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 *****************************************************************************/
36 #include "vlc_block.h"
37 #include "vlc_stream.h"
52 * struct XTag is kind of a union ... it normally represents a whole
53 * tag (and its children), but it could alternatively represent some
54 * PCDATA. Basically, if tag->pcdata is non-NULL, interpret only it and
55 * ignore the name, attributes and inner_tags.
67 typedef struct _XAttribute
73 typedef struct _XTagParser
75 int valid; /* boolean */
81 /*****************************************************************************
83 *****************************************************************************/
84 static int Open ( vlc_object_t * );
85 static void Close( vlc_object_t * );
88 set_description( _("Simple XML Parser") );
89 set_capability( "xml", 5 );
90 set_callbacks( Open, Close );
93 struct xml_reader_sys_t
95 XTag *p_root; /* Root tag */
96 XTag *p_curtag; /* Current tag */
97 XList *p_curattr; /* Current attribute */
101 static xml_reader_t *ReaderCreate( xml_t *, stream_t * );
102 static void ReaderDelete( xml_reader_t * );
103 static int ReaderRead( xml_reader_t * );
104 static int ReaderNodeType( xml_reader_t * );
105 static char *ReaderName( xml_reader_t * );
106 static char *ReaderValue( xml_reader_t * );
107 static int ReaderNextAttr( xml_reader_t * );
109 static int ReaderUseDTD ( xml_reader_t *, vlc_bool_t );
111 static void CatalogLoad( xml_t *, const char * );
112 static void CatalogAdd( xml_t *, const char *, const char *, const char * );
114 static XTag *xtag_new_parse( const char *, int );
115 static char *xtag_get_name( XTag * );
117 static char *xtag_get_pcdata( XTag * );
118 static char *xtag_get_attribute( XTag *, char * );
120 static XTag *xtag_first_child( XTag *, char * );
121 static XTag *xtag_next_child( XTag *, char * );
122 static XTag *xtag_free( XTag * );
123 static int xtag_snprint( char *, int, XTag * );
125 /*****************************************************************************
126 * Module initialization
127 *****************************************************************************/
128 static int Open( vlc_object_t *p_this )
130 xml_t *p_xml = (xml_t *)p_this;
132 p_xml->pf_reader_create = ReaderCreate;
133 p_xml->pf_reader_delete = ReaderDelete;
135 p_xml->pf_catalog_load = CatalogLoad;
136 p_xml->pf_catalog_add = CatalogAdd;
141 /*****************************************************************************
142 * Module deinitialization
143 *****************************************************************************/
144 static void Close( vlc_object_t *p_this )
150 /*****************************************************************************
151 * Catalogue functions
152 *****************************************************************************/
153 static void CatalogLoad( xml_t *p_xml, const char *psz_filename )
155 VLC_UNUSED(psz_filename);
156 msg_Dbg( p_xml, "catalog support not implemented" );
159 static void CatalogAdd( xml_t *p_xml, const char *psz_arg1,
160 const char *psz_arg2, const char *psz_filename )
162 VLC_UNUSED(p_xml); VLC_UNUSED(psz_arg1); VLC_UNUSED(psz_arg2);
163 VLC_UNUSED(psz_filename);
166 /*****************************************************************************
168 *****************************************************************************/
169 static xml_reader_t *ReaderCreate( xml_t *p_xml, stream_t *s )
171 xml_reader_t *p_reader;
172 char *p_buffer, *p_new;
173 int i_size, i_pos = 0, i_buffer = 2048;
176 /* Open and read file */
177 p_buffer = malloc( i_buffer );
178 if( p_buffer == NULL ) {
179 msg_Err( p_xml, "out of memory" );
183 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
187 p_new = realloc( p_buffer, i_buffer );
189 msg_Err( p_xml, "out of memory" );
195 p_buffer[ i_pos + i_size ] = 0; /* 0 terminated string */
197 if( i_pos + i_size == 0 )
199 msg_Dbg( p_xml, "empty XML" );
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) );
214 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
215 p_reader->p_sys->p_root = p_root;
216 p_reader->p_sys->p_curtag = NULL;
217 p_reader->p_sys->p_curattr = NULL;
218 p_reader->p_sys->b_endtag = VLC_FALSE;
219 p_reader->p_xml = p_xml;
221 p_reader->pf_read = ReaderRead;
222 p_reader->pf_node_type = ReaderNodeType;
223 p_reader->pf_name = ReaderName;
224 p_reader->pf_value = ReaderValue;
225 p_reader->pf_next_attr = ReaderNextAttr;
226 p_reader->pf_use_dtd = ReaderUseDTD;
231 static void ReaderDelete( xml_reader_t *p_reader )
233 xtag_free( p_reader->p_sys->p_root );
234 free( p_reader->p_sys );
238 static int ReaderUseDTD ( xml_reader_t *p_reader, vlc_bool_t b_use )
240 VLC_UNUSED(p_reader); VLC_UNUSED(b_use);
244 static int ReaderRead( xml_reader_t *p_reader )
248 if( !p_reader->p_sys->p_curtag )
250 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
256 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
258 p_reader->p_sys->p_curtag = p_child;
259 p_reader->p_sys->p_curattr = 0;
260 p_reader->p_sys->b_endtag = VLC_FALSE;
264 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
265 !p_reader->p_sys->b_endtag )
267 p_reader->p_sys->b_endtag = VLC_TRUE;
271 p_reader->p_sys->b_endtag = VLC_FALSE;
272 if( !p_reader->p_sys->p_curtag->parent ) return 0;
273 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
279 static int ReaderNodeType( xml_reader_t *p_reader )
281 if( p_reader->p_sys->p_curtag->name &&
282 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
283 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
284 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
285 return XML_READER_NONE;
288 static char *ReaderName( xml_reader_t *p_reader )
290 const char *psz_name;
292 if( !p_reader->p_sys->p_curattr )
294 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
296 printf( "TAG: %s\n", psz_name );
300 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
302 if( psz_name ) return strdup( psz_name );
306 static char *ReaderValue( xml_reader_t *p_reader )
308 const char *psz_name;
309 if( p_reader->p_sys->p_curtag->pcdata )
312 printf( "%s\n", p_reader->p_sys->p_curtag->pcdata );
314 return strdup( p_reader->p_sys->p_curtag->pcdata );
317 if( !p_reader->p_sys->p_curattr ) return 0;
320 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
321 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
324 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
326 if( psz_name ) return strdup( psz_name );
330 static int ReaderNextAttr( xml_reader_t *p_reader )
332 if( !p_reader->p_sys->p_curattr )
333 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
334 else if( p_reader->p_sys->p_curattr )
335 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
337 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
338 else return VLC_EGENERIC;
341 /*****************************************************************************
342 * XTAG parser functions
343 *****************************************************************************/
345 static XList *xlist_append( XList *list, void *data )
349 l = (XList *)malloc( sizeof(XList) );
350 l->prev = l->next = NULL;
353 if( list == NULL ) return l;
355 for( last = list; last; last = last->next )
356 if( last->next == NULL ) break;
358 if( last ) last->next = l;
363 static void xlist_free( XList *list )
367 for( l = list; l; l = ln )
374 /* Character classes */
376 #define X_WHITESPACE 1<<0
377 #define X_OPENTAG 1<<1
378 #define X_CLOSETAG 1<<2
379 #define X_DQUOTE 1<<3
380 #define X_SQUOTE 1<<4
387 static int xtag_cin( char c, int char_class )
389 if( char_class & X_WHITESPACE ) if( isspace(c) ) return VLC_TRUE;
390 if( char_class & X_OPENTAG ) if( c == '<' ) return VLC_TRUE;
391 if( char_class & X_CLOSETAG ) if( c == '>' ) return VLC_TRUE;
392 if( char_class & X_DQUOTE ) if( c == '"' ) return VLC_TRUE;
393 if( char_class & X_SQUOTE ) if( c == '\'' ) return VLC_TRUE;
394 if( char_class & X_EQUAL ) if( c == '=' ) return VLC_TRUE;
395 if( char_class & X_SLASH ) if( c == '/' ) return VLC_TRUE;
396 if( char_class & X_QMARK ) if( c == '?' ) return VLC_TRUE;
397 if( char_class & X_DASH ) if( c == '-' ) return VLC_TRUE;
398 if( char_class & X_EMARK ) if( c == '!' ) return VLC_TRUE;
403 static int xtag_index( XTagParser *parser, int char_class )
405 char *s = parser->start;
408 for( i = 0; s[i] && s != parser->end; i++ )
410 if( xtag_cin( s[i], char_class ) ) return i;
416 static void xtag_skip_over( XTagParser *parser, int char_class )
418 char *s = parser->start;
421 if( !parser->valid ) return;
423 for( i = 0; s[i] && s != parser->end; i++ )
425 if( !xtag_cin( s[i], char_class ) )
427 parser->start = &s[i];
435 static void xtag_skip_whitespace( XTagParser * parser )
437 xtag_skip_over( parser, X_WHITESPACE );
440 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
442 char *ret, *s = parser->start;
445 if( !parser->valid ) return NULL;
447 xi = xtag_index( parser, good_end | bad_end );
449 if( xi > 0 && xtag_cin (s[xi], good_end) )
451 ret = malloc( (xi+1) * sizeof(char) );
452 strncpy( ret, s, xi );
454 parser->start = &s[xi];
461 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
463 char *s = parser->start;
465 if( !parser->valid ) return VLC_FALSE;
467 if( !xtag_cin( s[0], char_class ) )
469 parser->valid = VLC_FALSE;
473 parser->start = &s[1];
478 static char *xtag_slurp_quoted( XTagParser *parser )
481 int quote = X_DQUOTE; /* quote char to match on */
484 if( !parser->valid ) return NULL;
486 xtag_skip_whitespace( parser );
490 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
492 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
496 for( xi = 0; s[xi]; xi++ )
498 if( xtag_cin( s[xi], quote ) )
500 if( !(xi > 1 && s[xi-1] == '\\') ) break;
504 ret = malloc( (xi+1) * sizeof(char) );
505 strncpy( ret, s, xi );
507 parser->start = &s[xi];
509 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
514 static XAttribute *xtag_parse_attribute( XTagParser *parser )
520 if( !parser->valid ) return NULL;
522 xtag_skip_whitespace( parser );
524 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
525 if( name == NULL ) return NULL;
527 xtag_skip_whitespace( parser );
530 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
533 printf( "xtag: attr failed EQUAL on <%s>\n", name );
538 xtag_skip_whitespace( parser );
540 value = xtag_slurp_quoted( parser );
545 printf ("Got NULL quoted attribute value\n");
550 attr = malloc( sizeof (*attr) );
557 parser->valid = VLC_FALSE;
561 static XTag *xtag_parse_tag( XTagParser *parser )
570 if( !parser->valid ) return NULL;
574 /* if this starts a comment tag, skip until end */
575 if( (parser->end - parser->start) > 7 &&
576 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
577 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
579 parser->start = s = &s[4];
580 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
582 parser->start = s = &s[xi+1];
583 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
585 parser->start = &s[2];
586 xtag_skip_whitespace( parser );
587 return xtag_parse_tag( parser );
593 /* ignore processing instructions '<?' ... '?>' */
594 if( (parser->end - parser->start) > 4 &&
595 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
597 parser->start = s = &s[2];
598 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
599 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
600 parser->start = &s[xi+2];
601 xtag_skip_whitespace( parser );
602 return xtag_parse_tag( parser );
608 /* ignore doctype '<!DOCTYPE' ... '>' */
609 if ( (parser->end - parser->start) > 8 &&
610 !strncmp( s, "<!DOCTYPE", 9 ) ) {
611 xi = xtag_index( parser, X_CLOSETAG );
613 parser->start = s = &s[xi+1];
614 xtag_skip_whitespace( parser );
615 return xtag_parse_tag( parser );
622 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
624 tag = malloc( sizeof(*tag) );
626 tag->pcdata = pcdata;
627 tag->parent = parser->current_tag;
628 tag->attributes = NULL;
629 tag->children = NULL;
630 tag->current_child = NULL;
635 /* if this starts a close tag, return NULL and let the parent take it */
636 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
639 /* parse CDATA content */
640 if ( (parser->end - parser->start) > 8 &&
641 !strncmp( s, "<![CDATA[", 9 ) ) {
642 parser->start = s = &s[9];
643 while (parser->end - s > 2) {
644 if (strncmp( s, "]]>", 3 ) == 0) {
645 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
646 if ( !(pcdata = malloc( sizeof(char)*(s - parser->start + 1))) ) return NULL;
647 strncpy( pcdata, parser->start, s - parser->start );
648 pcdata[s - parser->start]='\0';
649 parser->start = s = &s[3];
651 tag->pcdata = pcdata;
652 tag->parent = parser->current_tag;
653 tag->attributes = NULL;
654 tag->children = NULL;
655 tag->current_child = NULL;
665 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
667 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
668 if( name == NULL ) return NULL;
671 printf ("<%s ...\n", name);
674 tag = malloc( sizeof(*tag) );
677 tag->parent = parser->current_tag;
678 tag->attributes = NULL;
679 tag->children = NULL;
680 tag->current_child = NULL;
684 if( xtag_cin( s[0], X_WHITESPACE ) )
686 while( (attr = xtag_parse_attribute( parser )) != NULL )
688 tag->attributes = xlist_append( tag->attributes, attr );
692 xtag_skip_whitespace( parser );
696 if( xtag_cin( s[0], X_CLOSETAG ) )
698 parser->current_tag = tag;
700 xtag_assert_and_pass( parser, X_CLOSETAG );
702 while( (inner = xtag_parse_tag( parser ) ) != NULL )
704 tag->children = xlist_append( tag->children, inner );
707 parser->current_tag = tag->parent;
708 xtag_skip_whitespace( parser );
710 xtag_assert_and_pass( parser, X_OPENTAG );
711 xtag_assert_and_pass( parser, X_SLASH );
712 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
715 if( strcmp( name, tag->name ) )
718 printf ("got %s expected %s\n", name, tag->name);
720 parser->valid = VLC_FALSE;
725 xtag_skip_whitespace( parser );
726 xtag_assert_and_pass( parser, X_CLOSETAG );
727 xtag_skip_whitespace( parser );
731 xtag_assert_and_pass( parser, X_SLASH );
732 xtag_assert_and_pass( parser, X_CLOSETAG );
733 xtag_skip_whitespace( parser );
739 static XTag *xtag_free( XTag *xtag )
745 if( xtag == NULL ) return NULL;
747 if( xtag->name ) free( xtag->name );
748 if( xtag->pcdata ) free( xtag->pcdata );
750 for( l = xtag->attributes; l; l = l->next )
752 if( (attr = (XAttribute *)l->data) != NULL )
754 if( attr->name ) free( attr->name );
755 if( attr->value ) free( attr->value );
759 xlist_free( xtag->attributes );
761 for( l = xtag->children; l; l = l->next )
763 child = (XTag *)l->data;
766 xlist_free( xtag->children );
773 static XTag *xtag_new_parse( const char *s, int n )
776 XTag *tag, *ttag, *wrapper;
778 parser.valid = VLC_TRUE;
779 parser.current_tag = NULL;
780 parser.start = (char *)s;
782 if( n == -1 ) parser.end = NULL;
786 printf ("empty buffer");
790 else parser.end = (char *)&s[n];
792 /* can't have whitespace pcdata outside rootnode */
793 xtag_skip_whitespace( &parser );
795 tag = xtag_parse_tag( &parser );
800 printf ("invalid file");
806 if( (ttag = xtag_parse_tag( &parser )) != NULL )
814 wrapper = malloc( sizeof(XTag) );
815 wrapper->name = NULL;
816 wrapper->pcdata = NULL;
817 wrapper->parent = NULL;
818 wrapper->attributes = NULL;
819 wrapper->children = NULL;
820 wrapper->current_child = NULL;
822 wrapper->children = xlist_append( wrapper->children, tag );
823 wrapper->children = xlist_append( wrapper->children, ttag );
825 while( (ttag = xtag_parse_tag( &parser )) != NULL )
833 wrapper->children = xlist_append( wrapper->children, ttag );
841 static char *xtag_get_name( XTag *xtag )
843 return xtag ? xtag->name : NULL;
847 static char *xtag_get_pcdata( XTag *xtag )
852 if( xtag == NULL ) return NULL;
854 for( l = xtag->children; l; l = l->next )
856 child = (XTag *)l->data;
857 if( child->pcdata != NULL )
859 return child->pcdata;
866 static char *xtag_get_attribute( XTag *xtag, char *attribute )
871 if( xtag == NULL ) return NULL;
873 for( l = xtag->attributes; l; l = l->next )
875 if( (attr = (XAttribute *)l->data) != NULL )
877 if( !strcmp( attr->name, attribute ) ) return attr->value;
885 static XTag *xtag_first_child( XTag *xtag, char *name )
890 if( xtag == NULL ) return NULL;
891 if( (l = xtag->children) == NULL ) return NULL;
895 xtag->current_child = l;
896 return (XTag *)l->data;
899 for( ; l; l = l->next )
901 child = (XTag *)l->data;
903 if( !strcmp( child->name, name ) )
905 xtag->current_child = l;
910 xtag->current_child = NULL;
915 static XTag *xtag_next_child( XTag *xtag, char *name )
920 if( xtag == NULL ) return NULL;
922 if( (l = xtag->current_child) == NULL )
923 return xtag_first_child( xtag, name );
925 if( (l = l->next) == NULL ) return NULL;
929 xtag->current_child = l;
930 return (XTag *)l->data;
933 for( ; l; l = l->next )
935 child = (XTag *)l->data;
937 if( !strcmp( child->name, name ) )
939 xtag->current_child = l;
944 xtag->current_child = NULL;
950 * This snprints function takes a variable list of char *, the last of
951 * which must be NULL, and prints each in turn to buf.
952 * Returns C99-style total length that would have been written, even if
953 * this is larger than n.
955 static int xtag_snprints( char *buf, int n, ... )
959 int len, to_copy, total = 0;
963 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
967 if( (to_copy = __MIN(n, len) ) > 0 )
969 memcpy( buf, s, to_copy );
982 static int xtag_snprint( char *buf, int n, XTag *xtag )
990 buf += __MIN(n, N); \
996 if( n > 0 ) buf[0] = '\0';
1002 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
1010 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1013 for( l = xtag->attributes; l; l = l->next )
1015 attr = (XAttribute *)l->data;
1017 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1022 if( xtag->children == NULL )
1024 nn = xtag_snprints ( buf, n, "/>", NULL );
1030 nn = xtag_snprints( buf, n, ">", NULL );
1034 for( l = xtag->children; l; l = l->next )
1036 child = (XTag *)l->data;
1038 nn = xtag_snprint( buf, n, child );
1044 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );