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 * );
127 static int xtag_snprint( char *, int, XTag * );
130 /*****************************************************************************
131 * Module initialization
132 *****************************************************************************/
133 static int Open( vlc_object_t *p_this )
135 xml_t *p_xml = (xml_t *)p_this;
137 p_xml->pf_reader_create = ReaderCreate;
138 p_xml->pf_reader_delete = ReaderDelete;
140 p_xml->pf_catalog_load = CatalogLoad;
141 p_xml->pf_catalog_add = CatalogAdd;
146 /*****************************************************************************
147 * Module deinitialization
148 *****************************************************************************/
149 static void Close( vlc_object_t *p_this )
155 /*****************************************************************************
156 * Catalogue functions
157 *****************************************************************************/
158 static void CatalogLoad( xml_t *p_xml, const char *psz_filename )
160 VLC_UNUSED(psz_filename);
161 msg_Dbg( p_xml, "catalog support not implemented" );
164 static void CatalogAdd( xml_t *p_xml, const char *psz_arg1,
165 const char *psz_arg2, const char *psz_filename )
167 VLC_UNUSED(p_xml); VLC_UNUSED(psz_arg1); VLC_UNUSED(psz_arg2);
168 VLC_UNUSED(psz_filename);
171 /*****************************************************************************
173 *****************************************************************************/
174 static xml_reader_t *ReaderCreate( xml_t *p_xml, stream_t *s )
176 xml_reader_t *p_reader;
177 char *p_buffer, *p_new;
178 int i_size, i_pos = 0, i_buffer = 2048;
181 /* Open and read file */
182 p_buffer = malloc( i_buffer );
183 if( p_buffer == NULL )
186 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
190 p_new = realloc( p_buffer, i_buffer );
198 if( i_pos + i_size == 0 )
200 msg_Dbg( p_xml, "empty XML" );
204 p_buffer[ i_pos + i_size ] = '\0'; /* 0 terminated string */
206 p_root = xtag_new_parse( p_buffer, i_buffer );
209 msg_Warn( p_xml, "couldn't parse XML" );
215 p_reader = malloc( sizeof(xml_reader_t) );
218 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
219 if( !p_reader->p_sys )
224 p_reader->p_sys->p_root = p_root;
225 p_reader->p_sys->p_curtag = NULL;
226 p_reader->p_sys->p_curattr = NULL;
227 p_reader->p_sys->b_endtag = false;
228 p_reader->p_xml = p_xml;
230 p_reader->pf_read = ReaderRead;
231 p_reader->pf_node_type = ReaderNodeType;
232 p_reader->pf_name = ReaderName;
233 p_reader->pf_value = ReaderValue;
234 p_reader->pf_next_attr = ReaderNextAttr;
235 p_reader->pf_use_dtd = ReaderUseDTD;
240 static void ReaderDelete( xml_reader_t *p_reader )
242 xtag_free( p_reader->p_sys->p_root );
243 free( p_reader->p_sys );
247 static int ReaderUseDTD ( xml_reader_t *p_reader, bool b_use )
249 VLC_UNUSED(p_reader); VLC_UNUSED(b_use);
253 static int ReaderRead( xml_reader_t *p_reader )
257 if( !p_reader->p_sys->p_curtag )
259 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
265 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
267 p_reader->p_sys->p_curtag = p_child;
268 p_reader->p_sys->p_curattr = NULL;
269 p_reader->p_sys->b_endtag = false;
273 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
274 !p_reader->p_sys->b_endtag )
276 p_reader->p_sys->b_endtag = true;
280 p_reader->p_sys->b_endtag = false;
281 if( !p_reader->p_sys->p_curtag->parent ) return 0;
282 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
288 static int ReaderNodeType( xml_reader_t *p_reader )
290 if( p_reader->p_sys->p_curtag->name && p_reader->p_sys->b_endtag )
291 return XML_READER_ENDELEM;
292 if( p_reader->p_sys->p_curtag->name )
293 return XML_READER_STARTELEM;
294 if( p_reader->p_sys->p_curtag->pcdata )
295 return XML_READER_TEXT;
296 return XML_READER_NONE;
299 static char *ReaderName( xml_reader_t *p_reader )
301 const char *psz_name;
303 if( !p_reader->p_sys->p_curattr )
305 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
307 fprintf( stderr, "TAG: %s\n", psz_name );
311 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
313 return psz_name ? strdup( psz_name ) : NULL;
316 static char *ReaderValue( xml_reader_t *p_reader )
318 const char *psz_name;
319 if( p_reader->p_sys->p_curtag->pcdata )
322 fprintf( stderr, "%s\n", p_reader->p_sys->p_curtag->pcdata );
324 return strdup( p_reader->p_sys->p_curtag->pcdata );
327 if( !p_reader->p_sys->p_curattr ) return NULL;
330 fprintf( stderr, "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
331 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
334 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
336 return psz_name ? strdup( psz_name ) : NULL;
339 static int ReaderNextAttr( xml_reader_t *p_reader )
341 if( !p_reader->p_sys->p_curattr )
342 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
343 else if( p_reader->p_sys->p_curattr )
344 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
346 return p_reader->p_sys->p_curattr ? VLC_SUCCESS : VLC_EGENERIC;
349 /*****************************************************************************
350 * XTAG parser functions
351 *****************************************************************************/
353 static XList *xlist_append( XList *list, void *data )
357 l = (XList *)malloc( sizeof(XList) );
358 l->prev = l->next = NULL;
364 /* Find the last element */
374 static void xlist_free( XList *list )
378 for( l = list; l; l = ln )
385 /* Character classes */
387 #define X_WHITESPACE 1<<0
388 #define X_OPENTAG 1<<1
389 #define X_CLOSETAG 1<<2
390 #define X_DQUOTE 1<<3
391 #define X_SQUOTE 1<<4
398 static int xtag_cin( char c, int char_class )
400 if( char_class & X_WHITESPACE ) if( isspace(c) ) return true;
401 if( char_class & X_OPENTAG ) if( c == '<' ) return true;
402 if( char_class & X_CLOSETAG ) if( c == '>' ) return true;
403 if( char_class & X_DQUOTE ) if( c == '"' ) return true;
404 if( char_class & X_SQUOTE ) if( c == '\'' ) return true;
405 if( char_class & X_EQUAL ) if( c == '=' ) return true;
406 if( char_class & X_SLASH ) if( c == '/' ) return true;
407 if( char_class & X_QMARK ) if( c == '?' ) return true;
408 if( char_class & X_DASH ) if( c == '-' ) return true;
409 if( char_class & X_EMARK ) if( c == '!' ) return true;
414 static int xtag_index( XTagParser *parser, int char_class )
416 char *s = parser->start;
419 for( i = 0; s[i] && s != parser->end; i++ )
421 if( xtag_cin( s[i], char_class ) ) return i;
427 static void xtag_skip_over( XTagParser *parser, int char_class )
429 char *s = parser->start;
432 if( !parser->valid ) return;
434 for( i = 0; s[i] && s != parser->end; i++ )
436 if( !xtag_cin( s[i], char_class ) )
438 parser->start = &s[i];
446 static void xtag_skip_whitespace( XTagParser * parser )
448 xtag_skip_over( parser, X_WHITESPACE );
451 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
453 char *ret, *s = parser->start;
456 if( !parser->valid ) return NULL;
458 xi = xtag_index( parser, good_end | bad_end );
460 if( xi > 0 && xtag_cin (s[xi], good_end) )
462 ret = malloc( xi+1 );
463 strncpy( ret, s, xi );
465 parser->start = &s[xi];
472 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
474 char *s = parser->start;
476 if( !parser->valid ) return false;
478 if( !xtag_cin( s[0], char_class ) )
480 parser->valid = false;
484 parser->start = &s[1];
489 static char *xtag_slurp_quoted( XTagParser *parser )
492 int quote = X_DQUOTE; /* quote char to match on */
495 if( !parser->valid ) return NULL;
497 xtag_skip_whitespace( parser );
501 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
503 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
507 for( xi = 0; s[xi]; xi++ )
509 if( xtag_cin( s[xi], quote ) )
511 if( !(xi > 1 && s[xi-1] == '\\') ) break;
515 ret = malloc( xi+1 );
516 strncpy( ret, s, xi );
518 parser->start = &s[xi];
520 if( !xtag_assert_and_pass( parser, quote ) )
529 static XAttribute *xtag_parse_attribute( XTagParser *parser )
538 xtag_skip_whitespace( parser );
540 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
544 xtag_skip_whitespace( parser );
547 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
550 fprintf( stderr, "xtag: attr failed EQUAL on <%s>\n", name );
555 xtag_skip_whitespace( parser );
557 value = xtag_slurp_quoted( parser );
562 fprintf (stderr, "Got NULL quoted attribute value\n");
567 attr = malloc( sizeof (*attr) );
574 parser->valid = false;
578 static XTag *xtag_parse_tag( XTagParser *parser )
587 if( !parser->valid ) return NULL;
591 /* if this starts a comment tag, skip until end */
592 if( (parser->end - parser->start) > 7 &&
593 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
594 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
596 parser->start = s = &s[4];
597 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
599 parser->start = s = &s[xi+1];
600 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
602 parser->start = &s[2];
603 xtag_skip_whitespace( parser );
604 return xtag_parse_tag( parser );
610 /* ignore processing instructions '<?' ... '?>' */
611 if( (parser->end - parser->start) > 4 &&
612 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
614 parser->start = s = &s[2];
615 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
616 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
617 parser->start = &s[xi+2];
618 xtag_skip_whitespace( parser );
619 return xtag_parse_tag( parser );
625 /* ignore doctype '<!DOCTYPE' ... '>' */
626 if ( (parser->end - parser->start) > 8 &&
627 !strncmp( s, "<!DOCTYPE", 9 ) ) {
628 xi = xtag_index( parser, X_CLOSETAG );
630 parser->start = &s[xi+1];
631 xtag_skip_whitespace( parser );
632 return xtag_parse_tag( parser );
639 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
641 tag = malloc( sizeof(*tag) );
643 tag->pcdata = pcdata;
644 tag->parent = parser->current_tag;
645 tag->attributes = NULL;
646 tag->children = NULL;
647 tag->current_child = NULL;
652 /* if this starts a close tag, return NULL and let the parent take it */
653 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
656 /* parse CDATA content */
657 if ( (parser->end - parser->start) > 8 &&
658 !strncmp( s, "<![CDATA[", 9 ) ) {
659 parser->start = s = &s[9];
660 while (parser->end - s > 2) {
661 if (strncmp( s, "]]>", 3 ) == 0) {
662 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
663 if ( !(pcdata = malloc( s - parser->start + 1)) )
668 strncpy( pcdata, parser->start, s - parser->start );
669 pcdata[s - parser->start]='\0';
670 parser->start = &s[3];
672 tag->pcdata = pcdata;
673 tag->parent = parser->current_tag;
674 tag->attributes = NULL;
675 tag->children = NULL;
676 tag->current_child = NULL;
686 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
688 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
689 if( name == NULL ) return NULL;
692 fprintf (stderr, "<%s ...\n", name);
695 tag = malloc( sizeof(*tag) );
698 tag->parent = parser->current_tag;
699 tag->attributes = NULL;
700 tag->children = NULL;
701 tag->current_child = NULL;
705 if( xtag_cin( s[0], X_WHITESPACE ) )
707 while( (attr = xtag_parse_attribute( parser )) != NULL )
709 tag->attributes = xlist_append( tag->attributes, attr );
713 xtag_skip_whitespace( parser );
717 if( xtag_cin( s[0], X_CLOSETAG ) )
719 parser->current_tag = tag;
721 xtag_assert_and_pass( parser, X_CLOSETAG );
723 while( (inner = xtag_parse_tag( parser ) ) != NULL )
725 tag->children = xlist_append( tag->children, inner );
728 parser->current_tag = tag->parent;
729 xtag_skip_whitespace( parser );
731 xtag_assert_and_pass( parser, X_OPENTAG );
732 xtag_assert_and_pass( parser, X_SLASH );
733 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
736 if( strcmp( name, tag->name ) )
739 fprintf (stderr, "got %s expected %s\n", name, tag->name);
741 parser->valid = false;
746 xtag_skip_whitespace( parser );
747 xtag_assert_and_pass( parser, X_CLOSETAG );
748 xtag_skip_whitespace( parser );
752 xtag_assert_and_pass( parser, X_SLASH );
753 xtag_assert_and_pass( parser, X_CLOSETAG );
754 xtag_skip_whitespace( parser );
760 static void xtag_free( XTag *xtag )
770 free( xtag->pcdata );
772 for( l = xtag->attributes; l; l = l->next )
774 if( (attr = (XAttribute *)l->data) != NULL )
781 xlist_free( xtag->attributes );
783 for( l = xtag->children; l; l = l->next )
785 child = (XTag *)l->data;
788 xlist_free( xtag->children );
793 static XTag *xtag_new_parse( const char *s, int n )
796 XTag *tag, *ttag, *wrapper;
799 parser.current_tag = NULL;
800 parser.start = (char *)s;
802 if( n == -1 ) parser.end = NULL;
806 fprintf (stderr, "empty buffer\n");
810 else parser.end = (char *)&s[n];
812 /* can't have whitespace pcdata outside rootnode */
813 xtag_skip_whitespace( &parser );
815 tag = xtag_parse_tag( &parser );
820 fprintf (stderr, "invalid file\n");
826 if( (ttag = xtag_parse_tag( &parser )) != NULL )
834 wrapper = malloc( sizeof(XTag) );
835 wrapper->name = NULL;
836 wrapper->pcdata = NULL;
837 wrapper->parent = NULL;
838 wrapper->attributes = NULL;
839 wrapper->children = NULL;
840 wrapper->current_child = NULL;
842 wrapper->children = xlist_append( wrapper->children, tag );
843 wrapper->children = xlist_append( wrapper->children, ttag );
845 while( (ttag = xtag_parse_tag( &parser )) != NULL )
853 wrapper->children = xlist_append( wrapper->children, ttag );
861 static char *xtag_get_name( XTag *xtag )
863 return xtag ? xtag->name : NULL;
867 static char *xtag_get_pcdata( XTag *xtag )
872 if( xtag == NULL ) return NULL;
874 for( l = xtag->children; l; l = l->next )
876 child = (XTag *)l->data;
877 if( child->pcdata != NULL )
879 return child->pcdata;
886 static char *xtag_get_attribute( XTag *xtag, char *attribute )
891 if( xtag == NULL ) return NULL;
893 for( l = xtag->attributes; l; l = l->next )
895 if( (attr = (XAttribute *)l->data) != NULL )
897 if( !strcmp( attr->name, attribute ) ) return attr->value;
905 static XTag *xtag_first_child( XTag *xtag, char *name )
910 if( xtag == NULL ) return NULL;
911 if( (l = xtag->children) == NULL ) return NULL;
915 xtag->current_child = l;
916 return (XTag *)l->data;
919 for( ; l; l = l->next )
921 child = (XTag *)l->data;
923 if( !strcmp( child->name, name ) )
925 xtag->current_child = l;
930 xtag->current_child = NULL;
935 static XTag *xtag_next_child( XTag *xtag, char *name )
940 if( xtag == NULL ) return NULL;
942 if( (l = xtag->current_child) == NULL )
943 return xtag_first_child( xtag, name );
945 if( (l = l->next) == NULL ) return NULL;
949 xtag->current_child = l;
950 return (XTag *)l->data;
953 for( ; l; l = l->next )
955 child = (XTag *)l->data;
957 if( !strcmp( child->name, name ) )
959 xtag->current_child = l;
964 xtag->current_child = NULL;
971 * This snprints function takes a variable list of char *, the last of
972 * which must be NULL, and prints each in turn to buf.
973 * Returns C99-style total length that would have been written, even if
974 * this is larger than n.
976 static int xtag_snprints( char *buf, int n, ... )
980 int len, to_copy, total = 0;
984 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
988 if( (to_copy = __MIN(n, len) ) > 0 )
990 memcpy( buf, s, to_copy );
1003 static int xtag_snprint( char *buf, int n, XTag *xtag )
1005 int nn, written = 0;
1010 #define FORWARD(N) \
1011 buf += __MIN(n, N); \
1012 n = __MAX(n-N, 0); \
1017 if( n > 0 ) buf[0] = '\0';
1023 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
1031 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1034 for( l = xtag->attributes; l; l = l->next )
1036 attr = (XAttribute *)l->data;
1038 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1043 if( xtag->children == NULL )
1045 nn = xtag_snprints ( buf, n, "/>", NULL );
1051 nn = xtag_snprints( buf, n, ">", NULL );
1055 for( l = xtag->children; l; l = l->next )
1057 child = (XTag *)l->data;
1059 nn = xtag_snprint( buf, n, child );
1065 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );