10 * |
10 * |
11 * This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied |
11 * This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied |
12 * warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for details. |
12 * warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for details. |
13 * |
13 * |
14 * This script contains code originally found in MediaWiki (http://www.mediawiki.org). MediaWiki is also licensed under |
14 * This script contains code originally found in MediaWiki (http://www.mediawiki.org). MediaWiki is also licensed under |
15 * the GPLv2; see the file GPL included with this package for details. |
15 * the GPLv2 or later; see the file GPL included with this package for details. |
16 * |
16 * |
17 * We're using the MW parser because the Text_Wiki version simply refused to work under PHP 5.2.0. Porting this was |
17 * We're using the MW parser because the Text_Wiki version simply refused to work under PHP 5.2.0. Porting this was |
18 * _not_ easy. <leaves to get cup of coffee> |
18 * _not_ easy. <leaves to get cup of coffee> |
19 */ |
19 */ |
20 |
20 |
21 global $mStripState, $wgRandomKey; |
21 global $mStripState, $wgRandomKey; |
22 $mStripState = Array(); |
22 $mStripState = Array(); |
|
23 |
|
24 /** |
|
25 * emulate mediawiki parser, including stripping, etc. |
|
26 * |
|
27 * @param string $text the text to parse |
|
28 * @return string |
|
29 * @access public |
|
30 */ |
|
31 |
|
32 function process_tables( $text ) |
|
33 { |
|
34 // include some globals, do some parser stuff that would normally be done in the parent parser function |
|
35 global $mStripState; |
|
36 $x =& $mStripState; |
23 |
37 |
24 $attrib = '[a-zA-Z0-9]'; |
38 // parse the text |
25 $space = '[\x09\x0a\x0d\x20]'; |
39 $text = doTableStuff($text); |
|
40 |
|
41 return $text; |
|
42 } |
|
43 |
|
44 /** |
|
45 * parse the wiki syntax used to render tables |
|
46 * |
|
47 * @param string $t the text to parse |
|
48 * @return string |
|
49 * @access private |
|
50 */ |
|
51 function doTableStuff( $t ) { |
26 |
52 |
27 define( 'MW_CHAR_REFS_REGEX', |
53 $t = explode ( "\n" , $t ) ; |
28 '/&([A-Za-z0-9]+); |
54 $td = array () ; # Is currently a td tag open? |
29 |&\#([0-9]+); |
55 $ltd = array () ; # Was it TD or TH? |
30 |&\#x([0-9A-Za-z]+); |
56 $tr = array () ; # Is currently a tr tag open? |
31 |&\#X([0-9A-Za-z]+); |
57 $ltr = array () ; # tr attributes |
32 |(&)/x' ); |
58 $has_opened_tr = array(); # Did this table open a <tr> element? |
33 |
59 $indent_level = 0; # indent level of the table |
34 define( 'MW_ATTRIBS_REGEX', |
60 foreach ( $t AS $k => $x ) |
35 "/(?:^|$space)($attrib+) |
|
36 ($space*=$space* |
|
37 (?: |
|
38 # The attribute value: quoted or alone |
|
39 ".'"'."([^<".'"'."]*)".'"'." |
|
40 | '([^<']*)' |
|
41 | ([a-zA-Z0-9!#$%&()*,\\-.\\/:;<>?@[\\]^_`{|}~]+) |
|
42 | (\#[0-9a-fA-F]+) # Technically wrong, but lots of |
|
43 # colors are specified like this. |
|
44 # We'll be normalizing it. |
|
45 ) |
|
46 )?(?=$space|\$)/sx" ); |
|
47 |
|
48 /** |
|
49 * emulate mediawiki parser, including stripping, etc. |
|
50 * |
|
51 * @param string $text the text to parse |
|
52 * @return string |
|
53 * @access public |
|
54 */ |
|
55 |
|
56 function process_tables( $text ) |
|
57 { |
61 { |
58 // include some globals, do some parser stuff that would normally be done in the parent parser function |
62 $x = trim ( $x ) ; |
59 global $mStripState; |
63 $fc = substr ( $x , 0 , 1 ) ; |
60 $x =& $mStripState; |
64 if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) { |
61 //$text = mwStrip( $text, $x ); |
65 $indent_level = strlen( $matches[1] ); |
62 |
66 |
63 // parse the text |
67 $attributes = unstripForHTML( $matches[2] ); |
64 $text = doTableStuff($text); |
68 |
65 |
69 $t[$k] = str_repeat( '<dl><dd>', $indent_level ) . |
66 // Unstrip it |
70 '<_paragraph_bypass><table' . fixTagAttributes( $attributes, 'table' ) . '>' ; |
67 // $text = unstrip( $text, $mStripState ); |
71 array_push ( $td , false ) ; |
68 // $text = unstripNoWiki( $text, $mStripState ); |
72 array_push ( $ltd , '' ) ; |
69 //die('<pre>'.print_r($mStripState, true).'</pre>'); |
73 array_push ( $tr , false ) ; |
70 return $text; |
74 array_push ( $ltr , '' ) ; |
|
75 array_push ( $has_opened_tr, false ); |
|
76 } |
|
77 else if ( count ( $td ) == 0 ) { } # Don't do any of the following |
|
78 else if ( '|}' == substr ( $x , 0 , 2 ) ) { |
|
79 $z = "</table></_paragraph_bypass>" . substr ( $x , 2); |
|
80 $l = array_pop ( $ltd ) ; |
|
81 if ( !array_pop ( $has_opened_tr ) ) $z = "<tr><td></td></tr>" . $z ; |
|
82 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ; |
|
83 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ; |
|
84 array_pop ( $ltr ) ; |
|
85 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level ); |
|
86 } |
|
87 else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |--------------- |
|
88 $x = substr ( $x , 1 ) ; |
|
89 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ; |
|
90 $z = '' ; |
|
91 $l = array_pop ( $ltd ) ; |
|
92 array_pop ( $has_opened_tr ); |
|
93 array_push ( $has_opened_tr , true ) ; |
|
94 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ; |
|
95 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ; |
|
96 array_pop ( $ltr ) ; |
|
97 $t[$k] = $z ; |
|
98 array_push ( $tr , false ) ; |
|
99 array_push ( $td , false ) ; |
|
100 array_push ( $ltd , '' ) ; |
|
101 $attributes = unstripForHTML( $x ); |
|
102 array_push ( $ltr , fixTagAttributes( $attributes, 'tr' ) ) ; |
|
103 } |
|
104 else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption |
|
105 # $x is a table row |
|
106 if ( '|+' == substr ( $x , 0 , 2 ) ) { |
|
107 $fc = '+' ; |
|
108 $x = substr ( $x , 1 ) ; |
|
109 } |
|
110 $after = substr ( $x , 1 ) ; |
|
111 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ; |
|
112 |
|
113 // Split up multiple cells on the same line. |
|
114 // FIXME: This can result in improper nesting of tags processed |
|
115 // by earlier parser steps, but should avoid splitting up eg |
|
116 // attribute values containing literal "||". |
|
117 $after = wfExplodeMarkup( '||', $after ); |
|
118 |
|
119 $t[$k] = '' ; |
|
120 |
|
121 # Loop through each table cell |
|
122 foreach ( $after AS $theline ) |
|
123 { |
|
124 $z = '' ; |
|
125 if ( $fc != '+' ) |
|
126 { |
|
127 $tra = array_pop ( $ltr ) ; |
|
128 if ( !array_pop ( $tr ) ) $z = '<tr'.$tra.">\n" ; |
|
129 array_push ( $tr , true ) ; |
|
130 array_push ( $ltr , '' ) ; |
|
131 array_pop ( $has_opened_tr ); |
|
132 array_push ( $has_opened_tr , true ) ; |
|
133 } |
|
134 |
|
135 $l = array_pop ( $ltd ) ; |
|
136 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ; |
|
137 if ( $fc == '|' ) $l = 'td' ; |
|
138 else if ( $fc == '!' ) $l = 'th' ; |
|
139 else if ( $fc == '+' ) $l = 'caption' ; |
|
140 else $l = '' ; |
|
141 array_push ( $ltd , $l ) ; |
|
142 |
|
143 # Cell parameters |
|
144 $y = explode ( '|' , $theline , 2 ) ; |
|
145 # Note that a '|' inside an invalid link should not |
|
146 # be mistaken as delimiting cell parameters |
|
147 if ( strpos( $y[0], '[[' ) !== false ) { |
|
148 $y = array ($theline); |
|
149 } |
|
150 if ( count ( $y ) == 1 ) |
|
151 $y = "{$z}<{$l}>{$y[0]}" ; |
|
152 else { |
|
153 $attributes = unstripForHTML( $y[0] ); |
|
154 $y = "{$z}<{$l}".fixTagAttributes($attributes, $l).">{$y[1]}" ; |
|
155 } |
|
156 $t[$k] .= $y ; |
|
157 array_push ( $td , true ) ; |
|
158 } |
|
159 } |
71 } |
160 } |
72 |
161 |
73 /** |
162 # Closing open td, tr && table |
74 * parse the wiki syntax used to render tables |
163 while ( count ( $td ) > 0 ) |
75 * |
164 { |
76 * @param string $t the text to parse |
165 $l = array_pop ( $ltd ) ; |
77 * @return string |
166 if ( array_pop ( $td ) ) $t[] = '</td>' ; |
78 * @access private |
167 if ( array_pop ( $tr ) ) $t[] = '</tr>' ; |
79 */ |
168 if ( !array_pop ( $has_opened_tr ) ) $t[] = "<tr><td></td></tr>" ; |
80 function doTableStuff( $t ) { |
169 $t[] = '</table></_paragraph_bypass>' ; |
81 |
170 } |
82 $t = explode ( "\n" , $t ) ; |
|
83 $td = array () ; # Is currently a td tag open? |
|
84 $ltd = array () ; # Was it TD or TH? |
|
85 $tr = array () ; # Is currently a tr tag open? |
|
86 $ltr = array () ; # tr attributes |
|
87 $has_opened_tr = array(); # Did this table open a <tr> element? |
|
88 $indent_level = 0; # indent level of the table |
|
89 foreach ( $t AS $k => $x ) |
|
90 { |
|
91 $x = trim ( $x ) ; |
|
92 $fc = substr ( $x , 0 , 1 ) ; |
|
93 if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) { |
|
94 $indent_level = strlen( $matches[1] ); |
|
95 |
171 |
96 $attributes = unstripForHTML( $matches[2] ); |
172 $t = implode ( "\n" , $t ) ; |
97 |
|
98 $t[$k] = str_repeat( '<dl><dd>', $indent_level ) . |
|
99 '<nowiki><table' . fixTagAttributes( $attributes, 'table' ) . '></nowiki>' ; |
|
100 array_push ( $td , false ) ; |
|
101 array_push ( $ltd , '' ) ; |
|
102 array_push ( $tr , false ) ; |
|
103 array_push ( $ltr , '' ) ; |
|
104 array_push ( $has_opened_tr, false ); |
|
105 } |
|
106 else if ( count ( $td ) == 0 ) { } # Don't do any of the following |
|
107 else if ( '|}' == substr ( $x , 0 , 2 ) ) { |
|
108 $z = "<nowiki></table></nowiki>" . substr ( $x , 2); |
|
109 $l = array_pop ( $ltd ) ; |
|
110 if ( !array_pop ( $has_opened_tr ) ) $z = "<nowiki><tr><td></td></tr></nowiki>" . $z ; |
|
111 if ( array_pop ( $tr ) ) $z = '<nowiki></tr></nowiki>' . $z ; |
|
112 if ( array_pop ( $td ) ) $z = '<nowiki></'.$l.'></nowiki>' . $z ; |
|
113 array_pop ( $ltr ) ; |
|
114 $t[$k] = $z . str_repeat( '<nowiki></dd></dl></nowiki>', $indent_level ); |
|
115 } |
|
116 else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |--------------- |
|
117 $x = substr ( $x , 1 ) ; |
|
118 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ; |
|
119 $z = '' ; |
|
120 $l = array_pop ( $ltd ) ; |
|
121 array_pop ( $has_opened_tr ); |
|
122 array_push ( $has_opened_tr , true ) ; |
|
123 if ( array_pop ( $tr ) ) $z = '<nowiki></tr></nowiki>' . $z ; |
|
124 if ( array_pop ( $td ) ) $z = '<nowiki></'.$l.'></nowiki>' . $z ; |
|
125 array_pop ( $ltr ) ; |
|
126 $t[$k] = $z ; |
|
127 array_push ( $tr , false ) ; |
|
128 array_push ( $td , false ) ; |
|
129 array_push ( $ltd , '' ) ; |
|
130 $attributes = unstripForHTML( $x ); |
|
131 array_push ( $ltr , fixTagAttributes( $attributes, 'tr' ) ) ; |
|
132 } |
|
133 else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption |
|
134 # $x is a table row |
|
135 if ( '|+' == substr ( $x , 0 , 2 ) ) { |
|
136 $fc = '+' ; |
|
137 $x = substr ( $x , 1 ) ; |
|
138 } |
|
139 $after = substr ( $x , 1 ) ; |
|
140 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ; |
|
141 |
|
142 // Split up multiple cells on the same line. |
|
143 // FIXME: This can result in improper nesting of tags processed |
|
144 // by earlier parser steps, but should avoid splitting up eg |
|
145 // attribute values containing literal "||". |
|
146 $after = wfExplodeMarkup( '||', $after ); |
|
147 |
|
148 $t[$k] = '' ; |
|
149 |
|
150 # Loop through each table cell |
|
151 foreach ( $after AS $theline ) |
|
152 { |
|
153 $z = '' ; |
|
154 if ( $fc != '+' ) |
|
155 { |
|
156 $tra = array_pop ( $ltr ) ; |
|
157 if ( !array_pop ( $tr ) ) $z = '<nowiki><tr'.$tra."></nowiki>\n" ; |
|
158 array_push ( $tr , true ) ; |
|
159 array_push ( $ltr , '' ) ; |
|
160 array_pop ( $has_opened_tr ); |
|
161 array_push ( $has_opened_tr , true ) ; |
|
162 } |
|
163 |
|
164 $l = array_pop ( $ltd ) ; |
|
165 if ( array_pop ( $td ) ) $z = '<nowiki></'.$l.'></nowiki>' . $z ; |
|
166 if ( $fc == '|' ) $l = 'td' ; |
|
167 else if ( $fc == '!' ) $l = 'th' ; |
|
168 else if ( $fc == '+' ) $l = 'caption' ; |
|
169 else $l = '' ; |
|
170 array_push ( $ltd , $l ) ; |
|
171 |
|
172 # Cell parameters |
|
173 $y = explode ( '|' , $theline , 2 ) ; |
|
174 # Note that a '|' inside an invalid link should not |
|
175 # be mistaken as delimiting cell parameters |
|
176 if ( strpos( $y[0], '[[' ) !== false ) { |
|
177 $y = array ($theline); |
|
178 } |
|
179 if ( count ( $y ) == 1 ) |
|
180 $y = "{$z}<nowiki><{$l}></nowiki>{$y[0]}" ; |
|
181 else { |
|
182 $attributes = unstripForHTML( $y[0] ); |
|
183 $y = "{$z}<nowiki><{$l}".fixTagAttributes($attributes, $l)."></nowiki>{$y[1]}" ; |
|
184 } |
|
185 $t[$k] .= $y ; |
|
186 array_push ( $td , true ) ; |
|
187 } |
|
188 } |
|
189 } |
|
190 |
|
191 # Closing open td, tr && table |
|
192 while ( count ( $td ) > 0 ) |
|
193 { |
|
194 $l = array_pop ( $ltd ) ; |
|
195 if ( array_pop ( $td ) ) $t[] = '<nowiki></td></nowiki>' ; |
|
196 if ( array_pop ( $tr ) ) $t[] = '<nowiki></tr></nowiki>' ; |
|
197 if ( !array_pop ( $has_opened_tr ) ) $t[] = "<nowiki><tr><td></td></tr></nowiki>" ; |
|
198 $t[] = '<nowiki></table></nowiki>' ; |
|
199 } |
|
200 |
|
201 $t = implode ( "\n" , $t ) ; |
|
202 |
|
203 # special case: don't return empty table |
|
204 if($t == "<nowiki><table></nowiki>\n<nowiki><tr><td></td></tr></nowiki>\n<nowiki></table></nowiki>") |
|
205 $t = ''; |
|
206 return $t ; |
|
207 } |
|
208 |
173 |
209 /** |
174 # special case: don't return empty table |
210 * Take a tag soup fragment listing an HTML element's attributes |
175 if($t == "<table>\n<tr><td></td></tr>\n</table>") |
211 * and normalize it to well-formed XML, discarding unwanted attributes. |
176 $t = ''; |
212 * Output is safe for further wikitext processing, with escaping of |
177 return $t ; |
213 * values that could trigger problems. |
|
214 * |
|
215 * - Normalizes attribute names to lowercase |
|
216 * - Discards attributes not on a whitelist for the given element |
|
217 * - Turns broken or invalid entities into plaintext |
|
218 * - Double-quotes all attribute values |
|
219 * - Attributes without values are given the name as attribute |
|
220 * - Double attributes are discarded |
|
221 * - Unsafe style attributes are discarded |
|
222 * - Prepends space if there are attributes. |
|
223 * |
|
224 * @param string $text |
|
225 * @param string $element |
|
226 * @return string |
|
227 */ |
|
228 function fixTagAttributes( $text, $element ) { |
|
229 if( trim( $text ) == '' ) { |
|
230 return ''; |
|
231 } |
|
232 |
|
233 $stripped = validateTagAttributes( |
|
234 decodeTagAttributes( $text ), $element ); |
|
235 |
|
236 $attribs = array(); |
|
237 foreach( $stripped as $attribute => $value ) { |
|
238 $encAttribute = htmlspecialchars( $attribute ); |
|
239 $encValue = safeEncodeAttribute( $value ); |
|
240 |
|
241 $attribs[] = "$encAttribute=".'"'."$encValue".'"'.""; // " |
|
242 } |
|
243 return count( $attribs ) ? ' ' . implode( ' ', $attribs ) : ''; |
|
244 } |
|
245 |
|
246 /** |
|
247 * Encode an attribute value for HTML tags, with extra armoring |
|
248 * against further wiki processing. |
|
249 * @param $text |
|
250 * @return HTML-encoded text fragment |
|
251 */ |
|
252 function safeEncodeAttribute( $text ) { |
|
253 $encValue= encodeAttribute( $text ); |
|
254 |
|
255 # Templates and links may be expanded in later parsing, |
|
256 # creating invalid or dangerous output. Suppress this. |
|
257 $encValue = strtr( $encValue, array( |
|
258 '<' => '<', // This should never happen, |
|
259 '>' => '>', // we've received invalid input |
|
260 '"' => '"', // which should have been escaped. |
|
261 '{' => '{', |
|
262 '[' => '[', |
|
263 "''" => '''', |
|
264 'ISBN' => 'ISBN', |
|
265 'RFC' => 'RFC', |
|
266 'PMID' => 'PMID', |
|
267 '|' => '|', |
|
268 '__' => '__', |
|
269 ) ); |
|
270 |
|
271 return $encValue; |
|
272 } |
|
273 |
|
274 /** |
|
275 * Encode an attribute value for HTML output. |
|
276 * @param $text |
|
277 * @return HTML-encoded text fragment |
|
278 */ |
|
279 function encodeAttribute( $text ) { |
|
280 |
|
281 // In Enano 1.0.3, added this cheapo hack to keep ampersands |
|
282 // from being double-sanitized. Thanks to markybob from #deluge. |
|
283 |
|
284 // htmlspecialchars() the "manual" way |
|
285 $encValue = strtr( $text, array( |
|
286 '&' => '&', |
|
287 '"' => '"', |
|
288 '<' => '<', |
|
289 '>' => '>', |
|
290 ''' => "'" |
|
291 ) ); |
|
292 |
|
293 $encValue = strtr( $text, array( |
|
294 '&' => '&', |
|
295 '"' => '"', |
|
296 '<' => '<', |
|
297 '>' => '>', |
|
298 "'" => ''' |
|
299 ) ); |
|
300 |
|
301 |
|
302 // Whitespace is normalized during attribute decoding, |
|
303 // so if we've been passed non-spaces we must encode them |
|
304 // ahead of time or they won't be preserved. |
|
305 $encValue = strtr( $encValue, array( |
|
306 "\n" => ' ', |
|
307 "\r" => ' ', |
|
308 "\t" => '	', |
|
309 ) ); |
|
310 |
|
311 return $encValue; |
|
312 } |
|
313 |
|
314 function unstripForHTML( $text ) { |
|
315 global $mStripState; |
|
316 $text = unstrip( $text, $mStripState ); |
|
317 $text = unstripNoWiki( $text, $mStripState ); |
|
318 return $text; |
|
319 } |
|
320 |
|
321 /** |
|
322 * Always call this after unstrip() to preserve the order |
|
323 * |
|
324 * @private |
|
325 */ |
|
326 function unstripNoWiki( $text, &$state ) { |
|
327 if ( !isset( $state['nowiki'] ) ) { |
|
328 return $text; |
|
329 } |
|
330 |
|
331 # TODO: good candidate for FSS |
|
332 $text = strtr( $text, $state['nowiki'] ); |
|
333 |
|
334 return $text; |
|
335 } |
|
336 |
|
337 /** |
|
338 * Take an array of attribute names and values and normalize or discard |
|
339 * illegal values for the given element type. |
|
340 * |
|
341 * - Discards attributes not on a whitelist for the given element |
|
342 * - Unsafe style attributes are discarded |
|
343 * |
|
344 * @param array $attribs |
|
345 * @param string $element |
|
346 * @return array |
|
347 * |
|
348 * @todo Check for legal values where the DTD limits things. |
|
349 * @todo Check for unique id attribute :P |
|
350 */ |
|
351 function validateTagAttributes( $attribs, $element ) { |
|
352 $whitelist = array_flip( attributeWhitelist( $element ) ); |
|
353 $out = array(); |
|
354 foreach( $attribs as $attribute => $value ) { |
|
355 if( !isset( $whitelist[$attribute] ) ) { |
|
356 continue; |
|
357 } |
|
358 # Strip javascript "expression" from stylesheets. |
|
359 # http://msdn.microsoft.com/workshop/author/dhtml/overview/recalc.asp |
|
360 if( $attribute == 'style' ) { |
|
361 $value = checkCss( $value ); |
|
362 if( $value === false ) { |
|
363 # haxx0r |
|
364 continue; |
|
365 } |
|
366 } |
|
367 |
|
368 if ( $attribute === 'id' ) |
|
369 $value = escapeId( $value ); |
|
370 |
|
371 // If this attribute was previously set, override it. |
|
372 // Output should only have one attribute of each name. |
|
373 $out[$attribute] = $value; |
|
374 } |
|
375 return $out; |
|
376 } |
|
377 |
|
378 /** |
|
379 * Pick apart some CSS and check it for forbidden or unsafe structures. |
|
380 * Returns a sanitized string, or false if it was just too evil. |
|
381 * |
|
382 * Currently URL references, 'expression', 'tps' are forbidden. |
|
383 * |
|
384 * @param string $value |
|
385 * @return mixed |
|
386 */ |
|
387 function checkCss( $value ) { |
|
388 $stripped = decodeCharReferences( $value ); |
|
389 |
|
390 // Remove any comments; IE gets token splitting wrong |
|
391 $stripped = preg_replace( '!/\\*.*?\\*/!S', '', $stripped ); |
|
392 $value = $stripped; |
|
393 |
|
394 // ... and continue checks |
|
395 $stripped = preg_replace( '!\\\\([0-9A-Fa-f]{1,6})[ \\n\\r\\t\\f]?!e', |
|
396 'codepointToUtf8(hexdec("$1"))', $stripped ); |
|
397 $stripped = str_replace( '\\', '', $stripped ); |
|
398 if( preg_match( '/(expression|tps*:\/\/|url\\s*\().*/is', |
|
399 $stripped ) ) { |
|
400 # haxx0r |
|
401 return false; |
|
402 } |
|
403 |
|
404 return $value; |
|
405 } |
|
406 |
|
407 /** |
|
408 * Decode any character references, numeric or named entities, |
|
409 * in the text and return a UTF-8 string. |
|
410 * |
|
411 * @param string $text |
|
412 * @return string |
|
413 * @access public |
|
414 * @static |
|
415 */ |
|
416 function decodeCharReferences( $text ) { |
|
417 return preg_replace_callback( |
|
418 MW_CHAR_REFS_REGEX, |
|
419 'decodeCharReferencesCallback', |
|
420 $text ); |
|
421 } |
|
422 |
|
423 /** |
|
424 * Fetch the whitelist of acceptable attributes for a given |
|
425 * element name. |
|
426 * |
|
427 * @param string $element |
|
428 * @return array |
|
429 */ |
|
430 function attributeWhitelist( $element ) { |
|
431 static $list; |
|
432 if( !isset( $list ) ) { |
|
433 $list = setupAttributeWhitelist(); |
|
434 } |
|
435 return isset( $list[$element] ) |
|
436 ? $list[$element] |
|
437 : array(); |
|
438 } |
|
439 |
|
440 /** |
|
441 * @todo Document it a bit |
|
442 * @return array |
|
443 */ |
|
444 function setupAttributeWhitelist() { |
|
445 global $db, $session, $paths, $template, $plugins; |
|
446 $common = array( 'id', 'class', 'lang', 'dir', 'title', 'style' ); |
|
447 $block = array_merge( $common, array( 'align' ) ); |
|
448 $tablealign = array( 'align', 'char', 'charoff', 'valign' ); |
|
449 $tablecell = array( 'abbr', |
|
450 'axis', |
|
451 'headers', |
|
452 'scope', |
|
453 'rowspan', |
|
454 'colspan', |
|
455 'nowrap', # deprecated |
|
456 'width', # deprecated |
|
457 'height', # deprecated |
|
458 'bgcolor' # deprecated |
|
459 ); |
|
460 |
|
461 # Numbers refer to sections in HTML 4.01 standard describing the element. |
|
462 # See: http://www.w3.org/TR/html4/ |
|
463 $whitelist = array ( |
|
464 # 7.5.4 |
|
465 'div' => $block, |
|
466 'center' => $common, # deprecated |
|
467 'span' => $block, # ?? |
|
468 |
|
469 # 7.5.5 |
|
470 'h1' => $block, |
|
471 'h2' => $block, |
|
472 'h3' => $block, |
|
473 'h4' => $block, |
|
474 'h5' => $block, |
|
475 'h6' => $block, |
|
476 |
|
477 # 7.5.6 |
|
478 # address |
|
479 |
|
480 # 8.2.4 |
|
481 # bdo |
|
482 |
|
483 # 9.2.1 |
|
484 'em' => $common, |
|
485 'strong' => $common, |
|
486 'cite' => $common, |
|
487 # dfn |
|
488 'code' => $common, |
|
489 # samp |
|
490 # kbd |
|
491 'var' => $common, |
|
492 # abbr |
|
493 # acronym |
|
494 |
|
495 # 9.2.2 |
|
496 'blockquote' => array_merge( $common, array( 'cite' ) ), |
|
497 # q |
|
498 |
|
499 # 9.2.3 |
|
500 'sub' => $common, |
|
501 'sup' => $common, |
|
502 |
|
503 # 9.3.1 |
|
504 'p' => $block, |
|
505 |
|
506 # 9.3.2 |
|
507 'br' => array( 'id', 'class', 'title', 'style', 'clear' ), |
|
508 |
|
509 # 9.3.4 |
|
510 'pre' => array_merge( $common, array( 'width' ) ), |
|
511 |
|
512 # 9.4 |
|
513 'ins' => array_merge( $common, array( 'cite', 'datetime' ) ), |
|
514 'del' => array_merge( $common, array( 'cite', 'datetime' ) ), |
|
515 |
|
516 # 10.2 |
|
517 'ul' => array_merge( $common, array( 'type' ) ), |
|
518 'ol' => array_merge( $common, array( 'type', 'start' ) ), |
|
519 'li' => array_merge( $common, array( 'type', 'value' ) ), |
|
520 |
|
521 # 10.3 |
|
522 'dl' => $common, |
|
523 'dd' => $common, |
|
524 'dt' => $common, |
|
525 |
|
526 # 11.2.1 |
|
527 'table' => array_merge( $common, |
|
528 array( 'summary', 'width', 'border', 'frame', |
|
529 'rules', 'cellspacing', 'cellpadding', |
|
530 'align', 'bgcolor', |
|
531 ) ), |
|
532 |
|
533 # 11.2.2 |
|
534 'caption' => array_merge( $common, array( 'align' ) ), |
|
535 |
|
536 # 11.2.3 |
|
537 'thead' => array_merge( $common, $tablealign ), |
|
538 'tfoot' => array_merge( $common, $tablealign ), |
|
539 'tbody' => array_merge( $common, $tablealign ), |
|
540 |
|
541 # 11.2.4 |
|
542 'colgroup' => array_merge( $common, array( 'span', 'width' ), $tablealign ), |
|
543 'col' => array_merge( $common, array( 'span', 'width' ), $tablealign ), |
|
544 |
|
545 # 11.2.5 |
|
546 'tr' => array_merge( $common, array( 'bgcolor' ), $tablealign ), |
|
547 |
|
548 # 11.2.6 |
|
549 'td' => array_merge( $common, $tablecell, $tablealign ), |
|
550 'th' => array_merge( $common, $tablecell, $tablealign ), |
|
551 |
|
552 # 12.2 |
|
553 # added by dan |
|
554 'a' => array_merge( $common, array( 'href', 'name' ) ), |
|
555 |
|
556 # 13.2 |
|
557 # added by dan |
|
558 'img' => array_merge( $common, array( 'src', 'width', 'height', 'alt' ) ), |
|
559 |
|
560 # 15.2.1 |
|
561 'tt' => $common, |
|
562 'b' => $common, |
|
563 'i' => $common, |
|
564 'big' => $common, |
|
565 'small' => $common, |
|
566 'strike' => $common, |
|
567 's' => $common, |
|
568 'u' => $common, |
|
569 |
|
570 # 15.2.2 |
|
571 'font' => array_merge( $common, array( 'size', 'color', 'face' ) ), |
|
572 # basefont |
|
573 |
|
574 # 15.3 |
|
575 'hr' => array_merge( $common, array( 'noshade', 'size', 'width' ) ), |
|
576 |
|
577 # XHTML Ruby annotation text module, simple ruby only. |
|
578 # http://www.w3c.org/TR/ruby/ |
|
579 'ruby' => $common, |
|
580 # rbc |
|
581 # rtc |
|
582 'rb' => $common, |
|
583 'rt' => $common, #array_merge( $common, array( 'rbspan' ) ), |
|
584 'rp' => $common, |
|
585 |
|
586 # For compatibility with the XHTML parser. |
|
587 'nowiki' => array(), |
|
588 'noinclude' => array(), |
|
589 'nodisplay' => array(), |
|
590 'lang' => array('code'), |
|
591 |
|
592 # XHTML stuff |
|
593 'acronym' => $common |
|
594 ); |
|
595 |
|
596 // custom tags can be added by plugins |
|
597 $code = $plugins->setHook('html_attribute_whitelist'); |
|
598 foreach ( $code as $cmd ) |
|
599 { |
|
600 eval($cmd); |
|
601 } |
|
602 |
|
603 return $whitelist; |
|
604 } |
|
605 |
|
606 /** |
|
607 * Given a value escape it so that it can be used in an id attribute and |
|
608 * return it, this does not validate the value however (see first link) |
|
609 * |
|
610 * @link http://www.w3.org/TR/html401/types.html#type-name Valid characters |
|
611 * in the id and |
|
612 * name attributes |
|
613 * @link http://www.w3.org/TR/html401/struct/links.html#h-12.2.3 Anchors with the id attribute |
|
614 * |
|
615 * @bug 4461 |
|
616 * |
|
617 * @static |
|
618 * |
|
619 * @param string $id |
|
620 * @return string |
|
621 */ |
|
622 function escapeId( $id ) { |
|
623 static $replace = array( |
|
624 '%3A' => ':', |
|
625 '%' => '.' |
|
626 ); |
|
627 |
|
628 $id = urlencode( decodeCharReferences( strtr( $id, ' ', '_' ) ) ); |
|
629 |
|
630 return str_replace( array_keys( $replace ), array_values( $replace ), $id ); |
|
631 } |
|
632 |
|
633 /** |
|
634 * More or less "markup-safe" explode() |
|
635 * Ignores any instances of the separator inside <...> |
|
636 * @param string $separator |
|
637 * @param string $text |
|
638 * @return array |
|
639 */ |
|
640 function wfExplodeMarkup( $separator, $text ) { |
|
641 $placeholder = "\x00"; |
|
642 |
|
643 // Just in case... |
|
644 $text = str_replace( $placeholder, '', $text ); |
|
645 |
|
646 // Trim stuff |
|
647 $replacer = new ReplacerCallback( $separator, $placeholder ); |
|
648 $cleaned = preg_replace_callback( '/(<.*?>)/', array( $replacer, 'go' ), $text ); |
|
649 |
|
650 $items = explode( $separator, $cleaned ); |
|
651 foreach( $items as $i => $str ) { |
|
652 $items[$i] = str_replace( $placeholder, $separator, $str ); |
|
653 } |
|
654 |
|
655 return $items; |
|
656 } |
|
657 |
|
658 class ReplacerCallback { |
|
659 function ReplacerCallback( $from, $to ) { |
|
660 $this->from = $from; |
|
661 $this->to = $to; |
|
662 } |
|
663 |
|
664 function go( $matches ) { |
|
665 return str_replace( $this->from, $this->to, $matches[1] ); |
|
666 } |
|
667 } |
|
668 |
|
669 /** |
|
670 * Return an associative array of attribute names and values from |
|
671 * a partial tag string. Attribute names are forces to lowercase, |
|
672 * character references are decoded to UTF-8 text. |
|
673 * |
|
674 * @param string |
|
675 * @return array |
|
676 */ |
|
677 function decodeTagAttributes( $text ) { |
|
678 $attribs = array(); |
|
679 |
|
680 if( trim( $text ) == '' ) { |
|
681 return $attribs; |
|
682 } |
|
683 |
|
684 $pairs = array(); |
|
685 if( !preg_match_all( |
|
686 MW_ATTRIBS_REGEX, |
|
687 $text, |
|
688 $pairs, |
|
689 PREG_SET_ORDER ) ) { |
|
690 return $attribs; |
|
691 } |
|
692 |
|
693 foreach( $pairs as $set ) { |
|
694 $attribute = strtolower( $set[1] ); |
|
695 $value = getTagAttributeCallback( $set ); |
|
696 |
|
697 // Normalize whitespace |
|
698 $value = preg_replace( '/[\t\r\n ]+/', ' ', $value ); |
|
699 $value = trim( $value ); |
|
700 |
|
701 // Decode character references |
|
702 $attribs[$attribute] = decodeCharReferences( $value ); |
|
703 } |
|
704 return $attribs; |
|
705 } |
|
706 |
|
707 /** |
|
708 * Pick the appropriate attribute value from a match set from the |
|
709 * MW_ATTRIBS_REGEX matches. |
|
710 * |
|
711 * @param array $set |
|
712 * @return string |
|
713 * @access private |
|
714 */ |
|
715 function getTagAttributeCallback( $set ) { |
|
716 if( isset( $set[6] ) ) { |
|
717 # Illegal #XXXXXX color with no quotes. |
|
718 return $set[6]; |
|
719 } elseif( isset( $set[5] ) ) { |
|
720 # No quotes. |
|
721 return $set[5]; |
|
722 } elseif( isset( $set[4] ) ) { |
|
723 # Single-quoted |
|
724 return $set[4]; |
|
725 } elseif( isset( $set[3] ) ) { |
|
726 # Double-quoted |
|
727 return $set[3]; |
|
728 } elseif( !isset( $set[2] ) ) { |
|
729 # In XHTML, attributes must have a value. |
|
730 # For 'reduced' form, return explicitly the attribute name here. |
|
731 return $set[1]; |
|
732 } else { |
|
733 die_friendly('Parser error', "<p>Tag conditions not met. This should never happen and is a bug.</p>" ); |
|
734 } |
|
735 } |
|
736 |
|
737 /** |
|
738 * Strips and renders nowiki, pre, math, hiero |
|
739 * If $render is set, performs necessary rendering operations on plugins |
|
740 * Returns the text, and fills an array with data needed in unstrip() |
|
741 * If the $state is already a valid strip state, it adds to the state |
|
742 * |
|
743 * @param bool $stripcomments when set, HTML comments <!-- like this --> |
|
744 * will be stripped in addition to other tags. This is important |
|
745 * for section editing, where these comments cause confusion when |
|
746 * counting the sections in the wikisource |
|
747 * |
|
748 * @param array dontstrip contains tags which should not be stripped; |
|
749 * used to prevent stipping of <gallery> when saving (fixes bug 2700) |
|
750 * |
|
751 * @access private |
|
752 */ |
|
753 function mwStrip( $text, &$state, $stripcomments = false , $dontstrip = array () ) { |
|
754 global $wgRandomKey; |
|
755 $render = true; |
|
756 |
|
757 $wgRandomKey = "\x07UNIQ" . dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff)); |
|
758 $uniq_prefix =& $wgRandomKey; |
|
759 $commentState = array(); |
|
760 |
|
761 $elements = array( 'nowiki', 'gallery' ); |
|
762 |
|
763 # Removing $dontstrip tags from $elements list (currently only 'gallery', fixing bug 2700) |
|
764 foreach ( $elements AS $k => $v ) { |
|
765 if ( !in_array ( $v , $dontstrip ) ) continue; |
|
766 unset ( $elements[$k] ); |
|
767 } |
|
768 |
|
769 $matches = array(); |
|
770 $text = extractTagsAndParams( $elements, $text, $matches, $uniq_prefix ); |
|
771 |
|
772 foreach( $matches as $marker => $data ) { |
|
773 list( $element, $content, $params, $tag ) = $data; |
|
774 if( $render ) { |
|
775 $tagName = strtolower( $element ); |
|
776 switch( $tagName ) { |
|
777 case '!--': |
|
778 // Comment |
|
779 if( substr( $tag, -3 ) == '-->' ) { |
|
780 $output = $tag; |
|
781 } else { |
|
782 // Unclosed comment in input. |
|
783 // Close it so later stripping can remove it |
|
784 $output = "$tag-->"; |
|
785 } |
|
786 break; |
|
787 case 'html': |
|
788 if( $wgRawHtml ) { |
|
789 $output = $content; |
|
790 break; |
|
791 } |
|
792 // Shouldn't happen otherwise. :) |
|
793 case 'nowiki': |
|
794 $output = wfEscapeHTMLTagsOnly( $content ); |
|
795 break; |
|
796 default: |
|
797 } |
|
798 } else { |
|
799 // Just stripping tags; keep the source |
|
800 $output = $tag; |
|
801 } |
|
802 |
|
803 // Unstrip the output, because unstrip() is no longer recursive so |
|
804 // it won't do it itself |
|
805 $output = unstrip( $output, $state ); |
|
806 |
|
807 if( !$stripcomments && $element == '!--' ) { |
|
808 $commentState[$marker] = $output; |
|
809 } elseif ( $element == 'html' || $element == 'nowiki' ) { |
|
810 $state['nowiki'][$marker] = $output; |
|
811 } else { |
|
812 $state['general'][$marker] = $output; |
|
813 } |
|
814 } |
|
815 |
|
816 # Unstrip comments unless explicitly told otherwise. |
|
817 # (The comments are always stripped prior to this point, so as to |
|
818 # not invoke any extension tags / parser hooks contained within |
|
819 # a comment.) |
|
820 if ( !$stripcomments ) { |
|
821 // Put them all back and forget them |
|
822 $text = strtr( $text, $commentState ); |
|
823 } |
|
824 |
|
825 return $text; |
|
826 } |
|
827 |
|
828 /** |
|
829 * Replaces all occurrences of HTML-style comments and the given tags |
|
830 * in the text with a random marker and returns teh next text. The output |
|
831 * parameter $matches will be an associative array filled with data in |
|
832 * the form: |
|
833 * 'UNIQ-xxxxx' => array( |
|
834 * 'element', |
|
835 * 'tag content', |
|
836 * array( 'param' => 'x' ), |
|
837 * '<element param="x">tag content</element>' ) ) |
|
838 * |
|
839 * @param $elements list of element names. Comments are always extracted. |
|
840 * @param $text Source text string. |
|
841 * @param $uniq_prefix |
|
842 * |
|
843 * @access private |
|
844 * @static |
|
845 */ |
|
846 function extractTagsAndParams($elements, $text, &$matches, $uniq_prefix = ''){ |
|
847 static $n = 1; |
|
848 $stripped = ''; |
|
849 $matches = array(); |
|
850 |
|
851 $taglist = implode( '|', $elements ); |
|
852 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?>)|<(!--)/i"; |
|
853 |
|
854 while ( '' != $text ) { |
|
855 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE ); |
|
856 $stripped .= $p[0]; |
|
857 if( count( $p ) < 5 ) { |
|
858 break; |
|
859 } |
|
860 if( count( $p ) > 5 ) { |
|
861 // comment |
|
862 $element = $p[4]; |
|
863 $attributes = ''; |
|
864 $close = ''; |
|
865 $inside = $p[5]; |
|
866 } else { |
|
867 // tag |
|
868 $element = $p[1]; |
|
869 $attributes = $p[2]; |
|
870 $close = $p[3]; |
|
871 $inside = $p[4]; |
|
872 } |
|
873 |
|
874 $marker = "$uniq_prefix-$element-" . sprintf('%08X', $n++) . '-QINU'; |
|
875 $stripped .= $marker; |
|
876 |
|
877 if ( $close === '/>' ) { |
|
878 // Empty element tag, <tag /> |
|
879 $content = null; |
|
880 $text = $inside; |
|
881 $tail = null; |
|
882 } else { |
|
883 if( $element == '!--' ) { |
|
884 $end = '/(-->)/'; |
|
885 } else { |
|
886 $end = "/(<\\/$element\\s*>)/i"; |
|
887 } |
|
888 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE ); |
|
889 $content = $q[0]; |
|
890 if( count( $q ) < 3 ) { |
|
891 # No end tag -- let it run out to the end of the text. |
|
892 $tail = ''; |
|
893 $text = ''; |
|
894 } else { |
|
895 $tail = $q[1]; |
|
896 $text = $q[2]; |
|
897 } |
|
898 } |
|
899 |
|
900 $matches[$marker] = array( $element, |
|
901 $content, |
|
902 decodeTagAttributes( $attributes ), |
|
903 "<$element$attributes$close$content$tail" ); |
|
904 } |
|
905 return $stripped; |
|
906 } |
|
907 |
|
908 /** |
|
909 * Escape html tags |
|
910 * Basically replacing " > and < with HTML entities ( ", >, <) |
|
911 * |
|
912 * @param $in String: text that might contain HTML tags. |
|
913 * @return string Escaped string |
|
914 */ |
|
915 function wfEscapeHTMLTagsOnly( $in ) { |
|
916 return str_replace( |
|
917 array( '"', '>', '<' ), |
|
918 array( '"', '>', '<' ), |
|
919 $in ); |
|
920 } |
|
921 |
|
922 /** |
|
923 * Restores pre, math, and other extensions removed by strip() |
|
924 * |
|
925 * always call unstripNoWiki() after this one |
|
926 * @private |
|
927 */ |
|
928 function unstrip( $text, &$state ) { |
|
929 if ( !isset( $state['general'] ) ) { |
|
930 return $text; |
|
931 } |
|
932 |
|
933 # TODO: good candidate for FSS |
|
934 $text = strtr( $text, $state['general'] ); |
|
935 |
|
936 return $text; |
|
937 } |
|
938 |
|
939 /** |
|
940 * Return UTF-8 string for a codepoint if that is a valid |
|
941 * character reference, otherwise U+FFFD REPLACEMENT CHARACTER. |
|
942 * @param int $codepoint |
|
943 * @return string |
|
944 * @private |
|
945 */ |
|
946 function decodeChar( $codepoint ) { |
|
947 if( validateCodepoint( $codepoint ) ) { |
|
948 return codepointToUtf8( $codepoint ); |
|
949 } else { |
|
950 return UTF8_REPLACEMENT; |
|
951 } |
|
952 } |
|
953 |
|
954 /** |
|
955 * If the named entity is defined in the HTML 4.0/XHTML 1.0 DTD, |
|
956 * return the UTF-8 encoding of that character. Otherwise, returns |
|
957 * pseudo-entity source (eg &foo;) |
|
958 * |
|
959 * @param string $name |
|
960 * @return string |
|
961 */ |
|
962 function decodeEntity( $name ) { |
|
963 global $wgHtmlEntities; |
|
964 if( isset( $wgHtmlEntities[$name] ) ) { |
|
965 return codepointToUtf8( $wgHtmlEntities[$name] ); |
|
966 } else { |
|
967 return "&$name;"; |
|
968 } |
|
969 } |
|
970 |
|
971 /** |
|
972 * Returns true if a given Unicode codepoint is a valid character in XML. |
|
973 * @param int $codepoint |
|
974 * @return bool |
|
975 */ |
|
976 function validateCodepoint( $codepoint ) { |
|
977 return ($codepoint == 0x09) |
|
978 || ($codepoint == 0x0a) |
|
979 || ($codepoint == 0x0d) |
|
980 || ($codepoint >= 0x20 && $codepoint <= 0xd7ff) |
|
981 || ($codepoint >= 0xe000 && $codepoint <= 0xfffd) |
|
982 || ($codepoint >= 0x10000 && $codepoint <= 0x10ffff); |
|
983 } |
|
984 |
|
985 /** |
|
986 * Return UTF-8 sequence for a given Unicode code point. |
|
987 * May die if fed out of range data. |
|
988 * |
|
989 * @param $codepoint Integer: |
|
990 * @return String |
|
991 * @public |
|
992 */ |
|
993 function codepointToUtf8( $codepoint ) { |
|
994 if($codepoint < 0x80) return chr($codepoint); |
|
995 if($codepoint < 0x800) return chr($codepoint >> 6 & 0x3f | 0xc0) . |
|
996 chr($codepoint & 0x3f | 0x80); |
|
997 if($codepoint < 0x10000) return chr($codepoint >> 12 & 0x0f | 0xe0) . |
|
998 chr($codepoint >> 6 & 0x3f | 0x80) . |
|
999 chr($codepoint & 0x3f | 0x80); |
|
1000 if($codepoint < 0x110000) return chr($codepoint >> 18 & 0x07 | 0xf0) . |
|
1001 chr($codepoint >> 12 & 0x3f | 0x80) . |
|
1002 chr($codepoint >> 6 & 0x3f | 0x80) . |
|
1003 chr($codepoint & 0x3f | 0x80); |
|
1004 |
|
1005 echo "Asked for code outside of range ($codepoint)\n"; |
|
1006 die( -1 ); |
|
1007 } |
178 } |
1008 |
179 |
1009 /** |
|
1010 * @param string $matches |
|
1011 * @return string |
|
1012 */ |
|
1013 function decodeCharReferencesCallback( $matches ) { |
|
1014 if( $matches[1] != '' ) { |
|
1015 return decodeEntity( $matches[1] ); |
|
1016 } elseif( $matches[2] != '' ) { |
|
1017 return decodeChar( intval( $matches[2] ) ); |
|
1018 } elseif( $matches[3] != '' ) { |
|
1019 return decodeChar( hexdec( $matches[3] ) ); |
|
1020 } elseif( $matches[4] != '' ) { |
|
1021 return decodeChar( hexdec( $matches[4] ) ); |
|
1022 } |
|
1023 # Last case should be an ampersand by itself |
|
1024 return $matches[0]; |
|
1025 } |
|
1026 |
|
1027 ?> |
|