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 *****************************************************************************/
32 #include "vlc_block.h"
33 #include "vlc_stream.h"
48 * struct XTag is kind of a union ... it normally represents a whole
49 * tag (and its children), but it could alternatively represent some
50 * PCDATA. Basically, if tag->pcdata is non-NULL, interpret only it and
51 * ignore the name, attributes and inner_tags.
63 typedef struct _XAttribute
69 typedef struct _XTagParser
71 int valid; /* boolean */
77 /*****************************************************************************
79 *****************************************************************************/
80 static int Open ( vlc_object_t * );
81 static void Close( vlc_object_t * );
84 set_description( _("Simple XML Parser") );
85 set_capability( "xml", 5 );
86 set_callbacks( Open, Close );
89 struct xml_reader_sys_t
91 XTag *p_root; /* Root tag */
92 XTag *p_curtag; /* Current tag */
93 XList *p_curattr; /* Current attribute */
97 static xml_reader_t *ReaderCreate( xml_t *, stream_t * );
98 static void ReaderDelete( xml_reader_t * );
99 static int ReaderRead( xml_reader_t * );
100 static int ReaderNodeType( xml_reader_t * );
101 static char *ReaderName( xml_reader_t * );
102 static char *ReaderValue( xml_reader_t * );
103 static int ReaderNextAttr( xml_reader_t * );
105 static int ReaderUseDTD ( xml_reader_t *, vlc_bool_t );
107 static void CatalogLoad( xml_t *, const char * );
108 static void CatalogAdd( xml_t *, const char *, const char *, const char * );
110 static XTag *xtag_new_parse( const char *, int );
111 static char *xtag_get_name( XTag * );
112 static char *xtag_get_pcdata( XTag * );
113 static char *xtag_get_attribute( XTag *, char * );
114 static XTag *xtag_first_child( XTag *, char * );
115 static XTag *xtag_next_child( XTag *, char * );
116 static XTag *xtag_free( XTag * );
117 static int xtag_snprint( char *, int, XTag * );
119 /*****************************************************************************
120 * Module initialization
121 *****************************************************************************/
122 static int Open( vlc_object_t *p_this )
124 xml_t *p_xml = (xml_t *)p_this;
126 p_xml->pf_reader_create = ReaderCreate;
127 p_xml->pf_reader_delete = ReaderDelete;
129 p_xml->pf_catalog_load = CatalogLoad;
130 p_xml->pf_catalog_add = CatalogAdd;
135 /*****************************************************************************
136 * Module deinitialization
137 *****************************************************************************/
138 static void Close( vlc_object_t *p_this )
143 /*****************************************************************************
144 * Catalogue functions
145 *****************************************************************************/
146 static void CatalogLoad( xml_t *p_xml, const char *psz_filename )
148 msg_Dbg( p_xml, "catalog support not implemented" );
151 static void CatalogAdd( xml_t *p_xml, const char *psz_arg1,
152 const char *psz_arg2, const char *psz_filename )
156 /*****************************************************************************
158 *****************************************************************************/
159 static xml_reader_t *ReaderCreate( xml_t *p_xml, stream_t *s )
161 xml_reader_t *p_reader;
162 char *p_buffer, *p_new;
163 int i_size, i_pos = 0, i_buffer = 2048;
166 /* Open and read file */
167 p_buffer = malloc( i_buffer );
168 if( p_buffer == NULL ) {
169 msg_Err( p_xml, "out of memory" );
173 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
177 p_new = realloc( p_buffer, i_buffer );
179 msg_Err( p_xml, "out of memory" );
185 p_buffer[ i_pos + i_size ] = 0; /* 0 terminated string */
187 if( i_pos + i_size == 0 )
189 msg_Dbg( p_xml, "empty XML" );
194 p_root = xtag_new_parse( p_buffer, i_buffer );
197 msg_Warn( p_xml, "couldn't parse XML" );
203 p_reader = malloc( sizeof(xml_reader_t) );
204 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
205 p_reader->p_sys->p_root = p_root;
206 p_reader->p_sys->p_curtag = NULL;
207 p_reader->p_sys->p_curattr = NULL;
208 p_reader->p_sys->b_endtag = VLC_FALSE;
209 p_reader->p_xml = p_xml;
211 p_reader->pf_read = ReaderRead;
212 p_reader->pf_node_type = ReaderNodeType;
213 p_reader->pf_name = ReaderName;
214 p_reader->pf_value = ReaderValue;
215 p_reader->pf_next_attr = ReaderNextAttr;
216 p_reader->pf_use_dtd = ReaderUseDTD;
221 static void ReaderDelete( xml_reader_t *p_reader )
223 xtag_free( p_reader->p_sys->p_root );
224 free( p_reader->p_sys );
228 static int ReaderUseDTD ( xml_reader_t *p_reader, vlc_bool_t b_use )
233 static int ReaderRead( xml_reader_t *p_reader )
237 if( !p_reader->p_sys->p_curtag )
239 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
245 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
247 p_reader->p_sys->p_curtag = p_child;
248 p_reader->p_sys->p_curattr = 0;
249 p_reader->p_sys->b_endtag = VLC_FALSE;
253 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
254 !p_reader->p_sys->b_endtag )
256 p_reader->p_sys->b_endtag = VLC_TRUE;
260 p_reader->p_sys->b_endtag = VLC_FALSE;
261 if( !p_reader->p_sys->p_curtag->parent ) return 0;
262 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
268 static int ReaderNodeType( xml_reader_t *p_reader )
270 if( p_reader->p_sys->p_curtag->name &&
271 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
272 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
273 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
274 return XML_READER_NONE;
277 static char *ReaderName( xml_reader_t *p_reader )
279 const char *psz_name;
281 if( !p_reader->p_sys->p_curattr )
283 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
285 printf( "TAG: %s\n", psz_name );
289 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
291 if( psz_name ) return strdup( psz_name );
295 static char *ReaderValue( xml_reader_t *p_reader )
297 const char *psz_name;
298 if( p_reader->p_sys->p_curtag->pcdata )
301 printf( "%s\n", p_reader->p_sys->p_curtag->pcdata );
303 return strdup( p_reader->p_sys->p_curtag->pcdata );
306 if( !p_reader->p_sys->p_curattr ) return 0;
309 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
310 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
313 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
315 if( psz_name ) return strdup( psz_name );
319 static int ReaderNextAttr( xml_reader_t *p_reader )
321 if( !p_reader->p_sys->p_curattr )
322 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
323 else if( p_reader->p_sys->p_curattr )
324 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
326 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
327 else return VLC_EGENERIC;
330 /*****************************************************************************
331 * XTAG parser functions
332 *****************************************************************************/
334 static XList *xlist_append( XList *list, void *data )
338 l = (XList *)malloc( sizeof(XList) );
339 l->prev = l->next = NULL;
342 if( list == NULL ) return l;
344 for( last = list; last; last = last->next )
345 if( last->next == NULL ) break;
347 if( last ) last->next = l;
352 static void xlist_free( XList *list )
356 for( l = list; l; l = ln )
363 /* Character classes */
365 #define X_WHITESPACE 1<<0
366 #define X_OPENTAG 1<<1
367 #define X_CLOSETAG 1<<2
368 #define X_DQUOTE 1<<3
369 #define X_SQUOTE 1<<4
376 static int xtag_cin( char c, int char_class )
378 if( char_class & X_WHITESPACE ) if( isspace(c) ) return VLC_TRUE;
379 if( char_class & X_OPENTAG ) if( c == '<' ) return VLC_TRUE;
380 if( char_class & X_CLOSETAG ) if( c == '>' ) return VLC_TRUE;
381 if( char_class & X_DQUOTE ) if( c == '"' ) return VLC_TRUE;
382 if( char_class & X_SQUOTE ) if( c == '\'' ) return VLC_TRUE;
383 if( char_class & X_EQUAL ) if( c == '=' ) return VLC_TRUE;
384 if( char_class & X_SLASH ) if( c == '/' ) return VLC_TRUE;
385 if( char_class & X_QMARK ) if( c == '?' ) return VLC_TRUE;
386 if( char_class & X_DASH ) if( c == '-' ) return VLC_TRUE;
387 if( char_class & X_EMARK ) if( c == '!' ) return VLC_TRUE;
392 static int xtag_index( XTagParser *parser, int char_class )
394 char *s = parser->start;
397 for( i = 0; s[i] && s != parser->end; i++ )
399 if( xtag_cin( s[i], char_class ) ) return i;
405 static void xtag_skip_over( XTagParser *parser, int char_class )
407 char *s = parser->start;
410 if( !parser->valid ) return;
412 for( i = 0; s[i] && s != parser->end; i++ )
414 if( !xtag_cin( s[i], char_class ) )
416 parser->start = &s[i];
424 static void xtag_skip_whitespace( XTagParser * parser )
426 xtag_skip_over( parser, X_WHITESPACE );
429 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
431 char *ret, *s = parser->start;
434 if( !parser->valid ) return NULL;
436 xi = xtag_index( parser, good_end | bad_end );
438 if( xi > 0 && xtag_cin (s[xi], good_end) )
440 ret = malloc( (xi+1) * sizeof(char) );
441 strncpy( ret, s, xi );
443 parser->start = &s[xi];
450 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
452 char *s = parser->start;
454 if( !parser->valid ) return VLC_FALSE;
456 if( !xtag_cin( s[0], char_class ) )
458 parser->valid = VLC_FALSE;
462 parser->start = &s[1];
467 static char *xtag_slurp_quoted( XTagParser *parser )
470 int quote = X_DQUOTE; /* quote char to match on */
473 if( !parser->valid ) return NULL;
475 xtag_skip_whitespace( parser );
479 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
481 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
485 for( xi = 0; s[xi]; xi++ )
487 if( xtag_cin( s[xi], quote ) )
489 if( !(xi > 1 && s[xi-1] == '\\') ) break;
493 ret = malloc( (xi+1) * sizeof(char) );
494 strncpy( ret, s, xi );
496 parser->start = &s[xi];
498 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
503 static XAttribute *xtag_parse_attribute( XTagParser *parser )
509 if( !parser->valid ) return NULL;
511 xtag_skip_whitespace( parser );
513 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
514 if( name == NULL ) return NULL;
516 xtag_skip_whitespace( parser );
519 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
522 printf( "xtag: attr failed EQUAL on <%s>\n", name );
527 xtag_skip_whitespace( parser );
529 value = xtag_slurp_quoted( parser );
534 printf ("Got NULL quoted attribute value\n");
539 attr = malloc( sizeof (*attr) );
546 parser->valid = VLC_FALSE;
550 static XTag *xtag_parse_tag( XTagParser *parser )
559 if( !parser->valid ) return NULL;
563 /* if this starts a comment tag, skip until end */
564 if( (parser->end - parser->start) > 7 &&
565 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
566 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
568 parser->start = s = &s[4];
569 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
571 parser->start = s = &s[xi+1];
572 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
574 parser->start = &s[2];
575 xtag_skip_whitespace( parser );
576 return xtag_parse_tag( parser );
582 /* ignore processing instructions '<?' ... '?>' */
583 if( (parser->end - parser->start) > 4 &&
584 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
586 parser->start = s = &s[2];
587 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
588 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
589 parser->start = &s[xi+2];
590 xtag_skip_whitespace( parser );
591 return xtag_parse_tag( parser );
597 /* ignore doctype '<!DOCTYPE' ... '>' */
598 if ( (parser->end - parser->start) > 8 &&
599 !strncmp( s, "<!DOCTYPE", 9 ) ) {
600 xi = xtag_index( parser, X_CLOSETAG );
602 parser->start = s = &s[xi+1];
603 xtag_skip_whitespace( parser );
604 return xtag_parse_tag( parser );
611 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
613 tag = malloc( sizeof(*tag) );
615 tag->pcdata = pcdata;
616 tag->parent = parser->current_tag;
617 tag->attributes = NULL;
618 tag->children = NULL;
619 tag->current_child = NULL;
624 /* if this starts a close tag, return NULL and let the parent take it */
625 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
628 /* parse CDATA content */
629 if ( (parser->end - parser->start) > 8 &&
630 !strncmp( s, "<![CDATA[", 9 ) ) {
631 parser->start = s = &s[9];
632 while (parser->end - s > 2) {
633 if (strncmp( s, "]]>", 3 ) == 0) {
634 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
635 if ( !(pcdata = malloc( sizeof(char)*(s - parser->start + 1))) ) return NULL;
636 strncpy( pcdata, parser->start, s - parser->start );
637 pcdata[s - parser->start]='\0';
638 parser->start = s = &s[3];
640 tag->pcdata = pcdata;
641 tag->parent = parser->current_tag;
642 tag->attributes = NULL;
643 tag->children = NULL;
644 tag->current_child = NULL;
654 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
656 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
657 if( name == NULL ) return NULL;
660 printf ("<%s ...\n", name);
663 tag = malloc( sizeof(*tag) );
666 tag->parent = parser->current_tag;
667 tag->attributes = NULL;
668 tag->children = NULL;
669 tag->current_child = NULL;
673 if( xtag_cin( s[0], X_WHITESPACE ) )
675 while( (attr = xtag_parse_attribute( parser )) != NULL )
677 tag->attributes = xlist_append( tag->attributes, attr );
681 xtag_skip_whitespace( parser );
685 if( xtag_cin( s[0], X_CLOSETAG ) )
687 parser->current_tag = tag;
689 xtag_assert_and_pass( parser, X_CLOSETAG );
691 while( (inner = xtag_parse_tag( parser ) ) != NULL )
693 tag->children = xlist_append( tag->children, inner );
696 parser->current_tag = tag->parent;
697 xtag_skip_whitespace( parser );
699 xtag_assert_and_pass( parser, X_OPENTAG );
700 xtag_assert_and_pass( parser, X_SLASH );
701 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
704 if( strcmp( name, tag->name ) )
707 printf ("got %s expected %s\n", name, tag->name);
709 parser->valid = VLC_FALSE;
714 xtag_skip_whitespace( parser );
715 xtag_assert_and_pass( parser, X_CLOSETAG );
716 xtag_skip_whitespace( parser );
720 xtag_assert_and_pass( parser, X_SLASH );
721 xtag_assert_and_pass( parser, X_CLOSETAG );
722 xtag_skip_whitespace( parser );
728 static XTag *xtag_free( XTag *xtag )
734 if( xtag == NULL ) return NULL;
736 if( xtag->name ) free( xtag->name );
737 if( xtag->pcdata ) free( xtag->pcdata );
739 for( l = xtag->attributes; l; l = l->next )
741 if( (attr = (XAttribute *)l->data) != NULL )
743 if( attr->name ) free( attr->name );
744 if( attr->value ) free( attr->value );
748 xlist_free( xtag->attributes );
750 for( l = xtag->children; l; l = l->next )
752 child = (XTag *)l->data;
755 xlist_free( xtag->children );
762 static XTag *xtag_new_parse( const char *s, int n )
765 XTag *tag, *ttag, *wrapper;
767 parser.valid = VLC_TRUE;
768 parser.current_tag = NULL;
769 parser.start = (char *)s;
771 if( n == -1 ) parser.end = NULL;
775 printf ("empty buffer");
779 else parser.end = (char *)&s[n];
781 /* can't have whitespace pcdata outside rootnode */
782 xtag_skip_whitespace( &parser );
784 tag = xtag_parse_tag( &parser );
789 printf ("invalid file");
795 if( (ttag = xtag_parse_tag( &parser )) != NULL )
803 wrapper = malloc( sizeof(XTag) );
804 wrapper->name = NULL;
805 wrapper->pcdata = NULL;
806 wrapper->parent = NULL;
807 wrapper->attributes = NULL;
808 wrapper->children = NULL;
809 wrapper->current_child = NULL;
811 wrapper->children = xlist_append( wrapper->children, tag );
812 wrapper->children = xlist_append( wrapper->children, ttag );
814 while( (ttag = xtag_parse_tag( &parser )) != NULL )
822 wrapper->children = xlist_append( wrapper->children, ttag );
830 static char *xtag_get_name( XTag *xtag )
832 return xtag ? xtag->name : NULL;
835 static char *xtag_get_pcdata( XTag *xtag )
840 if( xtag == NULL ) return NULL;
842 for( l = xtag->children; l; l = l->next )
844 child = (XTag *)l->data;
845 if( child->pcdata != NULL )
847 return child->pcdata;
854 static char *xtag_get_attribute( XTag *xtag, char *attribute )
859 if( xtag == NULL ) return NULL;
861 for( l = xtag->attributes; l; l = l->next )
863 if( (attr = (XAttribute *)l->data) != NULL )
865 if( !strcmp( attr->name, attribute ) ) return attr->value;
872 static XTag *xtag_first_child( XTag *xtag, char *name )
877 if( xtag == NULL ) return NULL;
878 if( (l = xtag->children) == NULL ) return NULL;
882 xtag->current_child = l;
883 return (XTag *)l->data;
886 for( ; l; l = l->next )
888 child = (XTag *)l->data;
890 if( !strcmp( child->name, name ) )
892 xtag->current_child = l;
897 xtag->current_child = NULL;
902 static XTag *xtag_next_child( XTag *xtag, char *name )
907 if( xtag == NULL ) return NULL;
909 if( (l = xtag->current_child) == NULL )
910 return xtag_first_child( xtag, name );
912 if( (l = l->next) == NULL ) return NULL;
916 xtag->current_child = l;
917 return (XTag *)l->data;
920 for( ; l; l = l->next )
922 child = (XTag *)l->data;
924 if( !strcmp( child->name, name ) )
926 xtag->current_child = l;
931 xtag->current_child = NULL;
937 * This snprints function takes a variable list of char *, the last of
938 * which must be NULL, and prints each in turn to buf.
939 * Returns C99-style total length that would have been written, even if
940 * this is larger than n.
942 static int xtag_snprints( char *buf, int n, ... )
946 int len, to_copy, total = 0;
950 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
954 if( (to_copy = __MIN(n, len) ) > 0 )
956 memcpy( buf, s, to_copy );
969 static int xtag_snprint( char *buf, int n, XTag *xtag )
977 buf += __MIN(n, N); \
983 if( n > 0 ) buf[0] = '\0';
989 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
997 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1000 for( l = xtag->attributes; l; l = l->next )
1002 attr = (XAttribute *)l->data;
1004 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1009 if( xtag->children == NULL )
1011 nn = xtag_snprints ( buf, n, "/>", NULL );
1017 nn = xtag_snprints( buf, n, ">", NULL );
1021 for( l = xtag->children; l; l = l->next )
1023 child = (XTag *)l->data;
1025 nn = xtag_snprint( buf, n, child );
1031 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );