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" );
207 p_reader = malloc( sizeof(xml_reader_t) );
208 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
209 p_reader->p_sys->p_root = p_root;
210 p_reader->p_sys->p_curtag = NULL;
211 p_reader->p_sys->p_curattr = NULL;
212 p_reader->p_sys->b_endtag = VLC_FALSE;
213 p_reader->p_xml = p_xml;
215 p_reader->pf_read = ReaderRead;
216 p_reader->pf_node_type = ReaderNodeType;
217 p_reader->pf_name = ReaderName;
218 p_reader->pf_value = ReaderValue;
219 p_reader->pf_next_attr = ReaderNextAttr;
220 p_reader->pf_use_dtd = ReaderUseDTD;
225 static void ReaderDelete( xml_reader_t *p_reader )
227 xtag_free( p_reader->p_sys->p_root );
228 free( p_reader->p_sys );
232 static int ReaderUseDTD ( xml_reader_t *p_reader, vlc_bool_t b_use )
237 static int ReaderRead( xml_reader_t *p_reader )
241 if( !p_reader->p_sys->p_curtag )
243 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
249 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
251 p_reader->p_sys->p_curtag = p_child;
252 p_reader->p_sys->p_curattr = 0;
253 p_reader->p_sys->b_endtag = VLC_FALSE;
257 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
258 !p_reader->p_sys->b_endtag )
260 p_reader->p_sys->b_endtag = VLC_TRUE;
264 p_reader->p_sys->b_endtag = VLC_FALSE;
265 if( !p_reader->p_sys->p_curtag->parent ) return 0;
266 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
272 static int ReaderNodeType( xml_reader_t *p_reader )
274 if( p_reader->p_sys->p_curtag->name &&
275 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
276 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
277 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
278 return XML_READER_NONE;
281 static char *ReaderName( xml_reader_t *p_reader )
283 const char *psz_name;
285 if( !p_reader->p_sys->p_curattr )
287 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
289 printf( "TAG: %s\n", psz_name );
293 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
295 if( psz_name ) return strdup( psz_name );
299 static char *ReaderValue( xml_reader_t *p_reader )
301 const char *psz_name;
302 if( p_reader->p_sys->p_curtag->pcdata )
305 printf( "%s\n", p_reader->p_sys->p_curtag->pcdata );
307 return strdup( p_reader->p_sys->p_curtag->pcdata );
310 if( !p_reader->p_sys->p_curattr ) return 0;
313 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
314 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
317 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
319 if( psz_name ) return strdup( psz_name );
323 static int ReaderNextAttr( xml_reader_t *p_reader )
325 if( !p_reader->p_sys->p_curattr )
326 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
327 else if( p_reader->p_sys->p_curattr )
328 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
330 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
331 else return VLC_EGENERIC;
334 /*****************************************************************************
335 * XTAG parser functions
336 *****************************************************************************/
338 static XList *xlist_append( XList *list, void *data )
342 l = (XList *)malloc( sizeof(XList) );
343 l->prev = l->next = NULL;
346 if( list == NULL ) return l;
348 for( last = list; last; last = last->next )
349 if( last->next == NULL ) break;
351 if( last ) last->next = l;
356 static void xlist_free( XList *list )
360 for( l = list; l; l = ln )
367 /* Character classes */
369 #define X_WHITESPACE 1<<0
370 #define X_OPENTAG 1<<1
371 #define X_CLOSETAG 1<<2
372 #define X_DQUOTE 1<<3
373 #define X_SQUOTE 1<<4
380 static int xtag_cin( char c, int char_class )
382 if( char_class & X_WHITESPACE ) if( isspace(c) ) return VLC_TRUE;
383 if( char_class & X_OPENTAG ) if( c == '<' ) return VLC_TRUE;
384 if( char_class & X_CLOSETAG ) if( c == '>' ) return VLC_TRUE;
385 if( char_class & X_DQUOTE ) if( c == '"' ) return VLC_TRUE;
386 if( char_class & X_SQUOTE ) if( c == '\'' ) return VLC_TRUE;
387 if( char_class & X_EQUAL ) if( c == '=' ) return VLC_TRUE;
388 if( char_class & X_SLASH ) if( c == '/' ) return VLC_TRUE;
389 if( char_class & X_QMARK ) if( c == '?' ) return VLC_TRUE;
390 if( char_class & X_DASH ) if( c == '-' ) return VLC_TRUE;
391 if( char_class & X_EMARK ) if( c == '!' ) return VLC_TRUE;
396 static int xtag_index( XTagParser *parser, int char_class )
398 char *s = parser->start;
401 for( i = 0; s[i] && s != parser->end; i++ )
403 if( xtag_cin( s[i], char_class ) ) return i;
409 static void xtag_skip_over( XTagParser *parser, int char_class )
411 char *s = parser->start;
414 if( !parser->valid ) return;
416 for( i = 0; s[i] && s != parser->end; i++ )
418 if( !xtag_cin( s[i], char_class ) )
420 parser->start = &s[i];
428 static void xtag_skip_whitespace( XTagParser * parser )
430 xtag_skip_over( parser, X_WHITESPACE );
433 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
435 char *ret, *s = parser->start;
438 if( !parser->valid ) return NULL;
440 xi = xtag_index( parser, good_end | bad_end );
442 if( xi > 0 && xtag_cin (s[xi], good_end) )
444 ret = malloc( (xi+1) * sizeof(char) );
445 strncpy( ret, s, xi );
447 parser->start = &s[xi];
454 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
456 char *s = parser->start;
458 if( !parser->valid ) return VLC_FALSE;
460 if( !xtag_cin( s[0], char_class ) )
462 parser->valid = VLC_FALSE;
466 parser->start = &s[1];
471 static char *xtag_slurp_quoted( XTagParser *parser )
474 int quote = X_DQUOTE; /* quote char to match on */
477 if( !parser->valid ) return NULL;
479 xtag_skip_whitespace( parser );
483 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
485 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
489 for( xi = 0; s[xi]; xi++ )
491 if( xtag_cin( s[xi], quote ) )
493 if( !(xi > 1 && s[xi-1] == '\\') ) break;
497 ret = malloc( (xi+1) * sizeof(char) );
498 strncpy( ret, s, xi );
500 parser->start = &s[xi];
502 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
507 static XAttribute *xtag_parse_attribute( XTagParser *parser )
513 if( !parser->valid ) return NULL;
515 xtag_skip_whitespace( parser );
517 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
518 if( name == NULL ) return NULL;
520 xtag_skip_whitespace( parser );
523 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
526 printf( "xtag: attr failed EQUAL on <%s>\n", name );
531 xtag_skip_whitespace( parser );
533 value = xtag_slurp_quoted( parser );
538 printf ("Got NULL quoted attribute value\n");
543 attr = malloc( sizeof (*attr) );
550 parser->valid = VLC_FALSE;
554 static XTag *xtag_parse_tag( XTagParser *parser )
563 if( !parser->valid ) return NULL;
567 /* if this starts a comment tag, skip until end */
568 if( (parser->end - parser->start) > 7 &&
569 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
570 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
572 parser->start = s = &s[4];
573 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
575 parser->start = s = &s[xi+1];
576 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
578 parser->start = &s[2];
579 xtag_skip_whitespace( parser );
580 return xtag_parse_tag( parser );
586 /* ignore processing instructions '<?' ... '?>' */
587 if( (parser->end - parser->start) > 4 &&
588 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
590 parser->start = s = &s[2];
591 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
592 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
593 parser->start = &s[xi+2];
594 xtag_skip_whitespace( parser );
595 return xtag_parse_tag( parser );
601 /* ignore doctype '<!DOCTYPE' ... '>' */
602 if ( (parser->end - parser->start) > 8 &&
603 !strncmp( s, "<!DOCTYPE", 9 ) ) {
604 xi = xtag_index( parser, X_CLOSETAG );
606 parser->start = s = &s[xi+1];
607 xtag_skip_whitespace( parser );
608 return xtag_parse_tag( parser );
615 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
617 tag = malloc( sizeof(*tag) );
619 tag->pcdata = pcdata;
620 tag->parent = parser->current_tag;
621 tag->attributes = NULL;
622 tag->children = NULL;
623 tag->current_child = NULL;
628 /* if this starts a close tag, return NULL and let the parent take it */
629 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
632 /* parse CDATA content */
633 if ( (parser->end - parser->start) > 8 &&
634 !strncmp( s, "<![CDATA[", 9 ) ) {
635 parser->start = s = &s[9];
636 while (parser->end - s > 2) {
637 if (strncmp( s, "]]>", 3 ) == 0) {
638 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
639 if ( !(pcdata = malloc( sizeof(char)*(s - parser->start + 1))) ) return NULL;
640 strncpy( pcdata, parser->start, s - parser->start );
641 pcdata[s - parser->start]='\0';
642 parser->start = s = &s[3];
644 tag->pcdata = pcdata;
645 tag->parent = parser->current_tag;
646 tag->attributes = NULL;
647 tag->children = NULL;
648 tag->current_child = NULL;
658 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
660 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
661 if( name == NULL ) return NULL;
664 printf ("<%s ...\n", name);
667 tag = malloc( sizeof(*tag) );
670 tag->parent = parser->current_tag;
671 tag->attributes = NULL;
672 tag->children = NULL;
673 tag->current_child = NULL;
677 if( xtag_cin( s[0], X_WHITESPACE ) )
679 while( (attr = xtag_parse_attribute( parser )) != NULL )
681 tag->attributes = xlist_append( tag->attributes, attr );
685 xtag_skip_whitespace( parser );
689 if( xtag_cin( s[0], X_CLOSETAG ) )
691 parser->current_tag = tag;
693 xtag_assert_and_pass( parser, X_CLOSETAG );
695 while( (inner = xtag_parse_tag( parser ) ) != NULL )
697 tag->children = xlist_append( tag->children, inner );
700 parser->current_tag = tag->parent;
701 xtag_skip_whitespace( parser );
703 xtag_assert_and_pass( parser, X_OPENTAG );
704 xtag_assert_and_pass( parser, X_SLASH );
705 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
708 if( strcmp( name, tag->name ) )
711 printf ("got %s expected %s\n", name, tag->name);
713 parser->valid = VLC_FALSE;
718 xtag_skip_whitespace( parser );
719 xtag_assert_and_pass( parser, X_CLOSETAG );
720 xtag_skip_whitespace( parser );
724 xtag_assert_and_pass( parser, X_SLASH );
725 xtag_assert_and_pass( parser, X_CLOSETAG );
726 xtag_skip_whitespace( parser );
732 static XTag *xtag_free( XTag *xtag )
738 if( xtag == NULL ) return NULL;
740 if( xtag->name ) free( xtag->name );
741 if( xtag->pcdata ) free( xtag->pcdata );
743 for( l = xtag->attributes; l; l = l->next )
745 if( (attr = (XAttribute *)l->data) != NULL )
747 if( attr->name ) free( attr->name );
748 if( attr->value ) free( attr->value );
752 xlist_free( xtag->attributes );
754 for( l = xtag->children; l; l = l->next )
756 child = (XTag *)l->data;
759 xlist_free( xtag->children );
766 static XTag *xtag_new_parse( const char *s, int n )
769 XTag *tag, *ttag, *wrapper;
771 parser.valid = VLC_TRUE;
772 parser.current_tag = NULL;
773 parser.start = (char *)s;
775 if( n == -1 ) parser.end = NULL;
779 printf ("empty buffer");
783 else parser.end = (char *)&s[n];
785 /* can't have whitespace pcdata outside rootnode */
786 xtag_skip_whitespace( &parser );
788 tag = xtag_parse_tag( &parser );
793 printf ("invalid file");
799 if( (ttag = xtag_parse_tag( &parser )) != NULL )
807 wrapper = malloc( sizeof(XTag) );
808 wrapper->name = NULL;
809 wrapper->pcdata = NULL;
810 wrapper->parent = NULL;
811 wrapper->attributes = NULL;
812 wrapper->children = NULL;
813 wrapper->current_child = NULL;
815 wrapper->children = xlist_append( wrapper->children, tag );
816 wrapper->children = xlist_append( wrapper->children, ttag );
818 while( (ttag = xtag_parse_tag( &parser )) != NULL )
826 wrapper->children = xlist_append( wrapper->children, ttag );
834 static char *xtag_get_name( XTag *xtag )
836 return xtag ? xtag->name : NULL;
839 static char *xtag_get_pcdata( XTag *xtag )
844 if( xtag == NULL ) return NULL;
846 for( l = xtag->children; l; l = l->next )
848 child = (XTag *)l->data;
849 if( child->pcdata != NULL )
851 return child->pcdata;
858 static char *xtag_get_attribute( XTag *xtag, char *attribute )
863 if( xtag == NULL ) return NULL;
865 for( l = xtag->attributes; l; l = l->next )
867 if( (attr = (XAttribute *)l->data) != NULL )
869 if( !strcmp( attr->name, attribute ) ) return attr->value;
876 static XTag *xtag_first_child( XTag *xtag, char *name )
881 if( xtag == NULL ) return NULL;
882 if( (l = xtag->children) == NULL ) return NULL;
886 xtag->current_child = l;
887 return (XTag *)l->data;
890 for( ; l; l = l->next )
892 child = (XTag *)l->data;
894 if( !strcmp( child->name, name ) )
896 xtag->current_child = l;
901 xtag->current_child = NULL;
906 static XTag *xtag_next_child( XTag *xtag, char *name )
911 if( xtag == NULL ) return NULL;
913 if( (l = xtag->current_child) == NULL )
914 return xtag_first_child( xtag, name );
916 if( (l = l->next) == NULL ) return NULL;
920 xtag->current_child = l;
921 return (XTag *)l->data;
924 for( ; l; l = l->next )
926 child = (XTag *)l->data;
928 if( !strcmp( child->name, name ) )
930 xtag->current_child = l;
935 xtag->current_child = NULL;
941 * This snprints function takes a variable list of char *, the last of
942 * which must be NULL, and prints each in turn to buf.
943 * Returns C99-style total length that would have been written, even if
944 * this is larger than n.
946 static int xtag_snprints( char *buf, int n, ... )
950 int len, to_copy, total = 0;
954 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
958 if( (to_copy = __MIN(n, len) ) > 0 )
960 memcpy( buf, s, to_copy );
973 static int xtag_snprint( char *buf, int n, XTag *xtag )
981 buf += __MIN(n, N); \
987 if( n > 0 ) buf[0] = '\0';
993 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
1001 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1004 for( l = xtag->attributes; l; l = l->next )
1006 attr = (XAttribute *)l->data;
1008 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1013 if( xtag->children == NULL )
1015 nn = xtag_snprints ( buf, n, "/>", NULL );
1021 nn = xtag_snprints( buf, n, ">", NULL );
1025 for( l = xtag->children; l; l = l->next )
1027 child = (XTag *)l->data;
1029 nn = xtag_snprint( buf, n, child );
1035 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );