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 *****************************************************************************/
32 #include "vlc_block.h"
33 #include "vlc_stream.h"
50 * struct XTag is kind of a union ... it normally represents a whole
51 * tag (and its children), but it could alternatively represent some
52 * PCDATA. Basically, if tag->pcdata is non-NULL, interpret only it and
53 * ignore the name, attributes and inner_tags.
65 typedef struct _XAttribute
71 typedef struct _XTagParser
73 int valid; /* boolean */
79 /*****************************************************************************
81 *****************************************************************************/
82 static int Open ( vlc_object_t * );
83 static void Close( vlc_object_t * );
86 set_description( _("Simple XML Parser") );
87 set_capability( "xml", 5 );
88 set_callbacks( Open, Close );
91 struct xml_reader_sys_t
93 XTag *p_root; /* Root tag */
94 XTag *p_curtag; /* Current tag */
95 XList *p_curattr; /* Current attribute */
99 static xml_reader_t *ReaderCreate( xml_t *, stream_t * );
100 static void ReaderDelete( xml_reader_t * );
101 static int ReaderRead( xml_reader_t * );
102 static int ReaderNodeType( xml_reader_t * );
103 static char *ReaderName( xml_reader_t * );
104 static char *ReaderValue( xml_reader_t * );
105 static int ReaderNextAttr( xml_reader_t * );
107 static int ReaderUseDTD ( xml_reader_t *, vlc_bool_t );
109 static void CatalogLoad( xml_t *, const char * );
110 static void CatalogAdd( xml_t *, const char *, const char *, const char * );
112 static XTag *xtag_new_parse( const char *, int );
113 static char *xtag_get_name( XTag * );
114 static char *xtag_get_pcdata( XTag * );
115 static char *xtag_get_attribute( XTag *, char * );
116 static XTag *xtag_first_child( XTag *, char * );
117 static XTag *xtag_next_child( XTag *, char * );
118 static XTag *xtag_free( XTag * );
119 static int xtag_snprint( char *, int, XTag * );
121 /*****************************************************************************
122 * Module initialization
123 *****************************************************************************/
124 static int Open( vlc_object_t *p_this )
126 xml_t *p_xml = (xml_t *)p_this;
128 p_xml->pf_reader_create = ReaderCreate;
129 p_xml->pf_reader_delete = ReaderDelete;
131 p_xml->pf_catalog_load = CatalogLoad;
132 p_xml->pf_catalog_add = CatalogAdd;
137 /*****************************************************************************
138 * Module deinitialization
139 *****************************************************************************/
140 static void Close( vlc_object_t *p_this )
145 /*****************************************************************************
146 * Catalogue functions
147 *****************************************************************************/
148 static void CatalogLoad( xml_t *p_xml, const char *psz_filename )
150 msg_Dbg( p_xml, "catalog support not implemented" );
153 static void CatalogAdd( xml_t *p_xml, const char *psz_arg1,
154 const char *psz_arg2, const char *psz_filename )
158 /*****************************************************************************
160 *****************************************************************************/
161 static xml_reader_t *ReaderCreate( xml_t *p_xml, stream_t *s )
163 xml_reader_t *p_reader;
164 char *p_buffer, *p_new;
165 int i_size, i_pos = 0, i_buffer = 2048;
168 /* Open and read file */
169 p_buffer = malloc( i_buffer );
170 if( p_buffer == NULL ) {
171 msg_Err( p_xml, "out of memory" );
175 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
179 p_new = realloc( p_buffer, i_buffer );
181 msg_Err( p_xml, "out of memory" );
187 p_buffer[ i_pos + i_size ] = 0; /* 0 terminated string */
189 if( i_pos + i_size == 0 )
191 msg_Dbg( p_xml, "empty XML" );
196 p_root = xtag_new_parse( p_buffer, i_buffer );
199 msg_Warn( p_xml, "couldn't parse XML" );
205 p_reader = malloc( sizeof(xml_reader_t) );
206 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
207 p_reader->p_sys->p_root = p_root;
208 p_reader->p_sys->p_curtag = NULL;
209 p_reader->p_sys->p_curattr = NULL;
210 p_reader->p_sys->b_endtag = VLC_FALSE;
211 p_reader->p_xml = p_xml;
213 p_reader->pf_read = ReaderRead;
214 p_reader->pf_node_type = ReaderNodeType;
215 p_reader->pf_name = ReaderName;
216 p_reader->pf_value = ReaderValue;
217 p_reader->pf_next_attr = ReaderNextAttr;
218 p_reader->pf_use_dtd = ReaderUseDTD;
223 static void ReaderDelete( xml_reader_t *p_reader )
225 xtag_free( p_reader->p_sys->p_root );
226 free( p_reader->p_sys );
230 static int ReaderUseDTD ( xml_reader_t *p_reader, vlc_bool_t b_use )
235 static int ReaderRead( xml_reader_t *p_reader )
239 if( !p_reader->p_sys->p_curtag )
241 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
247 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
249 p_reader->p_sys->p_curtag = p_child;
250 p_reader->p_sys->p_curattr = 0;
251 p_reader->p_sys->b_endtag = VLC_FALSE;
255 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
256 !p_reader->p_sys->b_endtag )
258 p_reader->p_sys->b_endtag = VLC_TRUE;
262 p_reader->p_sys->b_endtag = VLC_FALSE;
263 if( !p_reader->p_sys->p_curtag->parent ) return 0;
264 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
270 static int ReaderNodeType( xml_reader_t *p_reader )
272 if( p_reader->p_sys->p_curtag->name &&
273 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
274 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
275 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
276 return XML_READER_NONE;
279 static char *ReaderName( xml_reader_t *p_reader )
281 const char *psz_name;
283 if( !p_reader->p_sys->p_curattr )
285 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
287 printf( "TAG: %s\n", psz_name );
291 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
293 if( psz_name ) return strdup( psz_name );
297 static char *ReaderValue( xml_reader_t *p_reader )
299 const char *psz_name;
300 if( p_reader->p_sys->p_curtag->pcdata )
303 printf( "%s\n", p_reader->p_sys->p_curtag->pcdata );
305 return strdup( p_reader->p_sys->p_curtag->pcdata );
308 if( !p_reader->p_sys->p_curattr ) return 0;
311 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
312 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
315 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
317 if( psz_name ) return strdup( psz_name );
321 static int ReaderNextAttr( xml_reader_t *p_reader )
323 if( !p_reader->p_sys->p_curattr )
324 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
325 else if( p_reader->p_sys->p_curattr )
326 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
328 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
329 else return VLC_EGENERIC;
332 /*****************************************************************************
333 * XTAG parser functions
334 *****************************************************************************/
336 static XList *xlist_append( XList *list, void *data )
340 l = (XList *)malloc( sizeof(XList) );
341 l->prev = l->next = NULL;
344 if( list == NULL ) return l;
346 for( last = list; last; last = last->next )
347 if( last->next == NULL ) break;
349 if( last ) last->next = l;
354 static void xlist_free( XList *list )
358 for( l = list; l; l = ln )
365 /* Character classes */
367 #define X_WHITESPACE 1<<0
368 #define X_OPENTAG 1<<1
369 #define X_CLOSETAG 1<<2
370 #define X_DQUOTE 1<<3
371 #define X_SQUOTE 1<<4
378 static int xtag_cin( char c, int char_class )
380 if( char_class & X_WHITESPACE ) if( isspace(c) ) return VLC_TRUE;
381 if( char_class & X_OPENTAG ) if( c == '<' ) return VLC_TRUE;
382 if( char_class & X_CLOSETAG ) if( c == '>' ) return VLC_TRUE;
383 if( char_class & X_DQUOTE ) if( c == '"' ) return VLC_TRUE;
384 if( char_class & X_SQUOTE ) if( c == '\'' ) return VLC_TRUE;
385 if( char_class & X_EQUAL ) if( c == '=' ) return VLC_TRUE;
386 if( char_class & X_SLASH ) if( c == '/' ) return VLC_TRUE;
387 if( char_class & X_QMARK ) if( c == '?' ) return VLC_TRUE;
388 if( char_class & X_DASH ) if( c == '-' ) return VLC_TRUE;
389 if( char_class & X_EMARK ) if( c == '!' ) return VLC_TRUE;
394 static int xtag_index( XTagParser *parser, int char_class )
396 char *s = parser->start;
399 for( i = 0; s[i] && s != parser->end; i++ )
401 if( xtag_cin( s[i], char_class ) ) return i;
407 static void xtag_skip_over( XTagParser *parser, int char_class )
409 char *s = parser->start;
412 if( !parser->valid ) return;
414 for( i = 0; s[i] && s != parser->end; i++ )
416 if( !xtag_cin( s[i], char_class ) )
418 parser->start = &s[i];
426 static void xtag_skip_whitespace( XTagParser * parser )
428 xtag_skip_over( parser, X_WHITESPACE );
431 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
433 char *ret, *s = parser->start;
436 if( !parser->valid ) return NULL;
438 xi = xtag_index( parser, good_end | bad_end );
440 if( xi > 0 && xtag_cin (s[xi], good_end) )
442 ret = malloc( (xi+1) * sizeof(char) );
443 strncpy( ret, s, xi );
445 parser->start = &s[xi];
452 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
454 char *s = parser->start;
456 if( !parser->valid ) return VLC_FALSE;
458 if( !xtag_cin( s[0], char_class ) )
460 parser->valid = VLC_FALSE;
464 parser->start = &s[1];
469 static char *xtag_slurp_quoted( XTagParser *parser )
472 int quote = X_DQUOTE; /* quote char to match on */
475 if( !parser->valid ) return NULL;
477 xtag_skip_whitespace( parser );
481 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
483 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
487 for( xi = 0; s[xi]; xi++ )
489 if( xtag_cin( s[xi], quote ) )
491 if( !(xi > 1 && s[xi-1] == '\\') ) break;
495 ret = malloc( (xi+1) * sizeof(char) );
496 strncpy( ret, s, xi );
498 parser->start = &s[xi];
500 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
505 static XAttribute *xtag_parse_attribute( XTagParser *parser )
511 if( !parser->valid ) return NULL;
513 xtag_skip_whitespace( parser );
515 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
516 if( name == NULL ) return NULL;
518 xtag_skip_whitespace( parser );
521 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
524 printf( "xtag: attr failed EQUAL on <%s>\n", name );
529 xtag_skip_whitespace( parser );
531 value = xtag_slurp_quoted( parser );
536 printf ("Got NULL quoted attribute value\n");
541 attr = malloc( sizeof (*attr) );
548 parser->valid = VLC_FALSE;
552 static XTag *xtag_parse_tag( XTagParser *parser )
561 if( !parser->valid ) return NULL;
565 /* if this starts a comment tag, skip until end */
566 if( (parser->end - parser->start) > 7 &&
567 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
568 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
570 parser->start = s = &s[4];
571 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
573 parser->start = s = &s[xi+1];
574 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
576 parser->start = &s[2];
577 xtag_skip_whitespace( parser );
578 return xtag_parse_tag( parser );
584 /* ignore processing instructions '<?' ... '?>' */
585 if( (parser->end - parser->start) > 4 &&
586 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
588 parser->start = s = &s[2];
589 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
590 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
591 parser->start = &s[xi+2];
592 xtag_skip_whitespace( parser );
593 return xtag_parse_tag( parser );
599 /* ignore doctype '<!DOCTYPE' ... '>' */
600 if ( (parser->end - parser->start) > 8 &&
601 !strncmp( s, "<!DOCTYPE", 9 ) ) {
602 xi = xtag_index( parser, X_CLOSETAG );
604 parser->start = s = &s[xi+1];
605 xtag_skip_whitespace( parser );
606 return xtag_parse_tag( parser );
613 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
615 tag = malloc( sizeof(*tag) );
617 tag->pcdata = pcdata;
618 tag->parent = parser->current_tag;
619 tag->attributes = NULL;
620 tag->children = NULL;
621 tag->current_child = NULL;
626 /* if this starts a close tag, return NULL and let the parent take it */
627 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
630 /* parse CDATA content */
631 if ( (parser->end - parser->start) > 8 &&
632 !strncmp( s, "<![CDATA[", 9 ) ) {
633 parser->start = s = &s[9];
634 while (parser->end - s > 2) {
635 if (strncmp( s, "]]>", 3 ) == 0) {
636 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
637 if ( !(pcdata = malloc( sizeof(char)*(s - parser->start + 1))) ) return NULL;
638 strncpy( pcdata, parser->start, s - parser->start );
639 pcdata[s - parser->start]='\0';
640 parser->start = s = &s[3];
642 tag->pcdata = pcdata;
643 tag->parent = parser->current_tag;
644 tag->attributes = NULL;
645 tag->children = NULL;
646 tag->current_child = NULL;
656 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
658 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
659 if( name == NULL ) return NULL;
662 printf ("<%s ...\n", name);
665 tag = malloc( sizeof(*tag) );
668 tag->parent = parser->current_tag;
669 tag->attributes = NULL;
670 tag->children = NULL;
671 tag->current_child = NULL;
675 if( xtag_cin( s[0], X_WHITESPACE ) )
677 while( (attr = xtag_parse_attribute( parser )) != NULL )
679 tag->attributes = xlist_append( tag->attributes, attr );
683 xtag_skip_whitespace( parser );
687 if( xtag_cin( s[0], X_CLOSETAG ) )
689 parser->current_tag = tag;
691 xtag_assert_and_pass( parser, X_CLOSETAG );
693 while( (inner = xtag_parse_tag( parser ) ) != NULL )
695 tag->children = xlist_append( tag->children, inner );
698 parser->current_tag = tag->parent;
699 xtag_skip_whitespace( parser );
701 xtag_assert_and_pass( parser, X_OPENTAG );
702 xtag_assert_and_pass( parser, X_SLASH );
703 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
706 if( strcmp( name, tag->name ) )
709 printf ("got %s expected %s\n", name, tag->name);
711 parser->valid = VLC_FALSE;
716 xtag_skip_whitespace( parser );
717 xtag_assert_and_pass( parser, X_CLOSETAG );
718 xtag_skip_whitespace( parser );
722 xtag_assert_and_pass( parser, X_SLASH );
723 xtag_assert_and_pass( parser, X_CLOSETAG );
724 xtag_skip_whitespace( parser );
730 static XTag *xtag_free( XTag *xtag )
736 if( xtag == NULL ) return NULL;
738 if( xtag->name ) free( xtag->name );
739 if( xtag->pcdata ) free( xtag->pcdata );
741 for( l = xtag->attributes; l; l = l->next )
743 if( (attr = (XAttribute *)l->data) != NULL )
745 if( attr->name ) free( attr->name );
746 if( attr->value ) free( attr->value );
750 xlist_free( xtag->attributes );
752 for( l = xtag->children; l; l = l->next )
754 child = (XTag *)l->data;
757 xlist_free( xtag->children );
764 static XTag *xtag_new_parse( const char *s, int n )
767 XTag *tag, *ttag, *wrapper;
769 parser.valid = VLC_TRUE;
770 parser.current_tag = NULL;
771 parser.start = (char *)s;
773 if( n == -1 ) parser.end = NULL;
777 printf ("empty buffer");
781 else parser.end = (char *)&s[n];
783 /* can't have whitespace pcdata outside rootnode */
784 xtag_skip_whitespace( &parser );
786 tag = xtag_parse_tag( &parser );
791 printf ("invalid file");
797 if( (ttag = xtag_parse_tag( &parser )) != NULL )
805 wrapper = malloc( sizeof(XTag) );
806 wrapper->name = NULL;
807 wrapper->pcdata = NULL;
808 wrapper->parent = NULL;
809 wrapper->attributes = NULL;
810 wrapper->children = NULL;
811 wrapper->current_child = NULL;
813 wrapper->children = xlist_append( wrapper->children, tag );
814 wrapper->children = xlist_append( wrapper->children, ttag );
816 while( (ttag = xtag_parse_tag( &parser )) != NULL )
824 wrapper->children = xlist_append( wrapper->children, ttag );
832 static char *xtag_get_name( XTag *xtag )
834 return xtag ? xtag->name : NULL;
837 static char *xtag_get_pcdata( XTag *xtag )
842 if( xtag == NULL ) return NULL;
844 for( l = xtag->children; l; l = l->next )
846 child = (XTag *)l->data;
847 if( child->pcdata != NULL )
849 return child->pcdata;
856 static char *xtag_get_attribute( XTag *xtag, char *attribute )
861 if( xtag == NULL ) return NULL;
863 for( l = xtag->attributes; l; l = l->next )
865 if( (attr = (XAttribute *)l->data) != NULL )
867 if( !strcmp( attr->name, attribute ) ) return attr->value;
874 static XTag *xtag_first_child( XTag *xtag, char *name )
879 if( xtag == NULL ) return NULL;
880 if( (l = xtag->children) == NULL ) return NULL;
884 xtag->current_child = l;
885 return (XTag *)l->data;
888 for( ; l; l = l->next )
890 child = (XTag *)l->data;
892 if( !strcmp( child->name, name ) )
894 xtag->current_child = l;
899 xtag->current_child = NULL;
904 static XTag *xtag_next_child( XTag *xtag, char *name )
909 if( xtag == NULL ) return NULL;
911 if( (l = xtag->current_child) == NULL )
912 return xtag_first_child( xtag, name );
914 if( (l = l->next) == NULL ) return NULL;
918 xtag->current_child = l;
919 return (XTag *)l->data;
922 for( ; l; l = l->next )
924 child = (XTag *)l->data;
926 if( !strcmp( child->name, name ) )
928 xtag->current_child = l;
933 xtag->current_child = NULL;
939 * This snprints function takes a variable list of char *, the last of
940 * which must be NULL, and prints each in turn to buf.
941 * Returns C99-style total length that would have been written, even if
942 * this is larger than n.
944 static int xtag_snprints( char *buf, int n, ... )
948 int len, to_copy, total = 0;
952 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
956 if( (to_copy = __MIN(n, len) ) > 0 )
958 memcpy( buf, s, to_copy );
971 static int xtag_snprint( char *buf, int n, XTag *xtag )
979 buf += __MIN(n, N); \
985 if( n > 0 ) buf[0] = '\0';
991 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
999 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1002 for( l = xtag->attributes; l; l = l->next )
1004 attr = (XAttribute *)l->data;
1006 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1011 if( xtag->children == NULL )
1013 nn = xtag_snprints ( buf, n, "/>", NULL );
1019 nn = xtag_snprints( buf, n, ">", NULL );
1023 for( l = xtag->children; l; l = l->next )
1025 child = (XTag *)l->data;
1027 nn = xtag_snprint( buf, n, child );
1033 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );