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>
39 #include <vlc_memory.h>
54 * struct XTag is kind of a union ... it normally represents a whole
55 * tag (and its children), but it could alternatively represent some
56 * PCDATA. Basically, if tag->pcdata is non-NULL, interpret only it and
57 * ignore the name, attributes and inner_tags.
69 typedef struct _XAttribute
75 typedef struct _XTagParser
77 int valid; /* boolean */
83 /*****************************************************************************
85 *****************************************************************************/
86 static int Open ( vlc_object_t * );
87 static void Close( vlc_object_t * );
90 set_description( N_("Simple XML Parser") )
91 set_capability( "xml", 5 )
92 set_callbacks( Open, Close )
95 struct xml_reader_sys_t
97 XTag *p_root; /* Root tag */
98 XTag *p_curtag; /* Current tag */
99 XList *p_curattr; /* Current attribute */
103 static xml_reader_t *ReaderCreate( xml_t *, stream_t * );
104 static void ReaderDelete( xml_reader_t * );
105 static int ReaderRead( xml_reader_t * );
106 static int ReaderNodeType( xml_reader_t * );
107 static char *ReaderName( xml_reader_t * );
108 static char *ReaderValue( xml_reader_t * );
109 static int ReaderNextAttr( xml_reader_t * );
111 static int ReaderUseDTD ( xml_reader_t *, bool );
113 static void CatalogLoad( xml_t *, const char * );
114 static void CatalogAdd( xml_t *, const char *, const char *, const char * );
116 static XTag *xtag_new_parse( const char *, int );
117 static char *xtag_get_name( XTag * );
119 static char *xtag_get_pcdata( XTag * );
120 static char *xtag_get_attribute( XTag *, char * );
122 static XTag *xtag_first_child( XTag *, char * );
123 static XTag *xtag_next_child( XTag *, char * );
124 static void xtag_free( XTag * );
126 static int xtag_snprint( char *, int, XTag * );
129 /*****************************************************************************
130 * Module initialization
131 *****************************************************************************/
132 static int Open( vlc_object_t *p_this )
134 xml_t *p_xml = (xml_t *)p_this;
136 p_xml->pf_reader_create = ReaderCreate;
137 p_xml->pf_reader_delete = ReaderDelete;
139 p_xml->pf_catalog_load = CatalogLoad;
140 p_xml->pf_catalog_add = CatalogAdd;
145 /*****************************************************************************
146 * Module deinitialization
147 *****************************************************************************/
148 static void Close( vlc_object_t *p_this )
154 /*****************************************************************************
155 * Catalogue functions
156 *****************************************************************************/
157 static void CatalogLoad( xml_t *p_xml, const char *psz_filename )
159 VLC_UNUSED(psz_filename);
160 msg_Dbg( p_xml, "catalog support not implemented" );
163 static void CatalogAdd( xml_t *p_xml, const char *psz_arg1,
164 const char *psz_arg2, const char *psz_filename )
166 VLC_UNUSED(p_xml); VLC_UNUSED(psz_arg1); VLC_UNUSED(psz_arg2);
167 VLC_UNUSED(psz_filename);
170 /*****************************************************************************
172 *****************************************************************************/
173 static xml_reader_t *ReaderCreate( xml_t *p_xml, stream_t *s )
175 xml_reader_t *p_reader;
177 int i_size, i_pos = 0, i_buffer = 2048;
180 /* Open and read file */
181 p_buffer = malloc( i_buffer );
182 if( p_buffer == NULL )
185 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
189 p_buffer = realloc_or_free( p_buffer, i_buffer );
193 if( i_pos + i_size == 0 )
195 msg_Dbg( p_xml, "empty XML" );
199 p_buffer[ i_pos + i_size ] = '\0'; /* 0 terminated string */
201 p_root = xtag_new_parse( p_buffer, i_buffer );
204 msg_Warn( p_xml, "couldn't parse XML" );
210 p_reader = malloc( sizeof(xml_reader_t) );
213 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
214 if( !p_reader->p_sys )
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 = NULL;
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 && p_reader->p_sys->b_endtag )
286 return XML_READER_ENDELEM;
287 if( p_reader->p_sys->p_curtag->name )
288 return XML_READER_STARTELEM;
289 if( p_reader->p_sys->p_curtag->pcdata )
290 return XML_READER_TEXT;
291 return XML_READER_NONE;
294 static char *ReaderName( xml_reader_t *p_reader )
296 const char *psz_name;
298 if( !p_reader->p_sys->p_curattr )
300 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
302 fprintf( stderr, "TAG: %s\n", psz_name );
306 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
308 return psz_name ? strdup( psz_name ) : NULL;
311 static char *ReaderValue( xml_reader_t *p_reader )
313 const char *psz_name;
314 if( p_reader->p_sys->p_curtag->pcdata )
317 fprintf( stderr, "%s\n", p_reader->p_sys->p_curtag->pcdata );
319 return strdup( p_reader->p_sys->p_curtag->pcdata );
322 if( !p_reader->p_sys->p_curattr ) return NULL;
325 fprintf( stderr, "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
326 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
329 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
331 return psz_name ? strdup( psz_name ) : NULL;
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 return p_reader->p_sys->p_curattr ? VLC_SUCCESS : VLC_EGENERIC;
344 /*****************************************************************************
345 * XTAG parser functions
346 *****************************************************************************/
348 static XList *xlist_append( XList *list, void *data )
352 l = (XList *)xmalloc( sizeof(XList) );
353 l->prev = l->next = NULL;
359 /* Find the last element */
369 static void xlist_free( XList *list )
373 for( l = list; l; l = ln )
380 /* Character classes */
382 #define X_WHITESPACE 1<<0
383 #define X_OPENTAG 1<<1
384 #define X_CLOSETAG 1<<2
385 #define X_DQUOTE 1<<3
386 #define X_SQUOTE 1<<4
393 static int xtag_cin( char c, int char_class )
395 if( char_class & X_WHITESPACE ) if( isspace(c) ) return true;
396 if( char_class & X_OPENTAG ) if( c == '<' ) return true;
397 if( char_class & X_CLOSETAG ) if( c == '>' ) return true;
398 if( char_class & X_DQUOTE ) if( c == '"' ) return true;
399 if( char_class & X_SQUOTE ) if( c == '\'' ) return true;
400 if( char_class & X_EQUAL ) if( c == '=' ) return true;
401 if( char_class & X_SLASH ) if( c == '/' ) return true;
402 if( char_class & X_QMARK ) if( c == '?' ) return true;
403 if( char_class & X_DASH ) if( c == '-' ) return true;
404 if( char_class & X_EMARK ) if( c == '!' ) return true;
409 static int xtag_index( XTagParser *parser, int char_class )
411 char *s = parser->start;
414 for( i = 0; s[i] && s != parser->end; i++ )
416 if( xtag_cin( s[i], char_class ) ) return i;
422 static void xtag_skip_over( XTagParser *parser, int char_class )
424 char *s = parser->start;
427 if( !parser->valid ) return;
429 for( i = 0; s[i] && s != parser->end; i++ )
431 if( !xtag_cin( s[i], char_class ) )
433 parser->start = &s[i];
441 static void xtag_skip_whitespace( XTagParser * parser )
443 xtag_skip_over( parser, X_WHITESPACE );
446 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
448 char *ret, *s = parser->start;
451 if( !parser->valid ) return NULL;
453 xi = xtag_index( parser, good_end | bad_end );
455 if( xi > 0 && xtag_cin (s[xi], good_end) )
457 ret = xmalloc( xi+1 );
458 strncpy( ret, s, xi );
460 parser->start = &s[xi];
467 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
469 char *s = parser->start;
471 if( !parser->valid ) return false;
473 if( !xtag_cin( s[0], char_class ) )
475 parser->valid = false;
479 parser->start = &s[1];
484 static char *xtag_slurp_quoted( XTagParser *parser )
487 int quote = X_DQUOTE; /* quote char to match on */
490 if( !parser->valid ) return NULL;
492 xtag_skip_whitespace( parser );
496 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
498 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
502 for( xi = 0; s[xi]; xi++ )
504 if( xtag_cin( s[xi], quote ) )
506 if( !(xi > 1 && s[xi-1] == '\\') ) break;
510 ret = xmalloc( xi+1 );
511 strncpy( ret, s, xi );
513 parser->start = &s[xi];
515 if( !xtag_assert_and_pass( parser, quote ) )
524 static XAttribute *xtag_parse_attribute( XTagParser *parser )
533 xtag_skip_whitespace( parser );
535 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
539 xtag_skip_whitespace( parser );
542 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
545 fprintf( stderr, "xtag: attr failed EQUAL on <%s>\n", name );
550 xtag_skip_whitespace( parser );
552 value = xtag_slurp_quoted( parser );
557 fprintf (stderr, "Got NULL quoted attribute value\n");
562 attr = xmalloc( sizeof (*attr) );
569 parser->valid = false;
573 static XTag *xtag_parse_tag( XTagParser *parser )
582 if( !parser->valid ) return NULL;
586 /* if this starts a comment tag, skip until end */
587 if( (parser->end - parser->start) > 7 &&
588 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
589 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
591 parser->start = s = &s[4];
592 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
594 parser->start = s = &s[xi+1];
595 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
597 parser->start = &s[2];
598 xtag_skip_whitespace( parser );
599 return xtag_parse_tag( parser );
605 /* ignore processing instructions '<?' ... '?>' */
606 if( (parser->end - parser->start) > 4 &&
607 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
609 parser->start = s = &s[2];
610 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
611 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
612 parser->start = &s[xi+2];
613 xtag_skip_whitespace( parser );
614 return xtag_parse_tag( parser );
620 /* ignore doctype '<!DOCTYPE' ... '>' */
621 if ( (parser->end - parser->start) > 8 &&
622 !strncmp( s, "<!DOCTYPE", 9 ) ) {
623 xi = xtag_index( parser, X_CLOSETAG );
625 parser->start = &s[xi+1];
626 xtag_skip_whitespace( parser );
627 return xtag_parse_tag( parser );
634 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
636 tag = xmalloc( sizeof(*tag) );
638 tag->pcdata = pcdata;
639 tag->parent = parser->current_tag;
640 tag->attributes = NULL;
641 tag->children = NULL;
642 tag->current_child = NULL;
647 /* if this starts a close tag, return NULL and let the parent take it */
648 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
651 /* parse CDATA content */
652 if ( (parser->end - parser->start) > 8 &&
653 !strncmp( s, "<![CDATA[", 9 ) ) {
654 parser->start = s = &s[9];
655 while (parser->end - s > 2) {
656 if (strncmp( s, "]]>", 3 ) == 0) {
657 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
658 if ( !(pcdata = malloc( s - parser->start + 1)) )
663 strncpy( pcdata, parser->start, s - parser->start );
664 pcdata[s - parser->start]='\0';
665 parser->start = &s[3];
667 tag->pcdata = pcdata;
668 tag->parent = parser->current_tag;
669 tag->attributes = NULL;
670 tag->children = NULL;
671 tag->current_child = NULL;
681 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
683 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
684 if( name == NULL ) return NULL;
687 fprintf (stderr, "<%s ...\n", name);
690 tag = xmalloc( sizeof(*tag) );
693 tag->parent = parser->current_tag;
694 tag->attributes = NULL;
695 tag->children = NULL;
696 tag->current_child = NULL;
700 if( xtag_cin( s[0], X_WHITESPACE ) )
702 while( (attr = xtag_parse_attribute( parser )) != NULL )
704 tag->attributes = xlist_append( tag->attributes, attr );
708 xtag_skip_whitespace( parser );
712 if( xtag_cin( s[0], X_CLOSETAG ) )
714 parser->current_tag = tag;
716 xtag_assert_and_pass( parser, X_CLOSETAG );
718 while( (inner = xtag_parse_tag( parser ) ) != NULL )
720 tag->children = xlist_append( tag->children, inner );
723 parser->current_tag = tag->parent;
724 xtag_skip_whitespace( parser );
726 xtag_assert_and_pass( parser, X_OPENTAG );
727 xtag_assert_and_pass( parser, X_SLASH );
728 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
731 if( strcmp( name, tag->name ) )
734 fprintf (stderr, "got %s expected %s\n", name, tag->name);
736 parser->valid = false;
741 xtag_skip_whitespace( parser );
742 xtag_assert_and_pass( parser, X_CLOSETAG );
743 xtag_skip_whitespace( parser );
747 xtag_assert_and_pass( parser, X_SLASH );
748 xtag_assert_and_pass( parser, X_CLOSETAG );
749 xtag_skip_whitespace( parser );
755 static void xtag_free( XTag *xtag )
765 free( xtag->pcdata );
767 for( l = xtag->attributes; l; l = l->next )
769 if( (attr = (XAttribute *)l->data) != NULL )
776 xlist_free( xtag->attributes );
778 for( l = xtag->children; l; l = l->next )
780 child = (XTag *)l->data;
783 xlist_free( xtag->children );
788 static XTag *xtag_new_parse( const char *s, int n )
791 XTag *tag, *ttag, *wrapper;
794 parser.current_tag = NULL;
795 parser.start = (char *)s;
797 if( n == -1 ) parser.end = NULL;
801 fprintf (stderr, "empty buffer\n");
805 else parser.end = (char *)&s[n];
807 /* can't have whitespace pcdata outside rootnode */
808 xtag_skip_whitespace( &parser );
810 tag = xtag_parse_tag( &parser );
815 fprintf (stderr, "invalid file\n");
821 if( (ttag = xtag_parse_tag( &parser )) != NULL )
829 wrapper = xmalloc( sizeof(XTag) );
830 wrapper->name = NULL;
831 wrapper->pcdata = NULL;
832 wrapper->parent = NULL;
833 wrapper->attributes = NULL;
834 wrapper->children = NULL;
835 wrapper->current_child = NULL;
837 wrapper->children = xlist_append( wrapper->children, tag );
838 wrapper->children = xlist_append( wrapper->children, ttag );
840 while( (ttag = xtag_parse_tag( &parser )) != NULL )
848 wrapper->children = xlist_append( wrapper->children, ttag );
856 static char *xtag_get_name( XTag *xtag )
858 return xtag ? xtag->name : NULL;
862 static char *xtag_get_pcdata( XTag *xtag )
867 if( xtag == NULL ) return NULL;
869 for( l = xtag->children; l; l = l->next )
871 child = (XTag *)l->data;
872 if( child->pcdata != NULL )
874 return child->pcdata;
881 static char *xtag_get_attribute( XTag *xtag, char *attribute )
886 if( xtag == NULL ) return NULL;
888 for( l = xtag->attributes; l; l = l->next )
890 if( (attr = (XAttribute *)l->data) != NULL )
892 if( !strcmp( attr->name, attribute ) ) return attr->value;
900 static XTag *xtag_first_child( XTag *xtag, char *name )
905 if( xtag == NULL ) return NULL;
906 if( (l = xtag->children) == NULL ) return NULL;
910 xtag->current_child = l;
911 return (XTag *)l->data;
914 for( ; l; l = l->next )
916 child = (XTag *)l->data;
918 if( !strcmp( child->name, name ) )
920 xtag->current_child = l;
925 xtag->current_child = NULL;
930 static XTag *xtag_next_child( XTag *xtag, char *name )
935 if( xtag == NULL ) return NULL;
937 if( (l = xtag->current_child) == NULL )
938 return xtag_first_child( xtag, name );
940 if( (l = l->next) == NULL ) return NULL;
944 xtag->current_child = l;
945 return (XTag *)l->data;
948 for( ; l; l = l->next )
950 child = (XTag *)l->data;
952 if( !strcmp( child->name, name ) )
954 xtag->current_child = l;
959 xtag->current_child = NULL;
966 * This snprints function takes a variable list of char *, the last of
967 * which must be NULL, and prints each in turn to buf.
968 * Returns C99-style total length that would have been written, even if
969 * this is larger than n.
971 static int xtag_snprints( char *buf, int n, ... )
975 int len, to_copy, total = 0;
979 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
983 if( (to_copy = __MIN(n, len) ) > 0 )
985 memcpy( buf, s, to_copy );
998 static int xtag_snprint( char *buf, int n, XTag *xtag )
1000 int nn, written = 0;
1005 #define FORWARD(N) \
1006 buf += __MIN(n, N); \
1007 n = __MAX(n-N, 0); \
1012 if( n > 0 ) buf[0] = '\0';
1018 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
1026 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1029 for( l = xtag->attributes; l; l = l->next )
1031 attr = (XAttribute *)l->data;
1033 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1038 if( xtag->children == NULL )
1040 nn = xtag_snprints ( buf, n, "/>", NULL );
1046 nn = xtag_snprints( buf, n, ">", NULL );
1050 for( l = xtag->children; l; l = l->next )
1052 child = (XTag *)l->data;
1054 nn = xtag_snprint( buf, n, child );
1060 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );