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 *****************************************************************************/
34 #include <vlc_plugin.h>
37 #include "vlc_block.h"
38 #include "vlc_stream.h"
53 * struct XTag is kind of a union ... it normally represents a whole
54 * tag (and its children), but it could alternatively represent some
55 * PCDATA. Basically, if tag->pcdata is non-NULL, interpret only it and
56 * ignore the name, attributes and inner_tags.
68 typedef struct _XAttribute
74 typedef struct _XTagParser
76 int valid; /* boolean */
82 /*****************************************************************************
84 *****************************************************************************/
85 static int Open ( vlc_object_t * );
86 static void Close( vlc_object_t * );
89 set_description( N_("Simple XML Parser") );
90 set_capability( "xml", 5 );
91 set_callbacks( Open, Close );
94 struct xml_reader_sys_t
96 XTag *p_root; /* Root tag */
97 XTag *p_curtag; /* Current tag */
98 XList *p_curattr; /* Current attribute */
102 static xml_reader_t *ReaderCreate( xml_t *, stream_t * );
103 static void ReaderDelete( xml_reader_t * );
104 static int ReaderRead( xml_reader_t * );
105 static int ReaderNodeType( xml_reader_t * );
106 static char *ReaderName( xml_reader_t * );
107 static char *ReaderValue( xml_reader_t * );
108 static int ReaderNextAttr( xml_reader_t * );
110 static int ReaderUseDTD ( xml_reader_t *, bool );
112 static void CatalogLoad( xml_t *, const char * );
113 static void CatalogAdd( xml_t *, const char *, const char *, const char * );
115 static XTag *xtag_new_parse( const char *, int );
116 static char *xtag_get_name( XTag * );
118 static char *xtag_get_pcdata( XTag * );
119 static char *xtag_get_attribute( XTag *, char * );
121 static XTag *xtag_first_child( XTag *, char * );
122 static XTag *xtag_next_child( XTag *, char * );
123 static XTag *xtag_free( XTag * );
124 static int xtag_snprint( char *, int, XTag * );
126 /*****************************************************************************
127 * Module initialization
128 *****************************************************************************/
129 static int Open( vlc_object_t *p_this )
131 xml_t *p_xml = (xml_t *)p_this;
133 p_xml->pf_reader_create = ReaderCreate;
134 p_xml->pf_reader_delete = ReaderDelete;
136 p_xml->pf_catalog_load = CatalogLoad;
137 p_xml->pf_catalog_add = CatalogAdd;
142 /*****************************************************************************
143 * Module deinitialization
144 *****************************************************************************/
145 static void Close( vlc_object_t *p_this )
151 /*****************************************************************************
152 * Catalogue functions
153 *****************************************************************************/
154 static void CatalogLoad( xml_t *p_xml, const char *psz_filename )
156 VLC_UNUSED(psz_filename);
157 msg_Dbg( p_xml, "catalog support not implemented" );
160 static void CatalogAdd( xml_t *p_xml, const char *psz_arg1,
161 const char *psz_arg2, const char *psz_filename )
163 VLC_UNUSED(p_xml); VLC_UNUSED(psz_arg1); VLC_UNUSED(psz_arg2);
164 VLC_UNUSED(psz_filename);
167 /*****************************************************************************
169 *****************************************************************************/
170 static xml_reader_t *ReaderCreate( xml_t *p_xml, stream_t *s )
172 xml_reader_t *p_reader;
173 char *p_buffer, *p_new;
174 int i_size, i_pos = 0, i_buffer = 2048;
177 /* Open and read file */
178 p_buffer = malloc( i_buffer );
179 if( p_buffer == NULL ) {
180 msg_Err( p_xml, "out of memory" );
184 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
188 p_new = realloc( p_buffer, i_buffer );
191 msg_Err( p_xml, "out of memory" );
197 p_buffer[ i_pos + i_size ] = 0; /* 0 terminated string */
199 if( i_pos + i_size == 0 )
201 msg_Dbg( p_xml, "empty XML" );
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) );
216 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
217 p_reader->p_sys->p_root = p_root;
218 p_reader->p_sys->p_curtag = NULL;
219 p_reader->p_sys->p_curattr = NULL;
220 p_reader->p_sys->b_endtag = false;
221 p_reader->p_xml = p_xml;
223 p_reader->pf_read = ReaderRead;
224 p_reader->pf_node_type = ReaderNodeType;
225 p_reader->pf_name = ReaderName;
226 p_reader->pf_value = ReaderValue;
227 p_reader->pf_next_attr = ReaderNextAttr;
228 p_reader->pf_use_dtd = ReaderUseDTD;
233 static void ReaderDelete( xml_reader_t *p_reader )
235 xtag_free( p_reader->p_sys->p_root );
236 free( p_reader->p_sys );
240 static int ReaderUseDTD ( xml_reader_t *p_reader, bool b_use )
242 VLC_UNUSED(p_reader); VLC_UNUSED(b_use);
246 static int ReaderRead( xml_reader_t *p_reader )
250 if( !p_reader->p_sys->p_curtag )
252 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
258 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
260 p_reader->p_sys->p_curtag = p_child;
261 p_reader->p_sys->p_curattr = 0;
262 p_reader->p_sys->b_endtag = false;
266 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
267 !p_reader->p_sys->b_endtag )
269 p_reader->p_sys->b_endtag = true;
273 p_reader->p_sys->b_endtag = false;
274 if( !p_reader->p_sys->p_curtag->parent ) return 0;
275 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
281 static int ReaderNodeType( xml_reader_t *p_reader )
283 if( p_reader->p_sys->p_curtag->name &&
284 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
285 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
286 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
287 return XML_READER_NONE;
290 static char *ReaderName( xml_reader_t *p_reader )
292 const char *psz_name;
294 if( !p_reader->p_sys->p_curattr )
296 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
298 printf( "TAG: %s\n", psz_name );
302 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
304 if( psz_name ) return strdup( psz_name );
308 static char *ReaderValue( xml_reader_t *p_reader )
310 const char *psz_name;
311 if( p_reader->p_sys->p_curtag->pcdata )
314 printf( "%s\n", p_reader->p_sys->p_curtag->pcdata );
316 return strdup( p_reader->p_sys->p_curtag->pcdata );
319 if( !p_reader->p_sys->p_curattr ) return 0;
322 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
323 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
326 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
328 if( psz_name ) return strdup( psz_name );
332 static int ReaderNextAttr( xml_reader_t *p_reader )
334 if( !p_reader->p_sys->p_curattr )
335 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
336 else if( p_reader->p_sys->p_curattr )
337 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
339 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
340 else return VLC_EGENERIC;
343 /*****************************************************************************
344 * XTAG parser functions
345 *****************************************************************************/
347 static XList *xlist_append( XList *list, void *data )
351 l = (XList *)malloc( sizeof(XList) );
352 l->prev = l->next = NULL;
355 if( list == NULL ) return l;
357 for( last = list; last; last = last->next )
358 if( last->next == NULL ) break;
360 if( last ) last->next = l;
365 static void xlist_free( XList *list )
369 for( l = list; l; l = ln )
376 /* Character classes */
378 #define X_WHITESPACE 1<<0
379 #define X_OPENTAG 1<<1
380 #define X_CLOSETAG 1<<2
381 #define X_DQUOTE 1<<3
382 #define X_SQUOTE 1<<4
389 static int xtag_cin( char c, int char_class )
391 if( char_class & X_WHITESPACE ) if( isspace(c) ) return true;
392 if( char_class & X_OPENTAG ) if( c == '<' ) return true;
393 if( char_class & X_CLOSETAG ) if( c == '>' ) return true;
394 if( char_class & X_DQUOTE ) if( c == '"' ) return true;
395 if( char_class & X_SQUOTE ) if( c == '\'' ) return true;
396 if( char_class & X_EQUAL ) if( c == '=' ) return true;
397 if( char_class & X_SLASH ) if( c == '/' ) return true;
398 if( char_class & X_QMARK ) if( c == '?' ) return true;
399 if( char_class & X_DASH ) if( c == '-' ) return true;
400 if( char_class & X_EMARK ) if( c == '!' ) return true;
405 static int xtag_index( XTagParser *parser, int char_class )
407 char *s = parser->start;
410 for( i = 0; s[i] && s != parser->end; i++ )
412 if( xtag_cin( s[i], char_class ) ) return i;
418 static void xtag_skip_over( XTagParser *parser, int char_class )
420 char *s = parser->start;
423 if( !parser->valid ) return;
425 for( i = 0; s[i] && s != parser->end; i++ )
427 if( !xtag_cin( s[i], char_class ) )
429 parser->start = &s[i];
437 static void xtag_skip_whitespace( XTagParser * parser )
439 xtag_skip_over( parser, X_WHITESPACE );
442 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
444 char *ret, *s = parser->start;
447 if( !parser->valid ) return NULL;
449 xi = xtag_index( parser, good_end | bad_end );
451 if( xi > 0 && xtag_cin (s[xi], good_end) )
453 ret = malloc( (xi+1) * sizeof(char) );
454 strncpy( ret, s, xi );
456 parser->start = &s[xi];
463 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
465 char *s = parser->start;
467 if( !parser->valid ) return false;
469 if( !xtag_cin( s[0], char_class ) )
471 parser->valid = false;
475 parser->start = &s[1];
480 static char *xtag_slurp_quoted( XTagParser *parser )
483 int quote = X_DQUOTE; /* quote char to match on */
486 if( !parser->valid ) return NULL;
488 xtag_skip_whitespace( parser );
492 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
494 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
498 for( xi = 0; s[xi]; xi++ )
500 if( xtag_cin( s[xi], quote ) )
502 if( !(xi > 1 && s[xi-1] == '\\') ) break;
506 ret = malloc( (xi+1) * sizeof(char) );
507 strncpy( ret, s, xi );
509 parser->start = &s[xi];
511 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
516 static XAttribute *xtag_parse_attribute( XTagParser *parser )
522 if( !parser->valid ) return NULL;
524 xtag_skip_whitespace( parser );
526 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
527 if( name == NULL ) return NULL;
529 xtag_skip_whitespace( parser );
532 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
535 printf( "xtag: attr failed EQUAL on <%s>\n", name );
540 xtag_skip_whitespace( parser );
542 value = xtag_slurp_quoted( parser );
547 printf ("Got NULL quoted attribute value\n");
552 attr = malloc( sizeof (*attr) );
559 parser->valid = false;
563 static XTag *xtag_parse_tag( XTagParser *parser )
572 if( !parser->valid ) return NULL;
576 /* if this starts a comment tag, skip until end */
577 if( (parser->end - parser->start) > 7 &&
578 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
579 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
581 parser->start = s = &s[4];
582 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
584 parser->start = s = &s[xi+1];
585 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
587 parser->start = &s[2];
588 xtag_skip_whitespace( parser );
589 return xtag_parse_tag( parser );
595 /* ignore processing instructions '<?' ... '?>' */
596 if( (parser->end - parser->start) > 4 &&
597 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
599 parser->start = s = &s[2];
600 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
601 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
602 parser->start = &s[xi+2];
603 xtag_skip_whitespace( parser );
604 return xtag_parse_tag( parser );
610 /* ignore doctype '<!DOCTYPE' ... '>' */
611 if ( (parser->end - parser->start) > 8 &&
612 !strncmp( s, "<!DOCTYPE", 9 ) ) {
613 xi = xtag_index( parser, X_CLOSETAG );
615 parser->start = s = &s[xi+1];
616 xtag_skip_whitespace( parser );
617 return xtag_parse_tag( parser );
624 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
626 tag = malloc( sizeof(*tag) );
628 tag->pcdata = pcdata;
629 tag->parent = parser->current_tag;
630 tag->attributes = NULL;
631 tag->children = NULL;
632 tag->current_child = NULL;
637 /* if this starts a close tag, return NULL and let the parent take it */
638 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
641 /* parse CDATA content */
642 if ( (parser->end - parser->start) > 8 &&
643 !strncmp( s, "<![CDATA[", 9 ) ) {
644 parser->start = s = &s[9];
645 while (parser->end - s > 2) {
646 if (strncmp( s, "]]>", 3 ) == 0) {
647 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
648 if ( !(pcdata = malloc( sizeof(char)*(s - parser->start + 1))) ) return NULL;
649 strncpy( pcdata, parser->start, s - parser->start );
650 pcdata[s - parser->start]='\0';
651 parser->start = s = &s[3];
653 tag->pcdata = pcdata;
654 tag->parent = parser->current_tag;
655 tag->attributes = NULL;
656 tag->children = NULL;
657 tag->current_child = NULL;
667 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
669 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
670 if( name == NULL ) return NULL;
673 printf ("<%s ...\n", name);
676 tag = malloc( sizeof(*tag) );
679 tag->parent = parser->current_tag;
680 tag->attributes = NULL;
681 tag->children = NULL;
682 tag->current_child = NULL;
686 if( xtag_cin( s[0], X_WHITESPACE ) )
688 while( (attr = xtag_parse_attribute( parser )) != NULL )
690 tag->attributes = xlist_append( tag->attributes, attr );
694 xtag_skip_whitespace( parser );
698 if( xtag_cin( s[0], X_CLOSETAG ) )
700 parser->current_tag = tag;
702 xtag_assert_and_pass( parser, X_CLOSETAG );
704 while( (inner = xtag_parse_tag( parser ) ) != NULL )
706 tag->children = xlist_append( tag->children, inner );
709 parser->current_tag = tag->parent;
710 xtag_skip_whitespace( parser );
712 xtag_assert_and_pass( parser, X_OPENTAG );
713 xtag_assert_and_pass( parser, X_SLASH );
714 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
717 if( strcmp( name, tag->name ) )
720 printf ("got %s expected %s\n", name, tag->name);
722 parser->valid = false;
727 xtag_skip_whitespace( parser );
728 xtag_assert_and_pass( parser, X_CLOSETAG );
729 xtag_skip_whitespace( parser );
733 xtag_assert_and_pass( parser, X_SLASH );
734 xtag_assert_and_pass( parser, X_CLOSETAG );
735 xtag_skip_whitespace( parser );
741 static XTag *xtag_free( XTag *xtag )
747 if( xtag == NULL ) return NULL;
750 free( xtag->pcdata );
752 for( l = xtag->attributes; l; l = l->next )
754 if( (attr = (XAttribute *)l->data) != NULL )
761 xlist_free( xtag->attributes );
763 for( l = xtag->children; l; l = l->next )
765 child = (XTag *)l->data;
768 xlist_free( xtag->children );
775 static XTag *xtag_new_parse( const char *s, int n )
778 XTag *tag, *ttag, *wrapper;
781 parser.current_tag = NULL;
782 parser.start = (char *)s;
784 if( n == -1 ) parser.end = NULL;
788 printf ("empty buffer");
792 else parser.end = (char *)&s[n];
794 /* can't have whitespace pcdata outside rootnode */
795 xtag_skip_whitespace( &parser );
797 tag = xtag_parse_tag( &parser );
802 printf ("invalid file");
808 if( (ttag = xtag_parse_tag( &parser )) != NULL )
816 wrapper = malloc( sizeof(XTag) );
817 wrapper->name = NULL;
818 wrapper->pcdata = NULL;
819 wrapper->parent = NULL;
820 wrapper->attributes = NULL;
821 wrapper->children = NULL;
822 wrapper->current_child = NULL;
824 wrapper->children = xlist_append( wrapper->children, tag );
825 wrapper->children = xlist_append( wrapper->children, ttag );
827 while( (ttag = xtag_parse_tag( &parser )) != NULL )
835 wrapper->children = xlist_append( wrapper->children, ttag );
843 static char *xtag_get_name( XTag *xtag )
845 return xtag ? xtag->name : NULL;
849 static char *xtag_get_pcdata( XTag *xtag )
854 if( xtag == NULL ) return NULL;
856 for( l = xtag->children; l; l = l->next )
858 child = (XTag *)l->data;
859 if( child->pcdata != NULL )
861 return child->pcdata;
868 static char *xtag_get_attribute( XTag *xtag, char *attribute )
873 if( xtag == NULL ) return NULL;
875 for( l = xtag->attributes; l; l = l->next )
877 if( (attr = (XAttribute *)l->data) != NULL )
879 if( !strcmp( attr->name, attribute ) ) return attr->value;
887 static XTag *xtag_first_child( XTag *xtag, char *name )
892 if( xtag == NULL ) return NULL;
893 if( (l = xtag->children) == NULL ) return NULL;
897 xtag->current_child = l;
898 return (XTag *)l->data;
901 for( ; l; l = l->next )
903 child = (XTag *)l->data;
905 if( !strcmp( child->name, name ) )
907 xtag->current_child = l;
912 xtag->current_child = NULL;
917 static XTag *xtag_next_child( XTag *xtag, char *name )
922 if( xtag == NULL ) return NULL;
924 if( (l = xtag->current_child) == NULL )
925 return xtag_first_child( xtag, name );
927 if( (l = l->next) == NULL ) return NULL;
931 xtag->current_child = l;
932 return (XTag *)l->data;
935 for( ; l; l = l->next )
937 child = (XTag *)l->data;
939 if( !strcmp( child->name, name ) )
941 xtag->current_child = l;
946 xtag->current_child = NULL;
952 * This snprints function takes a variable list of char *, the last of
953 * which must be NULL, and prints each in turn to buf.
954 * Returns C99-style total length that would have been written, even if
955 * this is larger than n.
957 static int xtag_snprints( char *buf, int n, ... )
961 int len, to_copy, total = 0;
965 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
969 if( (to_copy = __MIN(n, len) ) > 0 )
971 memcpy( buf, s, to_copy );
984 static int xtag_snprint( char *buf, int n, XTag *xtag )
992 buf += __MIN(n, N); \
998 if( n > 0 ) buf[0] = '\0';
1004 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
1012 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1015 for( l = xtag->attributes; l; l = l->next )
1017 attr = (XAttribute *)l->data;
1019 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1024 if( xtag->children == NULL )
1026 nn = xtag_snprints ( buf, n, "/>", NULL );
1032 nn = xtag_snprints( buf, n, ">", NULL );
1036 for( l = xtag->children; l; l = l->next )
1038 child = (XTag *)l->data;
1040 nn = xtag_snprint( buf, n, child );
1046 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );