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 void 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 )
184 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
188 p_new = realloc( p_buffer, i_buffer );
196 if( i_pos + i_size == 0 )
198 msg_Dbg( p_xml, "empty XML" );
202 p_buffer[ i_pos + i_size ] = '\0'; /* 0 terminated string */
204 p_root = xtag_new_parse( p_buffer, i_buffer );
207 msg_Warn( p_xml, "couldn't parse XML" );
213 p_reader = malloc( sizeof(xml_reader_t) );
216 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
217 if( !p_reader->p_sys )
222 p_reader->p_sys->p_root = p_root;
223 p_reader->p_sys->p_curtag = NULL;
224 p_reader->p_sys->p_curattr = NULL;
225 p_reader->p_sys->b_endtag = false;
226 p_reader->p_xml = p_xml;
228 p_reader->pf_read = ReaderRead;
229 p_reader->pf_node_type = ReaderNodeType;
230 p_reader->pf_name = ReaderName;
231 p_reader->pf_value = ReaderValue;
232 p_reader->pf_next_attr = ReaderNextAttr;
233 p_reader->pf_use_dtd = ReaderUseDTD;
238 static void ReaderDelete( xml_reader_t *p_reader )
240 xtag_free( p_reader->p_sys->p_root );
241 free( p_reader->p_sys );
245 static int ReaderUseDTD ( xml_reader_t *p_reader, bool b_use )
247 VLC_UNUSED(p_reader); VLC_UNUSED(b_use);
251 static int ReaderRead( xml_reader_t *p_reader )
255 if( !p_reader->p_sys->p_curtag )
257 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
263 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
265 p_reader->p_sys->p_curtag = p_child;
266 p_reader->p_sys->p_curattr = NULL;
267 p_reader->p_sys->b_endtag = false;
271 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
272 !p_reader->p_sys->b_endtag )
274 p_reader->p_sys->b_endtag = true;
278 p_reader->p_sys->b_endtag = false;
279 if( !p_reader->p_sys->p_curtag->parent ) return 0;
280 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
286 static int ReaderNodeType( xml_reader_t *p_reader )
288 if( p_reader->p_sys->p_curtag->name && p_reader->p_sys->b_endtag )
289 return XML_READER_ENDELEM;
290 if( p_reader->p_sys->p_curtag->name )
291 return XML_READER_STARTELEM;
292 if( p_reader->p_sys->p_curtag->pcdata )
293 return XML_READER_TEXT;
294 return XML_READER_NONE;
297 static char *ReaderName( xml_reader_t *p_reader )
299 const char *psz_name;
301 if( !p_reader->p_sys->p_curattr )
303 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
305 fprintf( stderr, "TAG: %s\n", psz_name );
309 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
311 return psz_name ? strdup( psz_name ) : NULL;
314 static char *ReaderValue( xml_reader_t *p_reader )
316 const char *psz_name;
317 if( p_reader->p_sys->p_curtag->pcdata )
320 fprintf( stderr, "%s\n", p_reader->p_sys->p_curtag->pcdata );
322 return strdup( p_reader->p_sys->p_curtag->pcdata );
325 if( !p_reader->p_sys->p_curattr ) return NULL;
328 fprintf( stderr, "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
329 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
332 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
334 return psz_name ? strdup( psz_name ) : NULL;
337 static int ReaderNextAttr( xml_reader_t *p_reader )
339 if( !p_reader->p_sys->p_curattr )
340 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
341 else if( p_reader->p_sys->p_curattr )
342 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
344 return p_reader->p_sys->p_curattr ? VLC_SUCCESS : VLC_EGENERIC;
347 /*****************************************************************************
348 * XTAG parser functions
349 *****************************************************************************/
351 static XList *xlist_append( XList *list, void *data )
355 l = (XList *)malloc( sizeof(XList) );
356 l->prev = l->next = NULL;
362 for( last = list; last; last = last->next )
368 if( last ) last->next = l;
373 static void xlist_free( XList *list )
377 for( l = list; l; l = ln )
384 /* Character classes */
386 #define X_WHITESPACE 1<<0
387 #define X_OPENTAG 1<<1
388 #define X_CLOSETAG 1<<2
389 #define X_DQUOTE 1<<3
390 #define X_SQUOTE 1<<4
397 static int xtag_cin( char c, int char_class )
399 if( char_class & X_WHITESPACE ) if( isspace(c) ) return true;
400 if( char_class & X_OPENTAG ) if( c == '<' ) return true;
401 if( char_class & X_CLOSETAG ) if( c == '>' ) return true;
402 if( char_class & X_DQUOTE ) if( c == '"' ) return true;
403 if( char_class & X_SQUOTE ) if( c == '\'' ) return true;
404 if( char_class & X_EQUAL ) if( c == '=' ) return true;
405 if( char_class & X_SLASH ) if( c == '/' ) return true;
406 if( char_class & X_QMARK ) if( c == '?' ) return true;
407 if( char_class & X_DASH ) if( c == '-' ) return true;
408 if( char_class & X_EMARK ) if( c == '!' ) return true;
413 static int xtag_index( XTagParser *parser, int char_class )
415 char *s = parser->start;
418 for( i = 0; s[i] && s != parser->end; i++ )
420 if( xtag_cin( s[i], char_class ) ) return i;
426 static void xtag_skip_over( XTagParser *parser, int char_class )
428 char *s = parser->start;
431 if( !parser->valid ) return;
433 for( i = 0; s[i] && s != parser->end; i++ )
435 if( !xtag_cin( s[i], char_class ) )
437 parser->start = &s[i];
445 static void xtag_skip_whitespace( XTagParser * parser )
447 xtag_skip_over( parser, X_WHITESPACE );
450 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
452 char *ret, *s = parser->start;
455 if( !parser->valid ) return NULL;
457 xi = xtag_index( parser, good_end | bad_end );
459 if( xi > 0 && xtag_cin (s[xi], good_end) )
461 ret = malloc( xi+1 );
462 strncpy( ret, s, xi );
464 parser->start = &s[xi];
471 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
473 char *s = parser->start;
475 if( !parser->valid ) return false;
477 if( !xtag_cin( s[0], char_class ) )
479 parser->valid = false;
483 parser->start = &s[1];
488 static char *xtag_slurp_quoted( XTagParser *parser )
491 int quote = X_DQUOTE; /* quote char to match on */
494 if( !parser->valid ) return NULL;
496 xtag_skip_whitespace( parser );
500 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
502 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
506 for( xi = 0; s[xi]; xi++ )
508 if( xtag_cin( s[xi], quote ) )
510 if( !(xi > 1 && s[xi-1] == '\\') ) break;
514 ret = malloc( xi+1 );
515 strncpy( ret, s, xi );
517 parser->start = &s[xi];
519 if( !xtag_assert_and_pass( parser, quote ) )
528 static XAttribute *xtag_parse_attribute( XTagParser *parser )
537 xtag_skip_whitespace( parser );
539 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
543 xtag_skip_whitespace( parser );
546 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
549 fprintf( stderr, "xtag: attr failed EQUAL on <%s>\n", name );
554 xtag_skip_whitespace( parser );
556 value = xtag_slurp_quoted( parser );
561 fprintf (stderr, "Got NULL quoted attribute value\n");
566 attr = malloc( sizeof (*attr) );
573 parser->valid = false;
577 static XTag *xtag_parse_tag( XTagParser *parser )
586 if( !parser->valid ) return NULL;
590 /* if this starts a comment tag, skip until end */
591 if( (parser->end - parser->start) > 7 &&
592 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
593 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
595 parser->start = s = &s[4];
596 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
598 parser->start = s = &s[xi+1];
599 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
601 parser->start = &s[2];
602 xtag_skip_whitespace( parser );
603 return xtag_parse_tag( parser );
609 /* ignore processing instructions '<?' ... '?>' */
610 if( (parser->end - parser->start) > 4 &&
611 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
613 parser->start = s = &s[2];
614 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
615 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
616 parser->start = &s[xi+2];
617 xtag_skip_whitespace( parser );
618 return xtag_parse_tag( parser );
624 /* ignore doctype '<!DOCTYPE' ... '>' */
625 if ( (parser->end - parser->start) > 8 &&
626 !strncmp( s, "<!DOCTYPE", 9 ) ) {
627 xi = xtag_index( parser, X_CLOSETAG );
629 parser->start = s = &s[xi+1];
630 xtag_skip_whitespace( parser );
631 return xtag_parse_tag( parser );
638 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
640 tag = malloc( sizeof(*tag) );
642 tag->pcdata = pcdata;
643 tag->parent = parser->current_tag;
644 tag->attributes = NULL;
645 tag->children = NULL;
646 tag->current_child = NULL;
651 /* if this starts a close tag, return NULL and let the parent take it */
652 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
655 /* parse CDATA content */
656 if ( (parser->end - parser->start) > 8 &&
657 !strncmp( s, "<![CDATA[", 9 ) ) {
658 parser->start = s = &s[9];
659 while (parser->end - s > 2) {
660 if (strncmp( s, "]]>", 3 ) == 0) {
661 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
662 if ( !(pcdata = malloc( s - parser->start + 1)) )
667 strncpy( pcdata, parser->start, s - parser->start );
668 pcdata[s - parser->start]='\0';
669 parser->start = s = &s[3];
671 tag->pcdata = pcdata;
672 tag->parent = parser->current_tag;
673 tag->attributes = NULL;
674 tag->children = NULL;
675 tag->current_child = NULL;
685 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
687 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
688 if( name == NULL ) return NULL;
691 fprintf (stderr, "<%s ...\n", name);
694 tag = malloc( sizeof(*tag) );
697 tag->parent = parser->current_tag;
698 tag->attributes = NULL;
699 tag->children = NULL;
700 tag->current_child = NULL;
704 if( xtag_cin( s[0], X_WHITESPACE ) )
706 while( (attr = xtag_parse_attribute( parser )) != NULL )
708 tag->attributes = xlist_append( tag->attributes, attr );
712 xtag_skip_whitespace( parser );
716 if( xtag_cin( s[0], X_CLOSETAG ) )
718 parser->current_tag = tag;
720 xtag_assert_and_pass( parser, X_CLOSETAG );
722 while( (inner = xtag_parse_tag( parser ) ) != NULL )
724 tag->children = xlist_append( tag->children, inner );
727 parser->current_tag = tag->parent;
728 xtag_skip_whitespace( parser );
730 xtag_assert_and_pass( parser, X_OPENTAG );
731 xtag_assert_and_pass( parser, X_SLASH );
732 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
735 if( strcmp( name, tag->name ) )
738 fprintf (stderr, "got %s expected %s\n", name, tag->name);
740 parser->valid = false;
745 xtag_skip_whitespace( parser );
746 xtag_assert_and_pass( parser, X_CLOSETAG );
747 xtag_skip_whitespace( parser );
751 xtag_assert_and_pass( parser, X_SLASH );
752 xtag_assert_and_pass( parser, X_CLOSETAG );
753 xtag_skip_whitespace( parser );
759 static void xtag_free( XTag *xtag )
769 free( xtag->pcdata );
771 for( l = xtag->attributes; l; l = l->next )
773 if( (attr = (XAttribute *)l->data) != NULL )
780 xlist_free( xtag->attributes );
782 for( l = xtag->children; l; l = l->next )
784 child = (XTag *)l->data;
787 xlist_free( xtag->children );
792 static XTag *xtag_new_parse( const char *s, int n )
795 XTag *tag, *ttag, *wrapper;
798 parser.current_tag = NULL;
799 parser.start = (char *)s;
801 if( n == -1 ) parser.end = NULL;
805 fprintf (stderr, "empty buffer\n");
809 else parser.end = (char *)&s[n];
811 /* can't have whitespace pcdata outside rootnode */
812 xtag_skip_whitespace( &parser );
814 tag = xtag_parse_tag( &parser );
819 fprintf (stderr, "invalid file\n");
825 if( (ttag = xtag_parse_tag( &parser )) != NULL )
833 wrapper = malloc( sizeof(XTag) );
834 wrapper->name = NULL;
835 wrapper->pcdata = NULL;
836 wrapper->parent = NULL;
837 wrapper->attributes = NULL;
838 wrapper->children = NULL;
839 wrapper->current_child = NULL;
841 wrapper->children = xlist_append( wrapper->children, tag );
842 wrapper->children = xlist_append( wrapper->children, ttag );
844 while( (ttag = xtag_parse_tag( &parser )) != NULL )
852 wrapper->children = xlist_append( wrapper->children, ttag );
860 static char *xtag_get_name( XTag *xtag )
862 return xtag ? xtag->name : NULL;
866 static char *xtag_get_pcdata( XTag *xtag )
871 if( xtag == NULL ) return NULL;
873 for( l = xtag->children; l; l = l->next )
875 child = (XTag *)l->data;
876 if( child->pcdata != NULL )
878 return child->pcdata;
885 static char *xtag_get_attribute( XTag *xtag, char *attribute )
890 if( xtag == NULL ) return NULL;
892 for( l = xtag->attributes; l; l = l->next )
894 if( (attr = (XAttribute *)l->data) != NULL )
896 if( !strcmp( attr->name, attribute ) ) return attr->value;
904 static XTag *xtag_first_child( XTag *xtag, char *name )
909 if( xtag == NULL ) return NULL;
910 if( (l = xtag->children) == NULL ) return NULL;
914 xtag->current_child = l;
915 return (XTag *)l->data;
918 for( ; l; l = l->next )
920 child = (XTag *)l->data;
922 if( !strcmp( child->name, name ) )
924 xtag->current_child = l;
929 xtag->current_child = NULL;
934 static XTag *xtag_next_child( XTag *xtag, char *name )
939 if( xtag == NULL ) return NULL;
941 if( (l = xtag->current_child) == NULL )
942 return xtag_first_child( xtag, name );
944 if( (l = l->next) == NULL ) return NULL;
948 xtag->current_child = l;
949 return (XTag *)l->data;
952 for( ; l; l = l->next )
954 child = (XTag *)l->data;
956 if( !strcmp( child->name, name ) )
958 xtag->current_child = l;
963 xtag->current_child = NULL;
969 * This snprints function takes a variable list of char *, the last of
970 * which must be NULL, and prints each in turn to buf.
971 * Returns C99-style total length that would have been written, even if
972 * this is larger than n.
974 static int xtag_snprints( char *buf, int n, ... )
978 int len, to_copy, total = 0;
982 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
986 if( (to_copy = __MIN(n, len) ) > 0 )
988 memcpy( buf, s, to_copy );
1001 static int xtag_snprint( char *buf, int n, XTag *xtag )
1003 int nn, written = 0;
1008 #define FORWARD(N) \
1009 buf += __MIN(n, N); \
1010 n = __MAX(n-N, 0); \
1015 if( n > 0 ) buf[0] = '\0';
1021 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
1029 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1032 for( l = xtag->attributes; l; l = l->next )
1034 attr = (XAttribute *)l->data;
1036 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1041 if( xtag->children == NULL )
1043 nn = xtag_snprints ( buf, n, "/>", NULL );
1049 nn = xtag_snprints( buf, n, ">", NULL );
1053 for( l = xtag->children; l; l = l->next )
1055 child = (XTag *)l->data;
1057 nn = xtag_snprint( buf, n, child );
1063 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );