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_block.h"
34 #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_category( CAT_ADVANCED );
89 set_subcategory( SUBCAT_ADVANCED_XML );
90 set_description( _("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 *, vlc_bool_t );
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 * );
118 static char *xtag_get_pcdata( XTag * );
119 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 )
149 /*****************************************************************************
150 * Catalogue functions
151 *****************************************************************************/
152 static void CatalogLoad( xml_t *p_xml, const char *psz_filename )
154 msg_Dbg( p_xml, "catalog support not implemented" );
157 static void CatalogAdd( xml_t *p_xml, const char *psz_arg1,
158 const char *psz_arg2, const char *psz_filename )
162 /*****************************************************************************
164 *****************************************************************************/
165 static xml_reader_t *ReaderCreate( xml_t *p_xml, stream_t *s )
167 xml_reader_t *p_reader;
168 char *p_buffer, *p_new;
169 int i_size, i_pos = 0, i_buffer = 2048;
172 /* Open and read file */
173 p_buffer = malloc( i_buffer );
174 if( p_buffer == NULL ) {
175 msg_Err( p_xml, "ENOMEM: alloc buffer" );
179 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
183 p_new = realloc( p_buffer, i_buffer );
185 msg_Err( p_xml, "ENOMEM: realloc buffer" );
191 p_buffer[ i_pos + i_size ] = 0; /* 0 terminated string */
193 if( i_pos + i_size == 0 )
195 msg_Dbg( p_xml, "empty xml" );
200 p_root = xtag_new_parse( p_buffer, i_buffer );
203 msg_Warn( p_xml, "couldn't parse xml" );
208 p_reader = malloc( sizeof(xml_reader_t) );
209 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
210 p_reader->p_sys->p_root = p_root;
211 p_reader->p_sys->p_curtag = NULL;
212 p_reader->p_sys->p_curattr = NULL;
213 p_reader->p_sys->b_endtag = VLC_FALSE;
214 p_reader->p_xml = p_xml;
216 p_reader->pf_read = ReaderRead;
217 p_reader->pf_node_type = ReaderNodeType;
218 p_reader->pf_name = ReaderName;
219 p_reader->pf_value = ReaderValue;
220 p_reader->pf_next_attr = ReaderNextAttr;
221 p_reader->pf_use_dtd = ReaderUseDTD;
226 static void ReaderDelete( xml_reader_t *p_reader )
228 xtag_free( p_reader->p_sys->p_root );
229 free( p_reader->p_sys );
233 static int ReaderUseDTD ( xml_reader_t *p_reader, vlc_bool_t b_use )
238 static int ReaderRead( xml_reader_t *p_reader )
242 if( !p_reader->p_sys->p_curtag )
244 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
250 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
252 p_reader->p_sys->p_curtag = p_child;
253 p_reader->p_sys->p_curattr = 0;
254 p_reader->p_sys->b_endtag = VLC_FALSE;
258 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
259 !p_reader->p_sys->b_endtag )
261 p_reader->p_sys->b_endtag = VLC_TRUE;
265 p_reader->p_sys->b_endtag = VLC_FALSE;
266 if( !p_reader->p_sys->p_curtag->parent ) return 0;
267 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
273 static int ReaderNodeType( xml_reader_t *p_reader )
275 if( p_reader->p_sys->p_curtag->name &&
276 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
277 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
278 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
279 return XML_READER_NONE;
282 static char *ReaderName( xml_reader_t *p_reader )
284 const char *psz_name;
286 if( !p_reader->p_sys->p_curattr )
288 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
290 printf( "TAG: %s\n", psz_name );
294 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
296 if( psz_name ) return strdup( psz_name );
300 static char *ReaderValue( xml_reader_t *p_reader )
302 const char *psz_name;
303 if( p_reader->p_sys->p_curtag->pcdata )
306 printf( "%s\n", p_reader->p_sys->p_curtag->pcdata );
308 return strdup( p_reader->p_sys->p_curtag->pcdata );
311 if( !p_reader->p_sys->p_curattr ) return 0;
314 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
315 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
318 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
320 if( psz_name ) return strdup( psz_name );
324 static int ReaderNextAttr( xml_reader_t *p_reader )
326 if( !p_reader->p_sys->p_curattr )
327 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
328 else if( p_reader->p_sys->p_curattr )
329 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
331 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
332 else return VLC_EGENERIC;
335 /*****************************************************************************
336 * XTAG parser functions
337 *****************************************************************************/
339 static XList *xlist_append( XList *list, void *data )
343 l = (XList *)malloc( sizeof(XList) );
344 l->prev = l->next = NULL;
347 if( list == NULL ) return l;
349 for( last = list; last; last = last->next )
350 if( last->next == NULL ) break;
352 if( last ) last->next = l;
357 static void xlist_free( XList *list )
361 for( l = list; l; l = ln )
368 /* Character classes */
370 #define X_WHITESPACE 1<<0
371 #define X_OPENTAG 1<<1
372 #define X_CLOSETAG 1<<2
373 #define X_DQUOTE 1<<3
374 #define X_SQUOTE 1<<4
381 static int xtag_cin( char c, int char_class )
383 if( char_class & X_WHITESPACE ) if( isspace(c) ) return VLC_TRUE;
384 if( char_class & X_OPENTAG ) if( c == '<' ) return VLC_TRUE;
385 if( char_class & X_CLOSETAG ) if( c == '>' ) return VLC_TRUE;
386 if( char_class & X_DQUOTE ) if( c == '"' ) return VLC_TRUE;
387 if( char_class & X_SQUOTE ) if( c == '\'' ) return VLC_TRUE;
388 if( char_class & X_EQUAL ) if( c == '=' ) return VLC_TRUE;
389 if( char_class & X_SLASH ) if( c == '/' ) return VLC_TRUE;
390 if( char_class & X_QMARK ) if( c == '?' ) return VLC_TRUE;
391 if( char_class & X_DASH ) if( c == '-' ) return VLC_TRUE;
392 if( char_class & X_EMARK ) if( c == '!' ) return VLC_TRUE;
397 static int xtag_index( XTagParser *parser, int char_class )
399 char *s = parser->start;
402 for( i = 0; s[i] && s != parser->end; i++ )
404 if( xtag_cin( s[i], char_class ) ) return i;
410 static void xtag_skip_over( XTagParser *parser, int char_class )
412 char *s = parser->start;
415 if( !parser->valid ) return;
417 for( i = 0; s[i] && s != parser->end; i++ )
419 if( !xtag_cin( s[i], char_class ) )
421 parser->start = &s[i];
429 static void xtag_skip_whitespace( XTagParser * parser )
431 xtag_skip_over( parser, X_WHITESPACE );
434 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
436 char *ret, *s = parser->start;
439 if( !parser->valid ) return NULL;
441 xi = xtag_index( parser, good_end | bad_end );
443 if( xi > 0 && xtag_cin (s[xi], good_end) )
445 ret = malloc( (xi+1) * sizeof(char) );
446 strncpy( ret, s, xi );
448 parser->start = &s[xi];
455 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
457 char *s = parser->start;
459 if( !parser->valid ) return VLC_FALSE;
461 if( !xtag_cin( s[0], char_class ) )
463 parser->valid = VLC_FALSE;
467 parser->start = &s[1];
472 static char *xtag_slurp_quoted( XTagParser *parser )
475 int quote = X_DQUOTE; /* quote char to match on */
478 if( !parser->valid ) return NULL;
480 xtag_skip_whitespace( parser );
484 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
486 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
490 for( xi = 0; s[xi]; xi++ )
492 if( xtag_cin( s[xi], quote ) )
494 if( !(xi > 1 && s[xi-1] == '\\') ) break;
498 ret = malloc( (xi+1) * sizeof(char) );
499 strncpy( ret, s, xi );
501 parser->start = &s[xi];
503 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
508 static XAttribute *xtag_parse_attribute( XTagParser *parser )
514 if( !parser->valid ) return NULL;
516 xtag_skip_whitespace( parser );
518 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
519 if( name == NULL ) return NULL;
521 xtag_skip_whitespace( parser );
524 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
527 printf( "xtag: attr failed EQUAL on <%s>\n", name );
532 xtag_skip_whitespace( parser );
534 value = xtag_slurp_quoted( parser );
539 printf ("Got NULL quoted attribute value\n");
544 attr = malloc( sizeof (*attr) );
551 parser->valid = VLC_FALSE;
555 static XTag *xtag_parse_tag( XTagParser *parser )
564 if( !parser->valid ) return NULL;
568 /* if this starts a comment tag, skip until end */
569 if( (parser->end - parser->start) > 7 &&
570 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
571 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
573 parser->start = s = &s[4];
574 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
576 parser->start = s = &s[xi+1];
577 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
579 parser->start = &s[2];
580 xtag_skip_whitespace( parser );
581 return xtag_parse_tag( parser );
587 /* ignore processing instructions '<?' ... '?>' */
588 if( (parser->end - parser->start) > 4 &&
589 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
591 parser->start = s = &s[2];
592 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
593 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
594 parser->start = &s[xi+2];
595 xtag_skip_whitespace( parser );
596 return xtag_parse_tag( parser );
602 /* ignore doctype '<!DOCTYPE' ... '>' */
603 if ( (parser->end - parser->start) > 8 &&
604 !strncmp( s, "<!DOCTYPE", 9 ) ) {
605 xi = xtag_index( parser, X_CLOSETAG );
607 parser->start = s = &s[xi+1];
608 xtag_skip_whitespace( parser );
609 return xtag_parse_tag( parser );
616 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
618 tag = malloc( sizeof(*tag) );
620 tag->pcdata = pcdata;
621 tag->parent = parser->current_tag;
622 tag->attributes = NULL;
623 tag->children = NULL;
624 tag->current_child = NULL;
629 /* if this starts a close tag, return NULL and let the parent take it */
630 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
633 /* parse CDATA content */
634 if ( (parser->end - parser->start) > 8 &&
635 !strncmp( s, "<![CDATA[", 9 ) ) {
636 parser->start = s = &s[9];
637 while (parser->end - s > 2) {
638 if (strncmp( s, "]]>", 3 ) == 0) {
639 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
640 if ( !(pcdata = malloc( sizeof(char)*(s - parser->start + 1))) ) return NULL;
641 strncpy( pcdata, parser->start, s - parser->start );
642 pcdata[s - parser->start]='\0';
643 parser->start = s = &s[3];
645 tag->pcdata = pcdata;
646 tag->parent = parser->current_tag;
647 tag->attributes = NULL;
648 tag->children = NULL;
649 tag->current_child = NULL;
659 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
661 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
662 if( name == NULL ) return NULL;
665 printf ("<%s ...\n", name);
668 tag = malloc( sizeof(*tag) );
671 tag->parent = parser->current_tag;
672 tag->attributes = NULL;
673 tag->children = NULL;
674 tag->current_child = NULL;
678 if( xtag_cin( s[0], X_WHITESPACE ) )
680 while( (attr = xtag_parse_attribute( parser )) != NULL )
682 tag->attributes = xlist_append( tag->attributes, attr );
686 xtag_skip_whitespace( parser );
690 if( xtag_cin( s[0], X_CLOSETAG ) )
692 parser->current_tag = tag;
694 xtag_assert_and_pass( parser, X_CLOSETAG );
696 while( (inner = xtag_parse_tag( parser ) ) != NULL )
698 tag->children = xlist_append( tag->children, inner );
701 parser->current_tag = tag->parent;
702 xtag_skip_whitespace( parser );
704 xtag_assert_and_pass( parser, X_OPENTAG );
705 xtag_assert_and_pass( parser, X_SLASH );
706 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
709 if( strcmp( name, tag->name ) )
712 printf ("got %s expected %s\n", name, tag->name);
714 parser->valid = VLC_FALSE;
719 xtag_skip_whitespace( parser );
720 xtag_assert_and_pass( parser, X_CLOSETAG );
721 xtag_skip_whitespace( parser );
725 xtag_assert_and_pass( parser, X_SLASH );
726 xtag_assert_and_pass( parser, X_CLOSETAG );
727 xtag_skip_whitespace( parser );
733 static XTag *xtag_free( XTag *xtag )
739 if( xtag == NULL ) return NULL;
741 if( xtag->name ) free( xtag->name );
742 if( xtag->pcdata ) free( xtag->pcdata );
744 for( l = xtag->attributes; l; l = l->next )
746 if( (attr = (XAttribute *)l->data) != NULL )
748 if( attr->name ) free( attr->name );
749 if( attr->value ) free( attr->value );
753 xlist_free( xtag->attributes );
755 for( l = xtag->children; l; l = l->next )
757 child = (XTag *)l->data;
760 xlist_free( xtag->children );
767 static XTag *xtag_new_parse( const char *s, int n )
770 XTag *tag, *ttag, *wrapper;
772 parser.valid = VLC_TRUE;
773 parser.current_tag = NULL;
774 parser.start = (char *)s;
776 if( n == -1 ) parser.end = NULL;
780 printf ("empty buffer");
784 else parser.end = (char *)&s[n];
786 /* can't have whitespace pcdata outside rootnode */
787 xtag_skip_whitespace( &parser );
789 tag = xtag_parse_tag( &parser );
794 printf ("invalid file");
800 if( (ttag = xtag_parse_tag( &parser )) != NULL )
808 wrapper = malloc( sizeof(XTag) );
809 wrapper->name = NULL;
810 wrapper->pcdata = NULL;
811 wrapper->parent = NULL;
812 wrapper->attributes = NULL;
813 wrapper->children = NULL;
814 wrapper->current_child = NULL;
816 wrapper->children = xlist_append( wrapper->children, tag );
817 wrapper->children = xlist_append( wrapper->children, ttag );
819 while( (ttag = xtag_parse_tag( &parser )) != NULL )
827 wrapper->children = xlist_append( wrapper->children, ttag );
835 static char *xtag_get_name( XTag *xtag )
837 return xtag ? xtag->name : NULL;
840 static char *xtag_get_pcdata( XTag *xtag )
845 if( xtag == NULL ) return NULL;
847 for( l = xtag->children; l; l = l->next )
849 child = (XTag *)l->data;
850 if( child->pcdata != NULL )
852 return child->pcdata;
859 static char *xtag_get_attribute( XTag *xtag, char *attribute )
864 if( xtag == NULL ) return NULL;
866 for( l = xtag->attributes; l; l = l->next )
868 if( (attr = (XAttribute *)l->data) != NULL )
870 if( !strcmp( attr->name, attribute ) ) return attr->value;
877 static XTag *xtag_first_child( XTag *xtag, char *name )
882 if( xtag == NULL ) return NULL;
883 if( (l = xtag->children) == NULL ) return NULL;
887 xtag->current_child = l;
888 return (XTag *)l->data;
891 for( ; l; l = l->next )
893 child = (XTag *)l->data;
895 if( !strcmp( child->name, name ) )
897 xtag->current_child = l;
902 xtag->current_child = NULL;
907 static XTag *xtag_next_child( XTag *xtag, char *name )
912 if( xtag == NULL ) return NULL;
914 if( (l = xtag->current_child) == NULL )
915 return xtag_first_child( xtag, name );
917 if( (l = l->next) == NULL ) return NULL;
921 xtag->current_child = l;
922 return (XTag *)l->data;
925 for( ; l; l = l->next )
927 child = (XTag *)l->data;
929 if( !strcmp( child->name, name ) )
931 xtag->current_child = l;
936 xtag->current_child = NULL;
942 * This snprints function takes a variable list of char *, the last of
943 * which must be NULL, and prints each in turn to buf.
944 * Returns C99-style total length that would have been written, even if
945 * this is larger than n.
947 static int xtag_snprints( char *buf, int n, ... )
951 int len, to_copy, total = 0;
955 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
959 if( (to_copy = __MIN(n, len) ) > 0 )
961 memcpy( buf, s, to_copy );
974 static int xtag_snprint( char *buf, int n, XTag *xtag )
982 buf += __MIN(n, N); \
988 if( n > 0 ) buf[0] = '\0';
994 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
1002 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1005 for( l = xtag->attributes; l; l = l->next )
1007 attr = (XAttribute *)l->data;
1009 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1014 if( xtag->children == NULL )
1016 nn = xtag_snprints ( buf, n, "/>", NULL );
1022 nn = xtag_snprints( buf, n, ">", NULL );
1026 for( l = xtag->children; l; l = l->next )
1028 child = (XTag *)l->data;
1030 nn = xtag_snprint( buf, n, child );
1036 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );