2  * Copyright (c) 2004-2005 Sergey Lyubka <valenok@gmail.com>
5  * "THE BEER-WARE LICENSE" (Revision 42):
52 	{"STARQ",	1, "o"},	/* Non-greedy STAR,  "*?"	*/
53 	{"PLUSQ",	1, "o"},	/* Non-greedy PLUS, "+?"	*/
56 	{"NONSPACE",	0, ""},		/* Match non-space, "\S"	*/
78  *	records the beginning of the matched substring (cap->ptr), CLOSE
79  *	sets the length (cap->len) for respective capture_number.
89  * STARQ, PLUSQ are non-greedy versions of STAR and PLUS.
121 	int	i, j, ch, op, pc;  in slre_dump()  local
123 	for (pc = 0; pc < r->code_size; pc++) {  in slre_dump()
125 		op = r->code[pc];  in slre_dump()
126 		(void) fprintf(fp, "%3d %s ", pc, opcodes[op].name);  in slre_dump()
131 				(void) fprintf(fp, "%d ", r->code[pc + 1]);  in slre_dump()
132 				pc++;  in slre_dump()
136 				    pc + r->code[pc + 1] - i);  in slre_dump()
137 				pc++;  in slre_dump()
140 				print_character_set(fp, r->data +  in slre_dump()
141 				    r->code[pc + 1], r->code[pc + 2]);  in slre_dump()
142 				pc += 2;  in slre_dump()
146 				for (j = 0; j < r->code[pc + 2]; j++) {  in slre_dump()
147 					ch = r->data[r->code[pc + 1] + j];  in slre_dump()
156 				pc += 2;  in slre_dump()
166 set_jump_offset(struct slre *r, int pc, int offset)  in set_jump_offset()  argument
168 	assert(offset < r->code_size);  in set_jump_offset()
170 	if (r->code_size - offset > 0xff)  in set_jump_offset()
171 		r->err_str = "Jump offset is too big";  in set_jump_offset()
173 		r->code[pc] = (unsigned char) (r->code_size - offset);  in set_jump_offset()
179 	if (r->code_size >= (int) (sizeof(r->code) / sizeof(r->code[0])))  in emit()
180 		r->err_str = "RE is too long (code overflow)";  in emit()
182 		r->code[r->code_size++] = (unsigned char) code;  in emit()
188 	if (r->data_size >= (int) sizeof(r->data))  in store_char_in_data()
189 		r->err_str = "RE is too long (data overflow)";  in store_char_in_data()
191 		r->data[r->data_size++] = ch;  in store_char_in_data()
197 	int	old_data_size = r->data_size;  in exact()
204 	emit(r, r->data_size - old_data_size);  in exact()
235 		res = (*re)[-1];  in get_escape_char()
245 	int	esc, old_data_size = r->data_size, op = ANYOF;  in anyof()
258 			emit(r, r->data_size - old_data_size);  in anyof()
272 			store_char_in_data(r, (*re)[-1]);  in anyof()
276 	r->err_str = "No closing ']' bracket";  in anyof()
283 	memmove(r->code + begin + shift, r->code + begin, r->code_size - begin);  in relocate()
284 	r->code_size += shift;  in relocate()
290 	if (r->code[prev] == EXACT && r->code[prev + 2] > 1) {  in quantifier()
291 		r->code[prev + 2]--;  in quantifier()
293 		emit(r, r->code[prev + 1] + r->code[prev + 2]);  in quantifier()
295 		prev = r->code_size - 3;  in quantifier()
298 	r->code[prev] = op;  in quantifier()
306 	emit(r, r->data_size);  in exact_one_char()
316 		set_jump_offset(r, fixup, fixup - 2);  in fixup_branch()
326 	level = r->num_caps;  in compile()
327 	branch_start = last_op = r->code_size;  in compile()
332 			(*re)--;  in compile()
343 			last_op = r->code_size;  in compile()
347 			last_op = r->code_size;  in compile()
351 			last_op = r->code_size;  in compile()
359 			last_op = r->code_size;  in compile()
360 			cap_no = ++r->num_caps;  in compile()
366 				r->err_str = "No closing bracket";  in compile()
374 			(*re)--;  in compile()
377 				r->err_str = "Unbalanced brackets";  in compile()
385 			op = (*re)[-1] == '*' ? STAR : PLUS;  in compile()
398 			r->code[branch_start] = BRANCH;  in compile()
401 			r->code[fixup] = 0xff;  in compile()
404 			(*re)--;  in compile()
405 			last_op = r->code_size;  in compile()
414 	r->err_str = NULL;  in slre_compile()
415 	r->code_size = r->data_size = r->num_caps = r->anchored = 0;  in slre_compile()
418 		r->anchored++;  in slre_compile()
426 	if (r->code[2] == BRANCH)  in slre_compile()
433 	return (r->err_str == NULL ? 1 : 0);  in slre_compile()
440 loop_greedy(const struct slre *r, int pc, const char *s, int len, int *ofs)  in loop_greedy()  argument
446 	while (match(r, pc + 2, s, len, ofs, NULL)) {  in loop_greedy()
448 		if (match(r, pc + r->code[pc + 1], s, len, ofs, NULL))  in loop_greedy()
457 loop_non_greedy(const struct slre *r, int pc, const char *s, int len, int *ofs)  in loop_non_greedy()  argument
461 	while (match(r, pc + 2, s, len, ofs, NULL)) {  in loop_non_greedy()
463 		if (match(r, pc + r->code[pc + 1], s, len, ofs, NULL))  in loop_non_greedy()
503 match(const struct slre *r, int pc, const char *s, int len,  in match()  argument
508 	while (res && r->code[pc] != END) {  in match()
510 		assert(pc < r->code_size);  in match()
511 		assert(pc < (int) (sizeof(r->code) / sizeof(r->code[0])));  in match()
513 		switch (r->code[pc]) {  in match()
516 			res = match(r, pc + 3, s, len, ofs, caps);  in match()
519 				res = match(r, pc + r->code[pc + 1],  in match()
522 			pc += r->code[pc + 2];  in match()
526 			n = r->code[pc + 2];	/* String length */  in match()
527 			if (n <= len - *ofs && !memcmp(s + *ofs, r->data +  in match()
528 			    r->code[pc + 1], n)) {  in match()
532 			pc += 3;  in match()
537 			if (!match(r, pc + 2, s, len, ofs, caps))  in match()
539 			pc += r->code[pc + 1];  in match()
543 			loop_greedy(r, pc, s, len, ofs);  in match()
544 			pc += r->code[pc + 1];  in match()
548 			loop_non_greedy(r, pc, s, len, ofs);  in match()
549 			pc += r->code[pc + 1];  in match()
552 			res = match(r, pc + 2, s, len, ofs, caps);  in match()
556 			loop_greedy(r, pc, s, len, ofs);  in match()
557 			pc += r->code[pc + 1];  in match()
560 			res = match(r, pc + 2, s, len, ofs, caps);  in match()
564 			loop_non_greedy(r, pc, s, len, ofs);  in match()
565 			pc += r->code[pc + 1];  in match()
573 			pc++;  in match()
582 			pc++;  in match()
590 			pc++;  in match()
598 			pc++;  in match()
603 				res = is_any_of(r->data + r->code[pc + 1],  in match()
604 					r->code[pc + 2], s, ofs);  in match()
605 			pc += 3;  in match()
610 				res = is_any_but(r->data + r->code[pc + 1],  in match()
611 					r->code[pc + 2], s, ofs);  in match()
612 			pc += 3;  in match()
616 			pc++;  in match()
620 			pc++;  in match()
624 				caps[r->code[pc + 1]].ptr = s + *ofs;  in match()
625 			pc += 2;  in match()
629 				caps[r->code[pc + 1]].len = (s + *ofs) -  in match()
630 				    caps[r->code[pc + 1]].ptr;  in match()
631 			pc += 2;  in match()
634 			pc++;  in match()
637 			printf("unknown cmd (%d) at %d\n", r->code[pc], pc);  in match()
652 	if (r->anchored) {  in slre_match()
697 		if ((len > 0) && (data[len-1] == '\n')) {  in main()
698 			data[len-1] = '\0';  in main()
699 			--len;  in main()
716 		printf("----------------------------------------------------\n");  in main()