1 /*****************************************************************************
2 * xlist.c : a trivial parser for XML-like tags
3 *****************************************************************************
4 * Copyright (C) 2003-2004 Commonwealth Scientific and Industrial Research
5 * Organisation (CSIRO) Australia
6 * Copyright (C) 2000-2004 the VideoLAN team
10 * Authors: Conrad Parker <Conrad.Parker@csiro.au>
11 * Andre Pang <Andre.Pang@csiro.au>
13 * This program is free software; you can redistribute it and/or modify
14 * it under the terms of the GNU General Public License as published by
15 * the Free Software Foundation; either version 2 of the License, or
16 * (at your option) any later version.
18 * This program is distributed in the hope that it will be useful,
19 * but WITHOUT ANY WARRANTY; without even the implied warranty of
20 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
21 * GNU General Public License for more details.
23 * You should have received a copy of the GNU General Public License
24 * along with this program; if not, write to the Free Software
25 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
26 *****************************************************************************/
46 #define MIN(a,b) ((a)<(b)?(a):(b))
49 #define MAX(a,b) ((a)>(b)?(a):(b))
51 typedef struct _XTag XTag;
52 typedef struct _XAttribute XAttribute;
53 typedef struct _XTagParser XTagParser;
56 * struct _XTag is kind of a union ... it normally represents a whole
57 * tag (and its children), but it could alternatively represent some
58 * PCDATA. Basically, if tag->pcdata is non-NULL, interpret only it and
59 * ignore the name, attributes and inner_tags.
67 XList * current_child;
76 int valid; /* boolean */
82 /* Character classes */
84 #define X_WHITESPACE 1<<0
85 #define X_OPENTAG 1<<1
86 #define X_CLOSETAG 1<<2
93 xtag_cin (char c, int char_class)
95 if (char_class & X_WHITESPACE)
96 if (isspace(c)) return TRUE;
98 if (char_class & X_OPENTAG)
99 if (c == '<') return TRUE;
101 if (char_class & X_CLOSETAG)
102 if (c == '>') return TRUE;
104 if (char_class & X_DQUOTE)
105 if (c == '"') return TRUE;
107 if (char_class & X_SQUOTE)
108 if (c == '\'') return TRUE;
110 if (char_class & X_EQUAL)
111 if (c == '=') return TRUE;
113 if (char_class & X_SLASH)
114 if (c == '/') return TRUE;
120 xtag_index (XTagParser * parser, int char_class)
127 for (i = 0; s[i] && s != parser->end; i++) {
128 if (xtag_cin(s[i], char_class)) return i;
135 xtag_skip_over (XTagParser * parser, int char_class)
140 if (!parser->valid) return;
142 s = (char *)parser->start;
144 for (i = 0; s[i] && s != parser->end; i++) {
145 if (!xtag_cin(s[i], char_class)) {
146 parser->start = &s[i];
155 xtag_skip_whitespace (XTagParser * parser)
157 xtag_skip_over (parser, X_WHITESPACE);
162 xtag_skip_to (XTagParser * parser, int char_class)
167 if (!parser->valid) return;
169 s = (char *)parser->start;
171 for (i = 0; s[i] && s != parser->end; i++) {
172 if (xtag_cin(s[i], char_class)) {
173 parser->start = &s[i];
183 xtag_slurp_to (XTagParser * parser, int good_end, int bad_end)
188 if (!parser->valid) return NULL;
192 xi = xtag_index (parser, good_end | bad_end);
194 if (xi > 0 && xtag_cin (s[xi], good_end)) {
195 ret = malloc ((xi+1) * sizeof(char));
196 strncpy (ret, s, xi);
198 parser->start = &s[xi];
206 xtag_assert_and_pass (XTagParser * parser, int char_class)
210 if (!parser->valid) return FALSE;
214 if (!xtag_cin (s[0], char_class)) {
215 parser->valid = FALSE;
219 parser->start = &s[1];
225 xtag_slurp_quoted (XTagParser * parser)
228 int quote = X_DQUOTE; /* quote char to match on */
231 if (!parser->valid) return NULL;
233 xtag_skip_whitespace (parser);
237 if (xtag_cin (s[0], X_SQUOTE)) quote = X_SQUOTE;
239 if (!xtag_assert_and_pass (parser, quote)) return NULL;
243 for (xi = 0; s[xi]; xi++) {
244 if (xtag_cin (s[xi], quote)) {
245 if (!(xi > 1 && s[xi-1] == '\\')) break;
249 ret = malloc ((xi+1) * sizeof(char));
250 strncpy (ret, s, xi);
252 parser->start = &s[xi];
254 if (!xtag_assert_and_pass (parser, quote)) return NULL;
260 xtag_parse_attribute (XTagParser * parser)
263 char * name, * value;
266 if (!parser->valid) return NULL;
268 xtag_skip_whitespace (parser);
270 name = xtag_slurp_to (parser, X_WHITESPACE | X_EQUAL, X_SLASH | X_CLOSETAG);
272 if (name == NULL) return NULL;
274 xtag_skip_whitespace (parser);
277 if (!xtag_assert_and_pass (parser, X_EQUAL)) {
279 printf ("xtag: attr failed EQUAL on <%s>\n", name);
284 xtag_skip_whitespace (parser);
286 value = xtag_slurp_quoted (parser);
290 printf ("Got NULL quoted attribute value\n");
295 attr = malloc (sizeof (*attr));
304 parser->valid = FALSE;
310 xtag_parse_tag (XTagParser * parser)
318 if (!parser->valid) return NULL;
320 if ((pcdata = xtag_slurp_to (parser, X_OPENTAG, X_NONE)) != NULL) {
321 tag = malloc (sizeof (*tag));
323 tag->pcdata = pcdata;
324 tag->parent = parser->current_tag;
325 tag->attributes = NULL;
326 tag->children = NULL;
327 tag->current_child = NULL;
334 /* if this starts a close tag, return NULL and let the parent take it */
335 if (xtag_cin (s[0], X_OPENTAG) && xtag_cin (s[1], X_SLASH))
338 if (!xtag_assert_and_pass (parser, X_OPENTAG)) return NULL;
340 name = xtag_slurp_to (parser, X_WHITESPACE | X_SLASH | X_CLOSETAG, X_NONE);
342 if (name == NULL) return NULL;
345 printf ("<%s ...\n", name);
348 tag = malloc (sizeof (*tag));
351 tag->parent = parser->current_tag;
352 tag->attributes = NULL;
353 tag->children = NULL;
354 tag->current_child = NULL;
358 if (xtag_cin (s[0], X_WHITESPACE)) {
359 while ((attr = xtag_parse_attribute (parser)) != NULL) {
360 tag->attributes = xlist_append (tag->attributes, attr);
364 xtag_skip_whitespace (parser);
368 if (xtag_cin (s[0], X_CLOSETAG)) {
369 parser->current_tag = tag;
371 xtag_assert_and_pass (parser, X_CLOSETAG);
373 while ((inner = xtag_parse_tag (parser)) != NULL) {
374 tag->children = xlist_append (tag->children, inner);
377 xtag_skip_whitespace (parser);
379 xtag_assert_and_pass (parser, X_OPENTAG);
380 xtag_assert_and_pass (parser, X_SLASH);
381 name = xtag_slurp_to (parser, X_WHITESPACE | X_CLOSETAG, X_NONE);
383 if (name && tag->name && strcmp (name, tag->name)) {
385 printf ("got %s expected %s\n", name, tag->name);
387 parser->valid = FALSE;
392 xtag_skip_whitespace (parser);
393 xtag_assert_and_pass (parser, X_CLOSETAG);
396 xtag_assert_and_pass (parser, X_SLASH);
397 xtag_assert_and_pass (parser, X_CLOSETAG);
405 xtag_free (XTag * xtag)
411 if (xtag == NULL) return NULL;
413 if (xtag->name) free (xtag->name);
414 if (xtag->pcdata) free (xtag->pcdata);
416 for (l = xtag->attributes; l; l = l->next) {
417 if ((attr = (XAttribute *)l->data) != NULL) {
418 if (attr->name) free (attr->name);
419 if (attr->value) free (attr->value);
423 xlist_free (xtag->attributes);
425 for (l = xtag->children; l; l = l->next) {
426 child = (XTag *)l->data;
429 xlist_free (xtag->children);
437 xtag_new_parse (const char * s, int n)
440 XTag * tag, * ttag, * wrapper;
443 parser.current_tag = NULL;
444 parser.start = (char *)s;
451 parser.end = (char *)&s[n];
453 tag = xtag_parse_tag (&parser);
460 if ((ttag = xtag_parse_tag (&parser)) != NULL) {
467 wrapper = malloc (sizeof (XTag));
468 wrapper->name = NULL;
469 wrapper->pcdata = NULL;
470 wrapper->parent = NULL;
471 wrapper->attributes = NULL;
472 wrapper->children = NULL;
473 wrapper->current_child = NULL;
475 wrapper->children = xlist_append (wrapper->children, tag);
476 wrapper->children = xlist_append (wrapper->children, ttag);
478 while ((ttag = xtag_parse_tag (&parser)) != NULL) {
485 wrapper->children = xlist_append (wrapper->children, ttag);
494 xtag_get_name (XTag * xtag)
496 return xtag ? xtag->name : NULL;
500 xtag_get_pcdata (XTag * xtag)
505 if (xtag == NULL) return NULL;
507 for (l = xtag->children; l; l = l->next) {
508 child = (XTag *)l->data;
509 if (child->pcdata != NULL) {
510 return child->pcdata;
518 xtag_get_attribute (XTag * xtag, char * attribute)
523 if (xtag == NULL) return NULL;
525 for (l = xtag->attributes; l; l = l->next) {
526 if ((attr = (XAttribute *)l->data) != NULL) {
527 if (attr->name && attribute && !strcmp (attr->name, attribute))
536 xtag_first_child (XTag * xtag, char * name)
541 if (xtag == NULL) return NULL;
543 if ((l = xtag->children) == NULL) return NULL;
546 xtag->current_child = l;
547 return (XTag *)l->data;
550 for (; l; l = l->next) {
551 child = (XTag *)l->data;
553 if (child->name && name && !strcmp(child->name, name)) {
554 xtag->current_child = l;
559 xtag->current_child = NULL;
565 xtag_next_child (XTag * xtag, char * name)
570 if (xtag == NULL) return NULL;
572 if ((l = xtag->current_child) == NULL)
573 return xtag_first_child (xtag, name);
575 if ((l = l->next) == NULL)
579 xtag->current_child = l;
580 return (XTag *)l->data;
583 for (; l; l = l->next) {
584 child = (XTag *)l->data;
586 if (child->name && name && !strcmp(child->name, name)) {
587 xtag->current_child = l;
592 xtag->current_child = NULL;
598 * This snprints function takes a variable list of char *, the last of
599 * which must be NULL, and prints each in turn to buf.
600 * Returns C99-style total length that would have been written, even if
601 * this is larger than n.
604 xtag_snprints (char * buf, int n, ...)
608 int len, to_copy, total = 0;
612 for (s = va_arg (ap, char *); s; s = va_arg (ap, char *)) {
615 if ((to_copy = MIN (n, len)) > 0) {
616 memcpy (buf, s, to_copy);
630 xtag_snprint (char * buf, int n, XTag * xtag)
643 if (n > 0) buf[0] = '\0';
648 nn = xtag_snprints (buf, n, xtag->pcdata, NULL);
655 nn = xtag_snprints (buf, n, "<", xtag->name, NULL);
658 for (l = xtag->attributes; l; l = l->next) {
659 attr = (XAttribute *)l->data;
661 nn = xtag_snprints (buf, n, " ", attr->name, "=\"", attr->value, "\"",
666 if (xtag->children == NULL) {
667 nn = xtag_snprints (buf, n, "/>", NULL);
673 nn = xtag_snprints (buf, n, ">", NULL);
677 for (l = xtag->children; l; l = l->next) {
678 child = (XTag *)l->data;
680 nn = xtag_snprint (buf, n, child);
685 nn = xtag_snprints (buf, n, "</", xtag->name, ">", NULL);