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 the VideoLAN team
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., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
27 *****************************************************************************/
33 #include <vlc_common.h>
34 #include <vlc_plugin.h>
37 #include <vlc_block.h>
38 #include <vlc_stream.h>
39 #include <vlc_memory.h>
54 * struct XTag is kind of a union ... it normally represents a whole
55 * tag (and its children), but it could alternatively represent some
56 * PCDATA. Basically, if tag->pcdata is non-NULL, interpret only it and
57 * ignore the name, attributes and inner_tags.
69 typedef struct _XAttribute
75 typedef struct _XTagParser
77 int valid; /* boolean */
83 /*****************************************************************************
85 *****************************************************************************/
86 static int Open ( vlc_object_t * );
87 static void Close( vlc_object_t * );
88 static int ReaderOpen( vlc_object_t * );
89 static void ReaderClose( vlc_object_t * );
92 set_description( N_("Simple XML Parser") )
93 set_capability( "xml", 5 )
94 set_callbacks( Open, Close )
97 set_capability( "xml reader", 5 )
98 set_callbacks( ReaderOpen, ReaderClose )
101 struct xml_reader_sys_t
103 XTag *p_root; /* Root tag */
104 XTag *p_curtag; /* Current tag */
105 XList *p_curattr; /* Current attribute */
109 static int ReaderRead( xml_reader_t * );
110 static int ReaderNodeType( xml_reader_t * );
111 static char *ReaderName( xml_reader_t * );
112 static char *ReaderValue( xml_reader_t * );
113 static int ReaderNextAttr( xml_reader_t * );
115 static int ReaderUseDTD ( xml_reader_t *, bool );
117 static void CatalogLoad( xml_t *, const char * );
118 static void CatalogAdd( xml_t *, const char *, const char *, const char * );
120 static XTag *xtag_new_parse( const char *, int );
121 static char *xtag_get_name( XTag * );
123 static char *xtag_get_pcdata( XTag * );
124 static char *xtag_get_attribute( XTag *, char * );
126 static XTag *xtag_first_child( XTag *, char * );
127 static XTag *xtag_next_child( XTag *, char * );
128 static void xtag_free( XTag * );
130 static int xtag_snprint( char *, int, XTag * );
133 /*****************************************************************************
134 * Module initialization
135 *****************************************************************************/
136 static int Open( vlc_object_t *p_this )
138 xml_t *p_xml = (xml_t *)p_this;
140 p_xml->pf_catalog_load = CatalogLoad;
141 p_xml->pf_catalog_add = CatalogAdd;
146 /*****************************************************************************
147 * Module deinitialization
148 *****************************************************************************/
149 static void Close( vlc_object_t *p_this )
155 /*****************************************************************************
156 * Catalogue functions
157 *****************************************************************************/
158 static void CatalogLoad( xml_t *p_xml, const char *psz_filename )
160 VLC_UNUSED(psz_filename);
161 msg_Dbg( p_xml, "catalog support not implemented" );
164 static void CatalogAdd( xml_t *p_xml, const char *psz_arg1,
165 const char *psz_arg2, const char *psz_filename )
167 VLC_UNUSED(p_xml); VLC_UNUSED(psz_arg1); VLC_UNUSED(psz_arg2);
168 VLC_UNUSED(psz_filename);
171 /*****************************************************************************
173 *****************************************************************************/
174 static int ReaderOpen( vlc_object_t *p_this )
176 xml_reader_t *p_reader = (xml_reader_t *)p_this;
177 stream_t *s = p_reader->p_stream;
179 int i_size, i_pos = 0, i_buffer = 2048;
182 /* Open and read file */
183 p_buffer = malloc( i_buffer );
184 if( p_buffer == NULL )
187 while( ( i_size = stream_Read( s, &p_buffer[i_pos], 2048 ) ) == 2048 )
191 p_buffer = realloc_or_free( p_buffer, i_buffer );
195 if( i_pos + i_size == 0 )
197 msg_Dbg( p_this, "empty XML" );
201 p_buffer[ i_pos + i_size ] = '\0'; /* 0 terminated string */
203 p_root = xtag_new_parse( p_buffer, i_buffer );
207 msg_Warn( p_this, "couldn't parse XML" );
211 p_reader->p_sys = malloc( sizeof(xml_reader_sys_t) );
212 if( !p_reader->p_sys )
217 p_reader->p_sys->p_root = p_root;
218 p_reader->p_sys->p_curtag = NULL;
219 p_reader->p_sys->p_curattr = NULL;
220 p_reader->p_sys->b_endtag = false;
222 p_reader->pf_read = ReaderRead;
223 p_reader->pf_node_type = ReaderNodeType;
224 p_reader->pf_name = ReaderName;
225 p_reader->pf_value = ReaderValue;
226 p_reader->pf_next_attr = ReaderNextAttr;
227 p_reader->pf_use_dtd = ReaderUseDTD;
232 static void ReaderClose( vlc_object_t *p_this )
234 xml_reader_t *p_reader = (vlc_object_t *)p_this;
236 xtag_free( p_reader->p_sys->p_root );
237 free( p_reader->p_sys );
240 static int ReaderUseDTD ( xml_reader_t *p_reader, bool b_use )
242 VLC_UNUSED(p_reader); VLC_UNUSED(b_use);
246 static int ReaderRead( xml_reader_t *p_reader )
250 if( !p_reader->p_sys->p_curtag )
252 p_reader->p_sys->p_curtag = p_reader->p_sys->p_root;
258 if( (p_child = xtag_next_child( p_reader->p_sys->p_curtag, 0 )) )
260 p_reader->p_sys->p_curtag = p_child;
261 p_reader->p_sys->p_curattr = NULL;
262 p_reader->p_sys->b_endtag = false;
266 if( p_reader->p_sys->p_curtag->name && /* no end tag for pcdata */
267 !p_reader->p_sys->b_endtag )
269 p_reader->p_sys->b_endtag = true;
273 p_reader->p_sys->b_endtag = false;
274 if( !p_reader->p_sys->p_curtag->parent ) return 0;
275 p_reader->p_sys->p_curtag = p_reader->p_sys->p_curtag->parent;
281 static int ReaderNodeType( xml_reader_t *p_reader )
283 if( p_reader->p_sys->p_curtag->name && p_reader->p_sys->b_endtag )
284 return XML_READER_ENDELEM;
285 if( p_reader->p_sys->p_curtag->name )
286 return XML_READER_STARTELEM;
287 if( p_reader->p_sys->p_curtag->pcdata )
288 return XML_READER_TEXT;
289 return XML_READER_NONE;
292 static char *ReaderName( xml_reader_t *p_reader )
294 const char *psz_name;
296 if( !p_reader->p_sys->p_curattr )
298 psz_name = xtag_get_name( p_reader->p_sys->p_curtag );
300 fprintf( stderr, "TAG: %s\n", psz_name );
304 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->name;
306 return psz_name ? strdup( psz_name ) : NULL;
309 static char *ReaderValue( xml_reader_t *p_reader )
311 const char *psz_name;
312 if( p_reader->p_sys->p_curtag->pcdata )
315 fprintf( stderr, "%s\n", p_reader->p_sys->p_curtag->pcdata );
317 return strdup( p_reader->p_sys->p_curtag->pcdata );
320 if( !p_reader->p_sys->p_curattr ) return NULL;
323 fprintf( stderr, "%s=%s\n", ((XAttribute *)p_reader->p_sys->p_curattr->data)->name,
324 ((XAttribute *)p_reader->p_sys->p_curattr->data)->value );
327 psz_name = ((XAttribute *)p_reader->p_sys->p_curattr->data)->value;
329 return psz_name ? strdup( psz_name ) : NULL;
332 static int ReaderNextAttr( xml_reader_t *p_reader )
334 if( !p_reader->p_sys->p_curattr )
335 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curtag->attributes;
336 else if( p_reader->p_sys->p_curattr )
337 p_reader->p_sys->p_curattr = p_reader->p_sys->p_curattr->next;
339 return p_reader->p_sys->p_curattr ? VLC_SUCCESS : VLC_EGENERIC;
342 /*****************************************************************************
343 * XTAG parser functions
344 *****************************************************************************/
346 static XList *xlist_append( XList *list, void *data )
350 l = (XList *)xmalloc( sizeof(XList) );
351 l->prev = l->next = NULL;
357 /* Find the last element */
367 static void xlist_free( XList *list )
371 for( l = list; l; l = ln )
378 /* Character classes */
380 #define X_WHITESPACE 1<<0
381 #define X_OPENTAG 1<<1
382 #define X_CLOSETAG 1<<2
383 #define X_DQUOTE 1<<3
384 #define X_SQUOTE 1<<4
391 static int xtag_cin( char c, int char_class )
393 if( char_class & X_WHITESPACE ) if( isspace(c) ) return true;
394 if( char_class & X_OPENTAG ) if( c == '<' ) return true;
395 if( char_class & X_CLOSETAG ) if( c == '>' ) return true;
396 if( char_class & X_DQUOTE ) if( c == '"' ) return true;
397 if( char_class & X_SQUOTE ) if( c == '\'' ) return true;
398 if( char_class & X_EQUAL ) if( c == '=' ) return true;
399 if( char_class & X_SLASH ) if( c == '/' ) return true;
400 if( char_class & X_QMARK ) if( c == '?' ) return true;
401 if( char_class & X_DASH ) if( c == '-' ) return true;
402 if( char_class & X_EMARK ) if( c == '!' ) return true;
407 static int xtag_index( XTagParser *parser, int char_class )
409 char *s = parser->start;
412 for( i = 0; s[i] && s != parser->end; i++ )
414 if( xtag_cin( s[i], char_class ) ) return i;
420 static void xtag_skip_over( XTagParser *parser, int char_class )
422 char *s = parser->start;
425 if( !parser->valid ) return;
427 for( i = 0; s[i] && s != parser->end; i++ )
429 if( !xtag_cin( s[i], char_class ) )
431 parser->start = &s[i];
439 static void xtag_skip_whitespace( XTagParser * parser )
441 xtag_skip_over( parser, X_WHITESPACE );
444 static char *xtag_slurp_to( XTagParser *parser, int good_end, int bad_end )
446 char *ret, *s = parser->start;
449 if( !parser->valid ) return NULL;
451 xi = xtag_index( parser, good_end | bad_end );
453 if( xi > 0 && xtag_cin (s[xi], good_end) )
455 ret = xmalloc( xi+1 );
456 strncpy( ret, s, xi );
458 parser->start = &s[xi];
465 static int xtag_assert_and_pass( XTagParser *parser, int char_class )
467 char *s = parser->start;
469 if( !parser->valid ) return false;
471 if( !xtag_cin( s[0], char_class ) )
473 parser->valid = false;
477 parser->start = &s[1];
482 static char *xtag_slurp_quoted( XTagParser *parser )
485 int quote = X_DQUOTE; /* quote char to match on */
488 if( !parser->valid ) return NULL;
490 xtag_skip_whitespace( parser );
494 if( xtag_cin( s[0], X_SQUOTE ) ) quote = X_SQUOTE;
496 if( !xtag_assert_and_pass( parser, quote ) ) return NULL;
500 for( xi = 0; s[xi]; xi++ )
502 if( xtag_cin( s[xi], quote ) )
504 if( !(xi > 1 && s[xi-1] == '\\') ) break;
508 ret = xmalloc( xi+1 );
509 strncpy( ret, s, xi );
511 parser->start = &s[xi];
513 if( !xtag_assert_and_pass( parser, quote ) )
522 static XAttribute *xtag_parse_attribute( XTagParser *parser )
531 xtag_skip_whitespace( parser );
533 name = xtag_slurp_to( parser, X_WHITESPACE|X_EQUAL, X_SLASH|X_CLOSETAG );
537 xtag_skip_whitespace( parser );
540 if( !xtag_assert_and_pass( parser, X_EQUAL ) )
543 fprintf( stderr, "xtag: attr failed EQUAL on <%s>\n", name );
548 xtag_skip_whitespace( parser );
550 value = xtag_slurp_quoted( parser );
555 fprintf (stderr, "Got NULL quoted attribute value\n");
560 attr = xmalloc( sizeof (*attr) );
567 parser->valid = false;
571 static XTag *xtag_parse_tag( XTagParser *parser )
580 if( !parser->valid ) return NULL;
584 /* if this starts a comment tag, skip until end */
585 if( (parser->end - parser->start) > 7 &&
586 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_EMARK ) &&
587 xtag_cin( s[2], X_DASH ) && xtag_cin( s[3], X_DASH ) )
589 parser->start = s = &s[4];
590 while( (xi = xtag_index( parser, X_DASH )) >= 0 )
592 parser->start = s = &s[xi+1];
593 if( xtag_cin( s[0], X_DASH ) && xtag_cin( s[1], X_CLOSETAG ) )
595 parser->start = &s[2];
596 xtag_skip_whitespace( parser );
597 return xtag_parse_tag( parser );
603 /* ignore processing instructions '<?' ... '?>' */
604 if( (parser->end - parser->start) > 4 &&
605 xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_QMARK ) )
607 parser->start = s = &s[2];
608 while ((xi = xtag_index( parser, X_QMARK )) >= 0) {
609 if (xtag_cin( s[xi+1], X_CLOSETAG )) {
610 parser->start = &s[xi+2];
611 xtag_skip_whitespace( parser );
612 return xtag_parse_tag( parser );
618 /* ignore doctype '<!DOCTYPE' ... '>' */
619 if ( (parser->end - parser->start) > 8 &&
620 !strncmp( s, "<!DOCTYPE", 9 ) ) {
621 xi = xtag_index( parser, X_CLOSETAG );
623 parser->start = &s[xi+1];
624 xtag_skip_whitespace( parser );
625 return xtag_parse_tag( parser );
632 if( (pcdata = xtag_slurp_to( parser, X_OPENTAG, X_NONE )) != NULL )
634 tag = xmalloc( sizeof(*tag) );
636 tag->pcdata = pcdata;
637 tag->parent = parser->current_tag;
638 tag->attributes = NULL;
639 tag->children = NULL;
640 tag->current_child = NULL;
645 /* if this starts a close tag, return NULL and let the parent take it */
646 if( xtag_cin( s[0], X_OPENTAG ) && xtag_cin( s[1], X_SLASH ) )
649 /* parse CDATA content */
650 if ( (parser->end - parser->start) > 8 &&
651 !strncmp( s, "<![CDATA[", 9 ) ) {
652 parser->start = s = &s[9];
653 while (parser->end - s > 2) {
654 if (strncmp( s, "]]>", 3 ) == 0) {
655 if ( !(tag = malloc( sizeof(*tag))) ) return NULL;
656 if ( !(pcdata = malloc( s - parser->start + 1)) )
661 strncpy( pcdata, parser->start, s - parser->start );
662 pcdata[s - parser->start]='\0';
663 parser->start = &s[3];
665 tag->pcdata = pcdata;
666 tag->parent = parser->current_tag;
667 tag->attributes = NULL;
668 tag->children = NULL;
669 tag->current_child = NULL;
679 if( !xtag_assert_and_pass( parser, X_OPENTAG ) ) return NULL;
681 name = xtag_slurp_to( parser, X_WHITESPACE|X_SLASH|X_CLOSETAG, X_NONE );
682 if( name == NULL ) return NULL;
685 fprintf (stderr, "<%s ...\n", name);
688 tag = xmalloc( sizeof(*tag) );
691 tag->parent = parser->current_tag;
692 tag->attributes = NULL;
693 tag->children = NULL;
694 tag->current_child = NULL;
698 if( xtag_cin( s[0], X_WHITESPACE ) )
700 while( (attr = xtag_parse_attribute( parser )) != NULL )
702 tag->attributes = xlist_append( tag->attributes, attr );
706 xtag_skip_whitespace( parser );
710 if( xtag_cin( s[0], X_CLOSETAG ) )
712 parser->current_tag = tag;
714 xtag_assert_and_pass( parser, X_CLOSETAG );
716 while( (inner = xtag_parse_tag( parser ) ) != NULL )
718 tag->children = xlist_append( tag->children, inner );
721 parser->current_tag = tag->parent;
722 xtag_skip_whitespace( parser );
724 xtag_assert_and_pass( parser, X_OPENTAG );
725 xtag_assert_and_pass( parser, X_SLASH );
726 name = xtag_slurp_to( parser, X_WHITESPACE | X_CLOSETAG, X_NONE );
729 if( strcmp( name, tag->name ) )
732 fprintf (stderr, "got %s expected %s\n", name, tag->name);
734 parser->valid = false;
739 xtag_skip_whitespace( parser );
740 xtag_assert_and_pass( parser, X_CLOSETAG );
741 xtag_skip_whitespace( parser );
745 xtag_assert_and_pass( parser, X_SLASH );
746 xtag_assert_and_pass( parser, X_CLOSETAG );
747 xtag_skip_whitespace( parser );
753 static void xtag_free( XTag *xtag )
763 free( xtag->pcdata );
765 for( l = xtag->attributes; l; l = l->next )
767 if( (attr = (XAttribute *)l->data) != NULL )
774 xlist_free( xtag->attributes );
776 for( l = xtag->children; l; l = l->next )
778 child = (XTag *)l->data;
781 xlist_free( xtag->children );
786 static XTag *xtag_new_parse( const char *s, int n )
789 XTag *tag, *ttag, *wrapper;
792 parser.current_tag = NULL;
793 parser.start = (char *)s;
795 if( n == -1 ) parser.end = NULL;
799 fprintf (stderr, "empty buffer\n");
803 else parser.end = (char *)&s[n];
805 /* can't have whitespace pcdata outside rootnode */
806 xtag_skip_whitespace( &parser );
808 tag = xtag_parse_tag( &parser );
813 fprintf (stderr, "invalid file\n");
819 if( (ttag = xtag_parse_tag( &parser )) != NULL )
827 wrapper = xmalloc( sizeof(XTag) );
828 wrapper->name = NULL;
829 wrapper->pcdata = NULL;
830 wrapper->parent = NULL;
831 wrapper->attributes = NULL;
832 wrapper->children = NULL;
833 wrapper->current_child = NULL;
835 wrapper->children = xlist_append( wrapper->children, tag );
836 wrapper->children = xlist_append( wrapper->children, ttag );
838 while( (ttag = xtag_parse_tag( &parser )) != NULL )
846 wrapper->children = xlist_append( wrapper->children, ttag );
854 static char *xtag_get_name( XTag *xtag )
856 return xtag ? xtag->name : NULL;
860 static char *xtag_get_pcdata( XTag *xtag )
865 if( xtag == NULL ) return NULL;
867 for( l = xtag->children; l; l = l->next )
869 child = (XTag *)l->data;
870 if( child->pcdata != NULL )
872 return child->pcdata;
879 static char *xtag_get_attribute( XTag *xtag, char *attribute )
884 if( xtag == NULL ) return NULL;
886 for( l = xtag->attributes; l; l = l->next )
888 if( (attr = (XAttribute *)l->data) != NULL )
890 if( !strcmp( attr->name, attribute ) ) return attr->value;
898 static XTag *xtag_first_child( XTag *xtag, char *name )
903 if( xtag == NULL ) return NULL;
904 if( (l = xtag->children) == NULL ) return NULL;
908 xtag->current_child = l;
909 return (XTag *)l->data;
912 for( ; l; l = l->next )
914 child = (XTag *)l->data;
916 if( !strcmp( child->name, name ) )
918 xtag->current_child = l;
923 xtag->current_child = NULL;
928 static XTag *xtag_next_child( XTag *xtag, char *name )
933 if( xtag == NULL ) return NULL;
935 if( (l = xtag->current_child) == NULL )
936 return xtag_first_child( xtag, name );
938 if( (l = l->next) == NULL ) return NULL;
942 xtag->current_child = l;
943 return (XTag *)l->data;
946 for( ; l; l = l->next )
948 child = (XTag *)l->data;
950 if( !strcmp( child->name, name ) )
952 xtag->current_child = l;
957 xtag->current_child = NULL;
964 * This snprints function takes a variable list of char *, the last of
965 * which must be NULL, and prints each in turn to buf.
966 * Returns C99-style total length that would have been written, even if
967 * this is larger than n.
969 static int xtag_snprints( char *buf, int n, ... )
973 int len, to_copy, total = 0;
977 for( s = va_arg( ap, char * ); s; s = va_arg( ap, char *) )
981 if( (to_copy = __MIN(n, len) ) > 0 )
983 memcpy( buf, s, to_copy );
996 static int xtag_snprint( char *buf, int n, XTag *xtag )
1003 #define FORWARD(N) \
1004 buf += __MIN(n, N); \
1005 n = __MAX(n-N, 0); \
1010 if( n > 0 ) buf[0] = '\0';
1016 nn = xtag_snprints( buf, n, xtag->pcdata, NULL );
1024 nn = xtag_snprints( buf, n, "<", xtag->name, NULL );
1027 for( l = xtag->attributes; l; l = l->next )
1029 attr = (XAttribute *)l->data;
1031 nn = xtag_snprints( buf, n, " ", attr->name, "=\"", attr->value,
1036 if( xtag->children == NULL )
1038 nn = xtag_snprints ( buf, n, "/>", NULL );
1044 nn = xtag_snprints( buf, n, ">", NULL );
1048 for( l = xtag->children; l; l = l->next )
1050 child = (XTag *)l->data;
1052 nn = xtag_snprint( buf, n, child );
1058 nn = xtag_snprints( buf, n, "</", xtag->name, ">", NULL );