1 /*****************************************************************************
2 * xtag.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>
12 * Gildas Bazin <gbazin@videolan.org>
14 * This program is free software; you can redistribute it and/or modify
15 * it under the terms of the GNU General Public License as published by
16 * the Free Software Foundation; either version 2 of the License, or
17 * (at your option) any later version.
19 * This program is distributed in the hope that it will be useful,
20 * but WITHOUT ANY WARRANTY; without even the implied warranty of
21 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
22 * GNU General Public License for more details.
24 * You should have received a copy of the GNU General Public License
25 * along with this program; if not, write to the Free Software
26 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
27 *****************************************************************************/
33 #include <vlc_common.h>
34 #include <vlc_plugin.h>
37 #include "vlc_block.h"
38 #include "vlc_stream.h"
55 * struct XTag is kind of a union ... it normally represents a whole
56 * tag (and its children), but it could alternatively represent some
57 * PCDATA. Basically, if tag->pcdata is non-NULL, interpret only it and
58 * ignore the name, attributes and inner_tags.
70 typedef struct _XAttribute
76 typedef struct _XTagParser
78 int valid; /* boolean */
84 /*****************************************************************************
86 *****************************************************************************/
87 static int Open ( vlc_object_t * );
88 static void Close( vlc_object_t * );
91 set_description( N_("Simple XML Parser") );
92 set_capability( "xml", 5 );
93 set_callbacks( Open, Close );
96 struct xml_reader_sys_t
98 XTag *p_root; /* Root tag */
99 XTag *p_curtag; /* Current tag */
100 XList *p_curattr; /* Current attribute */
104 static xml_reader_t *ReaderCreate( xml_t *, stream_t * );
105 static void ReaderDelete( xml_reader_t * );
106 static int ReaderRead( xml_reader_t * );
107 static int ReaderNodeType( xml_reader_t * );
108 static char *ReaderName( xml_reader_t * );
109 static char *ReaderValue( xml_reader_t * );
110 static int ReaderNextAttr( xml_reader_t * );
112 static int ReaderUseDTD ( xml_reader_t *, bool );
114 static void CatalogLoad( xml_t *, const char * );
115 static void CatalogAdd( xml_t *, const char *, const char *, const char * );
117 static XTag *xtag_new_parse( const char *, int );
118 static char *xtag_get_name( XTag * );
120 static char *xtag_get_pcdata( XTag * );
121 static char *xtag_get_attribute( XTag *, char * );
123 static XTag *xtag_first_child( XTag *, char * );
124 static XTag *xtag_next_child( XTag *, char * );
125 static XTag *xtag_free( XTag * );
126 static int xtag_snprint( char *, int, XTag * );
128 /*****************************************************************************
129 * Module initialization
130 *****************************************************************************/
131 static int Open( vlc_object_t *p_this )
133 xml_t *p_xml = (xml_t *)p_this;
135 p_xml->pf_reader_create = ReaderCreate;
136 p_xml->pf_reader_delete = ReaderDelete;
138 p_xml->pf_catalog_load = CatalogLoad;
139 p_xml->pf_catalog_add = CatalogAdd;
144 /*****************************************************************************
145 * Module deinitialization
146 *****************************************************************************/
147 static void Close( vlc_object_t *p_this )
153 /*****************************************************************************
154 * Catalogue functions
155 *****************************************************************************/
156 static void CatalogLoad( xml_t *p_xml, const char *psz_filename )
158 VLC_UNUSED(psz_filename);
159 msg_Dbg( p_xml, "catalog support not implemented" );
162 static void CatalogAdd( xml_t *p_xml, const char *psz_arg1,
163 const char *psz_arg2, const char *psz_filename )
165 VLC_UNUSED(p_xml); VLC_UNUSED(psz_arg1); VLC_UNUSED(psz_arg2);
166 VLC_UNUSED(psz_filename);
169 /*****************************************************************************
171 *****************************************************************************/
172 static xml_reader_t *ReaderCreate( xml_t *p_xml, stream_t *s )
174 xml_reader_t *p_reader;
175 char *p_buffer, *p_new;
176 int i_size, i_pos = 0, i_buffer = 2048;
179 /* Open and read file */
180 p_buffer = malloc( i_buffer );
181 if( p_buffer == NULL ) {
182 msg_Err( p_xml, "out of memory" );
186 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
190 p_new = realloc( p_buffer, i_buffer );
193 msg_Err( p_xml, "out of memory" );
199 p_buffer[ i_pos + i_size ] = 0; /* 0 terminated string */
201 if( i_pos + i_size == 0 )
203 msg_Dbg( p_xml, "empty XML" );
208 p_root = xtag_new_parse( p_buffer, i_buffer );
211 msg_Warn( p_xml, "couldn't parse XML" );
217 p_reader = malloc( sizeof(xml_reader_t) );
218 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
219 p_reader->p_sys->p_root = p_root;
220 p_reader->p_sys->p_curtag = NULL;
221 p_reader->p_sys->p_curattr = NULL;
222 p_reader->p_sys->b_endtag = false;
223 p_reader->p_xml = p_xml;
225 p_reader->pf_read = ReaderRead;
226 p_reader->pf_node_type = ReaderNodeType;
227 p_reader->pf_name = ReaderName;
228 p_reader->pf_value = ReaderValue;
229 p_reader->pf_next_attr = ReaderNextAttr;
230 p_reader->pf_use_dtd = ReaderUseDTD;
235 static void ReaderDelete( xml_reader_t *p_reader )
237 xtag_free( p_reader->p_sys->p_root );
238 free( p_reader->p_sys );
242 static int ReaderUseDTD ( xml_reader_t *p_reader, bool b_use )
244 VLC_UNUSED(p_reader); VLC_UNUSED(b_use);
248 static int ReaderRead( xml_reader_t *p_reader )
252 if( !p_reader->p_sys->p_curtag )
254 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
260 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
262 p_reader->p_sys->p_curtag = p_child;
263 p_reader->p_sys->p_curattr = 0;
264 p_reader->p_sys->b_endtag = false;
268 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
269 !p_reader->p_sys->b_endtag )
271 p_reader->p_sys->b_endtag = true;
275 p_reader->p_sys->b_endtag = false;
276 if( !p_reader->p_sys->p_curtag->parent ) return 0;
277 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
283 static int ReaderNodeType( xml_reader_t *p_reader )
285 if( p_reader->p_sys->p_curtag->name &&
286 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
287 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
288 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
289 return XML_READER_NONE;
292 static char *ReaderName( xml_reader_t *p_reader )
294 const char *psz_name;
296 if( !p_reader->p_sys->p_curattr )
298 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
300 printf( "TAG: %s\n", psz_name );
304 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
306 if( psz_name ) return strdup( psz_name );
310 static char *ReaderValue( xml_reader_t *p_reader )
312 const char *psz_name;
313 if( p_reader->p_sys->p_curtag->pcdata )
316 printf( "%s\n", p_reader->p_sys->p_curtag->pcdata );
318 return strdup( p_reader->p_sys->p_curtag->pcdata );
321 if( !p_reader->p_sys->p_curattr ) return 0;
324 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
325 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
328 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
330 if( psz_name ) return strdup( psz_name );
334 static int ReaderNextAttr( xml_reader_t *p_reader )
336 if( !p_reader->p_sys->p_curattr )
337 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
338 else if( p_reader->p_sys->p_curattr )
339 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
341 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
342 else return VLC_EGENERIC;
345 /*****************************************************************************
346 * XTAG parser functions
347 *****************************************************************************/
349 static XList *xlist_append( XList *list, void *data )
353 l = (XList *)malloc( sizeof(XList) );
354 l->prev = l->next = NULL;
357 if( list == NULL ) return l;
359 for( last = list; last; last = last->next )
360 if( last->next == NULL ) break;
362 if( last ) last->next = l;
367 static void xlist_free( XList *list )
371 for( l = list; l; l = ln )
378 /* Character classes */
380 #define X_WHITESPACE 1<<0
381 #define X_OPENTAG 1<<1
382 #define X_CLOSETAG 1<<2
383 #define X_DQUOTE 1<<3
384 #define X_SQUOTE 1<<4
391 static int xtag_cin( char c, int char_class )
393 if( char_class & X_WHITESPACE ) if( isspace(c) ) return true;
394 if( char_class & X_OPENTAG ) if( c == '<' ) return true;
395 if( char_class & X_CLOSETAG ) if( c == '>' ) return true;
396 if( char_class & X_DQUOTE ) if( c == '"' ) return true;
397 if( char_class & X_SQUOTE ) if( c == '\'' ) return true;
398 if( char_class & X_EQUAL ) if( c == '=' ) return true;
399 if( char_class & X_SLASH ) if( c == '/' ) return true;
400 if( char_class & X_QMARK ) if( c == '?' ) return true;
401 if( char_class & X_DASH ) if( c == '-' ) return true;
402 if( char_class & X_EMARK ) if( c == '!' ) return true;
407 static int xtag_index( XTagParser *parser, int char_class )
409 char *s = parser->start;
412 for( i = 0; s[i] && s != parser->end; i++ )
414 if( xtag_cin( s[i], char_class ) ) return i;
420 static void xtag_skip_over( XTagParser *parser, int char_class )
422 char *s = parser->start;
425 if( !parser->valid ) return;
427 for( i = 0; s[i] && s != parser->end; i++ )
429 if( !xtag_cin( s[i], char_class ) )
431 parser->start = &s[i];
439 static void xtag_skip_whitespace( XTagParser * parser )
441 xtag_skip_over( parser, X_WHITESPACE );
444 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
446 char *ret, *s = parser->start;
449 if( !parser->valid ) return NULL;
451 xi = xtag_index( parser, good_end | bad_end );
453 if( xi > 0 && xtag_cin (s[xi], good_end) )
455 ret = malloc( (xi+1) * sizeof(char) );
456 strncpy( ret, s, xi );
458 parser->start = &s[xi];
465 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
467 char *s = parser->start;
469 if( !parser->valid ) return false;
471 if( !xtag_cin( s[0], char_class ) )
473 parser->valid = false;
477 parser->start = &s[1];
482 static char *xtag_slurp_quoted( XTagParser *parser )
485 int quote = X_DQUOTE; /* quote char to match on */
488 if( !parser->valid ) return NULL;
490 xtag_skip_whitespace( parser );
494 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
496 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
500 for( xi = 0; s[xi]; xi++ )
502 if( xtag_cin( s[xi], quote ) )
504 if( !(xi > 1 && s[xi-1] == '\\') ) break;
508 ret = malloc( (xi+1) * sizeof(char) );
509 strncpy( ret, s, xi );
511 parser->start = &s[xi];
513 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
518 static XAttribute *xtag_parse_attribute( XTagParser *parser )
524 if( !parser->valid ) return NULL;
526 xtag_skip_whitespace( parser );
528 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
529 if( name == NULL ) return NULL;
531 xtag_skip_whitespace( parser );
534 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
537 printf( "xtag: attr failed EQUAL on <%s>\n", name );
542 xtag_skip_whitespace( parser );
544 value = xtag_slurp_quoted( parser );
549 printf ("Got NULL quoted attribute value\n");
554 attr = malloc( sizeof (*attr) );
561 parser->valid = false;
565 static XTag *xtag_parse_tag( XTagParser *parser )
574 if( !parser->valid ) return NULL;
578 /* if this starts a comment tag, skip until end */
579 if( (parser->end - parser->start) > 7 &&
580 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
581 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
583 parser->start = s = &s[4];
584 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
586 parser->start = s = &s[xi+1];
587 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
589 parser->start = &s[2];
590 xtag_skip_whitespace( parser );
591 return xtag_parse_tag( parser );
597 /* ignore processing instructions '<?' ... '?>' */
598 if( (parser->end - parser->start) > 4 &&
599 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
601 parser->start = s = &s[2];
602 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
603 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
604 parser->start = &s[xi+2];
605 xtag_skip_whitespace( parser );
606 return xtag_parse_tag( parser );
612 /* ignore doctype '<!DOCTYPE' ... '>' */
613 if ( (parser->end - parser->start) > 8 &&
614 !strncmp( s, "<!DOCTYPE", 9 ) ) {
615 xi = xtag_index( parser, X_CLOSETAG );
617 parser->start = s = &s[xi+1];
618 xtag_skip_whitespace( parser );
619 return xtag_parse_tag( parser );
626 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
628 tag = malloc( sizeof(*tag) );
630 tag->pcdata = pcdata;
631 tag->parent = parser->current_tag;
632 tag->attributes = NULL;
633 tag->children = NULL;
634 tag->current_child = NULL;
639 /* if this starts a close tag, return NULL and let the parent take it */
640 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
643 /* parse CDATA content */
644 if ( (parser->end - parser->start) > 8 &&
645 !strncmp( s, "<![CDATA[", 9 ) ) {
646 parser->start = s = &s[9];
647 while (parser->end - s > 2) {
648 if (strncmp( s, "]]>", 3 ) == 0) {
649 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
650 if ( !(pcdata = malloc( sizeof(char)*(s - parser->start + 1))) ) return NULL;
651 strncpy( pcdata, parser->start, s - parser->start );
652 pcdata[s - parser->start]='\0';
653 parser->start = s = &s[3];
655 tag->pcdata = pcdata;
656 tag->parent = parser->current_tag;
657 tag->attributes = NULL;
658 tag->children = NULL;
659 tag->current_child = NULL;
669 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
671 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
672 if( name == NULL ) return NULL;
675 printf ("<%s ...\n", name);
678 tag = malloc( sizeof(*tag) );
681 tag->parent = parser->current_tag;
682 tag->attributes = NULL;
683 tag->children = NULL;
684 tag->current_child = NULL;
688 if( xtag_cin( s[0], X_WHITESPACE ) )
690 while( (attr = xtag_parse_attribute( parser )) != NULL )
692 tag->attributes = xlist_append( tag->attributes, attr );
696 xtag_skip_whitespace( parser );
700 if( xtag_cin( s[0], X_CLOSETAG ) )
702 parser->current_tag = tag;
704 xtag_assert_and_pass( parser, X_CLOSETAG );
706 while( (inner = xtag_parse_tag( parser ) ) != NULL )
708 tag->children = xlist_append( tag->children, inner );
711 parser->current_tag = tag->parent;
712 xtag_skip_whitespace( parser );
714 xtag_assert_and_pass( parser, X_OPENTAG );
715 xtag_assert_and_pass( parser, X_SLASH );
716 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
719 if( strcmp( name, tag->name ) )
722 printf ("got %s expected %s\n", name, tag->name);
724 parser->valid = false;
729 xtag_skip_whitespace( parser );
730 xtag_assert_and_pass( parser, X_CLOSETAG );
731 xtag_skip_whitespace( parser );
735 xtag_assert_and_pass( parser, X_SLASH );
736 xtag_assert_and_pass( parser, X_CLOSETAG );
737 xtag_skip_whitespace( parser );
743 static XTag *xtag_free( XTag *xtag )
749 if( xtag == NULL ) return NULL;
752 free( xtag->pcdata );
754 for( l = xtag->attributes; l; l = l->next )
756 if( (attr = (XAttribute *)l->data) != NULL )
763 xlist_free( xtag->attributes );
765 for( l = xtag->children; l; l = l->next )
767 child = (XTag *)l->data;
770 xlist_free( xtag->children );
777 static XTag *xtag_new_parse( const char *s, int n )
780 XTag *tag, *ttag, *wrapper;
783 parser.current_tag = NULL;
784 parser.start = (char *)s;
786 if( n == -1 ) parser.end = NULL;
790 printf ("empty buffer");
794 else parser.end = (char *)&s[n];
796 /* can't have whitespace pcdata outside rootnode */
797 xtag_skip_whitespace( &parser );
799 tag = xtag_parse_tag( &parser );
804 printf ("invalid file");
810 if( (ttag = xtag_parse_tag( &parser )) != NULL )
818 wrapper = malloc( sizeof(XTag) );
819 wrapper->name = NULL;
820 wrapper->pcdata = NULL;
821 wrapper->parent = NULL;
822 wrapper->attributes = NULL;
823 wrapper->children = NULL;
824 wrapper->current_child = NULL;
826 wrapper->children = xlist_append( wrapper->children, tag );
827 wrapper->children = xlist_append( wrapper->children, ttag );
829 while( (ttag = xtag_parse_tag( &parser )) != NULL )
837 wrapper->children = xlist_append( wrapper->children, ttag );
845 static char *xtag_get_name( XTag *xtag )
847 return xtag ? xtag->name : NULL;
851 static char *xtag_get_pcdata( XTag *xtag )
856 if( xtag == NULL ) return NULL;
858 for( l = xtag->children; l; l = l->next )
860 child = (XTag *)l->data;
861 if( child->pcdata != NULL )
863 return child->pcdata;
870 static char *xtag_get_attribute( XTag *xtag, char *attribute )
875 if( xtag == NULL ) return NULL;
877 for( l = xtag->attributes; l; l = l->next )
879 if( (attr = (XAttribute *)l->data) != NULL )
881 if( !strcmp( attr->name, attribute ) ) return attr->value;
889 static XTag *xtag_first_child( XTag *xtag, char *name )
894 if( xtag == NULL ) return NULL;
895 if( (l = xtag->children) == NULL ) return NULL;
899 xtag->current_child = l;
900 return (XTag *)l->data;
903 for( ; l; l = l->next )
905 child = (XTag *)l->data;
907 if( !strcmp( child->name, name ) )
909 xtag->current_child = l;
914 xtag->current_child = NULL;
919 static XTag *xtag_next_child( XTag *xtag, char *name )
924 if( xtag == NULL ) return NULL;
926 if( (l = xtag->current_child) == NULL )
927 return xtag_first_child( xtag, name );
929 if( (l = l->next) == NULL ) return NULL;
933 xtag->current_child = l;
934 return (XTag *)l->data;
937 for( ; l; l = l->next )
939 child = (XTag *)l->data;
941 if( !strcmp( child->name, name ) )
943 xtag->current_child = l;
948 xtag->current_child = NULL;
954 * This snprints function takes a variable list of char *, the last of
955 * which must be NULL, and prints each in turn to buf.
956 * Returns C99-style total length that would have been written, even if
957 * this is larger than n.
959 static int xtag_snprints( char *buf, int n, ... )
963 int len, to_copy, total = 0;
967 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
971 if( (to_copy = __MIN(n, len) ) > 0 )
973 memcpy( buf, s, to_copy );
986 static int xtag_snprint( char *buf, int n, XTag *xtag )
994 buf += __MIN(n, N); \
1000 if( n > 0 ) buf[0] = '\0';
1006 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
1014 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1017 for( l = xtag->attributes; l; l = l->next )
1019 attr = (XAttribute *)l->data;
1021 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1026 if( xtag->children == NULL )
1028 nn = xtag_snprints ( buf, n, "/>", NULL );
1034 nn = xtag_snprints( buf, n, ">", NULL );
1038 for( l = xtag->children; l; l = l->next )
1040 child = (XTag *)l->data;
1042 nn = xtag_snprint( buf, n, child );
1048 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );