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;
168 xml_reader_sys_t *p_sys;
173 /* Open and read file */
175 i_size = stream_Size( s ) - stream_Tell( s );
176 p_buffer = malloc( i_size + 1 );
178 while( i_buffer < i_size )
180 msg_Dbg( p_xml, "got %d, want %d", i_buffer, i_size );
181 i_buffer += stream_Read( s, &p_buffer[i_buffer], i_size - i_buffer );
183 p_buffer[ i_buffer ] = 0;
187 msg_Dbg( p_xml, "empty xml" );
192 p_root = xtag_new_parse( p_buffer, i_buffer );
195 msg_Warn( p_xml, "couldn't parse xml" );
200 p_reader = malloc( sizeof(xml_reader_t) );
201 p_reader->p_sys = p_sys = malloc( sizeof(xml_reader_sys_t) );
202 p_reader->p_sys->p_root = p_root;
203 p_reader->p_sys->p_curtag = NULL;
204 p_reader->p_sys->p_curattr = NULL;
205 p_reader->p_sys->b_endtag = VLC_FALSE;
206 p_reader->p_xml = p_xml;
208 p_reader->pf_read = ReaderRead;
209 p_reader->pf_node_type = ReaderNodeType;
210 p_reader->pf_name = ReaderName;
211 p_reader->pf_value = ReaderValue;
212 p_reader->pf_next_attr = ReaderNextAttr;
213 p_reader->pf_use_dtd = ReaderUseDTD;
218 static void ReaderDelete( xml_reader_t *p_reader )
220 xtag_free( p_reader->p_sys->p_root );
221 free( p_reader->p_sys );
225 static int ReaderUseDTD ( xml_reader_t *p_reader, vlc_bool_t b_use )
230 static int ReaderRead( xml_reader_t *p_reader )
234 if( !p_reader->p_sys->p_curtag )
236 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
242 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
244 p_reader->p_sys->p_curtag = p_child;
245 p_reader->p_sys->p_curattr = 0;
246 p_reader->p_sys->b_endtag = VLC_FALSE;
250 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
251 !p_reader->p_sys->b_endtag )
253 p_reader->p_sys->b_endtag = VLC_TRUE;
257 p_reader->p_sys->b_endtag = VLC_FALSE;
258 if( !p_reader->p_sys->p_curtag->parent ) return 0;
259 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
265 static int ReaderNodeType( xml_reader_t *p_reader )
267 if( p_reader->p_sys->p_curtag->name &&
268 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
269 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
270 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
271 return XML_READER_NONE;
274 static char *ReaderName( xml_reader_t *p_reader )
276 const char *psz_name;
278 if( !p_reader->p_sys->p_curattr )
280 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
282 printf( "TAG: %s\n", psz_name );
286 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
288 if( psz_name ) return strdup( psz_name );
292 static char *ReaderValue( xml_reader_t *p_reader )
294 const char *psz_name;
295 if( p_reader->p_sys->p_curtag->pcdata )
298 printf( "%s\n", p_reader->p_sys->p_curtag->pcdata );
300 return strdup( p_reader->p_sys->p_curtag->pcdata );
303 if( !p_reader->p_sys->p_curattr ) return 0;
306 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
307 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
310 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
312 if( psz_name ) return strdup( psz_name );
316 static int ReaderNextAttr( xml_reader_t *p_reader )
318 if( !p_reader->p_sys->p_curattr )
319 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
320 else if( p_reader->p_sys->p_curattr )
321 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
323 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
324 else return VLC_EGENERIC;
327 /*****************************************************************************
328 * XTAG parser functions
329 *****************************************************************************/
331 static XList *xlist_append( XList *list, void *data )
335 l = (XList *)malloc( sizeof(XList) );
336 l->prev = l->next = NULL;
339 if( list == NULL ) return l;
341 for( last = list; last; last = last->next )
342 if( last->next == NULL ) break;
344 if( last ) last->next = l;
349 static void xlist_free( XList *list )
353 for( l = list; l; l = ln )
360 /* Character classes */
362 #define X_WHITESPACE 1<<0
363 #define X_OPENTAG 1<<1
364 #define X_CLOSETAG 1<<2
365 #define X_DQUOTE 1<<3
366 #define X_SQUOTE 1<<4
372 static int xtag_cin( char c, int char_class )
374 if( char_class & X_WHITESPACE ) if( isspace(c) ) return VLC_TRUE;
375 if( char_class & X_OPENTAG ) if( c == '<' ) return VLC_TRUE;
376 if( char_class & X_CLOSETAG ) if( c == '>' ) return VLC_TRUE;
377 if( char_class & X_DQUOTE ) if( c == '"' ) return VLC_TRUE;
378 if( char_class & X_SQUOTE ) if( c == '\'' ) return VLC_TRUE;
379 if( char_class & X_EQUAL ) if( c == '=' ) return VLC_TRUE;
380 if( char_class & X_SLASH ) if( c == '/' ) return VLC_TRUE;
381 if( char_class & X_QMARK ) if( c == '!' ) return VLC_TRUE;
382 if( char_class & X_DASH ) if( c == '-' ) return VLC_TRUE;
387 static int xtag_index( XTagParser *parser, int char_class )
389 char *s = parser->start;
392 for( i = 0; s[i] && s != parser->end; i++ )
394 if( xtag_cin( s[i], char_class ) ) return i;
400 static void xtag_skip_over( XTagParser *parser, int char_class )
402 char *s = parser->start;
405 if( !parser->valid ) return;
407 for( i = 0; s[i] && s != parser->end; i++ )
409 if( !xtag_cin( s[i], char_class ) )
411 parser->start = &s[i];
419 static void xtag_skip_whitespace( XTagParser * parser )
421 xtag_skip_over( parser, X_WHITESPACE );
424 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
426 char *ret, *s = parser->start;
429 if( !parser->valid ) return NULL;
431 xi = xtag_index( parser, good_end | bad_end );
433 if( xi > 0 && xtag_cin (s[xi], good_end) )
435 ret = malloc( (xi+1) * sizeof(char) );
436 strncpy( ret, s, xi );
438 parser->start = &s[xi];
445 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
447 char *s = parser->start;
449 if( !parser->valid ) return VLC_FALSE;
451 if( !xtag_cin( s[0], char_class ) )
453 parser->valid = VLC_FALSE;
457 parser->start = &s[1];
462 static char *xtag_slurp_quoted( XTagParser *parser )
465 int quote = X_DQUOTE; /* quote char to match on */
468 if( !parser->valid ) return NULL;
470 xtag_skip_whitespace( parser );
474 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
476 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
480 for( xi = 0; s[xi]; xi++ )
482 if( xtag_cin( s[xi], quote ) )
484 if( !(xi > 1 && s[xi-1] == '\\') ) break;
488 ret = malloc( (xi+1) * sizeof(char) );
489 strncpy( ret, s, xi );
491 parser->start = &s[xi];
493 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
498 static XAttribute *xtag_parse_attribute( XTagParser *parser )
504 if( !parser->valid ) return NULL;
506 xtag_skip_whitespace( parser );
508 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
509 if( name == NULL ) return NULL;
511 xtag_skip_whitespace( parser );
514 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
517 printf( "xtag: attr failed EQUAL on <%s>\n", name );
522 xtag_skip_whitespace( parser );
524 value = xtag_slurp_quoted( parser );
529 printf ("Got NULL quoted attribute value\n");
534 attr = malloc( sizeof (*attr) );
541 parser->valid = VLC_FALSE;
545 static XTag *xtag_parse_tag( XTagParser *parser )
553 if( !parser->valid ) return NULL;
555 #if 0 /* Do we really want all the whitespace pcdata ? */
556 xtag_skip_whitespace( parser );
559 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
561 tag = malloc( sizeof(*tag) );
563 tag->pcdata = pcdata;
564 tag->parent = parser->current_tag;
565 tag->attributes = NULL;
566 tag->children = NULL;
567 tag->current_child = NULL;
574 /* if this starts a close tag, return NULL and let the parent take it */
575 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
578 /* if this starts a comment tag, skip until end */
579 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) &&
580 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
584 parser->start = s = &s[4];
586 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
588 parser->start = s = &s[xi+1];
590 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
592 parser->start = &s[2];
593 xtag_skip_whitespace( parser );
594 return xtag_parse_tag( parser );
601 /* FIXME: if this starts a DOCTYPE tag, skip until end */
602 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
604 int xi = xtag_index( parser, X_CLOSETAG );
605 if( xi <= 0 ) return NULL;
607 parser->start = &s[xi+1];
608 xtag_skip_whitespace( parser );
609 return xtag_parse_tag( parser );
612 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
614 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
615 if( name == NULL ) return NULL;
618 printf ("<%s ...\n", name);
621 tag = malloc( sizeof(*tag) );
624 tag->parent = parser->current_tag;
625 tag->attributes = NULL;
626 tag->children = NULL;
627 tag->current_child = NULL;
631 if( xtag_cin( s[0], X_WHITESPACE ) )
633 while( (attr = xtag_parse_attribute( parser )) != NULL )
635 tag->attributes = xlist_append( tag->attributes, attr );
639 xtag_skip_whitespace( parser );
643 if( xtag_cin( s[0], X_CLOSETAG ) )
645 parser->current_tag = tag;
647 xtag_assert_and_pass( parser, X_CLOSETAG );
649 while( (inner = xtag_parse_tag( parser ) ) != NULL )
651 tag->children = xlist_append( tag->children, inner );
654 parser->current_tag = tag->parent;
655 xtag_skip_whitespace( parser );
657 xtag_assert_and_pass( parser, X_OPENTAG );
658 xtag_assert_and_pass( parser, X_SLASH );
659 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
662 if( strcmp( name, tag->name ) )
665 printf ("got %s expected %s\n", name, tag->name);
667 parser->valid = VLC_FALSE;
672 xtag_skip_whitespace( parser );
673 xtag_assert_and_pass( parser, X_CLOSETAG );
678 xtag_assert_and_pass( parser, X_SLASH );
679 xtag_assert_and_pass( parser, X_CLOSETAG );
685 static XTag *xtag_free( XTag *xtag )
691 if( xtag == NULL ) return NULL;
693 if( xtag->name ) free( xtag->name );
694 if( xtag->pcdata ) free( xtag->pcdata );
696 for( l = xtag->attributes; l; l = l->next )
698 if( (attr = (XAttribute *)l->data) != NULL )
700 if( attr->name ) free( attr->name );
701 if( attr->value ) free( attr->value );
705 xlist_free( xtag->attributes );
707 for( l = xtag->children; l; l = l->next )
709 child = (XTag *)l->data;
712 xlist_free( xtag->children );
719 static XTag *xtag_new_parse( const char *s, int n )
722 XTag *tag, *ttag, *wrapper;
724 parser.valid = VLC_TRUE;
725 parser.current_tag = NULL;
726 parser.start = (char *)s;
728 if( n == -1 ) parser.end = NULL;
732 printf ("empty buffer");
736 else parser.end = (char *)&s[n];
738 /* can't have whitespace pcdata outside rootnode */
739 xtag_skip_whitespace( &parser );
741 tag = xtag_parse_tag( &parser );
746 printf ("invalid file");
752 if( (ttag = xtag_parse_tag( &parser )) != NULL )
760 wrapper = malloc( sizeof(XTag) );
761 wrapper->name = NULL;
762 wrapper->pcdata = NULL;
763 wrapper->parent = NULL;
764 wrapper->attributes = NULL;
765 wrapper->children = NULL;
766 wrapper->current_child = NULL;
768 wrapper->children = xlist_append( wrapper->children, tag );
769 wrapper->children = xlist_append( wrapper->children, ttag );
771 while( (ttag = xtag_parse_tag( &parser )) != NULL )
779 wrapper->children = xlist_append( wrapper->children, ttag );
787 static char *xtag_get_name( XTag *xtag )
789 return xtag ? xtag->name : NULL;
792 static char *xtag_get_pcdata( XTag *xtag )
797 if( xtag == NULL ) return NULL;
799 for( l = xtag->children; l; l = l->next )
801 child = (XTag *)l->data;
802 if( child->pcdata != NULL )
804 return child->pcdata;
811 static char *xtag_get_attribute( XTag *xtag, char *attribute )
816 if( xtag == NULL ) return NULL;
818 for( l = xtag->attributes; l; l = l->next )
820 if( (attr = (XAttribute *)l->data) != NULL )
822 if( !strcmp( attr->name, attribute ) ) return attr->value;
829 static XTag *xtag_first_child( XTag *xtag, char *name )
834 if( xtag == NULL ) return NULL;
835 if( (l = xtag->children) == NULL ) return NULL;
839 xtag->current_child = l;
840 return (XTag *)l->data;
843 for( ; l; l = l->next )
845 child = (XTag *)l->data;
847 if( !strcmp( child->name, name ) )
849 xtag->current_child = l;
854 xtag->current_child = NULL;
859 static XTag *xtag_next_child( XTag *xtag, char *name )
864 if( xtag == NULL ) return NULL;
866 if( (l = xtag->current_child) == NULL )
867 return xtag_first_child( xtag, name );
869 if( (l = l->next) == NULL ) return NULL;
873 xtag->current_child = l;
874 return (XTag *)l->data;
877 for( ; l; l = l->next )
879 child = (XTag *)l->data;
881 if( !strcmp( child->name, name ) )
883 xtag->current_child = l;
888 xtag->current_child = NULL;
894 * This snprints function takes a variable list of char *, the last of
895 * which must be NULL, and prints each in turn to buf.
896 * Returns C99-style total length that would have been written, even if
897 * this is larger than n.
899 static int xtag_snprints( char *buf, int n, ... )
903 int len, to_copy, total = 0;
907 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
911 if( (to_copy = __MIN(n, len) ) > 0 )
913 memcpy( buf, s, to_copy );
926 static int xtag_snprint( char *buf, int n, XTag *xtag )
934 buf += __MIN(n, N); \
940 if( n > 0 ) buf[0] = '\0';
946 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
954 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
957 for( l = xtag->attributes; l; l = l->next )
959 attr = (XAttribute *)l->data;
961 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
966 if( xtag->children == NULL )
968 nn = xtag_snprints ( buf, n, "/>", NULL );
974 nn = xtag_snprints( buf, n, ">", NULL );
978 for( l = xtag->children; l; l = l->next )
980 child = (XTag *)l->data;
982 nn = xtag_snprint( buf, n, child );
988 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );