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_block.h"
34 #include "vlc_stream.h"
52 * struct XTag is kind of a union ... it normally represents a whole
53 * tag (and its children), but it could alternatively represent some
54 * PCDATA. Basically, if tag->pcdata is non-NULL, interpret only it and
55 * ignore the name, attributes and inner_tags.
67 typedef struct _XAttribute
73 typedef struct _XTagParser
75 int valid; /* boolean */
81 /*****************************************************************************
83 *****************************************************************************/
84 static int Open ( vlc_object_t * );
85 static void Close( vlc_object_t * );
88 set_description( _("Simple XML Parser") );
89 set_capability( "xml", 5 );
90 set_callbacks( Open, Close );
93 struct xml_reader_sys_t
95 XTag *p_root; /* Root tag */
96 XTag *p_curtag; /* Current tag */
97 XList *p_curattr; /* Current attribute */
101 static xml_reader_t *ReaderCreate( xml_t *, stream_t * );
102 static void ReaderDelete( xml_reader_t * );
103 static int ReaderRead( xml_reader_t * );
104 static int ReaderNodeType( xml_reader_t * );
105 static char *ReaderName( xml_reader_t * );
106 static char *ReaderValue( xml_reader_t * );
107 static int ReaderNextAttr( xml_reader_t * );
109 static int ReaderUseDTD ( xml_reader_t *, vlc_bool_t );
111 static void CatalogLoad( xml_t *, const char * );
112 static void CatalogAdd( xml_t *, const char *, const char *, const char * );
114 static XTag *xtag_new_parse( const char *, int );
115 static char *xtag_get_name( XTag * );
116 static char *xtag_get_pcdata( XTag * );
117 static char *xtag_get_attribute( XTag *, char * );
118 static XTag *xtag_first_child( XTag *, char * );
119 static XTag *xtag_next_child( XTag *, char * );
120 static XTag *xtag_free( XTag * );
121 static int xtag_snprint( char *, int, XTag * );
123 /*****************************************************************************
124 * Module initialization
125 *****************************************************************************/
126 static int Open( vlc_object_t *p_this )
128 xml_t *p_xml = (xml_t *)p_this;
130 p_xml->pf_reader_create = ReaderCreate;
131 p_xml->pf_reader_delete = ReaderDelete;
133 p_xml->pf_catalog_load = CatalogLoad;
134 p_xml->pf_catalog_add = CatalogAdd;
139 /*****************************************************************************
140 * Module deinitialization
141 *****************************************************************************/
142 static void Close( vlc_object_t *p_this )
147 /*****************************************************************************
148 * Catalogue functions
149 *****************************************************************************/
150 static void CatalogLoad( xml_t *p_xml, const char *psz_filename )
152 msg_Dbg( p_xml, "catalog support not implemented" );
155 static void CatalogAdd( xml_t *p_xml, const char *psz_arg1,
156 const char *psz_arg2, const char *psz_filename )
160 /*****************************************************************************
162 *****************************************************************************/
163 static xml_reader_t *ReaderCreate( xml_t *p_xml, stream_t *s )
165 xml_reader_t *p_reader;
166 char *p_buffer, *p_new;
167 int i_size, i_pos = 0, i_buffer = 2048;
170 /* Open and read file */
171 p_buffer = malloc( i_buffer );
172 if( p_buffer == NULL ) {
173 msg_Err( p_xml, "out of memory" );
177 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
181 p_new = realloc( p_buffer, i_buffer );
183 msg_Err( p_xml, "out of memory" );
189 p_buffer[ i_pos + i_size ] = 0; /* 0 terminated string */
191 if( i_pos + i_size == 0 )
193 msg_Dbg( p_xml, "empty XML" );
198 p_root = xtag_new_parse( p_buffer, i_buffer );
201 msg_Warn( p_xml, "couldn't parse XML" );
206 p_reader = malloc( sizeof(xml_reader_t) );
207 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
208 p_reader->p_sys->p_root = p_root;
209 p_reader->p_sys->p_curtag = NULL;
210 p_reader->p_sys->p_curattr = NULL;
211 p_reader->p_sys->b_endtag = VLC_FALSE;
212 p_reader->p_xml = p_xml;
214 p_reader->pf_read = ReaderRead;
215 p_reader->pf_node_type = ReaderNodeType;
216 p_reader->pf_name = ReaderName;
217 p_reader->pf_value = ReaderValue;
218 p_reader->pf_next_attr = ReaderNextAttr;
219 p_reader->pf_use_dtd = ReaderUseDTD;
224 static void ReaderDelete( xml_reader_t *p_reader )
226 xtag_free( p_reader->p_sys->p_root );
227 free( p_reader->p_sys );
231 static int ReaderUseDTD ( xml_reader_t *p_reader, vlc_bool_t b_use )
236 static int ReaderRead( xml_reader_t *p_reader )
240 if( !p_reader->p_sys->p_curtag )
242 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
248 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
250 p_reader->p_sys->p_curtag = p_child;
251 p_reader->p_sys->p_curattr = 0;
252 p_reader->p_sys->b_endtag = VLC_FALSE;
256 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
257 !p_reader->p_sys->b_endtag )
259 p_reader->p_sys->b_endtag = VLC_TRUE;
263 p_reader->p_sys->b_endtag = VLC_FALSE;
264 if( !p_reader->p_sys->p_curtag->parent ) return 0;
265 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
271 static int ReaderNodeType( xml_reader_t *p_reader )
273 if( p_reader->p_sys->p_curtag->name &&
274 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
275 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
276 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
277 return XML_READER_NONE;
280 static char *ReaderName( xml_reader_t *p_reader )
282 const char *psz_name;
284 if( !p_reader->p_sys->p_curattr )
286 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
288 printf( "TAG: %s\n", psz_name );
292 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
294 if( psz_name ) return strdup( psz_name );
298 static char *ReaderValue( xml_reader_t *p_reader )
300 const char *psz_name;
301 if( p_reader->p_sys->p_curtag->pcdata )
304 printf( "%s\n", p_reader->p_sys->p_curtag->pcdata );
306 return strdup( p_reader->p_sys->p_curtag->pcdata );
309 if( !p_reader->p_sys->p_curattr ) return 0;
312 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
313 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
316 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
318 if( psz_name ) return strdup( psz_name );
322 static int ReaderNextAttr( xml_reader_t *p_reader )
324 if( !p_reader->p_sys->p_curattr )
325 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
326 else if( p_reader->p_sys->p_curattr )
327 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
329 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
330 else return VLC_EGENERIC;
333 /*****************************************************************************
334 * XTAG parser functions
335 *****************************************************************************/
337 static XList *xlist_append( XList *list, void *data )
341 l = (XList *)malloc( sizeof(XList) );
342 l->prev = l->next = NULL;
345 if( list == NULL ) return l;
347 for( last = list; last; last = last->next )
348 if( last->next == NULL ) break;
350 if( last ) last->next = l;
355 static void xlist_free( XList *list )
359 for( l = list; l; l = ln )
366 /* Character classes */
368 #define X_WHITESPACE 1<<0
369 #define X_OPENTAG 1<<1
370 #define X_CLOSETAG 1<<2
371 #define X_DQUOTE 1<<3
372 #define X_SQUOTE 1<<4
379 static int xtag_cin( char c, int char_class )
381 if( char_class & X_WHITESPACE ) if( isspace(c) ) return VLC_TRUE;
382 if( char_class & X_OPENTAG ) if( c == '<' ) return VLC_TRUE;
383 if( char_class & X_CLOSETAG ) if( c == '>' ) return VLC_TRUE;
384 if( char_class & X_DQUOTE ) if( c == '"' ) return VLC_TRUE;
385 if( char_class & X_SQUOTE ) if( c == '\'' ) return VLC_TRUE;
386 if( char_class & X_EQUAL ) if( c == '=' ) return VLC_TRUE;
387 if( char_class & X_SLASH ) if( c == '/' ) return VLC_TRUE;
388 if( char_class & X_QMARK ) if( c == '?' ) return VLC_TRUE;
389 if( char_class & X_DASH ) if( c == '-' ) return VLC_TRUE;
390 if( char_class & X_EMARK ) if( c == '!' ) return VLC_TRUE;
395 static int xtag_index( XTagParser *parser, int char_class )
397 char *s = parser->start;
400 for( i = 0; s[i] && s != parser->end; i++ )
402 if( xtag_cin( s[i], char_class ) ) return i;
408 static void xtag_skip_over( XTagParser *parser, int char_class )
410 char *s = parser->start;
413 if( !parser->valid ) return;
415 for( i = 0; s[i] && s != parser->end; i++ )
417 if( !xtag_cin( s[i], char_class ) )
419 parser->start = &s[i];
427 static void xtag_skip_whitespace( XTagParser * parser )
429 xtag_skip_over( parser, X_WHITESPACE );
432 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
434 char *ret, *s = parser->start;
437 if( !parser->valid ) return NULL;
439 xi = xtag_index( parser, good_end | bad_end );
441 if( xi > 0 && xtag_cin (s[xi], good_end) )
443 ret = malloc( (xi+1) * sizeof(char) );
444 strncpy( ret, s, xi );
446 parser->start = &s[xi];
453 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
455 char *s = parser->start;
457 if( !parser->valid ) return VLC_FALSE;
459 if( !xtag_cin( s[0], char_class ) )
461 parser->valid = VLC_FALSE;
465 parser->start = &s[1];
470 static char *xtag_slurp_quoted( XTagParser *parser )
473 int quote = X_DQUOTE; /* quote char to match on */
476 if( !parser->valid ) return NULL;
478 xtag_skip_whitespace( parser );
482 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
484 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
488 for( xi = 0; s[xi]; xi++ )
490 if( xtag_cin( s[xi], quote ) )
492 if( !(xi > 1 && s[xi-1] == '\\') ) break;
496 ret = malloc( (xi+1) * sizeof(char) );
497 strncpy( ret, s, xi );
499 parser->start = &s[xi];
501 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
506 static XAttribute *xtag_parse_attribute( XTagParser *parser )
512 if( !parser->valid ) return NULL;
514 xtag_skip_whitespace( parser );
516 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
517 if( name == NULL ) return NULL;
519 xtag_skip_whitespace( parser );
522 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
525 printf( "xtag: attr failed EQUAL on <%s>\n", name );
530 xtag_skip_whitespace( parser );
532 value = xtag_slurp_quoted( parser );
537 printf ("Got NULL quoted attribute value\n");
542 attr = malloc( sizeof (*attr) );
549 parser->valid = VLC_FALSE;
553 static XTag *xtag_parse_tag( XTagParser *parser )
562 if( !parser->valid ) return NULL;
566 /* if this starts a comment tag, skip until end */
567 if( (parser->end - parser->start) > 7 &&
568 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
569 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
571 parser->start = s = &s[4];
572 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
574 parser->start = s = &s[xi+1];
575 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
577 parser->start = &s[2];
578 xtag_skip_whitespace( parser );
579 return xtag_parse_tag( parser );
585 /* ignore processing instructions '<?' ... '?>' */
586 if( (parser->end - parser->start) > 4 &&
587 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
589 parser->start = s = &s[2];
590 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
591 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
592 parser->start = &s[xi+2];
593 xtag_skip_whitespace( parser );
594 return xtag_parse_tag( parser );
600 /* ignore doctype '<!DOCTYPE' ... '>' */
601 if ( (parser->end - parser->start) > 8 &&
602 !strncmp( s, "<!DOCTYPE", 9 ) ) {
603 xi = xtag_index( parser, X_CLOSETAG );
605 parser->start = s = &s[xi+1];
606 xtag_skip_whitespace( parser );
607 return xtag_parse_tag( parser );
614 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
616 tag = malloc( sizeof(*tag) );
618 tag->pcdata = pcdata;
619 tag->parent = parser->current_tag;
620 tag->attributes = NULL;
621 tag->children = NULL;
622 tag->current_child = NULL;
627 /* if this starts a close tag, return NULL and let the parent take it */
628 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
631 /* parse CDATA content */
632 if ( (parser->end - parser->start) > 8 &&
633 !strncmp( s, "<![CDATA[", 9 ) ) {
634 parser->start = s = &s[9];
635 while (parser->end - s > 2) {
636 if (strncmp( s, "]]>", 3 ) == 0) {
637 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
638 if ( !(pcdata = malloc( sizeof(char)*(s - parser->start + 1))) ) return NULL;
639 strncpy( pcdata, parser->start, s - parser->start );
640 pcdata[s - parser->start]='\0';
641 parser->start = s = &s[3];
643 tag->pcdata = pcdata;
644 tag->parent = parser->current_tag;
645 tag->attributes = NULL;
646 tag->children = NULL;
647 tag->current_child = NULL;
657 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
659 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
660 if( name == NULL ) return NULL;
663 printf ("<%s ...\n", name);
666 tag = malloc( sizeof(*tag) );
669 tag->parent = parser->current_tag;
670 tag->attributes = NULL;
671 tag->children = NULL;
672 tag->current_child = NULL;
676 if( xtag_cin( s[0], X_WHITESPACE ) )
678 while( (attr = xtag_parse_attribute( parser )) != NULL )
680 tag->attributes = xlist_append( tag->attributes, attr );
684 xtag_skip_whitespace( parser );
688 if( xtag_cin( s[0], X_CLOSETAG ) )
690 parser->current_tag = tag;
692 xtag_assert_and_pass( parser, X_CLOSETAG );
694 while( (inner = xtag_parse_tag( parser ) ) != NULL )
696 tag->children = xlist_append( tag->children, inner );
699 parser->current_tag = tag->parent;
700 xtag_skip_whitespace( parser );
702 xtag_assert_and_pass( parser, X_OPENTAG );
703 xtag_assert_and_pass( parser, X_SLASH );
704 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
707 if( strcmp( name, tag->name ) )
710 printf ("got %s expected %s\n", name, tag->name);
712 parser->valid = VLC_FALSE;
717 xtag_skip_whitespace( parser );
718 xtag_assert_and_pass( parser, X_CLOSETAG );
719 xtag_skip_whitespace( parser );
723 xtag_assert_and_pass( parser, X_SLASH );
724 xtag_assert_and_pass( parser, X_CLOSETAG );
725 xtag_skip_whitespace( parser );
731 static XTag *xtag_free( XTag *xtag )
737 if( xtag == NULL ) return NULL;
739 if( xtag->name ) free( xtag->name );
740 if( xtag->pcdata ) free( xtag->pcdata );
742 for( l = xtag->attributes; l; l = l->next )
744 if( (attr = (XAttribute *)l->data) != NULL )
746 if( attr->name ) free( attr->name );
747 if( attr->value ) free( attr->value );
751 xlist_free( xtag->attributes );
753 for( l = xtag->children; l; l = l->next )
755 child = (XTag *)l->data;
758 xlist_free( xtag->children );
765 static XTag *xtag_new_parse( const char *s, int n )
768 XTag *tag, *ttag, *wrapper;
770 parser.valid = VLC_TRUE;
771 parser.current_tag = NULL;
772 parser.start = (char *)s;
774 if( n == -1 ) parser.end = NULL;
778 printf ("empty buffer");
782 else parser.end = (char *)&s[n];
784 /* can't have whitespace pcdata outside rootnode */
785 xtag_skip_whitespace( &parser );
787 tag = xtag_parse_tag( &parser );
792 printf ("invalid file");
798 if( (ttag = xtag_parse_tag( &parser )) != NULL )
806 wrapper = malloc( sizeof(XTag) );
807 wrapper->name = NULL;
808 wrapper->pcdata = NULL;
809 wrapper->parent = NULL;
810 wrapper->attributes = NULL;
811 wrapper->children = NULL;
812 wrapper->current_child = NULL;
814 wrapper->children = xlist_append( wrapper->children, tag );
815 wrapper->children = xlist_append( wrapper->children, ttag );
817 while( (ttag = xtag_parse_tag( &parser )) != NULL )
825 wrapper->children = xlist_append( wrapper->children, ttag );
833 static char *xtag_get_name( XTag *xtag )
835 return xtag ? xtag->name : NULL;
838 static char *xtag_get_pcdata( XTag *xtag )
843 if( xtag == NULL ) return NULL;
845 for( l = xtag->children; l; l = l->next )
847 child = (XTag *)l->data;
848 if( child->pcdata != NULL )
850 return child->pcdata;
857 static char *xtag_get_attribute( XTag *xtag, char *attribute )
862 if( xtag == NULL ) return NULL;
864 for( l = xtag->attributes; l; l = l->next )
866 if( (attr = (XAttribute *)l->data) != NULL )
868 if( !strcmp( attr->name, attribute ) ) return attr->value;
875 static XTag *xtag_first_child( XTag *xtag, char *name )
880 if( xtag == NULL ) return NULL;
881 if( (l = xtag->children) == NULL ) return NULL;
885 xtag->current_child = l;
886 return (XTag *)l->data;
889 for( ; l; l = l->next )
891 child = (XTag *)l->data;
893 if( !strcmp( child->name, name ) )
895 xtag->current_child = l;
900 xtag->current_child = NULL;
905 static XTag *xtag_next_child( XTag *xtag, char *name )
910 if( xtag == NULL ) return NULL;
912 if( (l = xtag->current_child) == NULL )
913 return xtag_first_child( xtag, name );
915 if( (l = l->next) == NULL ) return NULL;
919 xtag->current_child = l;
920 return (XTag *)l->data;
923 for( ; l; l = l->next )
925 child = (XTag *)l->data;
927 if( !strcmp( child->name, name ) )
929 xtag->current_child = l;
934 xtag->current_child = NULL;
940 * This snprints function takes a variable list of char *, the last of
941 * which must be NULL, and prints each in turn to buf.
942 * Returns C99-style total length that would have been written, even if
943 * this is larger than n.
945 static int xtag_snprints( char *buf, int n, ... )
949 int len, to_copy, total = 0;
953 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
957 if( (to_copy = __MIN(n, len) ) > 0 )
959 memcpy( buf, s, to_copy );
972 static int xtag_snprint( char *buf, int n, XTag *xtag )
980 buf += __MIN(n, N); \
986 if( n > 0 ) buf[0] = '\0';
992 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
1000 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1003 for( l = xtag->attributes; l; l = l->next )
1005 attr = (XAttribute *)l->data;
1007 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1012 if( xtag->children == NULL )
1014 nn = xtag_snprints ( buf, n, "/>", NULL );
1020 nn = xtag_snprints( buf, n, ">", NULL );
1024 for( l = xtag->children; l; l = l->next )
1026 child = (XTag *)l->data;
1028 nn = xtag_snprint( buf, n, child );
1034 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );