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 VideoLAN
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., 59 Temple Place - Suite 330, Boston, MA 02111, 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;
169 int i_pos,i_size,i_buffer = 5000;
172 /* Open and read file */
173 p_buffer = malloc( i_buffer + 1 );
174 if( p_buffer == NULL )
179 while( i_pos < i_buffer )
181 i_size = stream_Read( s, &p_buffer[i_pos], 5000 );
184 break; /* we're done */
188 p_buffer = realloc( p_buffer, i_buffer * sizeof( char *) );
191 p_buffer[ i_pos ] = 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
380 static int xtag_cin( char c, int char_class )
382 if( char_class & X_WHITESPACE ) if( isspace(c) ) return VLC_TRUE;
383 if( char_class & X_OPENTAG ) if( c == '<' ) return VLC_TRUE;
384 if( char_class & X_CLOSETAG ) if( c == '>' ) return VLC_TRUE;
385 if( char_class & X_DQUOTE ) if( c == '"' ) return VLC_TRUE;
386 if( char_class & X_SQUOTE ) if( c == '\'' ) return VLC_TRUE;
387 if( char_class & X_EQUAL ) if( c == '=' ) return VLC_TRUE;
388 if( char_class & X_SLASH ) if( c == '/' ) return VLC_TRUE;
389 if( char_class & X_QMARK ) if( c == '!' ) return VLC_TRUE;
390 if( char_class & X_DASH ) if( c == '-' ) return VLC_TRUE;
395 static int xtag_index( XTagParser *parser, int char_class )
397 char *s = parser->start;
400 for( i = 0; s[i] && s != parser->end; i++ )
402 if( xtag_cin( s[i], char_class ) ) return i;
408 static void xtag_skip_over( XTagParser *parser, int char_class )
410 char *s = parser->start;
413 if( !parser->valid ) return;
415 for( i = 0; s[i] && s != parser->end; i++ )
417 if( !xtag_cin( s[i], char_class ) )
419 parser->start = &s[i];
427 static void xtag_skip_whitespace( XTagParser * parser )
429 xtag_skip_over( parser, X_WHITESPACE );
432 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
434 char *ret, *s = parser->start;
437 if( !parser->valid ) return NULL;
439 xi = xtag_index( parser, good_end | bad_end );
441 if( xi > 0 && xtag_cin (s[xi], good_end) )
443 ret = malloc( (xi+1) * sizeof(char) );
444 strncpy( ret, s, xi );
446 parser->start = &s[xi];
453 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
455 char *s = parser->start;
457 if( !parser->valid ) return VLC_FALSE;
459 if( !xtag_cin( s[0], char_class ) )
461 parser->valid = VLC_FALSE;
465 parser->start = &s[1];
470 static char *xtag_slurp_quoted( XTagParser *parser )
473 int quote = X_DQUOTE; /* quote char to match on */
476 if( !parser->valid ) return NULL;
478 xtag_skip_whitespace( parser );
482 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
484 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
488 for( xi = 0; s[xi]; xi++ )
490 if( xtag_cin( s[xi], quote ) )
492 if( !(xi > 1 && s[xi-1] == '\\') ) break;
496 ret = malloc( (xi+1) * sizeof(char) );
497 strncpy( ret, s, xi );
499 parser->start = &s[xi];
501 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
506 static XAttribute *xtag_parse_attribute( XTagParser *parser )
512 if( !parser->valid ) return NULL;
514 xtag_skip_whitespace( parser );
516 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
517 if( name == NULL ) return NULL;
519 xtag_skip_whitespace( parser );
522 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
525 printf( "xtag: attr failed EQUAL on <%s>\n", name );
530 xtag_skip_whitespace( parser );
532 value = xtag_slurp_quoted( parser );
537 printf ("Got NULL quoted attribute value\n");
542 attr = malloc( sizeof (*attr) );
549 parser->valid = VLC_FALSE;
553 static XTag *xtag_parse_tag( XTagParser *parser )
561 if( !parser->valid ) return NULL;
563 #if 0 /* Do we really want all the whitespace pcdata ? */
564 xtag_skip_whitespace( parser );
567 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
569 tag = malloc( sizeof(*tag) );
571 tag->pcdata = pcdata;
572 tag->parent = parser->current_tag;
573 tag->attributes = NULL;
574 tag->children = NULL;
575 tag->current_child = NULL;
582 /* if this starts a close tag, return NULL and let the parent take it */
583 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
586 /* if this starts a comment tag, skip until end */
587 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) &&
588 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
592 parser->start = s = &s[4];
594 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
596 parser->start = s = &s[xi+1];
598 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
600 parser->start = &s[2];
601 xtag_skip_whitespace( parser );
602 return xtag_parse_tag( parser );
609 /* FIXME: if this starts a DOCTYPE tag, skip until end */
610 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
612 int xi = xtag_index( parser, X_CLOSETAG );
613 if( xi <= 0 ) return NULL;
615 parser->start = &s[xi+1];
616 xtag_skip_whitespace( parser );
617 return xtag_parse_tag( parser );
620 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
622 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
623 if( name == NULL ) return NULL;
626 printf ("<%s ...\n", name);
629 tag = malloc( sizeof(*tag) );
632 tag->parent = parser->current_tag;
633 tag->attributes = NULL;
634 tag->children = NULL;
635 tag->current_child = NULL;
639 if( xtag_cin( s[0], X_WHITESPACE ) )
641 while( (attr = xtag_parse_attribute( parser )) != NULL )
643 tag->attributes = xlist_append( tag->attributes, attr );
647 xtag_skip_whitespace( parser );
651 if( xtag_cin( s[0], X_CLOSETAG ) )
653 parser->current_tag = tag;
655 xtag_assert_and_pass( parser, X_CLOSETAG );
657 while( (inner = xtag_parse_tag( parser ) ) != NULL )
659 tag->children = xlist_append( tag->children, inner );
662 parser->current_tag = tag->parent;
663 xtag_skip_whitespace( parser );
665 xtag_assert_and_pass( parser, X_OPENTAG );
666 xtag_assert_and_pass( parser, X_SLASH );
667 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
670 if( strcmp( name, tag->name ) )
673 printf ("got %s expected %s\n", name, tag->name);
675 parser->valid = VLC_FALSE;
680 xtag_skip_whitespace( parser );
681 xtag_assert_and_pass( parser, X_CLOSETAG );
686 xtag_assert_and_pass( parser, X_SLASH );
687 xtag_assert_and_pass( parser, X_CLOSETAG );
693 static XTag *xtag_free( XTag *xtag )
699 if( xtag == NULL ) return NULL;
701 if( xtag->name ) free( xtag->name );
702 if( xtag->pcdata ) free( xtag->pcdata );
704 for( l = xtag->attributes; l; l = l->next )
706 if( (attr = (XAttribute *)l->data) != NULL )
708 if( attr->name ) free( attr->name );
709 if( attr->value ) free( attr->value );
713 xlist_free( xtag->attributes );
715 for( l = xtag->children; l; l = l->next )
717 child = (XTag *)l->data;
720 xlist_free( xtag->children );
727 static XTag *xtag_new_parse( const char *s, int n )
730 XTag *tag, *ttag, *wrapper;
732 parser.valid = VLC_TRUE;
733 parser.current_tag = NULL;
734 parser.start = (char *)s;
736 if( n == -1 ) parser.end = NULL;
740 printf ("empty buffer");
744 else parser.end = (char *)&s[n];
746 /* can't have whitespace pcdata outside rootnode */
747 xtag_skip_whitespace( &parser );
749 tag = xtag_parse_tag( &parser );
754 printf ("invalid file");
760 if( (ttag = xtag_parse_tag( &parser )) != NULL )
768 wrapper = malloc( sizeof(XTag) );
769 wrapper->name = NULL;
770 wrapper->pcdata = NULL;
771 wrapper->parent = NULL;
772 wrapper->attributes = NULL;
773 wrapper->children = NULL;
774 wrapper->current_child = NULL;
776 wrapper->children = xlist_append( wrapper->children, tag );
777 wrapper->children = xlist_append( wrapper->children, ttag );
779 while( (ttag = xtag_parse_tag( &parser )) != NULL )
787 wrapper->children = xlist_append( wrapper->children, ttag );
795 static char *xtag_get_name( XTag *xtag )
797 return xtag ? xtag->name : NULL;
800 static char *xtag_get_pcdata( XTag *xtag )
805 if( xtag == NULL ) return NULL;
807 for( l = xtag->children; l; l = l->next )
809 child = (XTag *)l->data;
810 if( child->pcdata != NULL )
812 return child->pcdata;
819 static char *xtag_get_attribute( XTag *xtag, char *attribute )
824 if( xtag == NULL ) return NULL;
826 for( l = xtag->attributes; l; l = l->next )
828 if( (attr = (XAttribute *)l->data) != NULL )
830 if( !strcmp( attr->name, attribute ) ) return attr->value;
837 static XTag *xtag_first_child( XTag *xtag, char *name )
842 if( xtag == NULL ) return NULL;
843 if( (l = xtag->children) == NULL ) return NULL;
847 xtag->current_child = l;
848 return (XTag *)l->data;
851 for( ; l; l = l->next )
853 child = (XTag *)l->data;
855 if( !strcmp( child->name, name ) )
857 xtag->current_child = l;
862 xtag->current_child = NULL;
867 static XTag *xtag_next_child( XTag *xtag, char *name )
872 if( xtag == NULL ) return NULL;
874 if( (l = xtag->current_child) == NULL )
875 return xtag_first_child( xtag, name );
877 if( (l = l->next) == NULL ) return NULL;
881 xtag->current_child = l;
882 return (XTag *)l->data;
885 for( ; l; l = l->next )
887 child = (XTag *)l->data;
889 if( !strcmp( child->name, name ) )
891 xtag->current_child = l;
896 xtag->current_child = NULL;
902 * This snprints function takes a variable list of char *, the last of
903 * which must be NULL, and prints each in turn to buf.
904 * Returns C99-style total length that would have been written, even if
905 * this is larger than n.
907 static int xtag_snprints( char *buf, int n, ... )
911 int len, to_copy, total = 0;
915 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
919 if( (to_copy = __MIN(n, len) ) > 0 )
921 memcpy( buf, s, to_copy );
934 static int xtag_snprint( char *buf, int n, XTag *xtag )
942 buf += __MIN(n, N); \
948 if( n > 0 ) buf[0] = '\0';
954 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
962 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
965 for( l = xtag->attributes; l; l = l->next )
967 attr = (XAttribute *)l->data;
969 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
974 if( xtag->children == NULL )
976 nn = xtag_snprints ( buf, n, "/>", NULL );
982 nn = xtag_snprints( buf, n, ">", NULL );
986 for( l = xtag->children; l; l = l->next )
988 child = (XTag *)l->data;
990 nn = xtag_snprint( buf, n, child );
996 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );