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_size, i_pos = 0, i_buffer = 2048;
172 /* Open and read file */
173 p_buffer = malloc( i_buffer );
174 if( p_buffer == NULL ) return NULL;
176 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
180 p_buffer = realloc( p_buffer, i_buffer );
182 p_buffer[ i_pos + i_size ] = 0; /* 0 terminated string */
184 if( i_pos + i_size == 0 )
186 msg_Dbg( p_xml, "empty xml" );
191 p_root = xtag_new_parse( p_buffer, i_buffer );
194 msg_Warn( p_xml, "couldn't parse xml" );
199 p_reader = malloc( sizeof(xml_reader_t) );
200 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
201 p_reader->p_sys->p_root = p_root;
202 p_reader->p_sys->p_curtag = NULL;
203 p_reader->p_sys->p_curattr = NULL;
204 p_reader->p_sys->b_endtag = VLC_FALSE;
205 p_reader->p_xml = p_xml;
207 p_reader->pf_read = ReaderRead;
208 p_reader->pf_node_type = ReaderNodeType;
209 p_reader->pf_name = ReaderName;
210 p_reader->pf_value = ReaderValue;
211 p_reader->pf_next_attr = ReaderNextAttr;
212 p_reader->pf_use_dtd = ReaderUseDTD;
217 static void ReaderDelete( xml_reader_t *p_reader )
219 xtag_free( p_reader->p_sys->p_root );
220 free( p_reader->p_sys );
224 static int ReaderUseDTD ( xml_reader_t *p_reader, vlc_bool_t b_use )
229 static int ReaderRead( xml_reader_t *p_reader )
233 if( !p_reader->p_sys->p_curtag )
235 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
241 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
243 p_reader->p_sys->p_curtag = p_child;
244 p_reader->p_sys->p_curattr = 0;
245 p_reader->p_sys->b_endtag = VLC_FALSE;
249 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
250 !p_reader->p_sys->b_endtag )
252 p_reader->p_sys->b_endtag = VLC_TRUE;
256 p_reader->p_sys->b_endtag = VLC_FALSE;
257 if( !p_reader->p_sys->p_curtag->parent ) return 0;
258 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
264 static int ReaderNodeType( xml_reader_t *p_reader )
266 if( p_reader->p_sys->p_curtag->name &&
267 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
268 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
269 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
270 return XML_READER_NONE;
273 static char *ReaderName( xml_reader_t *p_reader )
275 const char *psz_name;
277 if( !p_reader->p_sys->p_curattr )
279 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
281 printf( "TAG: %s\n", psz_name );
285 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
287 if( psz_name ) return strdup( psz_name );
291 static char *ReaderValue( xml_reader_t *p_reader )
293 const char *psz_name;
294 if( p_reader->p_sys->p_curtag->pcdata )
297 printf( "%s\n", p_reader->p_sys->p_curtag->pcdata );
299 return strdup( p_reader->p_sys->p_curtag->pcdata );
302 if( !p_reader->p_sys->p_curattr ) return 0;
305 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
306 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
309 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
311 if( psz_name ) return strdup( psz_name );
315 static int ReaderNextAttr( xml_reader_t *p_reader )
317 if( !p_reader->p_sys->p_curattr )
318 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
319 else if( p_reader->p_sys->p_curattr )
320 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
322 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
323 else return VLC_EGENERIC;
326 /*****************************************************************************
327 * XTAG parser functions
328 *****************************************************************************/
330 static XList *xlist_append( XList *list, void *data )
334 l = (XList *)malloc( sizeof(XList) );
335 l->prev = l->next = NULL;
338 if( list == NULL ) return l;
340 for( last = list; last; last = last->next )
341 if( last->next == NULL ) break;
343 if( last ) last->next = l;
348 static void xlist_free( XList *list )
352 for( l = list; l; l = ln )
359 /* Character classes */
361 #define X_WHITESPACE 1<<0
362 #define X_OPENTAG 1<<1
363 #define X_CLOSETAG 1<<2
364 #define X_DQUOTE 1<<3
365 #define X_SQUOTE 1<<4
371 static int xtag_cin( char c, int char_class )
373 if( char_class & X_WHITESPACE ) if( isspace(c) ) return VLC_TRUE;
374 if( char_class & X_OPENTAG ) if( c == '<' ) return VLC_TRUE;
375 if( char_class & X_CLOSETAG ) if( c == '>' ) return VLC_TRUE;
376 if( char_class & X_DQUOTE ) if( c == '"' ) return VLC_TRUE;
377 if( char_class & X_SQUOTE ) if( c == '\'' ) return VLC_TRUE;
378 if( char_class & X_EQUAL ) if( c == '=' ) return VLC_TRUE;
379 if( char_class & X_SLASH ) if( c == '/' ) return VLC_TRUE;
380 if( char_class & X_QMARK ) if( c == '!' ) return VLC_TRUE;
381 if( char_class & X_DASH ) if( c == '-' ) return VLC_TRUE;
386 static int xtag_index( XTagParser *parser, int char_class )
388 char *s = parser->start;
391 for( i = 0; s[i] && s != parser->end; i++ )
393 if( xtag_cin( s[i], char_class ) ) return i;
399 static void xtag_skip_over( XTagParser *parser, int char_class )
401 char *s = parser->start;
404 if( !parser->valid ) return;
406 for( i = 0; s[i] && s != parser->end; i++ )
408 if( !xtag_cin( s[i], char_class ) )
410 parser->start = &s[i];
418 static void xtag_skip_whitespace( XTagParser * parser )
420 xtag_skip_over( parser, X_WHITESPACE );
423 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
425 char *ret, *s = parser->start;
428 if( !parser->valid ) return NULL;
430 xi = xtag_index( parser, good_end | bad_end );
432 if( xi > 0 && xtag_cin (s[xi], good_end) )
434 ret = malloc( (xi+1) * sizeof(char) );
435 strncpy( ret, s, xi );
437 parser->start = &s[xi];
444 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
446 char *s = parser->start;
448 if( !parser->valid ) return VLC_FALSE;
450 if( !xtag_cin( s[0], char_class ) )
452 parser->valid = VLC_FALSE;
456 parser->start = &s[1];
461 static char *xtag_slurp_quoted( XTagParser *parser )
464 int quote = X_DQUOTE; /* quote char to match on */
467 if( !parser->valid ) return NULL;
469 xtag_skip_whitespace( parser );
473 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
475 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
479 for( xi = 0; s[xi]; xi++ )
481 if( xtag_cin( s[xi], quote ) )
483 if( !(xi > 1 && s[xi-1] == '\\') ) break;
487 ret = malloc( (xi+1) * sizeof(char) );
488 strncpy( ret, s, xi );
490 parser->start = &s[xi];
492 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
497 static XAttribute *xtag_parse_attribute( XTagParser *parser )
503 if( !parser->valid ) return NULL;
505 xtag_skip_whitespace( parser );
507 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
508 if( name == NULL ) return NULL;
510 xtag_skip_whitespace( parser );
513 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
516 printf( "xtag: attr failed EQUAL on <%s>\n", name );
521 xtag_skip_whitespace( parser );
523 value = xtag_slurp_quoted( parser );
528 printf ("Got NULL quoted attribute value\n");
533 attr = malloc( sizeof (*attr) );
540 parser->valid = VLC_FALSE;
544 static XTag *xtag_parse_tag( XTagParser *parser )
552 if( !parser->valid ) return NULL;
554 #if 0 /* Do we really want all the whitespace pcdata ? */
555 xtag_skip_whitespace( parser );
558 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
560 tag = malloc( sizeof(*tag) );
562 tag->pcdata = pcdata;
563 tag->parent = parser->current_tag;
564 tag->attributes = NULL;
565 tag->children = NULL;
566 tag->current_child = NULL;
573 /* if this starts a close tag, return NULL and let the parent take it */
574 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
577 /* if this starts a comment tag, skip until end */
578 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) &&
579 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
583 parser->start = s = &s[4];
585 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
587 parser->start = s = &s[xi+1];
589 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
591 parser->start = &s[2];
592 xtag_skip_whitespace( parser );
593 return xtag_parse_tag( parser );
600 /* FIXME: if this starts a DOCTYPE tag, skip until end */
601 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
603 int xi = xtag_index( parser, X_CLOSETAG );
604 if( xi <= 0 ) return NULL;
606 parser->start = &s[xi+1];
607 xtag_skip_whitespace( parser );
608 return xtag_parse_tag( parser );
611 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
613 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
614 if( name == NULL ) return NULL;
617 printf ("<%s ...\n", name);
620 tag = malloc( sizeof(*tag) );
623 tag->parent = parser->current_tag;
624 tag->attributes = NULL;
625 tag->children = NULL;
626 tag->current_child = NULL;
630 if( xtag_cin( s[0], X_WHITESPACE ) )
632 while( (attr = xtag_parse_attribute( parser )) != NULL )
634 tag->attributes = xlist_append( tag->attributes, attr );
638 xtag_skip_whitespace( parser );
642 if( xtag_cin( s[0], X_CLOSETAG ) )
644 parser->current_tag = tag;
646 xtag_assert_and_pass( parser, X_CLOSETAG );
648 while( (inner = xtag_parse_tag( parser ) ) != NULL )
650 tag->children = xlist_append( tag->children, inner );
653 parser->current_tag = tag->parent;
654 xtag_skip_whitespace( parser );
656 xtag_assert_and_pass( parser, X_OPENTAG );
657 xtag_assert_and_pass( parser, X_SLASH );
658 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
661 if( strcmp( name, tag->name ) )
664 printf ("got %s expected %s\n", name, tag->name);
666 parser->valid = VLC_FALSE;
671 xtag_skip_whitespace( parser );
672 xtag_assert_and_pass( parser, X_CLOSETAG );
677 xtag_assert_and_pass( parser, X_SLASH );
678 xtag_assert_and_pass( parser, X_CLOSETAG );
684 static XTag *xtag_free( XTag *xtag )
690 if( xtag == NULL ) return NULL;
692 if( xtag->name ) free( xtag->name );
693 if( xtag->pcdata ) free( xtag->pcdata );
695 for( l = xtag->attributes; l; l = l->next )
697 if( (attr = (XAttribute *)l->data) != NULL )
699 if( attr->name ) free( attr->name );
700 if( attr->value ) free( attr->value );
704 xlist_free( xtag->attributes );
706 for( l = xtag->children; l; l = l->next )
708 child = (XTag *)l->data;
711 xlist_free( xtag->children );
718 static XTag *xtag_new_parse( const char *s, int n )
721 XTag *tag, *ttag, *wrapper;
723 parser.valid = VLC_TRUE;
724 parser.current_tag = NULL;
725 parser.start = (char *)s;
727 if( n == -1 ) parser.end = NULL;
731 printf ("empty buffer");
735 else parser.end = (char *)&s[n];
737 /* can't have whitespace pcdata outside rootnode */
738 xtag_skip_whitespace( &parser );
740 tag = xtag_parse_tag( &parser );
745 printf ("invalid file");
751 if( (ttag = xtag_parse_tag( &parser )) != NULL )
759 wrapper = malloc( sizeof(XTag) );
760 wrapper->name = NULL;
761 wrapper->pcdata = NULL;
762 wrapper->parent = NULL;
763 wrapper->attributes = NULL;
764 wrapper->children = NULL;
765 wrapper->current_child = NULL;
767 wrapper->children = xlist_append( wrapper->children, tag );
768 wrapper->children = xlist_append( wrapper->children, ttag );
770 while( (ttag = xtag_parse_tag( &parser )) != NULL )
778 wrapper->children = xlist_append( wrapper->children, ttag );
786 static char *xtag_get_name( XTag *xtag )
788 return xtag ? xtag->name : NULL;
791 static char *xtag_get_pcdata( XTag *xtag )
796 if( xtag == NULL ) return NULL;
798 for( l = xtag->children; l; l = l->next )
800 child = (XTag *)l->data;
801 if( child->pcdata != NULL )
803 return child->pcdata;
810 static char *xtag_get_attribute( XTag *xtag, char *attribute )
815 if( xtag == NULL ) return NULL;
817 for( l = xtag->attributes; l; l = l->next )
819 if( (attr = (XAttribute *)l->data) != NULL )
821 if( !strcmp( attr->name, attribute ) ) return attr->value;
828 static XTag *xtag_first_child( XTag *xtag, char *name )
833 if( xtag == NULL ) return NULL;
834 if( (l = xtag->children) == NULL ) return NULL;
838 xtag->current_child = l;
839 return (XTag *)l->data;
842 for( ; l; l = l->next )
844 child = (XTag *)l->data;
846 if( !strcmp( child->name, name ) )
848 xtag->current_child = l;
853 xtag->current_child = NULL;
858 static XTag *xtag_next_child( XTag *xtag, char *name )
863 if( xtag == NULL ) return NULL;
865 if( (l = xtag->current_child) == NULL )
866 return xtag_first_child( xtag, name );
868 if( (l = l->next) == NULL ) return NULL;
872 xtag->current_child = l;
873 return (XTag *)l->data;
876 for( ; l; l = l->next )
878 child = (XTag *)l->data;
880 if( !strcmp( child->name, name ) )
882 xtag->current_child = l;
887 xtag->current_child = NULL;
893 * This snprints function takes a variable list of char *, the last of
894 * which must be NULL, and prints each in turn to buf.
895 * Returns C99-style total length that would have been written, even if
896 * this is larger than n.
898 static int xtag_snprints( char *buf, int n, ... )
902 int len, to_copy, total = 0;
906 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
910 if( (to_copy = __MIN(n, len) ) > 0 )
912 memcpy( buf, s, to_copy );
925 static int xtag_snprint( char *buf, int n, XTag *xtag )
933 buf += __MIN(n, N); \
939 if( n > 0 ) buf[0] = '\0';
945 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
953 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
956 for( l = xtag->attributes; l; l = l->next )
958 attr = (XAttribute *)l->data;
960 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
965 if( xtag->children == NULL )
967 nn = xtag_snprints ( buf, n, "/>", NULL );
973 nn = xtag_snprints( buf, n, ">", NULL );
977 for( l = xtag->children; l; l = l->next )
979 child = (XTag *)l->data;
981 nn = xtag_snprint( buf, n, child );
987 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );