2 # Copyright (C) 2009 Aryeh Gregor
3 # http://www.mediawiki.org/
5 # This program is free software; you can redistribute it and/or modify
6 # it under the terms of the GNU General Public License as published by
7 # the Free Software Foundation; either version 2 of the License, or
8 # (at your option) any later version.
10 # This program is distributed in the hope that it will be useful,
11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 # GNU General Public License for more details.
15 # You should have received a copy of the GNU General Public License along
16 # with this program; if not, write to the Free Software Foundation, Inc.,
17 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 # http://www.gnu.org/copyleft/gpl.html
21 * This class is a collection of static functions that serve two purposes:
23 * 1) Implement any algorithms specified by HTML5, or other HTML
24 * specifications, in a convenient and self-contained way.
26 * 2) Allow HTML elements to be conveniently and safely generated, like the
27 * current Xml class but a) less confused (Xml supports HTML-specific things,
28 * but only sometimes!) and b) not necessarily confined to XML-compatible
31 * There are two important configuration options this class uses:
33 * $wgHtml5: If this is set to false, then all output should be valid XHTML 1.0
35 * $wgWellFormedXml: If this is set to true, then all output should be
36 * well-formed XML (quotes on attributes, self-closing tags, etc.).
38 * This class is meant to be confined to utility functions that are called from
39 * trusted code paths. It does not do enforcement of policy like not allowing
43 # List of void elements from HTML5, section 9.1.2 as of 2009-08-10
44 private static $voidElements = array(
61 # Boolean attributes, which may have the value omitted entirely. Manually
62 # collected from the HTML5 spec as of 2009-08-10.
63 private static $boolAttribs = array(
87 * Returns an HTML element in a string. The major advantage here over
88 * manually typing out the HTML is that it will escape all attribute
89 * values. If you're hardcoding all the attributes, or there are none, you
90 * should probably type out the string yourself.
92 * This is quite similar to Xml::tags(), but it implements some useful
93 * HTML-specific logic. For instance, there is no $allowShortTag
94 * parameter: the closing tag is magically omitted if $element has an empty
95 * content model. If $wgWellFormedXml is false, then a few bytes will be
96 * shaved off the HTML output as well. In the future, other HTML-specific
97 * features might be added, like allowing arrays for the values of
98 * attributes like class= and media=.
100 * @param $element string The element's name, e.g., 'a'
101 * @param $attribs array Associative array of attributes, e.g., array(
102 * 'href' => 'http://www.mediawiki.org/' ). See expandAttributes() for
103 * further documentation.
104 * @param $contents string The raw HTML contents of the element: *not*
106 * @return string Raw HTML
108 public static function rawElement( $element, $attribs = array(), $contents = '' ) {
109 global $wgWellFormedXml;
110 $start = self::openElement( $element, $attribs );
111 if ( in_array( $element, self::$voidElements ) ) {
112 if ( $wgWellFormedXml ) {
114 return substr( $start, 0, -1 ) . ' />';
118 return "$start$contents</$element>";
123 * Identical to rawElement(), but HTML-escapes $contents (like
126 public static function element( $element, $attribs = array(), $contents = '' ) {
127 return self::rawElement( $element, $attribs, strtr( $contents, array(
128 # There's no point in escaping quotes, >, etc. in the contents of
136 * Identical to rawElement(), but has no third parameter and omits the end
137 * tag (and the self-closing / in XML mode for empty elements).
139 public static function openElement( $element, $attribs = array() ) {
141 $attribs = (array)$attribs;
142 # This is not required in HTML5, but let's do it anyway, for
143 # consistency and better compression.
144 $element = strtolower( $element );
146 # Remove HTML5-only attributes if we aren't doing HTML5, and disable
147 # form validation regardless (see bug 23769 and the more detailed
148 # comment in expandAttributes())
149 if ( $element == 'input' ) {
150 # Whitelist of types that don't cause validation. All except
151 # 'search' are valid in XHTML1.
165 if ( isset( $attribs['type'] )
166 && !in_array( $attribs['type'], $validTypes ) ) {
167 unset( $attribs['type'] );
169 if ( isset( $attribs['type'] ) && $attribs['type'] == 'search'
171 unset( $attribs['type'] );
173 # Here we're blacklisting some HTML5-only attributes...
174 $html5attribs = array(
186 foreach ( $html5attribs as $badAttr ) {
187 unset( $attribs[$badAttr] );
190 if ( !$wgHtml5 && $element == 'textarea' && isset( $attribs['maxlength'] ) ) {
191 unset( $attribs['maxlength'] );
194 return "<$element" . self::expandAttributes(
195 self::dropDefaults( $element, $attribs ) ) . '>';
199 * Given an element name and an associative array of element attributes,
200 * return an array that is functionally identical to the input array, but
201 * possibly smaller. In particular, attributes might be stripped if they
202 * are given their default values.
204 * This method is not guaranteed to remove all redundant attributes, only
205 * some common ones and some others selected arbitrarily at random. It
206 * only guarantees that the output array should be functionally identical
207 * to the input array (currently per the HTML 5 draft as of 2009-09-06).
209 * @param $element string Name of the element, e.g., 'a'
210 * @param $attribs array Associative array of attributes, e.g., array(
211 * 'href' => 'http://www.mediawiki.org/' ). See expandAttributes() for
212 * further documentation.
213 * @return array An array of attributes functionally identical to $attribs
215 private static function dropDefaults( $element, $attribs ) {
216 # Don't bother doing anything if we aren't outputting HTML5; it's too
217 # much of a pain to maintain two sets of defaults.
223 static $attribDefaults = array(
224 'area' => array( 'shape' => 'rect' ),
226 'formaction' => 'GET',
227 'formenctype' => 'application/x-www-form-urlencoded',
234 'command' => array( 'type' => 'command' ),
237 'autocomplete' => 'on',
238 'enctype' => 'application/x-www-form-urlencoded',
241 'formaction' => 'GET',
245 'keygen' => array( 'keytype' => 'rsa' ),
246 'link' => array( 'media' => 'all' ),
247 'menu' => array( 'type' => 'list' ),
248 # Note: the use of text/javascript here instead of other JavaScript
249 # MIME types follows the HTML5 spec.
250 'script' => array( 'type' => 'text/javascript' ),
253 'type' => 'text/css',
255 'textarea' => array( 'wrap' => 'soft' ),
258 $element = strtolower( $element );
260 foreach ( $attribs as $attrib => $value ) {
261 $lcattrib = strtolower( $attrib );
262 $value = strval( $value );
264 # Simple checks using $attribDefaults
265 if ( isset( $attribDefaults[$element][$lcattrib] ) &&
266 $attribDefaults[$element][$lcattrib] == $value ) {
267 unset( $attribs[$attrib] );
270 if ( $lcattrib == 'class' && $value == '' ) {
271 unset( $attribs[$attrib] );
276 if ( $element === 'link' && isset( $attribs['type'] )
277 && strval( $attribs['type'] ) == 'text/css' ) {
278 unset( $attribs['type'] );
280 if ( $element === 'select' && isset( $attribs['size'] ) ) {
281 if ( in_array( 'multiple', $attribs )
282 || ( isset( $attribs['multiple'] ) && $attribs['multiple'] !== false )
285 if ( strval( $attribs['size'] ) == '4' ) {
286 unset( $attribs['size'] );
290 if ( strval( $attribs['size'] ) == '1' ) {
291 unset( $attribs['size'] );
300 * Given an associative array of element attributes, generate a string
301 * to stick after the element name in HTML output. Like array( 'href' =>
302 * 'http://www.mediawiki.org/' ) becomes something like
303 * ' href="http://www.mediawiki.org"'. Again, this is like
304 * Xml::expandAttributes(), but it implements some HTML-specific logic.
305 * For instance, it will omit quotation marks if $wgWellFormedXml is false,
306 * and will treat boolean attributes specially.
308 * @param $attribs array Associative array of attributes, e.g., array(
309 * 'href' => 'http://www.mediawiki.org/' ). Values will be HTML-escaped.
310 * A value of false means to omit the attribute. For boolean attributes,
311 * you can omit the key, e.g., array( 'checked' ) instead of
312 * array( 'checked' => 'checked' ) or such.
313 * @return string HTML fragment that goes between element name and '>'
314 * (starting with a space if at least one attribute is output)
316 public static function expandAttributes( $attribs ) {
317 global $wgHtml5, $wgWellFormedXml;
320 $attribs = (array)$attribs;
321 foreach ( $attribs as $key => $value ) {
322 if ( $value === false ) {
326 # For boolean attributes, support array( 'foo' ) instead of
327 # requiring array( 'foo' => 'meaningless' ).
329 && in_array( strtolower( $value ), self::$boolAttribs ) ) {
333 # Not technically required in HTML5, but required in XHTML 1.0,
334 # and we'd like consistency and better compression anyway.
335 $key = strtolower( $key );
337 # Bug 23769: Blacklist all form validation attributes for now. Current
338 # (June 2010) WebKit has no UI, so the form just refuses to submit
339 # without telling the user why, which is much worse than failing
340 # server-side validation. Opera is the only other implementation at
341 # this time, and has ugly UI, so just kill the feature entirely until
342 # we have at least one good implementation.
343 if ( in_array( $key, array( 'max', 'min', 'pattern', 'required', 'step' ) ) ) {
347 # See the "Attributes" section in the HTML syntax part of HTML5,
348 # 9.1.2.3 as of 2009-08-10. Most attributes can have quotation
349 # marks omitted, but not all. (Although a literal " is not
350 # permitted, we don't check for that, since it will be escaped
353 # See also research done on further characters that need to be
354 # escaped: http://code.google.com/p/html5lib/issues/detail?id=93
355 $badChars = "\\x00- '=<>`/\x{00a0}\x{1680}\x{180e}\x{180F}\x{2000}\x{2001}"
356 . "\x{2002}\x{2003}\x{2004}\x{2005}\x{2006}\x{2007}\x{2008}\x{2009}"
357 . "\x{200A}\x{2028}\x{2029}\x{202F}\x{205F}\x{3000}";
358 if ( $wgWellFormedXml || $value === ''
359 || preg_match( "![$badChars]!u", $value ) ) {
365 if ( in_array( $key, self::$boolAttribs ) ) {
366 # In XHTML 1.0 Transitional, the value needs to be equal to the
367 # key. In HTML5, we can leave the value empty instead. If we
368 # don't need well-formed XML, we can omit the = entirely.
369 if ( !$wgWellFormedXml ) {
371 } elseif ( $wgHtml5 ) {
372 $ret .= " $key=\"\"";
374 $ret .= " $key=\"$key\"";
377 # Apparently we need to entity-encode \n, \r, \t, although the
378 # spec doesn't mention that. Since we're doing strtr() anyway,
379 # and we don't need <> escaped here, we may as well not call
380 # htmlspecialchars(). FIXME: verify that we actually need to
381 # escape \n\r\t here, and explain why, exactly.
383 # We could call Sanitizer::encodeAttribute() for this, but we
384 # don't because we're stubborn and like our marginal savings on
385 # byte size from not having to encode unnecessary quotes.
393 if ( $wgWellFormedXml ) {
394 # This is allowed per spec: <http://www.w3.org/TR/xml/#NT-AttValue>
395 # But reportedly it breaks some XML tools? FIXME: is this
399 $ret .= " $key=$quote" . strtr( $value, $map ) . $quote;
406 * Output a <script> tag with the given contents. TODO: do some useful
407 * escaping as well, like if $contents contains literal '</script>' or (for
408 * XML) literal "]]>".
410 * @param $contents string JavaScript
411 * @return string Raw HTML
413 public static function inlineScript( $contents ) {
414 global $wgHtml5, $wgJsMimeType, $wgWellFormedXml;
418 $attrs['type'] = $wgJsMimeType;
420 if ( $wgWellFormedXml && preg_match( '/[<&]/', $contents ) ) {
421 $contents = "/*<![CDATA[*/$contents/*]]>*/";
423 return self::rawElement( 'script', $attrs, $contents );
427 * Output a <script> tag linking to the given URL, e.g.,
428 * <script src=foo.js></script>.
431 * @return string Raw HTML
433 public static function linkedScript( $url ) {
434 global $wgHtml5, $wgJsMimeType;
436 $attrs = array( 'src' => $url );
438 $attrs['type'] = $wgJsMimeType;
440 return self::element( 'script', $attrs );
444 * Output a <style> tag with the given contents for the given media type
445 * (if any). TODO: do some useful escaping as well, like if $contents
446 * contains literal '</style>' (admittedly unlikely).
448 * @param $contents string CSS
449 * @param $media mixed A media type string, like 'screen'
450 * @return string Raw HTML
452 public static function inlineStyle( $contents, $media = 'all' ) {
453 global $wgWellFormedXml;
455 if ( $wgWellFormedXml && preg_match( '/[<&]/', $contents ) ) {
456 $contents = "/*<![CDATA[*/$contents/*]]>*/";
458 return self::rawElement( 'style', array(
459 'type' => 'text/css',
465 * Output a <link rel=stylesheet> linking to the given URL for the given
466 * media type (if any).
469 * @param $media mixed A media type string, like 'screen'
470 * @return string Raw HTML
472 public static function linkedStyle( $url, $media = 'all' ) {
473 return self::element( 'link', array(
474 'rel' => 'stylesheet',
476 'type' => 'text/css',
482 * Convenience function to produce an <input> element. This supports the
483 * new HTML5 input types and attributes, and will silently strip them if
486 * @param $name string name attribute
487 * @param $value mixed value attribute
488 * @param $type string type attribute
489 * @param $attribs array Associative array of miscellaneous extra
490 * attributes, passed to Html::element()
491 * @return string Raw HTML
493 public static function input( $name, $value = '', $type = 'text', $attribs = array() ) {
494 $attribs['type'] = $type;
495 $attribs['value'] = $value;
496 $attribs['name'] = $name;
498 return self::element( 'input', $attribs );
502 * Convenience function to produce an input element with type=hidden, like
505 * @param $name string name attribute
506 * @param $value string value attribute
507 * @param $attribs array Associative array of miscellaneous extra
508 * attributes, passed to Html::element()
509 * @return string Raw HTML
511 public static function hidden( $name, $value, $attribs = array() ) {
512 return self::input( $name, $value, 'hidden', $attribs );
516 * Convenience function to produce an <input> element. This supports leaving
517 * out the cols= and rows= which Xml requires and are required by HTML4/XHTML
518 * but not required by HTML5 and will silently set cols="" and rows="" if
519 * $wgHtml5 is false and cols and rows are omitted (HTML4 validates present
520 * but empty cols="" and rows="" as valid).
522 * @param $name string name attribute
523 * @param $value string value attribute
524 * @param $attribs array Associative array of miscellaneous extra
525 * attributes, passed to Html::element()
526 * @return string Raw HTML
528 public static function textarea( $name, $value = '', $attribs = array() ) {
530 $attribs['name'] = $name;
532 if ( !isset( $attribs['cols'] ) )
533 $attribs['cols'] = "";
534 if ( !isset( $attribs['rows'] ) )
535 $attribs['rows'] = "";
537 return self::element( 'textarea', $attribs, $value );