includes/Html.php

   1 <?php
   2 # Copyright (C) 2009 Aryeh Gregor
   3 # http://www.mediawiki.org/
   4 #
   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.
   9 #
  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.
  14 #
  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
  19
  20 /**
  21  * This class is a collection of static functions that serve two purposes:
  22  *
  23  * 1) Implement any algorithms specified by HTML 5, or other HTML
  24  * specifications, in a convenient and self-contained way.
  25  *
  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
  29  * output.
  30  *
  31  * There are two important configuration options this class uses:
  32  *
  33  * $wgHtml5: If this is set to false, then all output should be valid XHTML 1.0
  34  *     Transitional.
  35  * $wgWellFormedXml: If this is set to true, then all output should be
  36  *     well-formed XML (quotes on attributes, self-closing tags, etc.).
  37  *
  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
  40  * <a> elements.
  41  */
  42 class Html {
  43         # List of void elements from HTML 5, section 9.1.2 as of 2009-08-10
  44         private static $voidElements = array(
  45                 'area',
  46                 'base',
  47                 'br',
  48                 'col',
  49                 'command',
  50                 'embed',
  51                 'hr',
  52                 'img',
  53                 'input',
  54                 'keygen',
  55                 'link',
  56                 'meta',
  57                 'param',
  58                 'source',
  59         );
  60
  61         # Boolean attributes, which may have the value omitted entirely.  Manually
  62         # collected from the HTML 5 spec as of 2009-08-10.
  63         private static $boolAttribs = array(
  64                 'async',
  65                 'autobuffer',
  66                 'autofocus',
  67                 'autoplay',
  68                 'checked',
  69                 'controls',
  70                 'defer',
  71                 'disabled',
  72                 'formnovalidate',
  73                 'hidden',
  74                 'ismap',
  75                 'loop',
  76                 'multiple',
  77                 'novalidate',
  78                 'open',
  79                 'readonly',
  80                 'required',
  81                 'reversed',
  82                 'scoped',
  83                 'seamless',
  84         );
  85
  86         /**
  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.
  91          *
  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=.
  99          *
 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/' ).  Values will be HTML-escaped.
 103          *   A value of false means to omit the attribute.
 104          * @param $contents string The raw HTML contents of the element: *not*
 105          *   escaped!
 106          * @return string Raw HTML
 107          */
 108         public static function rawElement( $element, $attribs = array(), $contents = '' ) {
 109                 global $wgHtml5, $wgWellFormedXml;
 110                 # This is not required in HTML 5, but let's do it anyway, for
 111                 # consistency and better compression.
 112                 $element = strtolower( $element );
 113
 114                 # Element-specific hacks to slim down output and ensure validity
 115                 if ( $element == 'input' ) {
 116                         if ( !$wgHtml5 ) {
 117                                 # With $wgHtml5 off we want to validate as XHTML 1, so we
 118                                 # strip out any fancy HTML 5-only input types for now.
 119                                 #
 120                                 # Whitelist of valid types:
 121                                 $validTypes = array(
 122                                         'hidden',
 123                                         'text',
 124                                         'password',
 125                                         'checkbox',
 126                                         'radio',
 127                                         'file',
 128                                         'submit',
 129                                         'image',
 130                                         'reset',
 131                                         'button',
 132                                 );
 133                                 if ( isset( $attribs['type'] )
 134                                 && !in_array( $attribs['type'], $validTypes ) ) {
 135                                         # Fall back to type=text, the default
 136                                         unset( $attribs['type'] );
 137                                 }
 138                                 # Here we're blacklisting some HTML5-only attributes...
 139                                 $html5attribs = array(
 140                                         'autocomplete',
 141                                         'autofocus',
 142                                         'max',
 143                                         'min',
 144                                         'multiple',
 145                                         'pattern',
 146                                         'placeholder',
 147                                         'required',
 148                                         'step',
 149                                 );
 150                                 foreach ( $html5attribs as $badAttr ) {
 151                                         unset( $attribs[$badAttr] );
 152                                 }
 153                         }
 154                 }
 155
 156                 $start = "<$element" . self::expandAttributes(
 157                         self::dropDefaults( $element, $attribs ) );
 158                 if ( in_array( $element, self::$voidElements ) ) {
 159                         if ( $wgWellFormedXml ) {
 160                                 return "$start />";
 161                         }
 162                         return "$start>";
 163                 } else {
 164                         return "$start>$contents</$element>";
 165                 }
 166         }
 167
 168         /**
 169          * Identical to rawElement(), but HTML-escapes $contents (like
 170          * Xml::element()).
 171          */
 172         public static function element( $element, $attribs = array(), $contents = '' ) {
 173                 return self::rawElement( $element, $attribs, strtr( $contents, array(
 174                         # There's no point in escaping quotes, >, etc. in the contents of
 175                         # elements.
 176                         '&' => '&amp;',
 177                         '<' => '&lt;'
 178                 ) ) );
 179         }
 180
 181         /**
 182          * Given an element name and an associative array of element attributes,
 183          * return an array that is functionally identical to the input array, but
 184          * possibly smaller.  In particular, attributes might be stripped if they
 185          * are given their default values.
 186          *
 187          * This method is not guaranteed to remove all redundant attributes, only
 188          * some common ones and some others selected arbitrarily at random.  It
 189          * only guarantees that the output array should be functionally identical
 190          * to the input array (currently per the HTML 5 draft as of 2009-09-06).
 191          *
 192          * @param $element string Name of the element, e.g., 'a'
 193          * @param $attribs array  Associative array of attributes, e.g., array(
 194          *   'href' => 'http://www.mediawiki.org/' ).
 195          * @return array An array of attributes functionally identical to $attribs
 196          */
 197         private static function dropDefaults( $element, $attribs ) {
 198                 # Don't bother doing anything if we aren't outputting HTML5; it's too
 199                 # much of a pain to maintain two sets of defaults.
 200                 global $wgHtml5;
 201                 if ( !$wgHtml5 ) {
 202                         return $attribs;
 203                 }
 204
 205                 static $attribDefaults = array(
 206                         'area' => array( 'shape' => 'rect' ),
 207                         'button' => array(
 208                                 'formaction' => 'GET',
 209                                 'formenctype' => 'application/x-www-form-urlencoded',
 210                                 'type' => 'submit',
 211                         ),
 212                         'canvas' => array(
 213                                 'height' => '150',
 214                                 'width' => '300',
 215                         ),
 216                         'command' => array( 'type' => 'command' ),
 217                         'form' => array(
 218                                 'action' => 'GET',
 219                                 'autocomplete' => 'on',
 220                                 'enctype' => 'application/x-www-form-urlencoded',
 221                         ),
 222                         'input' => array(
 223                                 'formaction' => 'GET',
 224                                 'type' => 'text',
 225                                 'value' => '',
 226                         ),
 227                         'keygen' => array( 'keytype' => 'rsa' ),
 228                         'link' => array( 'media' => 'all' ),
 229                         'menu' => array( 'type' => 'list' ),
 230                         # Note: the use of text/javascript here instead of other JavaScript
 231                         # MIME types follows the HTML 5 spec.
 232                         'script' => array( 'type' => 'text/javascript' ),
 233                         'style' => array(
 234                                 'media' => 'all',
 235                                 'type' => 'text/css',
 236                         ),
 237                         'textarea' => array( 'wrap' => 'soft' ),
 238                 );
 239
 240                 $element = strtolower( $element );
 241
 242                 foreach ( $attribs as $attrib => $value ) {
 243                         $lcattrib = strtolower( $attrib );
 244                         $value = strval( $value );
 245
 246                         # Simple checks using $attribDefaults
 247                         if ( isset( $attribDefaults[$element][$lcattrib] ) &&
 248                         $attribDefaults[$element][$lcattrib] == $value ) {
 249                                 unset( $attribs[$attrib] );
 250                         }
 251
 252                         if ( $lcattrib == 'class' && $value == '' ) {
 253                                 unset( $attribs[$attrib] );
 254                         }
 255                 }
 256
 257                 # More subtle checks
 258                 if ( $element === 'link' && isset( $attribs['type'] )
 259                 && strval( $attribs['type'] ) == 'text/css' ) {
 260                         unset( $attribs['type'] );
 261                 }
 262                 if ( $element === 'select' && isset( $attribs['size'] ) ) {
 263                         if ( in_array( 'multiple', $attribs )
 264                                 || ( isset( $attribs['multiple'] ) && $attribs['multiple'] !== false )
 265                         ) {
 266                                 # A multi-select
 267                                 if ( strval( $attribs['size'] ) == '4' ) {
 268                                         unset( $attribs['size'] );
 269                                 }
 270                         } else {
 271                                 # Single select
 272                                 if ( strval( $attribs['size'] ) == '1' ) {
 273                                         unset( $attribs['size'] );
 274                                 }
 275                         }
 276                 }
 277
 278                 return $attribs;
 279         }
 280
 281         /**
 282          * Given an associative array of element attributes, generate a string
 283          * to stick after the element name in HTML output.  Like array( 'href' =>
 284          * 'http://www.mediawiki.org/' ) becomes something like
 285          * ' href="http://www.mediawiki.org"'.  Again, this is like
 286          * Xml::expandAttributes(), but it implements some HTML-specific logic.
 287          * For instance, it will omit quotation marks if $wgWellFormedXml is false,
 288          * and will treat boolean attributes specially.
 289          *
 290          * @param $attribs array Associative array of attributes, e.g., array(
 291          *   'href' => 'http://www.mediawiki.org/' ).  Values will be HTML-escaped.
 292          *   A value of false means to omit the attribute.
 293          * @return string HTML fragment that goes between element name and '>'
 294          *   (starting with a space if at least one attribute is output)
 295          */
 296         public static function expandAttributes( $attribs ) {
 297                 global $wgHtml5, $wgWellFormedXml;
 298
 299                 $ret = '';
 300                 $attribs = (array)$attribs;
 301                 foreach ( $attribs as $key => $value ) {
 302                         if ( $value === false ) {
 303                                 continue;
 304                         }
 305
 306                         # For boolean attributes, support array( 'foo' ) instead of
 307                         # requiring array( 'foo' => 'meaningless' ).
 308                         if ( is_int( $key )
 309                         && in_array( strtolower( $value ), self::$boolAttribs ) ) {
 310                                 $key = $value;
 311                         }
 312
 313                         # Not technically required in HTML 5, but required in XHTML 1.0,
 314                         # and we'd like consistency and better compression anyway.
 315                         $key = strtolower( $key );
 316
 317                         # See the "Attributes" section in the HTML syntax part of HTML 5,
 318                         # 9.1.2.3 as of 2009-08-10.  Most attributes can have quotation
 319                         # marks omitted, but not all.  (Although a literal " is not
 320                         # permitted, we don't check for that, since it will be escaped
 321                         # anyway.)
 322                         #
 323                         # See also research done on further characters that need to be
 324                         # escaped: http://code.google.com/p/html5lib/issues/detail?id=93
 325                         $badChars = "\\x00- '=<>`/\x{00a0}\x{1680}\x{180e}\x{180F}\x{2000}\x{2001}"
 326                                 . "\x{2002}\x{2003}\x{2004}\x{2005}\x{2006}\x{2007}\x{2008}\x{2009}"
 327                                 . "\x{200A}\x{2028}\x{2029}\x{202F}\x{205F}\x{3000}";
 328                         if ( $wgWellFormedXml || $value === ''
 329                         || preg_match( "![$badChars]!u", $value ) ) {
 330                                 $quote = '"';
 331                         } else {
 332                                 $quote = '';
 333                         }
 334
 335                         if ( in_array( $key, self::$boolAttribs ) ) {
 336                                 # In XHTML 1.0 Transitional, the value needs to be equal to the
 337                                 # key.  In HTML 5, we can leave the value empty instead.  If we
 338                                 # don't need well-formed XML, we can omit the = entirely.
 339                                 if ( !$wgWellFormedXml ) {
 340                                         $ret .= " $key";
 341                                 } elseif ( $wgHtml5 ) {
 342                                         $ret .= " $key=\"\"";
 343                                 } else {
 344                                         $ret .= " $key=\"$key\"";
 345                                 }
 346                         } else {
 347                                 # Apparently we need to entity-encode \n, \r, \t, although the
 348                                 # spec doesn't mention that.  Since we're doing strtr() anyway,
 349                                 # and we don't need <> escaped here, we may as well not call
 350                                 # htmlspecialchars().  FIXME: verify that we actually need to
 351                                 # escape \n\r\t here, and explain why, exactly.
 352                                 if ( $wgHtml5 ) {
 353                                         $ret .= " $key=$quote" . strtr( $value, array(
 354                                                 '&' => '&amp;',
 355                                                 '"' => '&quot;',
 356                                                 "\n" => '&#10;',
 357                                                 "\r" => '&#13;',
 358                                                 "\t" => '&#9;'
 359                                         ) ) . $quote;
 360                                 } else {
 361                                         $ret .= " $key=$quote" . Sanitizer::encodeAttribute( $value ) . $quote;
 362                                 }
 363                         }
 364                 }
 365                 return $ret;
 366         }
 367
 368         /**
 369          * Output a <script> tag with the given contents.  TODO: do some useful
 370          * escaping as well, like if $contents contains literal '</script>' or (for
 371          * XML) literal "]]>".
 372          *
 373          * @param $contents string JavaScript
 374          * @return string Raw HTML
 375          */
 376         public static function inlineScript( $contents ) {
 377                 global $wgHtml5, $wgJsMimeType, $wgWellFormedXml;
 378
 379                 $attrs = array();
 380                 if ( !$wgHtml5 ) {
 381                         $attrs['type'] = $wgJsMimeType;
 382                 }
 383                 if ( $wgWellFormedXml && preg_match( '/[<&]/', $contents ) ) {
 384                         $contents = "/*<![CDATA[*/$contents/*]]>*/";
 385                 }
 386                 return self::rawElement( 'script', $attrs, $contents );
 387         }
 388
 389         /**
 390          * Output a <script> tag linking to the given URL, e.g.,
 391          * <script src=foo.js></script>.
 392          *
 393          * @param $url string
 394          * @return string Raw HTML
 395          */
 396         public static function linkedScript( $url ) {
 397                 global $wgHtml5, $wgJsMimeType;
 398
 399                 $attrs = array( 'src' => $url );
 400                 if ( !$wgHtml5 ) {
 401                         $attrs['type'] = $wgJsMimeType;
 402                 }
 403                 return self::element( 'script', $attrs );
 404         }
 405
 406         /**
 407          * Output a <style> tag with the given contents for the given media type
 408          * (if any).  TODO: do some useful escaping as well, like if $contents
 409          * contains literal '</style>' (admittedly unlikely).
 410          *
 411          * @param $contents string CSS
 412          * @param $media mixed A media type string, like 'screen'
 413          * @return string Raw HTML
 414          */
 415         public static function inlineStyle( $contents, $media = 'all' ) {
 416                 global $wgWellFormedXml;
 417
 418                 if ( $wgWellFormedXml && preg_match( '/[<&]/', $contents ) ) {
 419                         $contents = "/*<![CDATA[*/$contents/*]]>*/";
 420                 }
 421                 return self::rawElement( 'style', array(
 422                         'type' => 'text/css',
 423                         'media' => $media,
 424                 ), $contents );
 425         }
 426
 427         /**
 428          * Output a <link rel=stylesheet> linking to the given URL for the given
 429          * media type (if any).
 430          *
 431          * @param $url string
 432          * @param $media mixed A media type string, like 'screen'
 433          * @return string Raw HTML
 434          */
 435         public static function linkedStyle( $url, $media = 'all' ) {
 436                 return self::element( 'link', array(
 437                         'rel' => 'stylesheet',
 438                         'href' => $url,
 439                         'type' => 'text/css',
 440                         'media' => $media,
 441                 ) );
 442         }
 443
 444         /**
 445          * Convenience function to produce an <input> element.  This supports the
 446          * new HTML 5 input types and attributes, and will silently strip them if
 447          * $wgHtml5 is false.
 448          *
 449          * @param $name    string name attribute
 450          * @param $value   mixed  value attribute
 451          * @param $type    string type attribute
 452          * @param $attribs array  Associative array of miscellaneous extra
 453          *   attributes, passed to Html::element()
 454          * @return string Raw HTML
 455          */
 456         public static function input( $name, $value = '', $type = 'text', $attribs = array() ) {
 457                 $attribs['type'] = $type;
 458                 $attribs['value'] = $value;
 459                 $attribs['name'] = $name;
 460
 461                 return self::element( 'input', $attribs );
 462         }
 463
 464         /**
 465          * Convenience function to produce an input element with type=hidden, like
 466          * Xml::hidden.
 467          *
 468          * @param $name    string name attribute
 469          * @param $value   string value attribute
 470          * @param $attribs array  Associative array of miscellaneous extra
 471          *   attributes, passed to Html::element()
 472          * @return string Raw HTML
 473          */
 474         public static function hidden( $name, $value, $attribs = array() ) {
 475                 return self::input( $name, $value, 'hidden', $attribs );
 476         }
 477 }