt>0 ){
1801: fprintf(err,"Valid command line options for \"%s\" are:\n",*a);
1802: OptPrint();
1803: exit(1);
1804: }
1805: return 0;
1806: }
1807:
1808: int OptNArgs(){
1809: int cnt = 0;
1810: int dashdash = 0;
1811: int i;
1812: if( argv!=0 && argv[0]!=0 ){
1813: for(i=1; argv[i]; i++){
1814: if( dashdash || !ISOPT(argv[i]) ) cnt++;
1815: if( strcmp(argv[i],"--")==0 ) dashdash = 1;
1816: }
1817: }
1818: return cnt;
1819: }
1820:
1821: char *OptArg(n)
1822: int n;
1823: {
1824: int i;
1825: i = argindex(n);
1826: return i>=0 ? argv[i] : 0;
1827: }
1828:
1829: void OptErr(n)
1830: int n;
1831: {
1832: int i;
1833: i = argindex(n);
1834: if( i>=0 ) errline(i,0,errstream);
1835: }
1836:
1837: void OptPrint(){
1838: int i;
1839: int max, len;
1840: max = 0;
1841: for(i=0; op[i].label; i++){
1842: len = strlen(op[i].label) + 1;
1843: switch( op[i].type ){
1844: case OPT_FLAG:
1845: case OPT_FFLAG:
1846: break;
1847: case OPT_INT:
1848: case OPT_FINT:
1849: len += 9; /* length of "<integer>" */
1850: break;
1851: case OPT_DBL:
1852: case OPT_FDBL:
1853: len += 6; /* length of "<real>" */
1854: break;
1855: case OPT_STR:
1856: case OPT_FSTR:
1857: len += 8; /* length of "<string>" */
1858: break;
1859: }
1860: if( len>max ) max = len;
1861: }
1862: for(i=0; op[i].label; i++){
1863: switch( op[i].type ){
1864: case OPT_FLAG:
1865: case OPT_FFLAG:
1866: fprintf(errstream," -%-*s %s\n",max,op[i].label,op[i].message);
1867: break;
1868: case OPT_INT:
1869: case OPT_FINT:
1870: fprintf(errstream," %s=<integer>%*s %s\n",op[i].label,
1871: (int)(max-strlen(op[i].label)-9),"",op[i].message);
1872: break;
1873: case OPT_DBL:
1874: case OPT_FDBL:
1875: fprintf(errstream," %s=<real>%*s %s\n",op[i].label,
1876: (int)(max-strlen(op[i].label)-6),"",op[i].message);
1877: break;
1878: case OPT_STR:
1879: case OPT_FSTR:
1880: fprintf(errstream," %s=<string>%*s %s\n",op[i].label,
1881: (int)(max-strlen(op[i].label)-8),"",op[i].message);
1882: break;
1883: }
1884: }
1885: }
1886: /*********************** From the file "parse.c" ****************************/
1887: /*
1888: ** Input file parser for the LEMON parser generator.
1889: */
1890:
1891: /* The state of the parser */
1892: struct pstate {
1893: char *filename; /* Name of the input file */
1894: int tokenlineno; /* Linenumber at which current token starts */
1895: int errorcnt; /* Number of errors so far */
1896: char *tokenstart; /* Text of current token */
1897: struct lemon *gp; /* Global state vector */
1898: enum e_state {
1899: INITIALIZE,
1900: WAITING_FOR_DECL_OR_RULE,
1901: WAITING_FOR_DECL_KEYWORD,
1902: WAITING_FOR_DECL_ARG,
1903: WAITING_FOR_PRECEDENCE_SYMBOL,
1904: WAITING_FOR_ARROW,
1905: IN_RHS,
1906: LHS_ALIAS_1,
1907: LHS_ALIAS_2,
1908: LHS_ALIAS_3,
1909: RHS_ALIAS_1,
1910: RHS_ALIAS_2,
1911: PRECEDENCE_MARK_1,
1912: PRECEDENCE_MARK_2,
1913: RESYNC_AFTER_RULE_ERROR,
1914: RESYNC_AFTER_DECL_ERROR,
1915: WAITING_FOR_DESTRUCTOR_SYMBOL,
1916: WAITING_FOR_DATATYPE_SYMBOL,
1917: WAITING_FOR_FALLBACK_ID
1918: } state; /* The state of the parser */
1919: struct symbol *fallback; /* The fallback token */
1920: struct symbol *lhs; /* Left-hand side of current rule */
1921: char *lhsalias; /* Alias for the LHS */
1922: int nrhs; /* Number of right-hand side symbols seen */
1923: struct symbol *rhs[MAXRHS]; /* RHS symbols */
1924: char *alias[MAXRHS]; /* Aliases for each RHS symbol (or NULL) */
1925: struct rule *prevrule; /* Previous rule parsed */
1926: char *declkeyword; /* Keyword of a declaration */
1927: char **declargslot; /* Where the declaration argument should be put */
1928: int *decllnslot; /* Where the declaration linenumber is put */
1929: enum e_assoc declassoc; /* Assign this association to decl arguments */
1930: int preccounter; /* Assign this precedence to decl arguments */
1931: struct rule *firstrule; /* Pointer to first rule in the grammar */
1932: struct rule *lastrule; /* Pointer to the most recently parsed rule */
1933: };
1934:
1935: /* Parse a single token */
1936: static void parseonetoken(psp)
1937: struct pstate *psp;
1938: {
1939: char *x;
1940: x = Strsafe(psp->tokenstart); /* Save the token permanently */
1941: #if 0
1942: printf("%s:%d: Token=[%s] state=%d\n",psp->filename,psp->tokenlineno,
1943: x,psp->state);
1944: #endif
1945: switch( psp->state ){
1946: case INITIALIZE:
1947: psp->prevrule = 0;
1948: psp->preccounter = 0;
1949: psp->firstrule = psp->lastrule = 0;
1950: psp->gp->nrule = 0;
1951: /* Fall thru to next case */
1952: case WAITING_FOR_DECL_OR_RULE:
1953: if( x[0]=='%' ){
1954: psp->state = WAITING_FOR_DECL_KEYWORD;
1955: }else if( islower(x[0]) ){
1956: psp->lhs = Symbol_new(x);
1957: psp->nrhs = 0;
1958: psp->lhsalias = 0;
1959: psp->state = WAITING_FOR_ARROW;
1960: }else if( x[0]=='{' ){
1961: if( psp->prevrule==0 ){
1962: ErrorMsg(psp->filename,psp->tokenlineno,
1963: "There is not prior rule opon which to attach the code \
1964: fragment which begins on this line.");
1965: psp->errorcnt++;
1966: }else if( psp->prevrule->code!=0 ){
1967: ErrorMsg(psp->filename,psp->tokenlineno,
1968: "Code fragment beginning on this line is not the first \
1969: to follow the previous rule.");
1970: psp->errorcnt++;
1971: }else{
1972: psp->prevrule->line = psp->tokenlineno;
1973: psp->prevrule->code = &x[1];
1974: }
1975: }else if( x[0]=='[' ){
1976: psp->state = PRECEDENCE_MARK_1;
1977: }else{
1978: ErrorMsg(psp->filename,psp->tokenlineno,
1979: "Token \"%s\" should be either \"%%\" or a nonterminal name.",
1980: x);
1981: psp->errorcnt++;
1982: }
1983: break;
1984: case PRECEDENCE_MARK_1:
1985: if( !isupper(x[0]) ){
1986: ErrorMsg(psp->filename,psp->tokenlineno,
1987: "The precedence symbol must be a terminal.");
1988: psp->errorcnt++;
1989: }else if( psp->prevrule==0 ){
1990: ErrorMsg(psp->filename,psp->tokenlineno,
1991: "There is no prior rule to assign precedence \"[%s]\".",x);
1992: psp->errorcnt++;
1993: }else if( psp->prevrule->precsym!=0 ){
1994: ErrorMsg(psp->filename,psp->tokenlineno,
1995: "Precedence mark on this line is not the first \
1996: to follow the previous rule.");
1997: psp->errorcnt++;
1998: }else{
1999: psp->prevrule->precsym = Symbol_new(x);
2000: }
2001: psp->state = PRECEDENCE_MARK_2;
2002: break;
2003: case PRECEDENCE_MARK_2:
2004: if( x[0]!=']' ){
2005: ErrorMsg(psp->filename,psp->tokenlineno,
2006: "Missing \"]\" on precedence mark.");
2007: psp->errorcnt++;
2008: }
2009: psp->state = WAITING_FOR_DECL_OR_RULE;
2010: break;
2011: case WAITING_FOR_ARROW:
2012: if( x[0]==':' && x[1]==':' && x[2]=='=' ){
2013: psp->state = IN_RHS;
2014: }else if( x[0]=='(' ){
2015: psp->state = LHS_ALIAS_1;
2016: }else{
2017: ErrorMsg(psp->filename,psp->tokenlineno,
2018: "Expected to see a \":\" following the LHS symbol \"%s\".",
2019: psp->lhs->name);
2020: psp->errorcnt++;
2021: psp->state = RESYNC_AFTER_RULE_ERROR;
2022: }
2023: break;
2024: case LHS_ALIAS_1:
2025: if( isalpha(x[0]) ){
2026: psp->lhsalias = x;
2027: psp->state = LHS_ALIAS_2;
2028: }else{
2029: ErrorMsg(psp->filename,psp->tokenlineno,
2030: "\"%s\" is not a valid alias for the LHS \"%s\"\n",
2031: x,psp->lhs->name);
2032: psp->errorcnt++;
2033: psp->state = RESYNC_AFTER_RULE_ERROR;
2034: }
2035: break;
2036: case LHS_ALIAS_2:
2037: if( x[0]==')' ){
2038: psp->state = LHS_ALIAS_3;
2039: }else{
2040: ErrorMsg(psp->filename,psp->tokenlineno,
2041: "Missing \")\" following LHS alias name \"%s\".",psp->lhsalias);
2042: psp->errorcnt++;
2043: psp->state = RESYNC_AFTER_RULE_ERROR;
2044: }
2045: break;
2046: case LHS_ALIAS_3:
2047: if( x[0]==':' && x[1]==':' && x[2]=='=' ){
2048: psp->state = IN_RHS;
2049: }else{
2050: ErrorMsg(psp->filename,psp->tokenlineno,
2051: "Missing \"->\" following: \"%s(%s)\".",
2052: psp->lhs->name,psp->lhsalias);
2053: psp->errorcnt++;
2054: psp->state = RESYNC_AFTER_RULE_ERROR;
2055: }
2056: break;
2057: case IN_RHS:
2058: if( x[0]=='.' ){
2059: struct rule *rp;
2060: rp = (struct rule *)malloc( sizeof(struct rule) +
2061: sizeof(struct symbol*)*psp->nrhs + sizeof(char*)*psp->nrhs );
2062: if( rp==0 ){
2063: ErrorMsg(psp->filename,psp->tokenlineno,
2064: "Can't allocate enough memory for this rule.");
2065: psp->errorcnt++;
2066: psp->prevrule = 0;
2067: }else{
2068: int i;
2069: rp->ruleline = psp->tokenlineno;
2070: rp->rhs = (struct symbol**)&rp[1];
2071: rp->rhsalias = (char**)&(rp->rhs[psp->nrhs]);
2072: for(i=0; i<psp->nrhs; i++){
2073: rp->rhs[i] = psp->rhs[i];
2074: rp->rhsalias[i] = psp->alias[i];
2075: }
2076: rp->lhs = psp->lhs;
2077: rp->lhsalias = psp->lhsalias;
2078: rp->nrhs = psp->nrhs;
2079: rp->code = 0;
2080: rp->precsym = 0;
2081: rp->index = psp->gp->nrule++;
2082: rp->nextlhs = rp->lhs->rule;
2083: rp->lhs->rule = rp;
2084: rp->next = 0;
2085: if( psp->firstrule==0 ){
2086: psp->firstrule = psp->lastrule = rp;
2087: }else{
2088: psp->lastrule->next = rp;
2089: psp->lastrule = rp;
2090: }
2091: psp->prevrule = rp;
2092: }
2093: psp->state = WAITING_FOR_DECL_OR_RULE;
2094: }else if( isalpha(x[0]) ){
2095: if( psp->nrhs>=MAXRHS ){
2096: ErrorMsg(psp->filename,psp->tokenlineno,
2097: "Too many symbol on RHS or rule beginning at \"%s\".",
2098: x);
2099: psp->errorcnt++;
2100: psp->state = RESYNC_AFTER_RULE_ERROR;
2101: }else{
2102: psp->rhs[psp->nrhs] = Symbol_new(x);
2103: psp->alias[psp->nrhs] = 0;
2104: psp->nrhs++;
2105: }
2106: }else if( x[0]=='(' && psp->nrhs>0 ){
2107: psp->state = RHS_ALIAS_1;
2108: }else{
2109: ErrorMsg(psp->filename,psp->tokenlineno,
2110: "Illegal character on RHS of rule: \"%s\".",x);
2111: psp->errorcnt++;
2112: psp->state = RESYNC_AFTER_RULE_ERROR;
2113: }
2114: break;
2115: case RHS_ALIAS_1:
2116: if( isalpha(x[0]) ){
2117: psp->alias[psp->nrhs-1] = x;
2118: psp->state = RHS_ALIAS_2;
2119: }else{
2120: ErrorMsg(psp->filename,psp->tokenlineno,
2121: "\"%s\" is not a valid alias for the RHS symbol \"%s\"\n",
2122: x,psp->rhs[psp->nrhs-1]->name);
2123: psp->errorcnt++;
2124: psp->state = RESYNC_AFTER_RULE_ERROR;
2125: }
2126: break;
2127: case RHS_ALIAS_2:
2128: if( x[0]==')' ){
2129: psp->state = IN_RHS;
2130: }else{
2131: ErrorMsg(psp->filename,psp->tokenlineno,
2132: "Missing \")\" following LHS alias name \"%s\".",psp->lhsalias);
2133: psp->errorcnt++;
2134: psp->state = RESYNC_AFTER_RULE_ERROR;
2135: }
2136: break;
2137: case WAITING_FOR_DECL_KEYWORD:
2138: if( isalpha(x[0]) ){
2139: psp->declkeyword = x;
2140: psp->declargslot = 0;
2141: psp->decllnslot = 0;
2142: psp->state = WAITING_FOR_DECL_ARG;
2143: if( strcmp(x,"name")==0 ){
2144: psp->declargslot = &(psp->gp->name);
2145: }else if( strcmp(x,"include")==0 ){
2146: psp->declargslot = &(psp->gp->include);
2147: psp->decllnslot = &psp->gp->includeln;
2148: }else if( strcmp(x,"code")==0 ){
2149: psp->declargslot = &(psp->gp->extracode);
2150: psp->decllnslot = &psp->gp->extracodeln;
2151: }else if( strcmp(x,"token_destructor")==0 ){
2152: psp->declargslot = &psp->gp->tokendest;
2153: psp->decllnslot = &psp->gp->tokendestln;
2154: }else if( strcmp(x,"default_destructor")==0 ){
2155: psp->declargslot = &psp->gp->vardest;
2156: psp->decllnslot = &psp->gp->vardestln;
2157: }else if( strcmp(x,"token_prefix")==0 ){
2158: psp->declargslot = &psp->gp->tokenprefix;
2159: }else if( strcmp(x,"syntax_error")==0 ){
2160: psp->declargslot = &(psp->gp->error);
2161: psp->decllnslot = &psp->gp->errorln;
2162: }else if( strcmp(x,"parse_accept")==0 ){
2163: psp->declargslot = &(psp->gp->accept);
2164: psp->decllnslot = &psp->gp->acceptln;
2165: }else if( strcmp(x,"parse_failure")==0 ){
2166: psp->declargslot = &(psp->gp->failure);
2167: psp->decllnslot = &psp->gp->failureln;
2168: }else if( strcmp(x,"stack_overflow")==0 ){
2169: psp->declargslot = &(psp->gp->overflow);
2170: psp->decllnslot = &psp->gp->overflowln;
2171: }else if( strcmp(x,"extra_argument")==0 ){
2172: psp->declargslot = &(psp->gp->arg);
2173: }else if( strcmp(x,"token_type")==0 ){
2174: psp->declargslot = &(psp->gp->tokentype);
2175: }else if( strcmp(x,"default_type")==0 ){
2176: psp->declargslot = &(psp->gp->vartype);
2177: }else if( strcmp(x,"stack_size")==0 ){
2178: psp->declargslot = &(psp->gp->stacksize);
2179: }else if( strcmp(x,"start_symbol")==0 ){
2180: psp->declargslot = &(psp->gp->start);
2181: }else if( strcmp(x,"left")==0 ){
2182: psp->preccounter++;
2183: psp->declassoc = LEFT;
2184: psp->state = WAITING_FOR_PRECEDENCE_SYMBOL;
2185: }else if( strcmp(x,"right")==0 ){
2186: psp->preccounter++;
2187: psp->declassoc = RIGHT;
2188: psp->state = WAITING_FOR_PRECEDENCE_SYMBOL;
2189: }else if( strcmp(x,"nonassoc")==0 ){
2190: psp->preccounter++;
2191: psp->declassoc = NONE;
2192: psp->state = WAITING_FOR_PRECEDENCE_SYMBOL;
2193: }else if( strcmp(x,"destructor")==0 ){
2194: psp->state = WAITING_FOR_DESTRUCTOR_SYMBOL;
2195: }else if( strcmp(x,"type")==0 ){
2196: psp->state = WAITING_FOR_DATATYPE_SYMBOL;
2197: }else if( strcmp(x,"fallback")==0 ){
2198: psp->fallback = 0;
2199: psp->state = WAITING_FOR_FALLBACK_ID;
2200: }else{
2201: ErrorMsg(psp->filename,psp->tokenlineno,
2202: "Unknown declaration keyword: \"%%%s\".",x);
2203: psp->errorcnt++;
2204: psp->state = RESYNC_AFTER_DECL_ERROR;
2205: }
2206: }else{
2207: ErrorMsg(psp->filename,psp->tokenlineno,
2208: "Illegal declaration keyword: \"%s\".",x);
2209: psp->errorcnt++;
2210: psp->state = RESYNC_AFTER_DECL_ERROR;
2211: }
2212: break;
2213: case WAITING_FOR_DESTRUCTOR_SYMBOL:
2214: if( !isalpha(x[0]) ){
2215: ErrorMsg(psp->filename,psp->tokenlineno,
2216: "Symbol name missing after %destructor keyword");
2217: psp->errorcnt++;
2218: psp->state = RESYNC_AFTER_DECL_ERROR;
2219: }else{
2220: struct symbol *sp = Symbol_new(x);
2221: psp->declargslot = &sp->destructor;
2222: psp->decllnslot = &sp->destructorln;
2223: psp->state = WAITING_FOR_DECL_ARG;
2224: }
2225: break;
2226: case WAITING_FOR_DATATYPE_SYMBOL:
2227: if( !isalpha(x[0]) ){
2228: ErrorMsg(psp->filename,psp->tokenlineno,
2229: "Symbol name missing after %destructor keyword");
2230: psp->errorcnt++;
2231: psp->state = RESYNC_AFTER_DECL_ERROR;
2232: }else{
2233: struct symbol *sp = Symbol_new(x);
2234: psp->declargslot = &sp->datatype;
2235: psp->decllnslot = 0;
2236: psp->state = WAITING_FOR_DECL_ARG;
2237: }
2238: break;
2239: case WAITING_FOR_PRECEDENCE_SYMBOL:
2240: if( x[0]=='.' ){
2241: psp->state = WAITING_FOR_DECL_OR_RULE;
2242: }else if( isupper(x[0]) ){
2243: struct symbol *sp;
2244: sp = Symbol_new(x);
2245: if( sp->prec>=0 ){
2246: ErrorMsg(psp->filename,psp->tokenlineno,
2247: "Symbol \"%s\" has already be given a precedence.",x);
2248: psp->errorcnt++;
2249: }else{
2250: sp->prec = psp->preccounter;
2251: sp->assoc = psp->declassoc;
2252: }
2253: }else{
2254: ErrorMsg(psp->filename,psp->tokenlineno,
2255: "Can't assign a precedence to \"%s\".",x);
2256: psp->errorcnt++;
2257: }
2258: break;
2259: case WAITING_FOR_DECL_ARG:
2260: if( (x[0]=='{' || x[0]=='\"' || isalnum(x[0])) ){
2261: if( *(psp->declargslot)!=0 ){
2262: ErrorMsg(psp->filename,psp->tokenlineno,
2263: "The argument \"%s\" to declaration \"%%%s\" is not the first.",
2264: x[0]=='\"' ? &x[1] : x,psp->declkeyword);
2265: psp->errorcnt++;
2266: psp->state = RESYNC_AFTER_DECL_ERROR;
2267: }else{
2268: *(psp->declargslot) = (x[0]=='\"' || x[0]=='{') ? &x[1] : x;
2269: if( psp->decllnslot ) *psp->decllnslot = psp->tokenlineno;
2270: psp->state = WAITING_FOR_DECL_OR_RULE;
2271: }
2272: }else{
2273: ErrorMsg(psp->filename,psp->tokenlineno,
2274: "Illegal argument to %%%s: %s",psp->declkeyword,x);
2275: psp->errorcnt++;
2276: psp->state = RESYNC_AFTER_DECL_ERROR;
2277: }
2278: break;
2279: case WAITING_FOR_FALLBACK_ID:
2280: if( x[0]=='.' ){
2281: psp->state = WAITING_FOR_DECL_OR_RULE;
2282: }else if( !isupper(x[0]) ){
2283: ErrorMsg(psp->filename, psp->tokenlineno,
2284: "%%fallback argument \"%s\" should be a token", x);
2285: psp->errorcnt++;
2286: }else{
2287: struct symbol *sp = Symbol_new(x);
2288: if( psp->fallback==0 ){
2289: psp->fallback = sp;
2290: }else if( sp->fallback ){
2291: ErrorMsg(psp->filename, psp->tokenlineno,
2292: "More than one fallback assigned to token %s", x);
2293: psp->errorcnt++;
2294: }else{
2295: sp->fallback = psp->fallback;
2296: psp->gp->has_fallback = 1;
2297: }
2298: }
2299: break;
2300: case RESYNC_AFTER_RULE_ERROR:
2301: /* if( x[0]=='.' ) psp->state = WAITING_FOR_DECL_OR_RULE;
2302: ** break; */
2303: case RESYNC_AFTER_DECL_ERROR:
2304: if( x[0]=='.' ) psp->state = WAITING_FOR_DECL_OR_RULE;
2305: if( x[0]=='%' ) psp->state = WAITING_FOR_DECL_KEYWORD;
2306: break;
2307: }
2308: }
2309:
2310: /* In spite of its name, this function is really a scanner. It read
2311: ** in the entire input file (all at once) then tokenizes it. Each
2312: ** token is passed to the function "parseonetoken" which builds all
2313: ** the appropriate data structures in the global state vector "gp".
2314: */
2315: struct pstate ps;
2316: void Parse(gp)
2317: struct lemon *gp;
2318: {
2319: FILE *fp;
2320: char *filebuf;
2321: size_t filesize;
2322: int lineno;
2323: int c;
2324: char *cp, *nextcp;
2325: int startline = 0;
2326:
2327: ps.gp = gp;
2328: ps.filename = gp->filename;
2329: ps.errorcnt = 0;
2330: ps.state = INITIALIZE;
2331:
2332: /* Begin by reading the input file */
2333: fp = fopen(ps.filename,"rb");
2334: if( fp==0 ){
2335: ErrorMsg(ps.filename,0,"Can't open this file for reading.");
2336: gp->errorcnt++;
2337: return;
2338: }
2339: fseek(fp,0,2);
2340: filesize = ftell(fp);
2341: rewind(fp);
2342: filebuf = (char *)malloc( filesize+1 );
2343: if( filebuf==0 ){
2344: ErrorMsg(ps.filename,0,"Can't allocate %d of memory to hold this file.",
2345: filesize+1);
2346: fclose(fp);
2347: gp->errorcnt++;
2348: return;
2349: }
2350: if( fread(filebuf,1,filesize,fp)!=filesize ){
2351: ErrorMsg(ps.filename,0,"Can't read in all %d bytes of this file.",
2352: filesize);
2353: free(filebuf);
2354: fclose(fp);
2355: gp->errorcnt++;
2356: return;
2357: }
2358: fclose(fp);
2359: filebuf[filesize] = 0;
2360:
2361: /* Now scan the text of the input file */
2362: lineno = 1;
2363: for(cp=filebuf; (c= *cp)!=0; ){
2364: if( c=='\n' ) lineno++; /* Keep track of the line number */
2365: if( isspace(c) ){ cp++; continue; } /* Skip all white space */
2366: if( c=='/' && cp[1]=='/' ){ /* Skip C++ style comments */
2367: cp+=2;
2368: while( (c= *cp)!=0 && c!='\n' ) cp++;
2369: continue;
2370: }
2371: if( c=='/' && cp[1]=='*' ){ /* Skip C style comments */
2372: cp+=2;
2373: while( (c= *cp)!=0 && (c!='/' || cp[-1]!='*') ){
2374: if( c=='\n' ) lineno++;
2375: cp++;
2376: }
2377: if( c ) cp++;
2378: continue;
2379: }
2380: ps.tokenstart = cp; /* Mark the beginning of the token */
2381: ps.tokenlineno = lineno; /* Linenumber on which token begins */
2382: if( c=='\"' ){ /* String literals */
2383: cp++;
2384: while( (c= *cp)!=0 && c!='\"' ){
2385: if( c=='\n' ) lineno++;
2386: cp++;
2387: }
2388: if( c==0 ){
2389: ErrorMsg(ps.filename,startline,
2390: "String starting on this line is not terminated before the end of the file.");
2391: ps.errorcnt++;
2392: nextcp = cp;
2393: }else{
2394: nextcp = cp+1;
2395: }
2396: }else if( c=='{' ){ /* A block of C code */
2397: int level;
2398: cp++;
2399: for(level=1; (c= *cp)!=0 && (level>1 || c!='}'); cp++){
2400: if( c=='\n' ) lineno++;
2401: else if( c=='{' ) level++;
2402: else if( c=='}' ) level--;
2403: else if( c=='/' && cp[1]=='*' ){ /* Skip comments */
2404: int prevc;
2405: cp = &cp[2];
2406: prevc = 0;
2407: while( (c= *cp)!=0 && (c!='/' || prevc!='*') ){
2408: if( c=='\n' ) lineno++;
2409: prevc = c;
2410: cp++;
2411: }
2412: }else if( c=='/' && cp[1]=='/' ){ /* Skip C++ style comments too */
2413: cp = &cp[2];
2414: while( (c= *cp)!=0 && c!='\n' ) cp++;
2415: if( c ) lineno++;
2416: }else if( c=='\'' || c=='\"' ){ /* String a character literals */
2417: int startchar, prevc;
2418: startchar = c;
2419: prevc = 0;
2420: for(cp++; (c= *cp)!=0 && (c!=startchar || prevc=='\\'); cp++){
2421: if( c=='\n' ) lineno++;
2422: if( prevc=='\\' ) prevc = 0;
2423: else prevc = c;
2424: }
2425: }
2426: }
2427: if( c==0 ){
2428: ErrorMsg(ps.filename,ps.tokenlineno,
2429: "C code starting on this line is not terminated before the end of the file.");
2430: ps.errorcnt++;
2431: nextcp = cp;
2432: }else{
2433: nextcp = cp+1;
2434: }
2435: }else if( isalnum(c) ){ /* Identifiers */
2436: while( (c= *cp)!=0 && (isalnum(c) || c=='_') ) cp++;
2437: nextcp = cp;
2438: }else if( c==':' && cp[1]==':' && cp[2]=='=' ){ /* The operator "::=" */
2439: cp += 3;
2440: nextcp = cp;
2441: }else{ /* All other (one character) operators */
2442: cp++;
2443: nextcp = cp;
2444: }
2445: c = *cp;
2446: *cp = 0; /* Null terminate the token */
2447: parseonetoken(&ps); /* Parse the token */
2448: *cp = c; /* Restore the buffer */
2449: cp = nextcp;
2450: }
2451: free(filebuf); /* Release the buffer after parsing */
2452: gp->rule = ps.firstrule;
2453: gp->errorcnt = ps.errorcnt;
2454: }
2455: /*************************** From the file "plink.c" *********************/
2456: /*
2457: ** Routines processing configuration follow-set propagation links
2458: ** in the LEMON parser generator.
2459: */
2460: static struct plink *plink_freelist = 0;
2461:
2462: /* Allocate a new plink */
2463: struct plink *Plink_new(){
2464: struct plink *new;
2465:
2466: if( plink_freelist==0 ){
2467: int i;
2468: int amt = 100;
2469: plink_freelist = (struct plink *)malloc( sizeof(struct plink)*amt );
2470: if( plink_freelist==0 ){
2471: fprintf(stderr,
2472: "Unable to allocate memory for a new follow-set propagation link.\n");
2473: exit(1);
2474: }
2475: for(i=0; i<amt-1; i++) plink_freelist[i].next = &plink_freelist[i+1];
2476: plink_freelist[amt-1].next = 0;
2477: }
2478: new = plink_freelist;
2479: plink_freelist = plink_freelist->next;
2480: return new;
2481: }
2482:
2483: /* Add a plink to a plink list */
2484: void Plink_add(plpp,cfp)
2485: struct plink **plpp;
2486: struct config *cfp;
2487: {
2488: struct plink *new;
2489: new = Plink_new();
2490: new->next = *plpp;
2491: *plpp = new;
2492: new->cfp = cfp;
2493: }
2494:
2495: /* Transfer every plink on the list "from" to the list "to" */
2496: void Plink_copy(to,from)
2497: struct plink **to;
2498: struct plink *from;
2499: {
2500: struct plink *nextpl;
2501: while( from ){
2502: nextpl = from->next;
2503: from->next = *to;
2504: *to = from;
2505: from = nextpl;
2506: }
2507: }
2508:
2509: /* Delete every plink on the list */
2510: void Plink_delete(plp)
2511: struct plink *plp;
2512: {
2513: struct plink *nextpl;
2514:
2515: while( plp ){
2516: nextpl = plp->next;
2517: plp->next = plink_freelist;
2518: plink_freelist = plp;
2519: plp = nextpl;
2520: }
2521: }
2522: /*********************** From the file "report.c" **************************/
2523: /*
2524: ** Procedures for generating reports and tables in the LEMON parser generator.
2525: */
2526:
2527: /* Generate a filename with the given suffix. Space to hold the
2528: ** name comes from malloc() and must be freed by the calling
2529: ** function.
2530: */
2531: PRIVATE char *file_makename(lemp,suffix)
2532: struct lemon *lemp;
2533: char *suffix;
2534: {
2535: char *name;
2536: char *cp;
2537:
2538: name = malloc( strlen(lemp->filename) + strlen(suffix) + 5 );
2539: if( name==0 ){
2540: fprintf(stderr,"Can't allocate space for a filename.\n");
2541: exit(1);
2542: }
2543: /* skip directory, JK */
2544: if (NULL == (cp = strrchr(lemp->filename, '/'))) {
2545: cp = lemp->filename;
2546: } else {
2547: cp++;
2548: }
2549: strcpy(name,cp);
2550: cp = strrchr(name,'.');
2551: if( cp ) *cp = 0;
2552: strcat(name,suffix);
2553: return name;
2554: }
2555:
2556: /* Open a file with a name based on the name of the input file,
2557: ** but with a different (specified) suffix, and return a pointer
2558: ** to the stream */
2559: PRIVATE FILE *file_open(lemp,suffix,mode)
2560: struct lemon *lemp;
2561: char *suffix;
2562: char *mode;
2563: {
2564: FILE *fp;
2565:
2566: if( lemp->outname ) free(lemp->outname);
2567: lemp->outname = file_makename(lemp, suffix);
2568: fp = fopen(lemp->outname,mode);
2569: if( fp==0 && *mode=='w' ){
2570: fprintf(stderr,"Can't open file \"%s\".\n",lemp->outname);
2571: lemp->errorcnt++;
2572: return 0;
2573: }
2574: return fp;
2575: }
2576:
2577: /* Duplicate the input file without comments and without actions
2578: ** on rules */
2579: void Reprint(lemp)
2580: struct lemon *lemp;
2581: {
2582: struct rule *rp;
2583: struct symbol *sp;
2584: int i, j, maxlen, len, ncolumns, skip;
2585: printf("// Reprint of input file \"%s\".\n// Symbols:\n",lemp->filename);
2586: maxlen = 10;
2587: for(i=0; i<lemp->nsymbol; i++){
2588: sp = lemp->symbols[i];
2589: len = strlen(sp->name);
2590: if( len>maxlen ) maxlen = len;
2591: }
2592: ncolumns = 76/(maxlen+5);
2593: if( ncolumns<1 ) ncolumns = 1;
2594: skip = (lemp->nsymbol + ncolumns - 1)/ncolumns;
2595: for(i=0; i<skip; i++){
2596: printf("//");
2597: for(j=i; j<lemp->nsymbol; j+=skip){
2598: sp = lemp->symbols[j];
2599: assert( sp->index==j );
2600: printf(" %3d %-*.*s",j,maxlen,maxlen,sp->name);
2601: }
2602: printf("\n");
2603: }
2604: for(rp=lemp->rule; rp; rp=rp->next){
2605: printf("%s",rp->lhs->name);
2606: /* if( rp->lhsalias ) printf("(%s)",rp->lhsalias); */
2607: printf(" ::=");
2608: for(i=0; i<rp->nrhs; i++){
2609: printf(" %s",rp->rhs[i]->name);
2610: /* if( rp->rhsalias[i] ) printf("(%s)",rp->rhsalias[i]); */
2611: }
2612: printf(".");
2613: if( rp->precsym ) printf(" [%s]",rp->precsym->name);
2614: /* if( rp->code ) printf("\n %s",rp->code); */
2615: printf("\n");
2616: }
2617: }
2618:
2619: PRIVATE void ConfigPrint(fp,cfp)
2620: FILE *fp;
2621: struct config *cfp;
2622: {
2623: struct rule *rp;
2624: int i;
2625: rp = cfp->rp;
2626: fprintf(fp,"%s ::=",rp->lhs->name);
2627: for(i=0; i<=rp->nrhs; i++){
2628: if( i==cfp->dot ) fprintf(fp," *");
2629: if( i==rp->nrhs ) break;
2630: fprintf(fp," %s",rp->rhs[i]->name);
2631: }
2632: }
2633:
2634: /* #define TEST */
2635: #ifdef TEST
2636: /* Print a set */
2637: PRIVATE void SetPrint(out,set,lemp)
2638: FILE *out;
2639: char *set;
2640: struct lemon *lemp;
2641: {
2642: int i;
2643: char *spacer;
2644: spacer = "";
2645: fprintf(out,"%12s[","");
2646: for(i=0; i<lemp->nterminal; i++){
2647: if( SetFind(set,i) ){
2648: fprintf(out,"%s%s",spacer,lemp->symbols[i]->name);
2649: spacer = " ";
2650: }
2651: }
2652: fprintf(out,"]\n");
2653: }
2654:
2655: /* Print a plink chain */
2656: void PlinkPrint(out,plp,tag)
2657: FILE *out;
2658: struct plink *plp;
2659: char *tag;
2660: {
2661: while( plp ){
2662: fprintf(out,"%12s%s (state %2d) ","",tag,plp->cfp->stp->index);
2663: ConfigPrint(out,plp->cfp);
2664: fprintf(out,"\n");
2665: plp = plp->next;
2666: }
2667: }
2668: #endif
2669:
2670: /* Print an action to the given file descriptor. Return FALSE if
2671: ** nothing was actually printed.
2672: */
2673: PRIVATE int PrintAction(struct action *ap, FILE *fp, int indent){
2674: int result = 1;
2675: switch( ap->type ){
2676: case SHIFT:
2677: fprintf(fp,"%*s shift %d",indent,ap->sp->name,ap->x.stp->index);
2678: break;
2679: case REDUCE:
2680: fprintf(fp,"%*s reduce %d",indent,ap->sp->name,ap->x.rp->index);
2681: break;
2682: case ACCEPT:
2683: fprintf(fp,"%*s accept",indent,ap->sp->name);
2684: break;
2685: case ERROR:
2686: fprintf(fp,"%*s error",indent,ap->sp->name);
2687: break;
2688: case CONFLICT:
2689: fprintf(fp,"%*s reduce %-3d ** Parsing conflict **",
2690: indent,ap->sp->name,ap->x.rp->index);
2691: break;
2692: case SH_RESOLVED:
2693: case RD_RESOLVED:
2694: case NOT_USED:
2695: result = 0;
2696: break;
2697: }
2698: return result;
2699: }
2700:
2701: /* Generate the "y.output" log file */
2702: void ReportOutput(lemp)
2703: struct lemon *lemp;
2704: {
2705: int i;
2706: struct state *stp;
2707: struct config *cfp;
2708: struct action *ap;
2709: FILE *fp;
2710:
2711: fp = file_open(lemp,".out","w");
2712: if( fp==0 ) return;
2713: fprintf(fp," \b");
2714: for(i=0; i<lemp->nstate; i++){
2715: stp = lemp->sorted[i];
2716: fprintf(fp,"State %d:\n",stp->index);
2717: if( lemp->basisflag ) cfp=stp->bp;
2718: else cfp=stp->cfp;
2719: while( cfp ){
2720: char buf[20];
2721: if( cfp->dot==cfp->rp->nrhs ){
2722: sprintf(buf,"(%d)",cfp->rp->index);
2723: fprintf(fp," %5s ",buf);
2724: }else{
2725: fprintf(fp," ");
2726: }
2727: ConfigPrint(fp,cfp);
2728: fprintf(fp,"\n");
2729: #ifdef TEST
2730: SetPrint(fp,cfp->fws,lemp);
2731: PlinkPrint(fp,cfp->fplp,"To ");
2732: PlinkPrint(fp,cfp->bplp,"From");
2733: #endif
2734: if( lemp->basisflag ) cfp=cfp->bp;
2735: else cfp=cfp->next;
2736: }
2737: fprintf(fp,"\n");
2738: for(ap=stp->ap; ap; ap=ap->next){
2739: if( PrintAction(ap,fp,30) ) fprintf(fp,"\n");
2740: }
2741: fprintf(fp,"\n");
2742: }
2743: fclose(fp);
2744: return;
2745: }
2746:
2747: extern int access();
2748: /* Search for the file "name" which is in the same directory as
2749: ** the exacutable */
2750: PRIVATE char *pathsearch(argv0,name,modemask)
2751: char *argv0;
2752: char *name;
2753: int modemask;
2754: {
2755: char *pathlist;
2756: char *path,*cp;
2757: char c;
2758:
2759: #ifdef __WIN32__
2760: cp = strrchr(argv0,'\\');
2761: #else
2762: cp = strrchr(argv0,'/');
2763: #endif
2764: if( cp ){
2765: c = *cp;
2766: *cp = 0;
2767: path = (char *)malloc( strlen(argv0) + strlen(name) + 2 );
2768: if( path ) sprintf(path,"%s/%s",argv0,name);
2769: *cp = c;
2770: }else{
2771: pathlist = getenv("PATH");
2772: if( pathlist==0 ) pathlist = ".:/bin:/usr/bin";
2773: path = (char *)malloc( strlen(pathlist)+strlen(name)+2 );
2774: if( path!=0 ){
2775: while( *pathlist ){
2776: cp = strchr(pathlist,':');
2777: if( cp==0 ) cp = &pathlist[strlen(pathlist)];
2778: c = *cp;
2779: *cp = 0;
2780: sprintf(path,"%s/%s",pathlist,name);
2781: *cp = c;
2782: if( c==0 ) pathlist = "";
2783: else pathlist = &cp[1];
2784: if( access(path,modemask)==0 ) break;
2785: }
2786: }
2787: }
2788: return path;
2789: }
2790:
2791: /* Given an action, compute the integer value for that action
2792: ** which is to be put in the action table of the generated machine.
2793: ** Return negative if no action should be generated.
2794: */
2795: PRIVATE int compute_action(lemp,ap)
2796: struct lemon *lemp;
2797: struct action *ap;
2798: {
2799: int act;
2800: switch( ap->type ){
2801: case SHIFT: act = ap->x.stp->index; break;
2802: case REDUCE: act = ap->x.rp->index + lemp->nstate; break;
2803: case ERROR: act = lemp->nstate + lemp->nrule; break;
2804: case ACCEPT: act = lemp->nstate + lemp->nrule + 1; break;
2805: default: act = -1; break;
2806: }
2807: return act;
2808: }
2809:
2810: #define LINESIZE 1000
2811: /* The next cluster of routines are for reading the template file
2812: ** and writing the results to the generated parser */
2813: /* The first function transfers data from "in" to "out" until
2814: ** a line is seen which begins with "%%". The line number is
2815: ** tracked.
2816: **
2817: ** if name!=0, then any word that begin with "Parse" is changed to
2818: ** begin with *name instead.
2819: */
2820: PRIVATE void tplt_xfer(name,in,out,lineno)
2821: char *name;
2822: FILE *in;
2823: FILE *out;
2824: int *lineno;
2825: {
2826: int i, iStart;
2827: char line[LINESIZE];
2828: while( fgets(line,LINESIZE,in) && (line[0]!='%' || line[1]!='%') ){
2829: (*lineno)++;
2830: iStart = 0;
2831: if( name ){
2832: for(i=0; line[i]; i++){
2833: if( line[i]=='P' && strncmp(&line[i],"Parse",5)==0
2834: && (i==0 || !isalpha(line[i-1]))
2835: ){
2836: if( i>iStart ) fprintf(out,"%.*s",i-iStart,&line[iStart]);
2837: fprintf(out,"%s",name);
2838: i += 4;
2839: iStart = i+1;
2840: }
2841: }
2842: }
2843: fprintf(out,"%s",&line[iStart]);
2844: }
2845: }
2846:
2847: /* The next function finds the template file and opens it, returning
2848: ** a pointer to the opened file. */
2849: PRIVATE FILE *tplt_open(lemp)
2850: struct lemon *lemp;
2851: {
2852:
2853: char buf[1000];
2854: FILE *in;
2855: char *tpltname;
2856: char *cp;
2857:
2858: cp = strrchr(lemp->filename,'.');
2859: if( cp ){
2860: sprintf(buf,"%.*s.lt",(int)(cp-lemp->filename),lemp->filename);
2861: }else{
2862: sprintf(buf,"%s.lt",lemp->filename);
2863: }
2864: if( access(buf,004)==0 ){
2865: tpltname = buf;
2866: }else if( access(lemp->tmplname,004)==0 ){
2867: tpltname = lemp->tmplname;
2868: }else{
2869: tpltname = pathsearch(lemp->argv0,lemp->tmplname,0);
2870: }
2871: if( tpltname==0 ){
2872: fprintf(stderr,"Can't find the parser driver template file \"%s\".\n",
2873: lemp->tmplname);
2874: lemp->errorcnt++;
2875: return 0;
2876: }
2877: in = fopen(tpltname,"r");
2878: if( in==0 ){
2879: fprintf(stderr,"Can't open the template file \"%s\".\n",lemp->tmplname);
2880: lemp->errorcnt++;
2881: return 0;
2882: }
2883: return in;
2884: }
2885:
2886: /* Print a string to the file and keep the linenumber up to date */
2887: PRIVATE void tplt_print(out,lemp,str,strln,lineno)
2888: FILE *out;
2889: struct lemon *lemp;
2890: char *str;
2891: int strln;
2892: int *lineno;
2893: {
2894: if( str==0 ) return;
2895: fprintf(out,"#line %d \"%s\"\n",strln,lemp->filename); (*lineno)++;
2896: while( *str ){
2897: if( *str=='\n' ) (*lineno)++;
2898: putc(*str,out);
2899: str++;
2900: }
2901: fprintf(out,"\n#line %d \"%s\"\n",*lineno+2,lemp->outname); (*lineno)+=2;
2902: return;
2903: }
2904:
2905: /*
2906: ** The following routine emits code for the destructor for the
2907: ** symbol sp
2908: */
2909: PRIVATE void emit_destructor_code(out,sp,lemp,lineno)
2910: FILE *out;
2911: struct symbol *sp;
2912: struct lemon *lemp;
2913: int *lineno;
2914: {
2915: char *cp = 0;
2916:
2917: int linecnt = 0;
2918: if( sp->type==TERMINAL ){
2919: cp = lemp->tokendest;
2920: if( cp==0 ) return;
2921: fprintf(out,"#line %d \"%s\"\n{",lemp->tokendestln,lemp->filename);
2922: }else if( sp->destructor ){
2923: cp = sp->destructor;
2924: fprintf(out,"#line %d \"%s\"\n{",sp->destructorln,lemp->filename);
2925: }else{
2926: cp = lemp->vardest;
2927: if( cp==0 ) return;
2928: fprintf(out,"#line %d \"%s\"\n{",lemp->vardestln,lemp->filename);
2929: }
2930: for(; *cp; cp++){
2931: if( *cp=='$' && cp[1]=='$' ){
2932: fprintf(out,"(yypminor->yy%d)",sp->dtnum);
2933: cp++;
2934: continue;
2935: }
2936: if( *cp=='\n' ) linecnt++;
2937: fputc(*cp,out);
2938: }
2939: (*lineno) += 3 + linecnt;
2940: fprintf(out,"}\n#line %d \"%s\"\n",*lineno,lemp->outname);
2941: return;
2942: }
2943:
2944: /*
2945: ** Return TRUE (non-zero) if the given symbol has a destructor.
2946: */
2947: PRIVATE int has_destructor(sp, lemp)
2948: struct symbol *sp;
2949: struct lemon *lemp;
2950: {
2951: int ret;
2952: if( sp->type==TERMINAL ){
2953: ret = lemp->tokendest!=0;
2954: }else{
2955: ret = lemp->vardest!=0 || sp->destructor!=0;
2956: }
2957: return ret;
2958: }
2959:
2960: /*
2961: ** Generate code which executes when the rule "rp" is reduced. Write
2962: ** the code to "out". Make sure lineno stays up-to-date.
2963: */
2964: PRIVATE void emit_code(out,rp,lemp,lineno)
2965: FILE *out;
2966: struct rule *rp;
2967: struct lemon *lemp;
2968: int *lineno;
2969: {
2970: char *cp, *xp;
2971: int linecnt = 0;
2972: int i;
2973: char lhsused = 0; /* True if the LHS element has been used */
2974: char used[MAXRHS]; /* True for each RHS element which is used */
2975:
2976: for(i=0; i<rp->nrhs; i++) used[i] = 0;
2977: lhsused = 0;
2978:
2979: /* Generate code to do the reduce action */
2980: if( rp->code ){
2981: fprintf(out,"#line %d \"%s\"\n{",rp->line,lemp->filename);
2982: for(cp=rp->code; *cp; cp++){
2983: if( isalpha(*cp) && (cp==rp->code || (!isalnum(cp[-1]) && cp[-1]!='_')) ){
2984: char saved;
2985: for(xp= &cp[1]; isalnum(*xp) || *xp=='_'; xp++);
2986: saved = *xp;
2987: *xp = 0;
2988: if( rp->lhsalias && strcmp(cp,rp->lhsalias)==0 ){
2989: fprintf(out,"yygotominor.yy%d",rp->lhs->dtnum);
2990: cp = xp;
2991: lhsused = 1;
2992: }else{
2993: for(i=0; i<rp->nrhs; i++){
2994: if( rp->rhsalias[i] && strcmp(cp,rp->rhsalias[i])==0 ){
2995: fprintf(out,"yymsp[%d].minor.yy%d",i-rp->nrhs+1,rp->rhs[i]->dtnum);
2996: cp = xp;
2997: used[i] = 1;
2998: break;
2999: }
3000: }
3001: }
3002: *xp = saved;
3003: }
3004: if( *cp=='\n' ) linecnt++;
3005: fputc(*cp,out);
3006: } /* End loop */
3007: (*lineno) += 3 + linecnt;
3008: fprintf(out,"}\n#line %d \"%s\"\n",*lineno,lemp->outname);
3009: } /* End if( rp->code ) */
3010:
3011: /* Check to make sure the LHS has been used */
3012: if( rp->lhsalias && !lhsused ){
3013: ErrorMsg(lemp->filename,rp->ruleline,
3014: "Label \"%s\" for \"%s(%s)\" is never used.",
3015: rp->lhsalias,rp->lhs->name,rp->lhsalias);
3016: lemp->errorcnt++;
3017: }
3018:
3019: /* Generate destructor code for RHS symbols which are not used in the
3020: ** reduce code */
3021: for(i=0; i<rp->nrhs; i++){
3022: if( rp->rhsalias[i] && !used[i] ){
3023: ErrorMsg(lemp->filename,rp->ruleline,
3024: "Label %s for \"%s(%s)\" is never used.",
3025: rp->rhsalias[i],rp->rhs[i]->name,rp->rhsalias[i]);
3026: lemp->errorcnt++;
3027: }else if( rp->rhsalias[i]==0 ){
3028: if( has_destructor(rp->rhs[i],lemp) ){
3029: fprintf(out," yy_destructor(%d,&yymsp[%d].minor);\n",
3030: rp->rhs[i]->index,i-rp->nrhs+1); (*lineno)++;
3031: }else{
3032: fprintf(out," /* No destructor defined for %s */\n",
3033: rp->rhs[i]->name);
3034: (*lineno)++;
3035: }
3036: }
3037: }
3038: return;
3039: }
3040:
3041: /*
3042: ** Print the definition of the union used for the parser's data stack.
3043: ** This union contains fields for every possible data type for tokens
3044: ** and nonterminals. In the process of computing and printing this
3045: ** union, also set the ".dtnum" field of every terminal and nonterminal
3046: ** symbol.
3047: */
3048: PRIVATE void print_stack_union(out,lemp,plineno,mhflag)
3049: FILE *out; /* The output stream */
3050: struct lemon *lemp; /* The main info structure for this parser */
3051: int *plineno; /* Pointer to the line number */
3052: int mhflag; /* True if generating makeheaders output */
3053: {
3054: int lineno; /* The line number of the output */
3055: char **types; /* A hash table of datatypes */
3056: int arraysize; /* Size of the "types" array */
3057: int maxdtlength; /* Maximum length of any ".datatype" field. */
3058: char *stddt; /* Standardized name for a datatype */
3059: int i,j; /* Loop counters */
3060: int hash; /* For hashing the name of a type */
3061: char *name; /* Name of the parser */
3062:
3063: /* Allocate and initialize types[] and allocate stddt[] */
3064: arraysize = lemp->nsymbol * 2;
3065: types = (char**)malloc( arraysize * sizeof(char*) );
3066: for(i=0; i<arraysize; i++) types[i] = 0;
3067: maxdtlength = 0;
3068: if( lemp->vartype ){
3069: maxdtlength = strlen(lemp->vartype);
3070: }
3071: for(i=0; i<lemp->nsymbol; i++){
3072: int len;
3073: struct symbol *sp = lemp->symbols[i];
3074: if( sp->datatype==0 ) continue;
3075: len = strlen(sp->datatype);
3076: if( len>maxdtlength ) maxdtlength = len;
3077: }
3078: stddt = (char*)malloc( maxdtlength*2 + 1 );
3079: if( types==0 || stddt==0 ){
3080: fprintf(stderr,"Out of memory.\n");
3081: exit(1);
3082: }
3083:
3084: /* Build a hash table of datatypes. The ".dtnum" field of each symbol
3085: ** is filled in with the hash index plus 1. A ".dtnum" value of 0 is
3086: ** used for terminal symbols. If there is no %default_type defined then
3087: ** 0 is also used as the .dtnum value for nonterminals which do not specify
3088: ** a datatype using the %type directive.
3089: */
3090: for(i=0; i<lemp->nsymbol; i++){
3091: struct symbol *sp = lemp->symbols[i];
3092: char *cp;
3093: if( sp==lemp->errsym ){
3094: sp->dtnum = arraysize+1;
3095: continue;
3096: }
3097: if( sp->type!=NONTERMINAL || (sp->datatype==0 && lemp->vartype==0) ){
3098: sp->dtnum = 0;
3099: continue;
3100: }
3101: cp = sp->datatype;
3102: if( cp==0 ) cp = lemp->vartype;
3103: j = 0;
3104: while( isspace(*cp) ) cp++;
3105: while( *cp ) stddt[j++] = *cp++;
3106: while( j>0 && isspace(stddt[j-1]) ) j--;
3107: stddt[j] = 0;
3108: hash = 0;
3109: for(j=0; stddt[j]; j++){
3110: hash = (unsigned int)hash*53u + (unsigned int) stddt[j];
3111: }
3112: hash = (hash & 0x7fffffff)%arraysize;
3113: while( types[hash] ){
3114: if( strcmp(types[hash],stddt)==0 ){
3115: sp->dtnum = hash + 1;
3116: break;
3117: }
3118: hash++;
3119: if( hash>=arraysize ) hash = 0;
3120: }
3121: if( types[hash]==0 ){
3122: sp->dtnum = hash + 1;
3123: types[hash] = (char*)malloc( strlen(stddt)+1 );
3124: if( types[hash]==0 ){
3125: fprintf(stderr,"Out of memory.\n");
3126: exit(1);
3127: }
3128: strcpy(types[hash],stddt);
3129: }
3130: }
3131:
3132: /* Print out the definition of YYTOKENTYPE and YYMINORTYPE */
3133: name = lemp->name ? lemp->name : "Parse";
3134: lineno = *plineno;
3135: if( mhflag ){ fprintf(out,"#if INTERFACE\n"); lineno++; }
3136: fprintf(out,"#define %sTOKENTYPE %s\n",name,
3137: lemp->tokentype?lemp->tokentype:"void*"); lineno++;
3138: if( mhflag ){ fprintf(out,"#endif\n"); lineno++; }
3139: fprintf(out,"typedef union {\n"); lineno++;
3140: fprintf(out," %sTOKENTYPE yy0;\n",name); lineno++;
3141: for(i=0; i<arraysize; i++){
3142: if( types[i]==0 ) continue;
3143: fprintf(out," %s yy%d;\n",types[i],i+1); lineno++;
3144: free(types[i]);
3145: }
3146: fprintf(out," int yy%d;\n",lemp->errsym->dtnum); lineno++;
3147: free(stddt);
3148: free(types);
3149: fprintf(out,"} YYMINORTYPE;\n"); lineno++;
3150: *plineno = lineno;
3151: }
3152:
3153: /*
3154: ** Return the name of a C datatype able to represent values between
3155: ** lwr and upr, inclusive.
3156: */
3157: static const char *minimum_size_type(int lwr, int upr){
3158: if( lwr>=0 ){
3159: if( upr<=255 ){
3160: return "unsigned char";
3161: }else if( upr<65535 ){
3162: return "unsigned short int";
3163: }else{
3164: return "unsigned int";
3165: }
3166: }else if( lwr>=-127 && upr<=127 ){
3167: return "signed char";
3168: }else if( lwr>=-32767 && upr<32767 ){
3169: return "short";
3170: }else{
3171: return "int";
3172: }
3173: }
3174:
3175: /*
3176: ** Each state contains a set of token transaction and a set of
3177: ** nonterminal transactions. Each of these sets makes an instance
3178: ** of the following structure. An array of these structures is used
3179: ** to order the creation of entries in the yy_action[] table.
3180: */
3181: struct axset {
3182: struct state *stp; /* A pointer to a state */
3183: int isTkn; /* True to use tokens. False for non-terminals */
3184: int nAction; /* Number of actions */
3185: };
3186:
3187: /*
3188: ** Compare to axset structures for sorting purposes
3189: */
3190: static int axset_compare(const void *a, const void *b){
3191: struct axset *p1 = (struct axset*)a;
3192: struct axset *p2 = (struct axset*)b;
3193: return p2->nAction - p1->nAction;
3194: }
3195:
3196: /* Generate C source code for the parser */
3197: void ReportTable(lemp, mhflag)
3198: struct lemon *lemp;
3199: int mhflag; /* Output in makeheaders format if true */
3200: {
3201: FILE *out, *in;
3202: char line[LINESIZE];
3203: int lineno;
3204: struct state *stp;
3205: struct action *ap;
3206: struct rule *rp;
3207: struct acttab *pActtab;
3208: int i, j, n;
3209: int mnTknOfst, mxTknOfst;
3210: int mnNtOfst, mxNtOfst;
3211: struct axset *ax;
3212: char *name;
3213:
3214: in = tplt_open(lemp);
3215: if( in==0 ) return;
3216: out = file_open(lemp,".c","w");
3217: if( out==0 ){
3218: fclose(in);
3219: return;
3220: }
3221: lineno = 1;
3222: tplt_xfer(lemp->name,in,out,&lineno);
3223:
3224: /* Generate the include code, if any */
3225: tplt_print(out,lemp,lemp->include,lemp->includeln,&lineno);
3226: if( mhflag ){
3227: name = file_makename(lemp, ".h");
3228: fprintf(out,"#include \"%s\"\n", name); lineno++;
3229: free(name);
3230: }
3231: tplt_xfer(lemp->name,in,out,&lineno);
3232:
3233: /* Generate #defines for all tokens */
3234: if( mhflag ){
3235: char *prefix;
3236: fprintf(out,"#if INTERFACE\n"); lineno++;
3237: if( lemp->tokenprefix ) prefix = lemp->tokenprefix;
3238: else prefix = "";
3239: for(i=1; i<lemp->nterminal; i++){
3240: fprintf(out,"#define %s%-30s %2d\n",prefix,lemp->symbols[i]->name,i);
3241: lineno++;
3242: }
3243: fprintf(out,"#endif\n"); lineno++;
3244: }
3245: tplt_xfer(lemp->name,in,out,&lineno);
3246:
3247: /* Generate the defines */
3248: fprintf(out,"/* \001 */\n");
3249: fprintf(out,"#define YYCODETYPE %s\n",
3250: minimum_size_type(0, lemp->nsymbol+5)); lineno++;
3251: fprintf(out,"#define YYNOCODE %d\n",lemp->nsymbol+1); lineno++;
3252: fprintf(out,"#define YYACTIONTYPE %s\n",
3253: minimum_size_type(0, lemp->nstate+lemp->nrule+5)); lineno++;
3254: print_stack_union(out,lemp,&lineno,mhflag);
3255: if( lemp->stacksize ){
3256: if( atoi(lemp->stacksize)<=0 ){
3257: ErrorMsg(lemp->filename,0,
3258: "Illegal stack size: [%s]. The stack size should be an integer constant.",
3259: lemp->stacksize);
3260: lemp->errorcnt++;
3261: lemp->stacksize = "100";
3262: }
3263: fprintf(out,"#define YYSTACKDEPTH %s\n",lemp->stacksize); lineno++;
3264: }else{
3265: fprintf(out,"#define YYSTACKDEPTH 100\n"); lineno++;
3266: }
3267: if( mhflag ){
3268: fprintf(out,"#if INTERFACE\n"); lineno++;
3269: }
3270: name = lemp->name ? lemp->name : "Parse";
3271: if( lemp->arg && lemp->arg[0] ){
3272: i = strlen(lemp->arg);
3273: while( i>=1 && isspace(lemp->arg[i-1]) ) i--;
3274: while( i>=1 && (isalnum(lemp->arg[i-1]) || lemp->arg[i-1]=='_') ) i--;
3275: fprintf(out,"#define %sARG_SDECL %s;\n",name,lemp->arg); lineno++;
3276: fprintf(out,"#define %sARG_PDECL ,%s\n",name,lemp->arg); lineno++;
3277: fprintf(out,"#define %sARG_FETCH %s = yypParser->%s\n",
3278: name,lemp->arg,&lemp->arg[i]); lineno++;
3279: fprintf(out,"#define %sARG_STORE yypParser->%s = %s\n",
3280: name,&lemp->arg[i],&lemp->arg[i]); lineno++;
3281: }else{
3282: fprintf(out,"#define %sARG_SDECL\n",name); lineno++;
3283: fprintf(out,"#define %sARG_PDECL\n",name); lineno++;
3284: fprintf(out,"#define %sARG_FETCH\n",name); lineno++;
3285: fprintf(out,"#define %sARG_STORE\n",name); lineno++;
3286: }
3287: if( mhflag ){
3288: fprintf(out,"#endif\n"); lineno++;
3289: }
3290: fprintf(out,"#define YYNSTATE %d\n",lemp->nstate); lineno++;
3291: fprintf(out,"#define YYNRULE %d\n",lemp->nrule); lineno++;
3292: fprintf(out,"#define YYERRORSYMBOL %d\n",lemp->errsym->index); lineno++;
3293: fprintf(out,"#define YYERRSYMDT yy%d\n",lemp->errsym->dtnum); lineno++;
3294: if( lemp->has_fallback ){
3295: fprintf(out,"#define YYFALLBACK 1\n"); lineno++;
3296: }
3297: tplt_xfer(lemp->name,in,out,&lineno);
3298:
3299: /* Generate the action table and its associates:
3300: **
3301: ** yy_action[] A single table containing all actions.
3302: ** yy_lookahead[] A table containing the lookahead for each entry in
3303: ** yy_action. Used to detect hash collisions.
3304: ** yy_shift_ofst[] For each state, the offset into yy_action for
3305: ** shifting terminals.
3306: ** yy_reduce_ofst[] For each state, the offset into yy_action for
3307: ** shifting non-terminals after a reduce.
3308: ** yy_default[] Default action for each state.
3309: */
3310:
3311: /* Compute the actions on all states and count them up */
3312: ax = malloc( sizeof(ax[0])*lemp->nstate*2 );
3313: if( ax==0 ){
3314: fprintf(stderr,"malloc failed\n");
3315: exit(1);
3316: }
3317: for(i=0; i<lemp->nstate; i++){
3318: stp = lemp->sorted[i];
3319: stp->nTknAct = stp->nNtAct = 0;
3320: stp->iDflt = lemp->nstate + lemp->nrule;
3321: stp->iTknOfst = NO_OFFSET;
3322: stp->iNtOfst = NO_OFFSET;
3323: for(ap=stp->ap; ap; ap=ap->next){
3324: if( compute_action(lemp,ap)>=0 ){
3325: if( ap->sp->index<lemp->nterminal ){
3326: stp->nTknAct++;
3327: }else if( ap->sp->index<lemp->nsymbol ){
3328: stp->nNtAct++;
3329: }else{
3330: stp->iDflt = compute_action(lemp, ap);
3331: }
3332: }
3333: }
3334: ax[i*2].stp = stp;
3335: ax[i*2].isTkn = 1;
3336: ax[i*2].nAction = stp->nTknAct;
3337: ax[i*2+1].stp = stp;
3338: ax[i*2+1].isTkn = 0;
3339: ax[i*2+1].nAction = stp->nNtAct;
3340: }
3341: mxTknOfst = mnTknOfst = 0;
3342: mxNtOfst = mnNtOfst = 0;
3343:
3344: /* Compute the action table. In order to try to keep the size of the
3345: ** action table to a minimum, the heuristic of placing the largest action
3346: ** sets first is used.
3347: */
3348: qsort(ax, lemp->nstate*2, sizeof(ax[0]), axset_compare);
3349: pActtab = acttab_alloc();
3350: for(i=0; i<lemp->nstate*2 && ax[i].nAction>0; i++){
3351: stp = ax[i].stp;
3352: if( ax[i].isTkn ){
3353: for(ap=stp->ap; ap; ap=ap->next){
3354: int action;
3355: if( ap->sp->index>=lemp->nterminal ) continue;
3356: action = compute_action(lemp, ap);
3357: if( action<0 ) continue;
3358: acttab_action(pActtab, ap->sp->index, action);
3359: }
3360: stp->iTknOfst = acttab_insert(pActtab);
3361: if( stp->iTknOfst<mnTknOfst ) mnTknOfst = stp->iTknOfst;
3362: if( stp->iTknOfst>mxTknOfst ) mxTknOfst = stp->iTknOfst;
3363: }else{
3364: for(ap=stp->ap; ap; ap=ap->next){
3365: int action;
3366: if( ap->sp->index<lemp->nterminal ) continue;
3367: if( ap->sp->index==lemp->nsymbol ) continue;
3368: action = compute_action(lemp, ap);
3369: if( action<0 ) continue;
3370: acttab_action(pActtab, ap->sp->index, action);
3371: }
3372: stp->iNtOfst = acttab_insert(pActtab);
3373: if( stp->iNtOfst<mnNtOfst ) mnNtOfst = stp->iNtOfst;
3374: if( stp->iNtOfst>mxNtOfst ) mxNtOfst = stp->iNtOfst;
3375: }
3376: }
3377: free(ax);
3378:
3379: /* Output the yy_action table */
3380: fprintf(out,"static YYACTIONTYPE yy_action[] = {\n"); lineno++;
3381: n = acttab_size(pActtab);
3382: for(i=j=0; i<n; i++){
3383: int action = acttab_yyaction(pActtab, i);
3384: if( action<0 ) action = lemp->nsymbol + lemp->nrule + 2;
3385: if( j==0 ) fprintf(out," /* %5d */ ", i);
3386: fprintf(out, " %4d,", action);
3387: if( j==9 || i==n-1 ){
3388: fprintf(out, "\n"); lineno++;
3389: j = 0;
3390: }else{
3391: j++;
3392: }
3393: }
3394: fprintf(out, "};\n"); lineno++;
3395:
3396: /* Output the yy_lookahead table */
3397: fprintf(out,"static YYCODETYPE yy_lookahead[] = {\n"); lineno++;
3398: for(i=j=0; i<n; i++){
3399: int la = acttab_yylookahead(pActtab, i);
3400: if( la<0 ) la = lemp->nsymbol;
3401: if( j==0 ) fprintf(out," /* %5d */ ", i);
3402: fprintf(out, " %4d,", la);
3403: if( j==9 || i==n-1 ){
3404: fprintf(out, "\n"); lineno++;
3405: j = 0;
3406: }else{
3407: j++;
3408: }
3409: }
3410: fprintf(out, "};\n"); lineno++;
3411:
3412: /* Output the yy_shift_ofst[] table */
3413: fprintf(out, "#define YY_SHIFT_USE_DFLT (%d)\n", mnTknOfst-1); lineno++;
3414: fprintf(out, "static %s yy_shift_ofst[] = {\n",
3415: minimum_size_type(mnTknOfst-1, mxTknOfst)); lineno++;
3416: n = lemp->nstate;
3417: for(i=j=0; i<n; i++){
3418: int ofst;
3419: stp = lemp->sorted[i];
3420: ofst = stp->iTknOfst;
3421: if( ofst==NO_OFFSET ) ofst = mnTknOfst - 1;
3422: if( j==0 ) fprintf(out," /* %5d */ ", i);
3423: fprintf(out, " %4d,", ofst);
3424: if( j==9 || i==n-1 ){
3425: fprintf(out, "\n"); lineno++;
3426: j = 0;
3427: }else{
3428: j++;
3429: }
3430: }
3431: fprintf(out, "};\n"); lineno++;
3432:
3433: /* Output the yy_reduce_ofst[] table */
3434: fprintf(out, "#define YY_REDUCE_USE_DFLT (%d)\n", mnNtOfst-1); lineno++;
3435: fprintf(out, "static %s yy_reduce_ofst[] = {\n",
3436: minimum_size_type(mnNtOfst-1, mxNtOfst)); lineno++;
3437: n = lemp->nstate;
3438: for(i=j=0; i<n; i++){
3439: int ofst;
3440: stp = lemp->sorted[i];
3441: ofst = stp->iNtOfst;
3442: if( ofst==NO_OFFSET ) ofst = mnNtOfst - 1;
3443: if( j==0 ) fprintf(out," /* %5d */ ", i);
3444: fprintf(out, " %4d,", ofst);
3445: if( j==9 || i==n-1 ){
3446: fprintf(out, "\n"); lineno++;
3447: j = 0;
3448: }else{
3449: j++;
3450: }
3451: }
3452: fprintf(out, "};\n"); lineno++;
3453:
3454: /* Output the default action table */
3455: fprintf(out, "static YYACTIONTYPE yy_default[] = {\n"); lineno++;
3456: n = lemp->nstate;
3457: for(i=j=0; i<n; i++){
3458: stp = lemp->sorted[i];
3459: if( j==0 ) fprintf(out," /* %5d */ ", i);
3460: fprintf(out, " %4d,", stp->iDflt);
3461: if( j==9 || i==n-1 ){
3462: fprintf(out, "\n"); lineno++;
3463: j = 0;
3464: }else{
3465: j++;
3466: }
3467: }
3468: fprintf(out, "};\n"); lineno++;
3469: tplt_xfer(lemp->name,in,out,&lineno);
3470:
3471: /* Generate the table of fallback tokens.
3472: */
3473: if( lemp->has_fallback ){
3474: for(i=0; i<lemp->nterminal; i++){
3475: struct symbol *p = lemp->symbols[i];
3476: if( p->fallback==0 ){
3477: fprintf(out, " 0, /* %10s => nothing */\n", p->name);
3478: }else{
3479: fprintf(out, " %3d, /* %10s => %s */\n", p->fallback->index,
3480: p->name, p->fallback->name);
3481: }
3482: lineno++;
3483: }
3484: }
3485: tplt_xfer(lemp->name, in, out, &lineno);
3486:
3487: /* Generate a table containing the symbolic name of every symbol
3488: */
3489: for(i=0; i<lemp->nsymbol; i++){
3490: sprintf(line,"\"%s\",",lemp->symbols[i]->name);
3491: fprintf(out," %-15s",line);
3492: if( (i&3)==3 ){ fprintf(out,"\n"); lineno++; }
3493: }
3494: if( (i&3)!=0 ){ fprintf(out,"\n"); lineno++; }
3495: tplt_xfer(lemp->name,in,out,&lineno);
3496:
3497: /* Generate a table containing a text string that describes every
3498: ** rule in the rule set of the grammer. This information is used
3499: ** when tracing REDUCE actions.
3500: */
3501: for(i=0, rp=lemp->rule; rp; rp=rp->next, i++){
3502: assert( rp->index==i );
3503: fprintf(out," /* %3d */ \"%s ::=", i, rp->lhs->name);
3504: for(j=0; j<rp->nrhs; j++) fprintf(out," %s",rp->rhs[j]->name);
3505: fprintf(out,"\",\n"); lineno++;
3506: }
3507: tplt_xfer(lemp->name,in,out,&lineno);
3508:
3509: /* Generate code which executes every time a symbol is popped from
3510: ** the stack while processing errors or while destroying the parser.
3511: ** (In other words, generate the %destructor actions)
3512: */
3513: if( lemp->tokendest ){
3514: for(i=0; i<lemp->nsymbol; i++){
3515: struct symbol *sp = lemp->symbols[i];
3516: if( sp==0 || sp->type!=TERMINAL ) continue;
3517: fprintf(out," case %d:\n",sp->index); lineno++;
3518: }
3519: for(i=0; i<lemp->nsymbol && lemp->symbols[i]->type!=TERMINAL; i++);
3520: if( i<lemp->nsymbol ){
3521: emit_destructor_code(out,lemp->symbols[i],lemp,&lineno);
3522: fprintf(out," break;\n"); lineno++;
3523: }
3524: }
3525: for(i=0; i<lemp->nsymbol; i++){
3526: struct symbol *sp = lemp->symbols[i];
3527: if( sp==0 || sp->type==TERMINAL || sp->destructor==0 ) continue;
3528: fprintf(out," case %d:\n",sp->index); lineno++;
3529: emit_destructor_code(out,lemp->symbols[i],lemp,&lineno);
3530: fprintf(out," break;\n"); lineno++;
3531: }
3532: if( lemp->vardest ){
3533: struct symbol *dflt_sp = 0;
3534: for(i=0; i<lemp->nsymbol; i++){
3535: struct symbol *sp = lemp->symbols[i];
3536: if( sp==0 || sp->type==TERMINAL ||
3537: sp->index<=0 || sp->destructor!=0 ) continue;
3538: fprintf(out," case %d:\n",sp->index); lineno++;
3539: dflt_sp = sp;
3540: }
3541: if( dflt_sp!=0 ){
3542: emit_destructor_code(out,dflt_sp,lemp,&lineno);
3543: fprintf(out," break;\n"); lineno++;
3544: }
3545: }
3546: tplt_xfer(lemp->name,in,out,&lineno);
3547:
3548: /* Generate code which executes whenever the parser stack overflows */
3549: tplt_print(out,lemp,lemp->overflow,lemp->overflowln,&lineno);
3550: tplt_xfer(lemp->name,in,out,&lineno);
3551:
3552: /* Generate the table of rule information
3553: **
3554: ** Note: This code depends on the fact that rules are number
3555: ** sequentually beginning with 0.
3556: */
3557: for(rp=lemp->rule; rp; rp=rp->next){
3558: fprintf(out," { %d, %d },\n",rp->lhs->index,rp->nrhs); lineno++;
3559: }
3560: tplt_xfer(lemp->name,in,out,&lineno);
3561:
3562: /* Generate code which execution during each REDUCE action */
3563: for(rp=lemp->rule; rp; rp=rp->next){
3564: fprintf(out," case %d:\n",rp->index); lineno++;
3565: emit_code(out,rp,lemp,&lineno);
3566: fprintf(out," break;\n"); lineno++;
3567: }
3568: tplt_xfer(lemp->name,in,out,&lineno);
3569:
3570: /* Generate code which executes if a parse fails */
3571: tplt_print(out,lemp,lemp->failure,lemp->failureln,&lineno);
3572: tplt_xfer(lemp->name,in,out,&lineno);
3573:
3574: /* Generate code which executes when a syntax error occurs */
3575: tplt_print(out,lemp,lemp->error,lemp->errorln,&lineno);
3576: tplt_xfer(lemp->name,in,out,&lineno);
3577:
3578: /* Generate code which executes when the parser accepts its input */
3579: tplt_print(out,lemp,lemp->accept,lemp->acceptln,&lineno);
3580: tplt_xfer(lemp->name,in,out,&lineno);
3581:
3582: /* Append any addition code the user desires */
3583: tplt_print(out,lemp,lemp->extracode,lemp->extracodeln,&lineno);
3584:
3585: fclose(in);
3586: fclose(out);
3587: return;
3588: }
3589:
3590: /* Generate a header file for the parser */
3591: void ReportHeader(lemp)
3592: struct lemon *lemp;
3593: {
3594: FILE *out, *in;
3595: char *prefix;
3596: char line[LINESIZE];
3597: char pattern[LINESIZE];
3598: int i;
3599:
3600: if( lemp->tokenprefix ) prefix = lemp->tokenprefix;
3601: else prefix = "";
3602: in = file_open(lemp,".h","r");
3603: if( in ){
3604: for(i=1; i<lemp->nterminal && fgets(line,LINESIZE,in); i++){
3605: sprintf(pattern,"#define %s%-30s %2d\n",prefix,lemp->symbols[i]->name,i);
3606: if( strcmp(line,pattern) ) break;
3607: }
3608: fclose(in);
3609: if( i==lemp->nterminal ){
3610: /* No change in the file. Don't rewrite it. */
3611: return;
3612: }
3613: }
3614: out = file_open(lemp,".h","w");
3615: if( out ){
3616: for(i=1; i<lemp->nterminal; i++){
3617: fprintf(out,"#define %s%-30s %2d\n",prefix,lemp->symbols[i]->name,i);
3618: }
3619: fclose(out);
3620: }
3621: return;
3622: }
3623:
3624: /* Reduce the size of the action tables, if possible, by making use
3625: ** of defaults.
3626: **
3627: ** In this version, we take the most frequent REDUCE action and make
3628: ** it the default. Only default a reduce if there are more than one.
3629: */
3630: void CompressTables(lemp)
3631: struct lemon *lemp;
3632: {
3633: struct state *stp;
3634: struct action *ap, *ap2;
3635: struct rule *rp, *rp2, *rbest;
3636: int nbest, n;
3637: int i;
3638:
3639: for(i=0; i<lemp->nstate; i++){
3640: stp = lemp->sorted[i];
3641: nbest = 0;
3642: rbest = 0;
3643:
3644: for(ap=stp->ap; ap; ap=ap->next){
3645: if( ap->type!=REDUCE ) continue;
3646: rp = ap->x.rp;
3647: if( rp==rbest ) continue;
3648: n = 1;
3649: for(ap2=ap->next; ap2; ap2=ap2->next){
3650: if( ap2->type!=REDUCE ) continue;
3651: rp2 = ap2->x.rp;
3652: if( rp2==rbest ) continue;
3653: if( rp2==rp ) n++;
3654: }
3655: if( n>nbest ){
3656: nbest = n;
3657: rbest = rp;
3658: }
3659: }
3660:
3661: /* Do not make a default if the number of rules to default
3662: ** is not at least 2 */
3663: if( nbest<2 ) continue;
3664:
3665:
3666: /* Combine matching REDUCE actions into a single default */
3667: for(ap=stp->ap; ap; ap=ap->next){
3668: if( ap->type==REDUCE && ap->x.rp==rbest ) break;
3669: }
3670: assert( ap );
3671: ap->sp = Symbol_new("{default}");
3672: for(ap=ap->next; ap; ap=ap->next){
3673: if( ap->type==REDUCE && ap->x.rp==rbest ) ap->type = NOT_USED;
3674: }
3675: stp->ap = Action_sort(stp->ap);
3676: }
3677: }
3678:
3679: /***************** From the file "set.c" ************************************/
3680: /*
3681: ** Set manipulation routines for the LEMON parser generator.
3682: */
3683:
3684: static int global_size = 0;
3685:
3686: /* Set the set size */
3687: void SetSize(n)
3688: int n;
3689: {
3690: global_size = n+1;
3691: }
3692:
3693: /* Allocate a new set */
3694: char *SetNew(){
3695: char *s;
3696: int i;
3697: s = (char*)malloc( global_size );
3698: if( s==0 ){
3699: memory_error();
3700: }
3701: for(i=0; i<global_size; i++) s[i] = 0;
3702: return s;
3703: }
3704:
3705: /* Deallocate a set */
3706: void SetFree(s)
3707: char *s;
3708: {
3709: free(s);
3710: }
3711:
3712: /* Add a new element to the set. Return TRUE if the element was added
3713: ** and FALSE if it was already there. */
3714: int SetAdd(s,e)
3715: char *s;
3716: int e;
3717: {
3718: int rv;
3719: rv = s[e];
3720: s[e] = 1;
3721: return !rv;
3722: }
3723:
3724: /* Add every element of s2 to s1. Return TRUE if s1 changes. */
3725: int SetUnion(s1,s2)
3726: char *s1;
3727: char *s2;
3728: {
3729: int i, progress;
3730: progress = 0;
3731: for(i=0; i<global_size; i++){
3732: if( s2[i]==0 ) continue;
3733: if( s1[i]==0 ){
3734: progress = 1;
3735: s1[i] = 1;
3736: }
3737: }
3738: return progress;
3739: }
3740: /********************** From the file "table.c" ****************************/
3741: /*
3742: ** All code in this file has been automatically generated
3743: ** from a specification in the file
3744: ** "table.q"
3745: ** by the associative array code building program "aagen".
3746: ** Do not edit this file! Instead, edit the specification
3747: ** file, then rerun aagen.
3748: */
3749: /*
3750: ** Code for processing tables in the LEMON parser generator.
3751: */
3752:
3753: PRIVATE int strhash(x)
3754: char *x;
3755: {
3756: unsigned int h = 0;
3757: while( *x) h = h*13u + (unsigned int) *(x++);
3758: return h;
3759: }
3760:
3761: /* Works like strdup, sort of. Save a string in malloced memory, but
3762: ** keep strings in a table so that the same string is not in more
3763: ** than one place.
3764: */
3765: char *Strsafe(y)
3766: char *y;
3767: {
3768: char *z;
3769:
3770: z = Strsafe_find(y);
3771: if( z==0 && (z=malloc( strlen(y)+1 ))!=0 ){
3772: strcpy(z,y);
3773: Strsafe_insert(z);
3774: }
3775: MemoryCheck(z);
3776: return z;
3777: }
3778:
3779: /* There is one instance of the following structure for each
3780: ** associative array of type "x1".
3781: */
3782: struct s_x1 {
3783: int size; /* The number of available slots. */
3784: /* Must be a power of 2 greater than or */
3785: /* equal to 1 */
3786: int count; /* Number of currently slots filled */
3787: struct s_x1node *tbl; /* The data stored here */
3788: struct s_x1node **ht; /* Hash table for lookups */
3789: };
3790:
3791: /* There is one instance of this structure for every data element
3792: ** in an associative array of type "x1".
3793: */
3794: typedef struct s_x1node {
3795: char *data; /* The data */
3796: struct s_x1node *next; /* Next entry with the same hash */
3797: struct s_x1node **from; /* Previous link */
3798: } x1node;
3799:
3800: /* There is only one instance of the array, which is the following */
3801: static struct s_x1 *x1a;
3802:
3803: /* Allocate a new associative array */
3804: void Strsafe_init(){
3805: if( x1a ) return;
3806: x1a = (struct s_x1*)malloc( sizeof(struct s_x1) );
3807: if( x1a ){
3808: x1a->size = 1024;
3809: x1a->count = 0;
3810: x1a->tbl = (x1node*)malloc(
3811: (sizeof(x1node) + sizeof(x1node*))*1024 );
3812: if( x1a->tbl==0 ){
3813: free(x1a);
3814: x1a = 0;
3815: }else{
3816: int i;
3817: x1a->ht = (x1node**)&(x1a->tbl[1024]);
3818: for(i=0; i<1024; i++) x1a->ht[i] = 0;
3819: }
3820: }
3821: }
3822: /* Insert a new record into the array. Return TRUE if successful.
3823: ** Prior data with the same key is NOT overwritten */
3824: int Strsafe_insert(data)
3825: char *data;
3826: {
3827: x1node *np;
3828: int h;
3829: int ph;
3830:
3831: if( x1a==0 ) return 0;
3832: ph = strhash(data);
3833: h = ph & (x1a->size-1);
3834: np = x1a->ht[h];
3835: while( np ){
3836: if( strcmp(np->data,data)==0 ){
3837: /* An existing entry with the same key is found. */
3838: /* Fail because overwrite is not allows. */
3839: return 0;
3840: }
3841: np = np->next;
3842: }
3843: if( x1a->count>=x1a->size ){
3844: /* Need to make the hash table bigger */
3845: int i,size;
3846: struct s_x1 array;
3847: array.size = size = x1a->size*2;
3848: array.count = x1a->count;
3849: array.tbl = (x1node*)malloc(
3850: (sizeof(x1node) + sizeof(x1node*))*size );
3851: if( array.tbl==0 ) return 0; /* Fail due to malloc failure */
3852: array.ht = (x1node**)&(array.tbl[size]);
3853: for(i=0; i<size; i++) array.ht[i] = 0;
3854: for(i=0; i<x1a->count; i++){
3855: x1node *oldnp, *newnp;
3856: oldnp = &(x1a->tbl[i]);
3857: h = strhash(oldnp->data) & (size-1);
3858: newnp = &(array.tbl[i]);
3859: if( array.ht[h] ) array.ht[h]->from = &(newnp->next);
3860: newnp->next = array.ht[h];
3861: newnp->data = oldnp->data;
3862: newnp->from = &(array.ht[h]);
3863: array.ht[h] = newnp;
3864: }
3865: free(x1a->tbl);
3866: *x1a = array;
3867: }
3868: /* Insert the new data */
3869: h = ph & (x1a->size-1);
3870: np = &(x1a->tbl[x1a->count++]);
3871: np->data = data;
3872: if( x1a->ht[h] ) x1a->ht[h]->from = &(np->next);
3873: np->next = x1a->ht[h];
3874: x1a->ht[h] = np;
3875: np->from = &(x1a->ht[h]);
3876: return 1;
3877: }
3878:
3879: /* Return a pointer to data assigned to the given key. Return NULL
3880: ** if no such key. */
3881: char *Strsafe_find(key)
3882: char *key;
3883: {
3884: int h;
3885: x1node *np;
3886:
3887: if( x1a==0 ) return 0;
3888: h = strhash(key) & (x1a->size-1);
3889: np = x1a->ht[h];
3890: while( np ){
3891: if( strcmp(np->data,key)==0 ) break;
3892: np = np->next;
3893: }
3894: return np ? np->data : 0;
3895: }
3896:
3897: /* Return a pointer to the (terminal or nonterminal) symbol "x".
3898: ** Create a new symbol if this is the first time "x" has been seen.
3899: */
3900: struct symbol *Symbol_new(x)
3901: char *x;
3902: {
3903: struct symbol *sp;
3904:
3905: sp = Symbol_find(x);
3906: if( sp==0 ){
3907: sp = (struct symbol *)malloc( sizeof(struct symbol) );
3908: MemoryCheck(sp);
3909: sp->name = Strsafe(x);
3910: sp->type = isupper(*x) ? TERMINAL : NONTERMINAL;
3911: sp->rule = 0;
3912: sp->fallback = 0;
3913: sp->prec = -1;
3914: sp->assoc = UNK;
3915: sp->firstset = 0;
3916: sp->lambda = Bo_FALSE;
3917: sp->destructor = 0;
3918: sp->datatype = 0;
3919: Symbol_insert(sp,sp->name);
3920: }
3921: return sp;
3922: }
3923:
3924: /* Compare two symbols for working purposes
3925: **
3926: ** Symbols that begin with upper case letters (terminals or tokens)
3927: ** must sort before symbols that begin with lower case letters
3928: ** (non-terminals). Other than that, the order does not matter.
3929: **
3930: ** We find experimentally that leaving the symbols in their original
3931: ** order (the order they appeared in the grammar file) gives the
3932: ** smallest parser tables in SQLite.
3933: */
3934: int Symbolcmpp(struct symbol **a, struct symbol **b){
3935: int i1 = (**a).index + 10000000*((**a).name[0]>'Z');
3936: int i2 = (**b).index + 10000000*((**b).name[0]>'Z');
3937: return i1-i2;
3938: }
3939:
3940: /* There is one instance of the following structure for each
3941: ** associative array of type "x2".
3942: */
3943: struct s_x2 {
3944: int size; /* The number of available slots. */
3945: /* Must be a power of 2 greater than or */
3946: /* equal to 1 */
3947: int count; /* Number of currently slots filled */
3948: struct s_x2node *tbl; /* The data stored here */
3949: struct s_x2node **ht; /* Hash table for lookups */
3950: };
3951:
3952: /* There is one instance of this structure for every data element
3953: ** in an associative array of type "x2".
3954: */
3955: typedef struct s_x2node {
3956: struct symbol *data; /* The data */
3957: char *key; /* The key */
3958: struct s_x2node *next; /* Next entry with the same hash */
3959: struct s_x2node **from; /* Previous link */
3960: } x2node;
3961:
3962: /* There is only one instance of the array, which is the following */
3963: static struct s_x2 *x2a;
3964:
3965: /* Allocate a new associative array */
3966: void Symbol_init(){
3967: if( x2a ) return;
3968: x2a = (struct s_x2*)malloc( sizeof(struct s_x2) );
3969: if( x2a ){
3970: x2a->size = 128;
3971: x2a->count = 0;
3972: x2a->tbl = (x2node*)malloc(
3973: (sizeof(x2node) + sizeof(x2node*))*128 );
3974: if( x2a->tbl==0 ){
3975: free(x2a);
3976: x2a = 0;
3977: }else{
3978: int i;
3979: x2a->ht = (x2node**)&(x2a->tbl[128]);
3980: for(i=0; i<128; i++) x2a->ht[i] = 0;
3981: }
3982: }
3983: }
3984: /* Insert a new record into the array. Return TRUE if successful.
3985: ** Prior data with the same key is NOT overwritten */
3986: int Symbol_insert(data,key)
3987: struct symbol *data;
3988: char *key;
3989: {
3990: x2node *np;
3991: int h;
3992: int ph;
3993:
3994: if( x2a==0 ) return 0;
3995: ph = strhash(key);
3996: h = ph & (x2a->size-1);
3997: np = x2a->ht[h];
3998: while( np ){
3999: if( strcmp(np->key,key)==0 ){
4000: /* An existing entry with the same key is found. */
4001: /* Fail because overwrite is not allows. */
4002: return 0;
4003: }
4004: np = np->next;
4005: }
4006: if( x2a->count>=x2a->size ){
4007: /* Need to make the hash table bigger */
4008: int i,size;
4009: struct s_x2 array;
4010: array.size = size = x2a->size*2;
4011: array.count = x2a->count;
4012: array.tbl = (x2node*)malloc(
4013: (sizeof(x2node) + sizeof(x2node*))*size );
4014: if( array.tbl==0 ) return 0; /* Fail due to malloc failure */
4015: array.ht = (x2node**)&(array.tbl[size]);
4016: for(i=0; i<size; i++) array.ht[i] = 0;
4017: for(i=0; i<x2a->count; i++){
4018: x2node *oldnp, *newnp;
4019: oldnp = &(x2a->tbl[i]);
4020: h = strhash(oldnp->key) & (size-1);
4021: newnp = &(array.tbl[i]);
4022: if( array.ht[h] ) array.ht[h]->from = &(newnp->next);
4023: newnp->next = array.ht[h];
4024: newnp->key = oldnp->key;
4025: newnp->data = oldnp->data;
4026: newnp->from = &(array.ht[h]);
4027: array.ht[h] = newnp;
4028: }
4029: free(x2a->tbl);
4030: *x2a = array;
4031: }
4032: /* Insert the new data */
4033: h = ph & (x2a->size-1);
4034: np = &(x2a->tbl[x2a->count++]);
4035: np->key = key;
4036: np->data = data;
4037: if( x2a->ht[h] ) x2a->ht[h]->from = &(np->next);
4038: np->next = x2a->ht[h];
4039: x2a->ht[h] = np;
4040: np->from = &(x2a->ht[h]);
4041: return 1;
4042: }
4043:
4044: /* Return a pointer to data assigned to the given key. Return NULL
4045: ** if no such key. */
4046: struct symbol *Symbol_find(key)
4047: char *key;
4048: {
4049: int h;
4050: x2node *np;
4051:
4052: if( x2a==0 ) return 0;
4053: h = strhash(key) & (x2a->size-1);
4054: np = x2a->ht[h];
4055: while( np ){
4056: if( strcmp(np->key,key)==0 ) break;
4057: np = np->next;
4058: }
4059: return np ? np->data : 0;
4060: }
4061:
4062: /* Return the n-th data. Return NULL if n is out of range. */
4063: struct symbol *Symbol_Nth(n)
4064: int n;
4065: {
4066: struct symbol *data;
4067: if( x2a && n>0 && n<=x2a->count ){
4068: data = x2a->tbl[n-1].data;
4069: }else{
4070: data = 0;
4071: }
4072: return data;
4073: }
4074:
4075: /* Return the size of the array */
4076: int Symbol_count()
4077: {
4078: return x2a ? x2a->count : 0;
4079: }
4080:
4081: /* Return an array of pointers to all data in the table.
4082: ** The array is obtained from malloc. Return NULL if memory allocation
4083: ** problems, or if the array is empty. */
4084: struct symbol **Symbol_arrayof()
4085: {
4086: struct symbol **array;
4087: int i,size;
4088: if( x2a==0 ) return 0;
4089: size = x2a->count;
4090: array = (struct symbol **)malloc( sizeof(struct symbol *)*size );
4091: if( array ){
4092: for(i=0; i<size; i++) array[i] = x2a->tbl[i].data;
4093: }
4094: return array;
4095: }
4096:
4097: /* Compare two configurations */
4098: int Configcmp(a,b)
4099: struct config *a;
4100: struct config *b;
4101: {
4102: int x;
4103: x = a->rp->index - b->rp->index;
4104: if( x==0 ) x = a->dot - b->dot;
4105: return x;
4106: }
4107:
4108: /* Compare two states */
4109: PRIVATE int statecmp(a,b)
4110: struct config *a;
4111: struct config *b;
4112: {
4113: int rc;
4114: for(rc=0; rc==0 && a && b; a=a->bp, b=b->bp){
4115: rc = a->rp->index - b->rp->index;
4116: if( rc==0 ) rc = a->dot - b->dot;
4117: }
4118: if( rc==0 ){
4119: if( a ) rc = 1;
4120: if( b ) rc = -1;
4121: }
4122: return rc;
4123: }
4124:
4125: /* Hash a state */
4126: PRIVATE int statehash(a)
4127: struct config *a;
4128: {
4129: unsigned int h=0;
4130: while( a ){
4131: h = h*571u + (unsigned int)a->rp->index*37u + (unsigned int)a->dot;
4132: a = a->bp;
4133: }
4134: return h;
4135: }
4136:
4137: /* Allocate a new state structure */
4138: struct state *State_new()
4139: {
4140: struct state *new;
4141: new = (struct state *)malloc( sizeof(struct state) );
4142: MemoryCheck(new);
4143: return new;
4144: }
4145:
4146: /* There is one instance of the following structure for each
4147: ** associative array of type "x3".
4148: */
4149: struct s_x3 {
4150: int size; /* The number of available slots. */
4151: /* Must be a power of 2 greater than or */
4152: /* equal to 1 */
4153: int count; /* Number of currently slots filled */
4154: struct s_x3node *tbl; /* The data stored here */
4155: struct s_x3node **ht; /* Hash table for lookups */
4156: };
4157:
4158: /* There is one instance of this structure for every data element
4159: ** in an associative array of type "x3".
4160: */
4161: typedef struct s_x3node {
4162: struct state *data; /* The data */
4163: struct config *key; /* The key */
4164: struct s_x3node *next; /* Next entry with the same hash */
4165: struct s_x3node **from; /* Previous link */
4166: } x3node;
4167:
4168: /* There is only one instance of the array, which is the following */
4169: static struct s_x3 *x3a;
4170:
4171: /* Allocate a new associative array */
4172: void State_init(){
4173: if( x3a ) return;
4174: x3a = (struct s_x3*)malloc( sizeof(struct s_x3) );
4175: if( x3a ){
4176: x3a->size = 128;
4177: x3a->count = 0;
4178: x3a->tbl = (x3node*)malloc(
4179: (sizeof(x3node) + sizeof(x3node*))*128 );
4180: if( x3a->tbl==0 ){
4181: free(x3a);
4182: x3a = 0;
4183: }else{
4184: int i;
4185: x3a->ht = (x3node**)&(x3a->tbl[128]);
4186: for(i=0; i<128; i++) x3a->ht[i] = 0;
4187: }
4188: }
4189: }
4190: /* Insert a new record into the array. Return TRUE if successful.
4191: ** Prior data with the same key is NOT overwritten */
4192: int State_insert(data,key)
4193: struct state *data;
4194: struct config *key;
4195: {
4196: x3node *np;
4197: int h;
4198: int ph;
4199:
4200: if( x3a==0 ) return 0;
4201: ph = statehash(key);
4202: h = ph & (x3a->size-1);
4203: np = x3a->ht[h];
4204: while( np ){
4205: if( statecmp(np->key,key)==0 ){
4206: /* An existing entry with the same key is found. */
4207: /* Fail because overwrite is not allows. */
4208: return 0;
4209: }
4210: np = np->next;
4211: }
4212: if( x3a->count>=x3a->size ){
4213: /* Need to make the hash table bigger */
4214: int i,size;
4215: struct s_x3 array;
4216: array.size = size = x3a->size*2;
4217: array.count = x3a->count;
4218: array.tbl = (x3node*)malloc(
4219: (sizeof(x3node) + sizeof(x3node*))*size );
4220: if( array.tbl==0 ) return 0; /* Fail due to malloc failure */
4221: array.ht = (x3node**)&(array.tbl[size]);
4222: for(i=0; i<size; i++) array.ht[i] = 0;
4223: for(i=0; i<x3a->count; i++){
4224: x3node *oldnp, *newnp;
4225: oldnp = &(x3a->tbl[i]);
4226: h = statehash(oldnp->key) & (size-1);
4227: newnp = &(array.tbl[i]);
4228: if( array.ht[h] ) array.ht[h]->from = &(newnp->next);
4229: newnp->next = array.ht[h];
4230: newnp->key = oldnp->key;
4231: newnp->data = oldnp->data;
4232: newnp->from = &(array.ht[h]);
4233: array.ht[h] = newnp;
4234: }
4235: free(x3a->tbl);
4236: *x3a = array;
4237: }
4238: /* Insert the new data */
4239: h = ph & (x3a->size-1);
4240: np = &(x3a->tbl[x3a->count++]);
4241: np->key = key;
4242: np->data = data;
4243: if( x3a->ht[h] ) x3a->ht[h]->from = &(np->next);
4244: np->next = x3a->ht[h];
4245: x3a->ht[h] = np;
4246: np->from = &(x3a->ht[h]);
4247: return 1;
4248: }
4249:
4250: /* Return a pointer to data assigned to the given key. Return NULL
4251: ** if no such key. */
4252: struct state *State_find(key)
4253: struct config *key;
4254: {
4255: int h;
4256: x3node *np;
4257:
4258: if( x3a==0 ) return 0;
4259: h = statehash(key) & (x3a->size-1);
4260: np = x3a->ht[h];
4261: while( np ){
4262: if( statecmp(np->key,key)==0 ) break;
4263: np = np->next;
4264: }
4265: return np ? np->data : 0;
4266: }
4267:
4268: /* Return an array of pointers to all data in the table.
4269: ** The array is obtained from malloc. Return NULL if memory allocation
4270: ** problems, or if the array is empty. */
4271: struct state **State_arrayof()
4272: {
4273: struct state **array;
4274: int i,size;
4275: if( x3a==0 ) return 0;
4276: size = x3a->count;
4277: array = (struct state **)malloc( sizeof(struct state *)*size );
4278: if( array ){
4279: for(i=0; i<size; i++) array[i] = x3a->tbl[i].data;
4280: }
4281: return array;
4282: }
4283:
4284: /* Hash a configuration */
4285: PRIVATE int confighash(a)
4286: struct config *a;
4287: {
4288: int h=0;
4289: h = h*571 + a->rp->index*37 + a->dot;
4290: return h;
4291: }
4292:
4293: /* There is one instance of the following structure for each
4294: ** associative array of type "x4".
4295: */
4296: struct s_x4 {
4297: int size; /* The number of available slots. */
4298: /* Must be a power of 2 greater than or */
4299: /* equal to 1 */
4300: int count; /* Number of currently slots filled */
4301: struct s_x4node *tbl; /* The data stored here */
4302: struct s_x4node **ht; /* Hash table for lookups */
4303: };
4304:
4305: /* There is one instance of this structure for every data element
4306: ** in an associative array of type "x4".
4307: */
4308: typedef struct s_x4node {
4309: struct config *data; /* The data */
4310: struct s_x4node *next; /* Next entry with the same hash */
4311: struct s_x4node **from; /* Previous link */
4312: } x4node;
4313:
4314: /* There is only one instance of the array, which is the following */
4315: static struct s_x4 *x4a;
4316:
4317: /* Allocate a new associative array */
4318: void Configtable_init(){
4319: if( x4a ) return;
4320: x4a = (struct s_x4*)malloc( sizeof(struct s_x4) );
4321: if( x4a ){
4322: x4a->size = 64;
4323: x4a->count = 0;
4324: x4a->tbl = (x4node*)malloc(
4325: (sizeof(x4node) + sizeof(x4node*))*64 );
4326: if( x4a->tbl==0 ){
4327: free(x4a);
4328: x4a = 0;
4329: }else{
4330: int i;
4331: x4a->ht = (x4node**)&(x4a->tbl[64]);
4332: for(i=0; i<64; i++) x4a->ht[i] = 0;
4333: }
4334: }
4335: }
4336: /* Insert a new record into the array. Return TRUE if successful.
4337: ** Prior data with the same key is NOT overwritten */
4338: int Configtable_insert(data)
4339: struct config *data;
4340: {
4341: x4node *np;
4342: int h;
4343: int ph;
4344:
4345: if( x4a==0 ) return 0;
4346: ph = confighash(data);
4347: h = ph & (x4a->size-1);
4348: np = x4a->ht[h];
4349: while( np ){
4350: if( Configcmp(np->data,data)==0 ){
4351: /* An existing entry with the same key is found. */
4352: /* Fail because overwrite is not allows. */
4353: return 0;
4354: }
4355: np = np->next;
4356: }
4357: if( x4a->count>=x4a->size ){
4358: /* Need to make the hash table bigger */
4359: int i,size;
4360: struct s_x4 array;
4361: array.size = size = x4a->size*2;
4362: array.count = x4a->count;
4363: array.tbl = (x4node*)malloc(
4364: (sizeof(x4node) + sizeof(x4node*))*size );
4365: if( array.tbl==0 ) return 0; /* Fail due to malloc failure */
4366: array.ht = (x4node**)&(array.tbl[size]);
4367: for(i=0; i<size; i++) array.ht[i] = 0;
4368: for(i=0; i<x4a->count; i++){
4369: x4node *oldnp, *newnp;
4370: oldnp = &(x4a->tbl[i]);
4371: h = confighash(oldnp->data) & (size-1);
4372: newnp = &(array.tbl[i]);
4373: if( array.ht[h] ) array.ht[h]->from = &(newnp->next);
4374: newnp->next = array.ht[h];
4375: newnp->data = oldnp->data;
4376: newnp->from = &(array.ht[h]);
4377: array.ht[h] = newnp;
4378: }
4379: free(x4a->tbl);
4380: *x4a = array;
4381: }
4382: /* Insert the new data */
4383: h = ph & (x4a->size-1);
4384: np = &(x4a->tbl[x4a->count++]);
4385: np->data = data;
4386: if( x4a->ht[h] ) x4a->ht[h]->from = &(np->next);
4387: np->next = x4a->ht[h];
4388: x4a->ht[h] = np;
4389: np->from = &(x4a->ht[h]);
4390: return 1;
4391: }
4392:
4393: /* Return a pointer to data assigned to the given key. Return NULL
4394: ** if no such key. */
4395: struct config *Configtable_find(key)
4396: struct config *key;
4397: {
4398: int h;
4399: x4node *np;
4400:
4401: if( x4a==0 ) return 0;
4402: h = confighash(key) & (x4a->size-1);
4403: np = x4a->ht[h];
4404: while( np ){
4405: if( Configcmp(np->data,key)==0 ) break;
4406: np = np->next;
4407: }
4408: return np ? np->data : 0;
4409: }
4410:
4411: /* Remove all data from the table. Pass each data to the function "f"
4412: ** as it is removed. ("f" may be null to avoid this step.) */
4413: void Configtable_clear(f)
4414: int(*f)(/* struct config * */);
4415: {
4416: int i;
4417: if( x4a==0 || x4a->count==0 ) return;
4418: if( f ) for(i=0; i<x4a->count; i++) (*f)(x4a->tbl[i].data);
4419: for(i=0; i<x4a->size; i++) x4a->ht[i] = 0;
4420: x4a->count = 0;
4421: return;
4422: }
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>
![[BACK]](/icons/cvsweb/back.gif){kind=icon}
Return to lemon.c CVS log ![[TXT]](/icons/cvsweb/text.gif){kind=icon}
![[DIR]](/icons/cvsweb/dir.gif){kind=icon}
Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / lighttpd / src