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_size = stream_Size( s ) - stream_Tell( s );
174 p_buffer = malloc( i_size + 1 );
176 while( i_buffer < i_size )
178 msg_Dbg( p_xml, "got %d, want %d", i_buffer, i_size );
179 i_buffer += stream_Read( s, &p_buffer[i_buffer], i_size - i_buffer );
181 p_buffer[ i_buffer ] = 0;
185 msg_Dbg( p_xml, "empty xml" );
190 p_root = xtag_new_parse( p_buffer, i_buffer );
193 msg_Warn( p_xml, "couldn't parse xml" );
198 p_reader = malloc( sizeof(xml_reader_t) );
199 p_reader->p_sys = p_sys = malloc( sizeof(xml_reader_sys_t) );
200 p_reader->p_sys->p_root = p_root;
201 p_reader->p_sys->p_curtag = NULL;
202 p_reader->p_sys->p_curattr = NULL;
203 p_reader->p_sys->b_endtag = VLC_FALSE;
204 p_reader->p_xml = p_xml;
206 p_reader->pf_read = ReaderRead;
207 p_reader->pf_node_type = ReaderNodeType;
208 p_reader->pf_name = ReaderName;
209 p_reader->pf_value = ReaderValue;
210 p_reader->pf_next_attr = ReaderNextAttr;
215 static void ReaderDelete( xml_reader_t *p_reader )
217 xtag_free( p_reader->p_sys->p_root );
218 free( p_reader->p_sys );
222 static int ReaderRead( xml_reader_t *p_reader )
226 if( !p_reader->p_sys->p_curtag )
228 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
234 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
236 p_reader->p_sys->p_curtag = p_child;
237 p_reader->p_sys->p_curattr = 0;
238 p_reader->p_sys->b_endtag = VLC_FALSE;
242 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
243 !p_reader->p_sys->b_endtag )
245 p_reader->p_sys->b_endtag = VLC_TRUE;
249 p_reader->p_sys->b_endtag = VLC_FALSE;
250 if( !p_reader->p_sys->p_curtag->parent ) return 0;
251 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
257 static int ReaderNodeType( xml_reader_t *p_reader )
259 if( p_reader->p_sys->p_curtag->name &&
260 p_reader->p_sys->b_endtag ) return XML_READER_ENDELEM;
261 if( p_reader->p_sys->p_curtag->name ) return XML_READER_STARTELEM;
262 if( p_reader->p_sys->p_curtag->pcdata ) return XML_READER_TEXT;
263 return XML_READER_NONE;
266 static char *ReaderName( xml_reader_t *p_reader )
268 const char *psz_name;
270 if( !p_reader->p_sys->p_curattr )
272 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
274 printf( "TAG: %s\n", psz_name );
278 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
280 if( psz_name ) return strdup( psz_name );
284 static char *ReaderValue( xml_reader_t *p_reader )
286 const char *psz_name;
287 if( p_reader->p_sys->p_curtag->pcdata )
290 printf( "%s\n", p_reader->p_sys->p_curtag->pcdata );
292 return strdup( p_reader->p_sys->p_curtag->pcdata );
295 if( !p_reader->p_sys->p_curattr ) return 0;
298 printf( "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
299 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
302 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
304 if( psz_name ) return strdup( psz_name );
308 static int ReaderNextAttr( xml_reader_t *p_reader )
310 if( !p_reader->p_sys->p_curattr )
311 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
312 else if( p_reader->p_sys->p_curattr )
313 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
315 if( p_reader->p_sys->p_curattr ) return VLC_SUCCESS;
316 else return VLC_EGENERIC;
319 /*****************************************************************************
320 * XTAG parser functions
321 *****************************************************************************/
323 static XList *xlist_append( XList *list, void *data )
327 l = (XList *)malloc( sizeof(XList) );
328 l->prev = l->next = NULL;
331 if( list == NULL ) return l;
333 for( last = list; last; last = last->next )
334 if( last->next == NULL ) break;
336 if( last ) last->next = l;
341 static void xlist_free( XList *list )
345 for( l = list; l; l = ln )
352 /* Character classes */
354 #define X_WHITESPACE 1<<0
355 #define X_OPENTAG 1<<1
356 #define X_CLOSETAG 1<<2
357 #define X_DQUOTE 1<<3
358 #define X_SQUOTE 1<<4
364 static int xtag_cin( char c, int char_class )
366 if( char_class & X_WHITESPACE ) if( isspace(c) ) return VLC_TRUE;
367 if( char_class & X_OPENTAG ) if( c == '<' ) return VLC_TRUE;
368 if( char_class & X_CLOSETAG ) if( c == '>' ) return VLC_TRUE;
369 if( char_class & X_DQUOTE ) if( c == '"' ) return VLC_TRUE;
370 if( char_class & X_SQUOTE ) if( c == '\'' ) return VLC_TRUE;
371 if( char_class & X_EQUAL ) if( c == '=' ) return VLC_TRUE;
372 if( char_class & X_SLASH ) if( c == '/' ) return VLC_TRUE;
373 if( char_class & X_QMARK ) if( c == '!' ) return VLC_TRUE;
374 if( char_class & X_DASH ) if( c == '-' ) return VLC_TRUE;
379 static int xtag_index( XTagParser *parser, int char_class )
381 char *s = parser->start;
384 for( i = 0; s[i] && s != parser->end; i++ )
386 if( xtag_cin( s[i], char_class ) ) return i;
392 static void xtag_skip_over( XTagParser *parser, int char_class )
394 char *s = parser->start;
397 if( !parser->valid ) return;
399 for( i = 0; s[i] && s != parser->end; i++ )
401 if( !xtag_cin( s[i], char_class ) )
403 parser->start = &s[i];
411 static void xtag_skip_whitespace( XTagParser * parser )
413 xtag_skip_over( parser, X_WHITESPACE );
416 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
418 char *ret, *s = parser->start;
421 if( !parser->valid ) return NULL;
423 xi = xtag_index( parser, good_end | bad_end );
425 if( xi > 0 && xtag_cin (s[xi], good_end) )
427 ret = malloc( (xi+1) * sizeof(char) );
428 strncpy( ret, s, xi );
430 parser->start = &s[xi];
437 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
439 char *s = parser->start;
441 if( !parser->valid ) return VLC_FALSE;
443 if( !xtag_cin( s[0], char_class ) )
445 parser->valid = VLC_FALSE;
449 parser->start = &s[1];
454 static char *xtag_slurp_quoted( XTagParser *parser )
457 int quote = X_DQUOTE; /* quote char to match on */
460 if( !parser->valid ) return NULL;
462 xtag_skip_whitespace( parser );
466 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
468 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
472 for( xi = 0; s[xi]; xi++ )
474 if( xtag_cin( s[xi], quote ) )
476 if( !(xi > 1 && s[xi-1] == '\\') ) break;
480 ret = malloc( (xi+1) * sizeof(char) );
481 strncpy( ret, s, xi );
483 parser->start = &s[xi];
485 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
490 static XAttribute *xtag_parse_attribute( XTagParser *parser )
496 if( !parser->valid ) return NULL;
498 xtag_skip_whitespace( parser );
500 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
501 if( name == NULL ) return NULL;
503 xtag_skip_whitespace( parser );
506 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
509 printf( "xtag: attr failed EQUAL on <%s>\n", name );
514 xtag_skip_whitespace( parser );
516 value = xtag_slurp_quoted( parser );
521 printf ("Got NULL quoted attribute value\n");
526 attr = malloc( sizeof (*attr) );
533 parser->valid = VLC_FALSE;
537 static XTag *xtag_parse_tag( XTagParser *parser )
545 if( !parser->valid ) return NULL;
547 #if 0 /* Do we really want all the whitespace pcdata ? */
548 xtag_skip_whitespace( parser );
551 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
553 tag = malloc( sizeof(*tag) );
555 tag->pcdata = pcdata;
556 tag->parent = parser->current_tag;
557 tag->attributes = NULL;
558 tag->children = NULL;
559 tag->current_child = NULL;
566 /* if this starts a close tag, return NULL and let the parent take it */
567 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
570 /* if this starts a comment tag, skip until end */
571 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) &&
572 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
576 parser->start = s = &s[4];
578 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
580 parser->start = s = &s[xi+1];
582 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
584 parser->start = &s[2];
585 xtag_skip_whitespace( parser );
586 return xtag_parse_tag( parser );
593 /* FIXME: if this starts a DOCTYPE tag, skip until end */
594 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
596 int xi = xtag_index( parser, X_CLOSETAG );
597 if( xi <= 0 ) return NULL;
599 parser->start = &s[xi+1];
600 xtag_skip_whitespace( parser );
601 return xtag_parse_tag( parser );
604 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
606 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
607 if( name == NULL ) return NULL;
610 printf ("<%s ...\n", name);
613 tag = malloc( sizeof(*tag) );
616 tag->parent = parser->current_tag;
617 tag->attributes = NULL;
618 tag->children = NULL;
619 tag->current_child = NULL;
623 if( xtag_cin( s[0], X_WHITESPACE ) )
625 while( (attr = xtag_parse_attribute( parser )) != NULL )
627 tag->attributes = xlist_append( tag->attributes, attr );
631 xtag_skip_whitespace( parser );
635 if( xtag_cin( s[0], X_CLOSETAG ) )
637 parser->current_tag = tag;
639 xtag_assert_and_pass( parser, X_CLOSETAG );
641 while( (inner = xtag_parse_tag( parser ) ) != NULL )
643 tag->children = xlist_append( tag->children, inner );
646 parser->current_tag = tag->parent;
647 xtag_skip_whitespace( parser );
649 xtag_assert_and_pass( parser, X_OPENTAG );
650 xtag_assert_and_pass( parser, X_SLASH );
651 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
654 if( strcmp( name, tag->name ) )
657 printf ("got %s expected %s\n", name, tag->name);
659 parser->valid = VLC_FALSE;
664 xtag_skip_whitespace( parser );
665 xtag_assert_and_pass( parser, X_CLOSETAG );
670 xtag_assert_and_pass( parser, X_SLASH );
671 xtag_assert_and_pass( parser, X_CLOSETAG );
677 static XTag *xtag_free( XTag *xtag )
683 if( xtag == NULL ) return NULL;
685 if( xtag->name ) free( xtag->name );
686 if( xtag->pcdata ) free( xtag->pcdata );
688 for( l = xtag->attributes; l; l = l->next )
690 if( (attr = (XAttribute *)l->data) != NULL )
692 if( attr->name ) free( attr->name );
693 if( attr->value ) free( attr->value );
697 xlist_free( xtag->attributes );
699 for( l = xtag->children; l; l = l->next )
701 child = (XTag *)l->data;
704 xlist_free( xtag->children );
711 static XTag *xtag_new_parse( const char *s, int n )
714 XTag *tag, *ttag, *wrapper;
716 parser.valid = VLC_TRUE;
717 parser.current_tag = NULL;
718 parser.start = (char *)s;
720 if( n == -1 ) parser.end = NULL;
724 printf ("empty buffer");
728 else parser.end = (char *)&s[n];
730 /* can't have whitespace pcdata outside rootnode */
731 xtag_skip_whitespace( &parser );
733 tag = xtag_parse_tag( &parser );
738 printf ("invalid file");
744 if( (ttag = xtag_parse_tag( &parser )) != NULL )
752 wrapper = malloc( sizeof(XTag) );
753 wrapper->name = NULL;
754 wrapper->pcdata = NULL;
755 wrapper->parent = NULL;
756 wrapper->attributes = NULL;
757 wrapper->children = NULL;
758 wrapper->current_child = NULL;
760 wrapper->children = xlist_append( wrapper->children, tag );
761 wrapper->children = xlist_append( wrapper->children, ttag );
763 while( (ttag = xtag_parse_tag( &parser )) != NULL )
771 wrapper->children = xlist_append( wrapper->children, ttag );
779 static char *xtag_get_name( XTag *xtag )
781 return xtag ? xtag->name : NULL;
784 static char *xtag_get_pcdata( XTag *xtag )
789 if( xtag == NULL ) return NULL;
791 for( l = xtag->children; l; l = l->next )
793 child = (XTag *)l->data;
794 if( child->pcdata != NULL )
796 return child->pcdata;
803 static char *xtag_get_attribute( XTag *xtag, char *attribute )
808 if( xtag == NULL ) return NULL;
810 for( l = xtag->attributes; l; l = l->next )
812 if( (attr = (XAttribute *)l->data) != NULL )
814 if( !strcmp( attr->name, attribute ) ) return attr->value;
821 static XTag *xtag_first_child( XTag *xtag, char *name )
826 if( xtag == NULL ) return NULL;
827 if( (l = xtag->children) == NULL ) return NULL;
831 xtag->current_child = l;
832 return (XTag *)l->data;
835 for( ; l; l = l->next )
837 child = (XTag *)l->data;
839 if( !strcmp( child->name, name ) )
841 xtag->current_child = l;
846 xtag->current_child = NULL;
851 static XTag *xtag_next_child( XTag *xtag, char *name )
856 if( xtag == NULL ) return NULL;
858 if( (l = xtag->current_child) == NULL )
859 return xtag_first_child( xtag, name );
861 if( (l = l->next) == NULL ) return NULL;
865 xtag->current_child = l;
866 return (XTag *)l->data;
869 for( ; l; l = l->next )
871 child = (XTag *)l->data;
873 if( !strcmp( child->name, name ) )
875 xtag->current_child = l;
880 xtag->current_child = NULL;
886 * This snprints function takes a variable list of char *, the last of
887 * which must be NULL, and prints each in turn to buf.
888 * Returns C99-style total length that would have been written, even if
889 * this is larger than n.
891 static int xtag_snprints( char *buf, int n, ... )
895 int len, to_copy, total = 0;
899 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
903 if( (to_copy = __MIN(n, len) ) > 0 )
905 memcpy( buf, s, to_copy );
918 static int xtag_snprint( char *buf, int n, XTag *xtag )
926 buf += __MIN(n, N); \
932 if( n > 0 ) buf[0] = '\0';
938 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
946 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
949 for( l = xtag->attributes; l; l = l->next )
951 attr = (XAttribute *)l->data;
953 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
958 if( xtag->children == NULL )
960 nn = xtag_snprints ( buf, n, "/>", NULL );
966 nn = xtag_snprints( buf, n, ">", NULL );
970 for( l = xtag->children; l; l = l->next )
972 child = (XTag *)l->data;
974 nn = xtag_snprint( buf, n, child );
980 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );