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 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 * );
116 static char *xtag_get_pcdata( XTag * );
117 static char *xtag_get_attribute( XTag *, char * );
118 static XTag *xtag_first_child( XTag *, char * );
119 static XTag *xtag_next_child( XTag *, char * );
120 static XTag *xtag_free( XTag * );
121 static int xtag_snprint( char *, int, XTag * );
123 /*****************************************************************************
124 * Module initialization
125 *****************************************************************************/
126 static int Open( vlc_object_t *p_this )
128 xml_t *p_xml = (xml_t *)p_this;
130 p_xml->pf_reader_create = ReaderCreate;
131 p_xml->pf_reader_delete = ReaderDelete;
133 p_xml->pf_catalog_load = CatalogLoad;
134 p_xml->pf_catalog_add = CatalogAdd;
139 /*****************************************************************************
140 * Module deinitialization
141 *****************************************************************************/
142 static void Close( vlc_object_t *p_this )
147 /*****************************************************************************
148 * Catalogue functions
149 *****************************************************************************/
150 static void CatalogLoad( xml_t *p_xml, const char *psz_filename )
152 msg_Dbg( p_xml, "catalog support not implemented" );
155 static void CatalogAdd( xml_t *p_xml, const char *psz_arg1,
156 const char *psz_arg2, const char *psz_filename )
160 /*****************************************************************************
162 *****************************************************************************/
163 static xml_reader_t *ReaderCreate( xml_t *p_xml, stream_t *s )
165 xml_reader_t *p_reader;
166 xml_reader_sys_t *p_sys;
171 /* Open and read file */
173 i_buffer = stream_Size( s );
174 p_buffer = malloc( i_buffer + 1 );
175 i_buffer = stream_Read( s, p_buffer, i_buffer );
176 p_buffer[ i_buffer ] = 0;
180 msg_Dbg( p_xml, "empty xml" );
185 p_root = xtag_new_parse( p_buffer, i_buffer );
188 msg_Warn( p_xml, "couldn't parse xml" );
193 p_reader = malloc( sizeof(xml_reader_t) );
194 p_reader->p_sys = p_sys = malloc( sizeof(xml_reader_sys_t) );
195 p_reader->p_sys->p_root = p_root;
196 p_reader->p_sys->p_curtag = NULL;
197 p_reader->p_sys->p_curattr = NULL;
198 p_reader->p_sys->b_endtag = VLC_FALSE;
199 p_reader->p_xml = p_xml;
201 p_reader->pf_read = ReaderRead;
202 p_reader->pf_node_type = ReaderNodeType;
203 p_reader->pf_name = ReaderName;
204 p_reader->pf_value = ReaderValue;
205 p_reader->pf_next_attr = ReaderNextAttr;
210 static void ReaderDelete( xml_reader_t *p_reader )
212 xtag_free( p_reader->p_sys->p_root );
213 free( p_reader->p_sys );
217 static int ReaderRead( xml_reader_t *p_reader )
221 if( !p_reader->p_sys->p_curtag )
223 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
229 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
231 p_reader->p_sys->p_curtag = p_child;
232 p_reader->p_sys->p_curattr = 0;
233 p_reader->p_sys->b_endtag = VLC_FALSE;
237 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
238 !p_reader->p_sys->b_endtag )
240 p_reader->p_sys->b_endtag = VLC_TRUE;
244 p_reader->p_sys->b_endtag = VLC_FALSE;
245 if( !p_reader->p_sys->p_curtag->parent ) return 0;
246 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
252 static int ReaderNodeType( xml_reader_t *p_reader )
254 if( p_reader->p_sys->p_curtag->name &&
255 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
256 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
257 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
258 return XML_READER_NONE;
261 static char *ReaderName( xml_reader_t *p_reader )
263 const char *psz_name;
265 if( !p_reader->p_sys->p_curattr )
267 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
269 printf( "TAG: %s\n", psz_name );
273 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
275 if( psz_name ) return strdup( psz_name );
279 static char *ReaderValue( xml_reader_t *p_reader )
281 const char *psz_name;
283 if( !p_reader->p_sys->p_curattr ) return 0;
286 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
287 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
290 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
292 if( psz_name ) return strdup( psz_name );
296 static int ReaderNextAttr( xml_reader_t *p_reader )
298 if( !p_reader->p_sys->p_curattr )
299 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
300 else if( p_reader->p_sys->p_curattr )
301 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
303 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
304 else return VLC_EGENERIC;
307 /*****************************************************************************
308 * XTAG parser functions
309 *****************************************************************************/
311 static XList *xlist_append( XList *list, void *data )
315 l = (XList *)malloc( sizeof(XList) );
316 l->prev = l->next = NULL;
319 if( list == NULL ) return l;
321 for( last = list; last; last = last->next )
322 if( last->next == NULL ) break;
324 if( last ) last->next = l;
329 static void xlist_free( XList *list )
333 for( l = list; l; l = ln )
340 /* Character classes */
342 #define X_WHITESPACE 1<<0
343 #define X_OPENTAG 1<<1
344 #define X_CLOSETAG 1<<2
345 #define X_DQUOTE 1<<3
346 #define X_SQUOTE 1<<4
352 static int xtag_cin( char c, int char_class )
354 if( char_class & X_WHITESPACE ) if( isspace(c) ) return VLC_TRUE;
355 if( char_class & X_OPENTAG ) if( c == '<' ) return VLC_TRUE;
356 if( char_class & X_CLOSETAG ) if( c == '>' ) return VLC_TRUE;
357 if( char_class & X_DQUOTE ) if( c == '"' ) return VLC_TRUE;
358 if( char_class & X_SQUOTE ) if( c == '\'' ) return VLC_TRUE;
359 if( char_class & X_EQUAL ) if( c == '=' ) return VLC_TRUE;
360 if( char_class & X_SLASH ) if( c == '/' ) return VLC_TRUE;
361 if( char_class & X_QMARK ) if( c == '!' ) return VLC_TRUE;
362 if( char_class & X_DASH ) if( c == '-' ) return VLC_TRUE;
367 static int xtag_index( XTagParser *parser, int char_class )
369 char *s = parser->start;
372 for( i = 0; s[i] && s != parser->end; i++ )
374 if( xtag_cin( s[i], char_class ) ) return i;
380 static void xtag_skip_over( XTagParser *parser, int char_class )
382 char *s = parser->start;
385 if( !parser->valid ) return;
387 for( i = 0; s[i] && s != parser->end; i++ )
389 if( !xtag_cin( s[i], char_class ) )
391 parser->start = &s[i];
399 static void xtag_skip_whitespace( XTagParser * parser )
401 xtag_skip_over( parser, X_WHITESPACE );
404 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
406 char *ret, *s = parser->start;
409 if( !parser->valid ) return NULL;
411 xi = xtag_index( parser, good_end | bad_end );
413 if( xi > 0 && xtag_cin (s[xi], good_end) )
415 ret = malloc( (xi+1) * sizeof(char) );
416 strncpy( ret, s, xi );
418 parser->start = &s[xi];
425 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
427 char *s = parser->start;
429 if( !parser->valid ) return VLC_FALSE;
431 if( !xtag_cin( s[0], char_class ) )
433 parser->valid = VLC_FALSE;
437 parser->start = &s[1];
442 static char *xtag_slurp_quoted( XTagParser *parser )
445 int quote = X_DQUOTE; /* quote char to match on */
448 if( !parser->valid ) return NULL;
450 xtag_skip_whitespace( parser );
454 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
456 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
460 for( xi = 0; s[xi]; xi++ )
462 if( xtag_cin( s[xi], quote ) )
464 if( !(xi > 1 && s[xi-1] == '\\') ) break;
468 ret = malloc( (xi+1) * sizeof(char) );
469 strncpy( ret, s, xi );
471 parser->start = &s[xi];
473 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
478 static XAttribute *xtag_parse_attribute( XTagParser *parser )
484 if( !parser->valid ) return NULL;
486 xtag_skip_whitespace( parser );
488 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
489 if( name == NULL ) return NULL;
491 xtag_skip_whitespace( parser );
494 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
497 printf( "xtag: attr failed EQUAL on <%s>\n", name );
502 xtag_skip_whitespace( parser );
504 value = xtag_slurp_quoted( parser );
509 printf ("Got NULL quoted attribute value\n");
514 attr = malloc( sizeof (*attr) );
521 parser->valid = VLC_FALSE;
525 static XTag *xtag_parse_tag( XTagParser *parser )
533 if( !parser->valid ) return NULL;
535 #if 0 /* Do we really want all the whitespace pcdata ? */
536 xtag_skip_whitespace( parser );
539 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
541 tag = malloc( sizeof(*tag) );
543 tag->pcdata = pcdata;
544 tag->parent = parser->current_tag;
545 tag->attributes = NULL;
546 tag->children = NULL;
547 tag->current_child = NULL;
554 /* if this starts a close tag, return NULL and let the parent take it */
555 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
558 /* if this starts a comment tag, skip until end */
559 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) &&
560 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
564 parser->start = s = &s[4];
566 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
568 parser->start = s = &s[xi+1];
570 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
572 parser->start = &s[2];
573 xtag_skip_whitespace( parser );
574 return xtag_parse_tag( parser );
581 /* FIXME: if this starts a DOCTYPE tag, skip until end */
582 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
584 int xi = xtag_index( parser, X_CLOSETAG );
585 if( xi <= 0 ) return NULL;
587 parser->start = &s[xi+1];
588 xtag_skip_whitespace( parser );
589 return xtag_parse_tag( parser );
592 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
594 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
595 if( name == NULL ) return NULL;
598 printf ("<%s ...\n", name);
601 tag = malloc( sizeof(*tag) );
604 tag->parent = parser->current_tag;
605 tag->attributes = NULL;
606 tag->children = NULL;
607 tag->current_child = NULL;
611 if( xtag_cin( s[0], X_WHITESPACE ) )
613 while( (attr = xtag_parse_attribute( parser )) != NULL )
615 tag->attributes = xlist_append( tag->attributes, attr );
619 xtag_skip_whitespace( parser );
623 if( xtag_cin( s[0], X_CLOSETAG ) )
625 parser->current_tag = tag;
627 xtag_assert_and_pass( parser, X_CLOSETAG );
629 while( (inner = xtag_parse_tag( parser ) ) != NULL )
631 tag->children = xlist_append( tag->children, inner );
634 parser->current_tag = tag->parent;
635 xtag_skip_whitespace( parser );
637 xtag_assert_and_pass( parser, X_OPENTAG );
638 xtag_assert_and_pass( parser, X_SLASH );
639 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
642 if( strcmp( name, tag->name ) )
645 printf ("got %s expected %s\n", name, tag->name);
647 parser->valid = VLC_FALSE;
652 xtag_skip_whitespace( parser );
653 xtag_assert_and_pass( parser, X_CLOSETAG );
658 xtag_assert_and_pass( parser, X_SLASH );
659 xtag_assert_and_pass( parser, X_CLOSETAG );
665 static XTag *xtag_free( XTag *xtag )
671 if( xtag == NULL ) return NULL;
673 if( xtag->name ) free( xtag->name );
674 if( xtag->pcdata ) free( xtag->pcdata );
676 for( l = xtag->attributes; l; l = l->next )
678 if( (attr = (XAttribute *)l->data) != NULL )
680 if( attr->name ) free( attr->name );
681 if( attr->value ) free( attr->value );
685 xlist_free( xtag->attributes );
687 for( l = xtag->children; l; l = l->next )
689 child = (XTag *)l->data;
692 xlist_free( xtag->children );
699 static XTag *xtag_new_parse( const char *s, int n )
702 XTag *tag, *ttag, *wrapper;
704 parser.valid = VLC_TRUE;
705 parser.current_tag = NULL;
706 parser.start = (char *)s;
708 if( n == -1 ) parser.end = NULL;
709 else if( n == 0 ) return NULL;
710 else parser.end = (char *)&s[n];
712 tag = xtag_parse_tag( &parser );
720 if( (ttag = xtag_parse_tag( &parser )) != NULL )
728 wrapper = malloc( sizeof(XTag) );
729 wrapper->name = NULL;
730 wrapper->pcdata = NULL;
731 wrapper->parent = NULL;
732 wrapper->attributes = NULL;
733 wrapper->children = NULL;
734 wrapper->current_child = NULL;
736 wrapper->children = xlist_append( wrapper->children, tag );
737 wrapper->children = xlist_append( wrapper->children, ttag );
739 while( (ttag = xtag_parse_tag( &parser )) != NULL )
747 wrapper->children = xlist_append( wrapper->children, ttag );
755 static char *xtag_get_name( XTag *xtag )
757 return xtag ? xtag->name : NULL;
760 static char *xtag_get_pcdata( XTag *xtag )
765 if( xtag == NULL ) return NULL;
767 for( l = xtag->children; l; l = l->next )
769 child = (XTag *)l->data;
770 if( child->pcdata != NULL )
772 return child->pcdata;
779 static char *xtag_get_attribute( XTag *xtag, char *attribute )
784 if( xtag == NULL ) return NULL;
786 for( l = xtag->attributes; l; l = l->next )
788 if( (attr = (XAttribute *)l->data) != NULL )
790 if( !strcmp( attr->name, attribute ) ) return attr->value;
797 static XTag *xtag_first_child( XTag *xtag, char *name )
802 if( xtag == NULL ) return NULL;
803 if( (l = xtag->children) == NULL ) return NULL;
807 xtag->current_child = l;
808 return (XTag *)l->data;
811 for( ; l; l = l->next )
813 child = (XTag *)l->data;
815 if( !strcmp( child->name, name ) )
817 xtag->current_child = l;
822 xtag->current_child = NULL;
827 static XTag *xtag_next_child( XTag *xtag, char *name )
832 if( xtag == NULL ) return NULL;
834 if( (l = xtag->current_child) == NULL )
835 return xtag_first_child( xtag, name );
837 if( (l = l->next) == NULL ) return NULL;
841 xtag->current_child = l;
842 return (XTag *)l->data;
845 for( ; l; l = l->next )
847 child = (XTag *)l->data;
849 if( !strcmp( child->name, name ) )
851 xtag->current_child = l;
856 xtag->current_child = NULL;
862 * This snprints function takes a variable list of char *, the last of
863 * which must be NULL, and prints each in turn to buf.
864 * Returns C99-style total length that would have been written, even if
865 * this is larger than n.
867 static int xtag_snprints( char *buf, int n, ... )
871 int len, to_copy, total = 0;
875 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
879 if( (to_copy = __MIN(n, len) ) > 0 )
881 memcpy( buf, s, to_copy );
894 static int xtag_snprint( char *buf, int n, XTag *xtag )
902 buf += __MIN(n, N); \
908 if( n > 0 ) buf[0] = '\0';
914 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
922 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
925 for( l = xtag->attributes; l; l = l->next )
927 attr = (XAttribute *)l->data;
929 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
934 if( xtag->children == NULL )
936 nn = xtag_snprints ( buf, n, "/>", NULL );
942 nn = xtag_snprints( buf, n, ">", NULL );
946 for( l = xtag->children; l; l = l->next )
948 child = (XTag *)l->data;
950 nn = xtag_snprint( buf, n, child );
956 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );