Sync changes from latest snowball compiler version
[xapian.git] / xapian-core / languages / compiler / analyser.c
blobc7fd3f83eab1f4ecd41bc9ed7fa41199523bcca8
2 #include <stdio.h> /* printf etc */
3 #include <stdlib.h> /* exit */
4 #include <string.h> /* memmove */
5 #include "header.h"
7 typedef enum {
8 e_token_omitted = 0,
9 e_unexpected_token = 1,
10 e_string_omitted = 2,
11 e_unexpected_token_in_among = 3,
12 /* For codes above here, report "after " t->previous_token after the error. */
13 e_unresolved_substring = 14,
14 e_not_allowed_inside_reverse = 15,
15 e_empty_grouping = 16,
16 e_already_backwards = 17,
17 e_empty_among = 18,
18 e_adjacent_bracketed_in_among = 19,
19 e_substring_preceded_by_substring = 20,
20 /* For codes below here, tokeniser->b is printed before the error. */
21 e_redeclared = 30,
22 e_undeclared = 31,
23 e_declared_as_different_mode = 32,
24 e_not_of_type_x = 33,
25 e_not_of_type_string_or_integer = 34,
26 e_misplaced = 35,
27 e_redefined = 36,
28 e_misused = 37
29 } error_code;
31 /* recursive usage: */
33 static void read_program_(struct analyser * a, int terminator);
34 static struct node * read_C(struct analyser * a);
35 static struct node * C_style(struct analyser * a, const char * s, int token);
38 static void print_node_(struct node * p, int n, const char * s) {
40 int i;
41 for (i = 0; i < n; i++) fputs(i == n - 1 ? s : " ", stdout);
42 printf("%s ", name_of_token(p->type));
43 if (p->name) report_b(stdout, p->name->b);
44 if (p->literalstring) {
45 printf("'");
46 report_b(stdout, p->literalstring);
47 printf("'");
49 printf("\n");
50 if (p->AE) print_node_(p->AE, n+1, "# ");
51 if (p->left) print_node_(p->left, n+1, " ");
52 if (p->right) print_node_(p->right, n, " ");
53 if (p->aux) print_node_(p->aux, n+1, "@ ");
56 extern void print_program(struct analyser * a) {
57 print_node_(a->program, 0, " ");
60 static struct node * new_node(struct analyser * a, int type) {
61 NEW(node, p);
62 p->next = a->nodes; a->nodes = p;
63 p->left = 0;
64 p->right = 0;
65 p->aux = 0;
66 p->AE = 0;
67 p->name = 0;
68 p->literalstring = 0;
69 p->mode = a->mode;
70 p->line_number = a->tokeniser->line_number;
71 p->type = type;
72 return p;
75 static const char * name_of_mode(int n) {
76 switch (n) {
77 case m_backward: return "string backward";
78 case m_forward: return "string forward";
79 /* case m_integer: return "integer"; */
81 fprintf(stderr, "Invalid mode %d in name_of_mode()\n", n);
82 exit(1);
85 static const char * name_of_type(int n) {
86 switch (n) {
87 case 's': return "string";
88 case 'i': return "integer";
89 case 'r': return "routine";
90 case 'R': return "routine or grouping";
91 case 'g': return "grouping";
93 fprintf(stderr, "Invalid type %d in name_of_type()\n", n);
94 exit(1);
97 static const char * name_of_name_type(int code) {
98 switch (code) {
99 case t_string: return "string";
100 case t_boolean: return "boolean";
101 case t_integer: return "integer";
102 case t_routine: return "routine";
103 case t_external: return "external";
104 case t_grouping: return "grouping";
106 fprintf(stderr, "Invalid type code %d in name_of_name_type()\n", code);
107 exit(1);
110 static void count_error(struct analyser * a) {
111 struct tokeniser * t = a->tokeniser;
112 if (t->error_count >= 20) { fprintf(stderr, "... etc\n"); exit(1); }
113 t->error_count++;
116 static void error2(struct analyser * a, error_code n, int x) {
117 struct tokeniser * t = a->tokeniser;
118 count_error(a);
119 fprintf(stderr, "%s:%d: ", t->file, t->line_number);
120 if ((int)n >= (int)e_redeclared) report_b(stderr, t->b);
121 switch (n) {
122 case e_token_omitted:
123 fprintf(stderr, "%s omitted", name_of_token(t->omission)); break;
124 case e_unexpected_token_in_among:
125 fprintf(stderr, "in among(...), ");
126 /* fall through */
127 case e_unexpected_token:
128 fprintf(stderr, "unexpected %s", name_of_token(t->token));
129 if (t->token == c_number) fprintf(stderr, " %d", t->number);
130 if (t->token == c_name) {
131 fprintf(stderr, " ");
132 report_b(stderr, t->b);
133 } break;
134 case e_string_omitted:
135 fprintf(stderr, "string omitted"); break;
137 case e_unresolved_substring:
138 fprintf(stderr, "unresolved substring on line %d", x); break;
139 case e_not_allowed_inside_reverse:
140 fprintf(stderr, "%s not allowed inside reverse(...)", name_of_token(t->token)); break;
141 case e_empty_grouping:
142 fprintf(stderr, "empty grouping"); break;
143 case e_already_backwards:
144 fprintf(stderr, "backwards used when already in this mode"); break;
145 case e_empty_among:
146 fprintf(stderr, "empty among(...)"); break;
147 case e_adjacent_bracketed_in_among:
148 fprintf(stderr, "two adjacent bracketed expressions in among(...)"); break;
149 case e_substring_preceded_by_substring:
150 fprintf(stderr, "substring preceded by another substring on line %d", x); break;
152 case e_redeclared:
153 fprintf(stderr, " re-declared"); break;
154 case e_undeclared:
155 fprintf(stderr, " undeclared"); break;
156 case e_declared_as_different_mode:
157 fprintf(stderr, " declared as %s mode; used as %s mode",
158 name_of_mode(a->mode), name_of_mode(x)); break;
159 case e_not_of_type_x:
160 fprintf(stderr, " not of type %s", name_of_type(x)); break;
161 case e_not_of_type_string_or_integer:
162 fprintf(stderr, " not of type string or integer"); break;
163 case e_misplaced:
164 fprintf(stderr, " misplaced"); break;
165 case e_redefined:
166 fprintf(stderr, " redefined"); break;
167 case e_misused:
168 fprintf(stderr, " mis-used as %s mode",
169 name_of_mode(x)); break;
171 if ((int)n < (int)e_unresolved_substring && t->previous_token > 0)
172 fprintf(stderr, " after %s", name_of_token(t->previous_token));
173 fprintf(stderr, "\n");
176 static void error(struct analyser * a, error_code n) { error2(a, n, 0); }
178 static void error3(struct analyser * a, struct node * p, symbol * b) {
179 count_error(a);
180 fprintf(stderr, "%s:%d: among(...) has repeated string '", a->tokeniser->file, p->line_number);
181 report_b(stderr, b);
182 fprintf(stderr, "'\n");
185 static void error3a(struct analyser * a, struct node * p) {
186 count_error(a);
187 fprintf(stderr, "%s:%d: previously seen here\n", a->tokeniser->file, p->line_number);
190 static void error4(struct analyser * a, struct name * q) {
191 count_error(a);
192 fprintf(stderr, "%s:%d: ", a->tokeniser->file, q->used->line_number);
193 report_b(stderr, q->b);
194 fprintf(stderr, " undefined\n");
197 static void omission_error(struct analyser * a, int n) {
198 a->tokeniser->omission = n;
199 error(a, e_token_omitted);
202 static int check_token(struct analyser * a, int code) {
203 struct tokeniser * t = a->tokeniser;
204 if (t->token != code) { omission_error(a, code); return false; }
205 return true;
208 static int get_token(struct analyser * a, int code) {
209 struct tokeniser * t = a->tokeniser;
210 read_token(t);
212 int x = check_token(a, code);
213 if (!x) t->token_held = true;
214 return x;
218 static struct name * look_for_name(struct analyser * a) {
219 symbol * q = a->tokeniser->b;
220 struct name * p;
221 for (p = a->names; p; p = p->next) {
222 symbol * b = p->b;
223 int n = SIZE(b);
224 if (n == SIZE(q) && memcmp(q, b, n * sizeof(symbol)) == 0) {
225 p->referenced = true;
226 return p;
229 return 0;
232 static struct name * find_name(struct analyser * a) {
233 struct name * p = look_for_name(a);
234 if (p == 0) error(a, e_undeclared);
235 return p;
238 static void check_routine_mode(struct analyser * a, struct name * p, int mode) {
239 if (p->mode < 0) p->mode = mode; else
240 if (p->mode != mode) error2(a, e_misused, mode);
243 static void check_name_type(struct analyser * a, struct name * p, int type) {
244 switch (type) {
245 case 's':
246 if (p->type == t_string) return;
247 break;
248 case 'i':
249 if (p->type == t_integer) return;
250 break;
251 case 'b':
252 if (p->type == t_boolean) return;
253 break;
254 case 'R':
255 if (p->type == t_grouping) return;
256 /* FALLTHRU */
257 case 'r':
258 if (p->type == t_routine || p->type == t_external) return;
259 break;
260 case 'g':
261 if (p->type == t_grouping) return;
262 break;
264 error2(a, e_not_of_type_x, type);
267 static void read_names(struct analyser * a, int type) {
268 struct tokeniser * t = a->tokeniser;
269 if (!get_token(a, c_bra)) return;
270 while (true) {
271 int token = read_token(t);
272 switch (token) {
273 case c_len: {
274 /* Context-sensitive token - once declared as a name, it loses
275 * its special meaning, for compatibility with older versions
276 * of snowball.
278 static const symbol c_len_lit[] = {
279 'l', 'e', 'n'
281 MOVE_TO_B(t->b, c_len_lit);
282 goto handle_as_name;
284 case c_lenof: {
285 /* Context-sensitive token - once declared as a name, it loses
286 * its special meaning, for compatibility with older versions
287 * of snowball.
289 static const symbol c_lenof_lit[] = {
290 'l', 'e', 'n', 'o', 'f'
292 MOVE_TO_B(t->b, c_lenof_lit);
293 goto handle_as_name;
295 case c_name:
296 handle_as_name:
297 if (look_for_name(a) != 0) error(a, e_redeclared); else {
298 NEW(name, p);
299 p->b = copy_b(t->b);
300 p->type = type;
301 p->mode = -1; /* routines, externals */
302 p->count = a->name_count[type];
303 p->referenced = false;
304 p->used_in_among = false;
305 p->used = 0;
306 p->local_to = 0;
307 p->grouping = 0;
308 p->definition = 0;
309 p->declaration_line_number = t->line_number;
310 a->name_count[type]++;
311 p->next = a->names;
312 a->names = p;
313 if (token != c_name) {
314 disable_token(t, token);
317 break;
318 default:
319 if (!check_token(a, c_ket)) t->token_held = true;
320 return;
325 static symbol * new_literalstring(struct analyser * a) {
326 NEW(literalstring, p);
327 p->b = copy_b(a->tokeniser->b);
328 p->next = a->literalstrings;
329 a->literalstrings = p;
330 return p->b;
333 static int read_AE_test(struct analyser * a) {
335 struct tokeniser * t = a->tokeniser;
336 switch (read_token(t)) {
337 case c_assign: return c_mathassign;
338 case c_plusassign:
339 case c_minusassign:
340 case c_multiplyassign:
341 case c_divideassign:
342 case c_eq:
343 case c_ne:
344 case c_gr:
345 case c_ge:
346 case c_ls:
347 case c_le: return t->token;
348 default: error(a, e_unexpected_token); t->token_held = true; return c_eq;
352 static int binding(int t) {
353 switch (t) {
354 case c_plus: case c_minus: return 1;
355 case c_multiply: case c_divide: return 2;
356 default: return -2;
360 static void mark_used_in(struct analyser * a, struct name * q, struct node * p) {
361 if (!q->used) {
362 q->used = p;
363 q->local_to = a->program_end->name;
364 } else if (q->local_to) {
365 if (q->local_to != a->program_end->name) {
366 /* Used in more than one routine/external. */
367 q->local_to = NULL;
372 static void name_to_node(struct analyser * a, struct node * p, int type) {
373 struct name * q = find_name(a);
374 if (q) {
375 check_name_type(a, q, type);
376 mark_used_in(a, q, p);
378 p->name = q;
381 static struct node * read_AE(struct analyser * a, int B) {
382 struct tokeniser * t = a->tokeniser;
383 struct node * p;
384 struct node * q;
385 switch (read_token(t)) {
386 case c_minus: /* monadic */
387 q = read_AE(a, 100);
388 if (q->type == c_neg) {
389 /* Optimise away double negation, which avoids generators
390 * having to worry about generating "--" (decrement operator
391 * in many languages).
393 p = q->right;
394 /* Don't free q, it's in the linked list a->nodes. */
395 break;
397 p = new_node(a, c_neg);
398 p->right = q;
399 break;
400 case c_bra:
401 p = read_AE(a, 0);
402 get_token(a, c_ket);
403 break;
404 case c_name:
405 p = new_node(a, c_name);
406 name_to_node(a, p, 'i');
407 break;
408 case c_maxint:
409 case c_minint:
410 a->int_limits_used = true;
411 /* fall through */
412 case c_cursor:
413 case c_limit:
414 case c_len:
415 case c_size:
416 p = new_node(a, t->token);
417 break;
418 case c_number:
419 p = new_node(a, c_number);
420 p->number = t->number;
421 break;
422 case c_lenof:
423 case c_sizeof:
424 p = C_style(a, "s", t->token);
425 break;
426 default:
427 error(a, e_unexpected_token);
428 t->token_held = true;
429 return 0;
431 while (true) {
432 int token = read_token(t);
433 int b = binding(token);
434 if (binding(token) <= B) {
435 t->token_held = true;
436 return p;
438 q = new_node(a, token);
439 q->left = p;
440 q->right = read_AE(a, b);
441 p = q;
445 static struct node * read_C_connection(struct analyser * a, struct node * q, int op) {
446 struct tokeniser * t = a->tokeniser;
447 struct node * p = new_node(a, op);
448 struct node * p_end = q;
449 p->left = q;
450 do {
451 q = read_C(a);
452 p_end->right = q; p_end = q;
453 } while (read_token(t) == op);
454 t->token_held = true;
455 return p;
458 static struct node * read_C_list(struct analyser * a) {
459 struct tokeniser * t = a->tokeniser;
460 struct node * p = new_node(a, c_bra);
461 struct node * p_end = 0;
462 while (true) {
463 int token = read_token(t);
464 if (token == c_ket) return p;
465 if (token < 0) { omission_error(a, c_ket); return p; }
466 t->token_held = true;
468 struct node * q = read_C(a);
469 while (true) {
470 token = read_token(t);
471 if (token != c_and && token != c_or) {
472 t->token_held = true;
473 break;
475 q = read_C_connection(a, q, token);
477 if (p_end == 0) p->left = q; else p_end->right = q;
478 p_end = q;
483 static struct node * C_style(struct analyser * a, const char * s, int token) {
484 int i;
485 struct node * p = new_node(a, token);
486 for (i = 0; s[i] != 0; i++) switch (s[i]) {
487 case 'C':
488 p->left = read_C(a); continue;
489 case 'D':
490 p->aux = read_C(a); continue;
491 case 'A':
492 p->AE = read_AE(a, 0); continue;
493 case 'f':
494 get_token(a, c_for); continue;
495 case 'S':
497 int str_token = read_token(a->tokeniser);
498 if (str_token == c_name) name_to_node(a, p, 's'); else
499 if (str_token == c_literalstring) p->literalstring = new_literalstring(a);
500 else error(a, e_string_omitted);
502 continue;
503 case 'b':
504 case 's':
505 case 'i':
506 if (get_token(a, c_name)) name_to_node(a, p, s[i]);
507 continue;
509 return p;
512 static struct node * read_literalstring(struct analyser * a) {
513 struct node * p = new_node(a, c_literalstring);
514 p->literalstring = new_literalstring(a);
515 return p;
518 static void reverse_b(symbol * b) {
519 int i = 0; int j = SIZE(b) - 1;
520 while (i < j) {
521 int ch1 = b[i]; int ch2 = b[j];
522 b[i++] = ch2; b[j--] = ch1;
526 static int compare_amongvec(const void *pv, const void *qv) {
527 const struct amongvec * p = (const struct amongvec*)pv;
528 const struct amongvec * q = (const struct amongvec*)qv;
529 symbol * b_p = p->b; int p_size = p->size;
530 symbol * b_q = q->b; int q_size = q->size;
531 int smaller_size = p_size < q_size ? p_size : q_size;
532 int i;
533 for (i = 0; i < smaller_size; i++)
534 if (b_p[i] != b_q[i]) return b_p[i] - b_q[i];
535 if (p_size - q_size)
536 return p_size - q_size;
537 return p->p->line_number - q->p->line_number;
540 static void make_among(struct analyser * a, struct node * p, struct node * substring) {
542 NEW(among, x);
543 NEWVEC(amongvec, v, p->number);
544 struct node * q = p->left;
545 struct amongvec * w0 = v;
546 struct amongvec * w1 = v;
547 int result = 1;
549 int direction = substring != 0 ? substring->mode : p->mode;
550 int backward = direction == m_backward;
552 if (a->amongs == 0) a->amongs = x; else a->amongs_end->next = x;
553 a->amongs_end = x;
554 x->next = 0;
555 x->b = v;
556 x->number = a->among_count++;
557 x->function_count = 0;
558 x->starter = 0;
560 if (q->type == c_bra) { x->starter = q; q = q->right; }
562 while (q) {
563 if (q->type == c_literalstring) {
564 symbol * b = q->literalstring;
565 w1->b = b; /* pointer to case string */
566 w1->p = q; /* pointer to corresponding node */
567 w1->size = SIZE(b); /* number of characters in string */
568 w1->i = -1; /* index of longest substring */
569 w1->result = -1; /* number of corresponding case expression */
570 if (q->left) {
571 struct name * function = q->left->name;
572 w1->function = function;
573 function->used_in_among = true;
574 check_routine_mode(a, function, direction);
575 x->function_count++;
576 } else {
577 w1->function = 0;
579 w1++;
581 else
582 if (q->left == 0) /* empty command: () */
583 w0 = w1;
584 else {
585 while (w0 != w1) {
586 w0->p = q;
587 w0->result = result;
588 w0++;
590 result++;
592 q = q->right;
594 if (w1-v != p->number) { fprintf(stderr, "oh! %d %d\n", (int)(w1-v), p->number); exit(1); }
595 if (backward) for (w0 = v; w0 < w1; w0++) reverse_b(w0->b);
596 qsort(v, w1 - v, sizeof(struct amongvec), compare_amongvec);
598 /* the following loop is O(n squared) */
599 for (w0 = w1 - 1; w0 >= v; w0--) {
600 symbol * b = w0->b;
601 int size = w0->size;
602 struct amongvec * w;
604 for (w = w0 - 1; w >= v; w--) {
605 if (w->size < size && memcmp(w->b, b, w->size * sizeof(symbol)) == 0) {
606 w0->i = w - v; /* fill in index of longest substring */
607 break;
611 if (backward) for (w0 = v; w0 < w1; w0++) reverse_b(w0->b);
613 for (w0 = v; w0 < w1 - 1; w0++)
614 if (w0->size == (w0 + 1)->size &&
615 memcmp(w0->b, (w0 + 1)->b, w0->size * sizeof(symbol)) == 0) {
616 error3(a, (w0 + 1)->p, (w0 + 1)->b);
617 error3a(a, w0->p);
620 x->literalstring_count = p->number;
621 x->command_count = result - 1;
622 p->among = x;
624 x->substring = substring;
625 if (substring != 0) substring->among = x;
626 if (x->command_count != 0 || x->starter != 0) a->amongvar_needed = true;
629 static struct node * read_among(struct analyser * a) {
630 struct tokeniser * t = a->tokeniser;
631 struct node * p = new_node(a, c_among);
632 struct node * p_end = 0;
633 int previous_token = -1;
634 struct node * substring = a->substring;
636 a->substring = 0;
637 p->number = 0; /* counts the number of literals */
638 if (!get_token(a, c_bra)) return p;
639 while (true) {
640 struct node * q;
641 int token = read_token(t);
642 switch (token) {
643 case c_literalstring:
644 q = read_literalstring(a);
645 if (read_token(t) == c_name) {
646 struct node * r = new_node(a, c_name);
647 name_to_node(a, r, 'r');
648 q->left = r;
650 else t->token_held = true;
651 p->number++; break;
652 case c_bra:
653 if (previous_token == c_bra) error(a, e_adjacent_bracketed_in_among);
654 q = read_C_list(a); break;
655 default:
656 error(a, e_unexpected_token_in_among);
657 previous_token = token;
658 continue;
659 case c_ket:
660 if (p->number == 0) error(a, e_empty_among);
661 if (t->error_count == 0) make_among(a, p, substring);
662 return p;
664 previous_token = token;
665 if (p_end == 0) p->left = q; else p_end->right = q;
666 p_end = q;
670 static struct node * read_substring(struct analyser * a) {
672 struct node * p = new_node(a, c_substring);
673 if (a->substring != 0) error2(a, e_substring_preceded_by_substring, a->substring->line_number);
674 a->substring = p;
675 return p;
678 static void check_modifyable(struct analyser * a) {
679 if (!a->modifyable) error(a, e_not_allowed_inside_reverse);
682 static struct node * read_C(struct analyser * a) {
683 struct tokeniser * t = a->tokeniser;
684 int token = read_token(t);
685 switch (token) {
686 case c_bra:
687 return read_C_list(a);
688 case c_backwards:
690 int mode = a->mode;
691 if (a->mode == m_backward) error(a, e_already_backwards); else a->mode = m_backward;
692 { struct node * p = C_style(a, "C", token);
693 a->mode = mode;
694 return p;
697 case c_reverse:
699 int mode = a->mode;
700 int modifyable = a->modifyable;
701 a->modifyable = false;
702 a->mode = mode == m_forward ? m_backward : m_forward;
704 struct node * p = C_style(a, "C", token);
705 a->mode = mode;
706 a->modifyable = modifyable;
707 return p;
710 case c_not:
711 case c_try:
712 case c_fail:
713 case c_test:
714 case c_do:
715 case c_goto:
716 case c_gopast:
717 case c_repeat:
718 return C_style(a, "C", token);
719 case c_loop:
720 case c_atleast:
721 return C_style(a, "AC", token);
722 case c_setmark:
723 return C_style(a, "i", token);
724 case c_tomark:
725 case c_atmark:
726 case c_hop:
727 return C_style(a, "A", token);
728 case c_delete:
729 check_modifyable(a);
730 /* fall through */
731 case c_next:
732 case c_tolimit:
733 case c_atlimit:
734 case c_leftslice:
735 case c_rightslice:
736 case c_true:
737 case c_false:
738 case c_debug:
739 return C_style(a, "", token);
740 case c_assignto:
741 case c_sliceto:
742 check_modifyable(a);
743 return C_style(a, "s", token);
744 case c_assign:
745 case c_insert:
746 case c_attach:
747 case c_slicefrom:
748 check_modifyable(a);
749 return C_style(a, "S", token);
750 case c_setlimit:
751 return C_style(a, "CfD", token);
752 case c_set:
753 case c_unset:
754 return C_style(a, "b", token);
755 case c_dollar:
756 get_token(a, c_name);
758 struct node * p;
759 struct name * q = find_name(a);
760 int mode = a->mode;
761 int modifyable = a->modifyable;
762 switch (q ? q->type : t_string)
763 /* above line was: switch (q->type) - bug #1 fix 7/2/2003 */
765 default:
766 error(a, e_not_of_type_string_or_integer);
767 /* Handle $foo for unknown 'foo' as string since
768 * that's more common and so less likely to cause
769 * an error avalanche. */
770 /* fall through */
771 case t_string:
772 a->mode = m_forward;
773 a->modifyable = true;
774 p = new_node(a, c_dollar);
775 p->left = read_C(a); break;
776 case t_integer:
777 /* a->mode = m_integer; */
778 p = new_node(a, read_AE_test(a));
779 p->AE = read_AE(a, 0); break;
781 if (q) mark_used_in(a, q, p);
782 p->name = q;
783 a->mode = mode;
784 a->modifyable = modifyable;
785 return p;
787 case c_name:
789 struct name * q = find_name(a);
790 struct node * p = new_node(a, c_name);
791 if (q) {
792 mark_used_in(a, q, p);
793 switch (q->type) {
794 case t_boolean:
795 p->type = c_booltest; break;
796 case t_integer:
797 error(a, e_misplaced); /* integer name misplaced */
798 case t_string:
799 break;
800 case t_routine:
801 case t_external:
802 p->type = c_call;
803 check_routine_mode(a, q, a->mode);
804 break;
805 case t_grouping:
806 p->type = c_grouping; break;
809 p->name = q;
810 return p;
812 case c_non:
814 struct node * p = new_node(a, token);
815 read_token(t);
816 if (t->token == c_minus) read_token(t);
817 if (!check_token(a, c_name)) { omission_error(a, c_name); return p; }
818 name_to_node(a, p, 'g');
819 return p;
821 case c_literalstring:
822 return read_literalstring(a);
823 case c_among: return read_among(a);
824 case c_substring: return read_substring(a);
825 default: error(a, e_unexpected_token); return 0;
829 static int next_symbol(symbol * p, symbol * W, int utf8) {
830 if (utf8) {
831 int ch;
832 int j = get_utf8(p, & ch);
833 W[0] = ch; return j;
834 } else {
835 W[0] = p[0]; return 1;
839 static symbol * alter_grouping(symbol * p, symbol * q, int style, int utf8) {
840 int j = 0;
841 symbol W[1];
842 int width;
843 if (style == c_plus) {
844 while (j < SIZE(q)) {
845 width = next_symbol(q + j, W, utf8);
846 p = add_to_b(p, 1, W);
847 j += width;
849 } else {
850 while (j < SIZE(q)) {
851 int i;
852 width = next_symbol(q + j, W, utf8);
853 for (i = 0; i < SIZE(p); i++) {
854 if (p[i] == W[0]) {
855 memmove(p + i, p + i + 1, (SIZE(p) - i - 1) * sizeof(symbol));
856 SIZE(p)--;
859 j += width;
862 return p;
865 static void read_define_grouping(struct analyser * a, struct name * q) {
866 struct tokeniser * t = a->tokeniser;
867 int style = c_plus;
869 NEW(grouping, p);
870 if (a->groupings == 0) a->groupings = p; else a->groupings_end->next = p;
871 a->groupings_end = p;
872 if (q) q->grouping = p;
873 p->next = 0;
874 p->name = q;
875 p->number = q ? q->count : 0;
876 p->line_number = a->tokeniser->line_number;
877 p->b = create_b(0);
878 while (true) {
879 switch (read_token(t)) {
880 case c_name:
882 struct name * r = find_name(a);
883 if (r) {
884 check_name_type(a, r, 'g');
885 p->b = alter_grouping(p->b, r->grouping->b, style, false);
888 break;
889 case c_literalstring:
890 p->b = alter_grouping(p->b, t->b, style, (a->encoding == ENC_UTF8));
891 break;
892 default: error(a, e_unexpected_token); return;
894 switch (read_token(t)) {
895 case c_plus:
896 case c_minus: style = t->token; break;
897 default: goto label0;
900 label0:
902 int i;
903 int max = 0;
904 int min = 1<<16;
905 for (i = 0; i < SIZE(p->b); i++) {
906 if (p->b[i] > max) max = p->b[i];
907 if (p->b[i] < min) min = p->b[i];
909 p->largest_ch = max;
910 p->smallest_ch = min;
911 if (min == 1<<16) error(a, e_empty_grouping);
913 t->token_held = true; return;
917 static void read_define_routine(struct analyser * a, struct name * q) {
918 struct node * p = new_node(a, c_define);
919 a->amongvar_needed = false;
920 if (q) {
921 check_name_type(a, q, 'R');
922 if (q->definition != 0) error(a, e_redefined);
923 if (q->mode < 0) q->mode = a->mode; else
924 if (q->mode != a->mode) error2(a, e_declared_as_different_mode, q->mode);
926 p->name = q;
927 if (a->program == 0) a->program = p; else a->program_end->right = p;
928 a->program_end = p;
929 get_token(a, c_as);
930 p->left = read_C(a);
931 if (q) q->definition = p->left;
933 if (a->substring != 0) {
934 error2(a, e_unresolved_substring, a->substring->line_number);
935 a->substring = 0;
937 p->amongvar_needed = a->amongvar_needed;
940 static void read_define(struct analyser * a) {
941 if (get_token(a, c_name)) {
942 struct name * q = find_name(a);
943 int type;
944 if (q) {
945 type = q->type;
946 } else {
947 /* No declaration, so sniff next token - if it is 'as' then parse
948 * as a routine, otherwise as a grouping.
950 if (read_token(a->tokeniser) == c_as) {
951 type = t_routine;
952 } else {
953 type = t_grouping;
955 a->tokeniser->token_held = true;
958 if (type == t_grouping) {
959 read_define_grouping(a, q);
960 } else {
961 read_define_routine(a, q);
966 static void read_backwardmode(struct analyser * a) {
967 int mode = a->mode;
968 a->mode = m_backward;
969 if (get_token(a, c_bra)) {
970 read_program_(a, c_ket);
971 check_token(a, c_ket);
973 a->mode = mode;
976 static void read_program_(struct analyser * a, int terminator) {
977 struct tokeniser * t = a->tokeniser;
978 while (true) {
979 switch (read_token(t)) {
980 case c_strings: read_names(a, t_string); break;
981 case c_booleans: read_names(a, t_boolean); break;
982 case c_integers: read_names(a, t_integer); break;
983 case c_routines: read_names(a, t_routine); break;
984 case c_externals: read_names(a, t_external); break;
985 case c_groupings: read_names(a, t_grouping); break;
986 case c_define: read_define(a); break;
987 case c_backwardmode:read_backwardmode(a); break;
988 case c_ket:
989 if (terminator == c_ket) return;
990 /* fall through */
991 default:
992 error(a, e_unexpected_token); break;
993 case -1:
994 if (terminator >= 0) omission_error(a, c_ket);
995 return;
1000 extern void read_program(struct analyser * a) {
1001 read_program_(a, -1);
1003 struct name * q = a->names;
1004 while (q) {
1005 switch (q->type) {
1006 case t_external: case t_routine:
1007 if (q->used && q->definition == 0) error4(a, q);
1008 break;
1009 case t_grouping:
1010 if (q->used && q->grouping == 0) error4(a, q);
1011 break;
1013 q = q->next;
1017 if (a->tokeniser->error_count == 0) {
1018 struct name * q = a->names;
1019 while (q) {
1020 if (!q->referenced) {
1021 fprintf(stderr, "%s:%d: warning: %s '",
1022 a->tokeniser->file,
1023 q->declaration_line_number,
1024 name_of_name_type(q->type));
1025 report_b(stderr, q->b);
1026 if (q->type == t_routine ||
1027 q->type == t_external ||
1028 q->type == t_grouping) {
1029 fprintf(stderr, "' declared but not defined\n");
1030 } else {
1031 fprintf(stderr, "' defined but not used\n");
1033 } else if (!q->used &&
1034 (q->type == t_routine || q->type == t_grouping)) {
1035 int line_num;
1036 if (q->type == t_routine) {
1037 line_num = q->definition->line_number;
1038 } else {
1039 line_num = q->grouping->line_number;
1041 fprintf(stderr, "%s:%d: warning: %s '",
1042 a->tokeniser->file,
1043 line_num,
1044 name_of_name_type(q->type));
1045 report_b(stderr, q->b);
1046 fprintf(stderr, "' defined but not used\n");
1048 q = q->next;
1053 extern struct analyser * create_analyser(struct tokeniser * t) {
1054 NEW(analyser, a);
1055 a->tokeniser = t;
1056 a->nodes = 0;
1057 a->names = 0;
1058 a->literalstrings = 0;
1059 a->program = 0;
1060 a->amongs = 0;
1061 a->among_count = 0;
1062 a->groupings = 0;
1063 a->mode = m_forward;
1064 a->modifyable = true;
1065 { int i; for (i = 0; i < t_size; i++) a->name_count[i] = 0; }
1066 a->substring = 0;
1067 a->int_limits_used = false;
1068 return a;
1071 extern void close_analyser(struct analyser * a) {
1073 struct node * q = a->nodes;
1074 while (q) {
1075 struct node * q_next = q->next;
1076 FREE(q);
1077 q = q_next;
1081 struct name * q = a->names;
1082 while (q) {
1083 struct name * q_next = q->next;
1084 lose_b(q->b); FREE(q);
1085 q = q_next;
1089 struct literalstring * q = a->literalstrings;
1090 while (q) {
1091 struct literalstring * q_next = q->next;
1092 lose_b(q->b); FREE(q);
1093 q = q_next;
1097 struct among * q = a->amongs;
1098 while (q) {
1099 struct among * q_next = q->next;
1100 FREE(q->b); FREE(q);
1101 q = q_next;
1105 struct grouping * q = a->groupings;
1106 while (q) {
1107 struct grouping * q_next = q->next;
1108 lose_b(q->b); FREE(q);
1109 q = q_next;
1112 FREE(a);