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                 $attribs = (array)$attribs;
 111                 # This is not required in HTML 5, but let's do it anyway, for
 112                 # consistency and better compression.
 113                 $element = strtolower( $element );
 114
 115                 # Element-specific hacks to slim down output and ensure validity
 116                 if ( $element == 'input' ) {
 117                         if ( !$wgHtml5 ) {
 118                                 # With $wgHtml5 off we want to validate as XHTML 1, so we
 119                                 # strip out any fancy HTML 5-only input types for now.
 120                                 #
 121                                 # Whitelist of valid types:
 122                                 $validTypes = array(
 123                                         'hidden',
 124                                         'text',
 125                                         'password',
 126                                         'checkbox',
 127                                         'radio',
 128                                         'file',
 129                                         'submit',
 130                                         'image',
 131                                         'reset',
 132                                         'button',
 133                                 );
 134                                 if ( isset( $attribs['type'] )
 135                                 && !in_array( $attribs['type'], $validTypes ) ) {
 136                                         # Fall back to type=text, the default
 137                                         unset( $attribs['type'] );
 138                                 }
 139                                 # Here we're blacklisting some HTML5-only attributes...
 140                                 $html5attribs = array(
 141                                         'autocomplete',
 142                                         'autofocus',
 143                                         'max',
 144                                         'min',
 145                                         'multiple',
 146                                         'pattern',
 147                                         'placeholder',
 148                                         'required',
 149                                         'step',
 150                                 );
 151                                 foreach ( $html5attribs as $badAttr ) {
 152                                         unset( $attribs[$badAttr] );
 153                                 }
 154                         }
 155                 }
 156
 157                 $start = "<$element" . self::expandAttributes(
 158                         self::dropDefaults( $element, $attribs ) );
 159                 if ( in_array( $element, self::$voidElements ) ) {
 160                         if ( $wgWellFormedXml ) {
 161                                 return "$start />";
 162                         }
 163                         return "$start>";
 164                 } else {
 165                         return "$start>$contents</$element>";
 166                 }
 167         }
 168
 169         /**
 170          * Identical to rawElement(), but HTML-escapes $contents (like
 171          * Xml::element()).
 172          */
 173         public static function element( $element, $attribs = array(), $contents = '' ) {
 174                 return self::rawElement( $element, $attribs, strtr( $contents, array(
 175                         # There's no point in escaping quotes, >, etc. in the contents of
 176                         # elements.
 177                         '&' => '&amp;',
 178                         '<' => '&lt;'
 179                 ) ) );
 180         }
 181
 182         /**
 183          * Given an element name and an associative array of element attributes,
 184          * return an array that is functionally identical to the input array, but
 185          * possibly smaller.  In particular, attributes might be stripped if they
 186          * are given their default values.
 187          *
 188          * This method is not guaranteed to remove all redundant attributes, only
 189          * some common ones and some others selected arbitrarily at random.  It
 190          * only guarantees that the output array should be functionally identical
 191          * to the input array (currently per the HTML 5 draft as of 2009-09-06).
 192          *
 193          * @param $element string Name of the element, e.g., 'a'
 194          * @param $attribs array  Associative array of attributes, e.g., array(
 195          *   'href' => 'http://www.mediawiki.org/' ).
 196          * @return array An array of attributes functionally identical to $attribs
 197          */
 198         private static function dropDefaults( $element, $attribs ) {
 199                 # Don't bother doing anything if we aren't outputting HTML5; it's too
 200                 # much of a pain to maintain two sets of defaults.
 201                 global $wgHtml5;
 202                 if ( !$wgHtml5 ) {
 203                         return $attribs;
 204                 }
 205
 206                 static $attribDefaults = array(
 207                         'area' => array( 'shape' => 'rect' ),
 208                         'button' => array(
 209                                 'formaction' => 'GET',
 210                                 'formenctype' => 'application/x-www-form-urlencoded',
 211                                 'type' => 'submit',
 212                         ),
 213                         'canvas' => array(
 214                                 'height' => '150',
 215                                 'width' => '300',
 216                         ),
 217                         'command' => array( 'type' => 'command' ),
 218                         'form' => array(
 219                                 'action' => 'GET',
 220                                 'autocomplete' => 'on',
 221                                 'enctype' => 'application/x-www-form-urlencoded',
 222                         ),
 223                         'input' => array(
 224                                 'formaction' => 'GET',
 225                                 'type' => 'text',
 226                                 'value' => '',
 227                         ),
 228                         'keygen' => array( 'keytype' => 'rsa' ),
 229                         'link' => array( 'media' => 'all' ),
 230                         'menu' => array( 'type' => 'list' ),
 231                         # Note: the use of text/javascript here instead of other JavaScript
 232                         # MIME types follows the HTML 5 spec.
 233                         'script' => array( 'type' => 'text/javascript' ),
 234                         'style' => array(
 235                                 'media' => 'all',
 236                                 'type' => 'text/css',
 237                         ),
 238                         'textarea' => array( 'wrap' => 'soft' ),
 239                 );
 240
 241                 $element = strtolower( $element );
 242
 243                 foreach ( $attribs as $attrib => $value ) {
 244                         $lcattrib = strtolower( $attrib );
 245                         $value = strval( $value );
 246
 247                         # Simple checks using $attribDefaults
 248                         if ( isset( $attribDefaults[$element][$lcattrib] ) &&
 249                         $attribDefaults[$element][$lcattrib] == $value ) {
 250                                 unset( $attribs[$attrib] );
 251                         }
 252
 253                         if ( $lcattrib == 'class' && $value == '' ) {
 254                                 unset( $attribs[$attrib] );
 255                         }
 256                 }
 257
 258                 # More subtle checks
 259                 if ( $element === 'link' && isset( $attribs['type'] )
 260                 && strval( $attribs['type'] ) == 'text/css' ) {
 261                         unset( $attribs['type'] );
 262                 }
 263                 if ( $element === 'select' && isset( $attribs['size'] ) ) {
 264                         if ( in_array( 'multiple', $attribs )
 265                                 || ( isset( $attribs['multiple'] ) && $attribs['multiple'] !== false )
 266                         ) {
 267                                 # A multi-select
 268                                 if ( strval( $attribs['size'] ) == '4' ) {
 269                                         unset( $attribs['size'] );
 270                                 }
 271                         } else {
 272                                 # Single select
 273                                 if ( strval( $attribs['size'] ) == '1' ) {
 274                                         unset( $attribs['size'] );
 275                                 }
 276                         }
 277                 }
 278
 279                 return $attribs;
 280         }
 281
 282         /**
 283          * Given an associative array of element attributes, generate a string
 284          * to stick after the element name in HTML output.  Like array( 'href' =>
 285          * 'http://www.mediawiki.org/' ) becomes something like
 286          * ' href="http://www.mediawiki.org"'.  Again, this is like
 287          * Xml::expandAttributes(), but it implements some HTML-specific logic.
 288          * For instance, it will omit quotation marks if $wgWellFormedXml is false,
 289          * and will treat boolean attributes specially.
 290          *
 291          * @param $attribs array Associative array of attributes, e.g., array(
 292          *   'href' => 'http://www.mediawiki.org/' ).  Values will be HTML-escaped.
 293          *   A value of false means to omit the attribute.
 294          * @return string HTML fragment that goes between element name and '>'
 295          *   (starting with a space if at least one attribute is output)
 296          */
 297         public static function expandAttributes( $attribs ) {
 298                 global $wgHtml5, $wgWellFormedXml;
 299
 300                 $ret = '';
 301                 $attribs = (array)$attribs;
 302                 foreach ( $attribs as $key => $value ) {
 303                         if ( $value === false ) {
 304                                 continue;
 305                         }
 306
 307                         # For boolean attributes, support array( 'foo' ) instead of
 308                         # requiring array( 'foo' => 'meaningless' ).
 309                         if ( is_int( $key )
 310                         && in_array( strtolower( $value ), self::$boolAttribs ) ) {
 311                                 $key = $value;
 312                         }
 313
 314                         # Not technically required in HTML 5, but required in XHTML 1.0,
 315                         # and we'd like consistency and better compression anyway.
 316                         $key = strtolower( $key );
 317
 318                         # See the "Attributes" section in the HTML syntax part of HTML 5,
 319                         # 9.1.2.3 as of 2009-08-10.  Most attributes can have quotation
 320                         # marks omitted, but not all.  (Although a literal " is not
 321                         # permitted, we don't check for that, since it will be escaped
 322                         # anyway.)
 323                         #
 324                         # See also research done on further characters that need to be
 325                         # escaped: http://code.google.com/p/html5lib/issues/detail?id=93
 326                         $badChars = "\\x00- '=<>`/\x{00a0}\x{1680}\x{180e}\x{180F}\x{2000}\x{2001}"
 327                                 . "\x{2002}\x{2003}\x{2004}\x{2005}\x{2006}\x{2007}\x{2008}\x{2009}"
 328                                 . "\x{200A}\x{2028}\x{2029}\x{202F}\x{205F}\x{3000}";
 329                         if ( $wgWellFormedXml || $value === ''
 330                         || preg_match( "![$badChars]!u", $value ) ) {
 331                                 $quote = '"';
 332                         } else {
 333                                 $quote = '';
 334                         }
 335
 336                         if ( in_array( $key, self::$boolAttribs ) ) {
 337                                 # In XHTML 1.0 Transitional, the value needs to be equal to the
 338                                 # key.  In HTML 5, we can leave the value empty instead.  If we
 339                                 # don't need well-formed XML, we can omit the = entirely.
 340                                 if ( !$wgWellFormedXml ) {
 341                                         $ret .= " $key";
 342                                 } elseif ( $wgHtml5 ) {
 343                                         $ret .= " $key=\"\"";
 344                                 } else {
 345                                         $ret .= " $key=\"$key\"";
 346                                 }
 347                         } else {
 348                                 # Apparently we need to entity-encode \n, \r, \t, although the
 349                                 # spec doesn't mention that.  Since we're doing strtr() anyway,
 350                                 # and we don't need <> escaped here, we may as well not call
 351                                 # htmlspecialchars().  FIXME: verify that we actually need to
 352                                 # escape \n\r\t here, and explain why, exactly.
 353                                 if ( $wgHtml5 ) {
 354                                         $ret .= " $key=$quote" . strtr( $value, array(
 355                                                 '&' => '&amp;',
 356                                                 '"' => '&quot;',
 357                                                 "\n" => '&#10;',
 358                                                 "\r" => '&#13;',
 359                                                 "\t" => '&#9;'
 360                                         ) ) . $quote;
 361                                 } else {
 362                                         $ret .= " $key=$quote" . Sanitizer::encodeAttribute( $value ) . $quote;
 363                                 }
 364                         }
 365                 }
 366                 return $ret;
 367         }
 368
 369         /**
 370          * Output a <script> tag with the given contents.  TODO: do some useful
 371          * escaping as well, like if $contents contains literal '</script>' or (for
 372          * XML) literal "]]>".
 373          *
 374          * @param $contents string JavaScript
 375          * @return string Raw HTML
 376          */
 377         public static function inlineScript( $contents ) {
 378                 global $wgHtml5, $wgJsMimeType, $wgWellFormedXml;
 379
 380                 $attrs = array();
 381                 if ( !$wgHtml5 ) {
 382                         $attrs['type'] = $wgJsMimeType;
 383                 }
 384                 if ( $wgWellFormedXml && preg_match( '/[<&]/', $contents ) ) {
 385                         $contents = "/*<![CDATA[*/$contents/*]]>*/";
 386                 }
 387                 return self::rawElement( 'script', $attrs, $contents );
 388         }
 389
 390         /**
 391          * Output a <script> tag linking to the given URL, e.g.,
 392          * <script src=foo.js></script>.
 393          *
 394          * @param $url string
 395          * @return string Raw HTML
 396          */
 397         public static function linkedScript( $url ) {
 398                 global $wgHtml5, $wgJsMimeType;
 399
 400                 $attrs = array( 'src' => $url );
 401                 if ( !$wgHtml5 ) {
 402                         $attrs['type'] = $wgJsMimeType;
 403                 }
 404                 return self::element( 'script', $attrs );
 405         }
 406
 407         /**
 408          * Output a <style> tag with the given contents for the given media type
 409          * (if any).  TODO: do some useful escaping as well, like if $contents
 410          * contains literal '</style>' (admittedly unlikely).
 411          *
 412          * @param $contents string CSS
 413          * @param $media mixed A media type string, like 'screen'
 414          * @return string Raw HTML
 415          */
 416         public static function inlineStyle( $contents, $media = 'all' ) {
 417                 global $wgWellFormedXml;
 418
 419                 if ( $wgWellFormedXml && preg_match( '/[<&]/', $contents ) ) {
 420                         $contents = "/*<![CDATA[*/$contents/*]]>*/";
 421                 }
 422                 return self::rawElement( 'style', array(
 423                         'type' => 'text/css',
 424                         'media' => $media,
 425                 ), $contents );
 426         }
 427
 428         /**
 429          * Output a <link rel=stylesheet> linking to the given URL for the given
 430          * media type (if any).
 431          *
 432          * @param $url string
 433          * @param $media mixed A media type string, like 'screen'
 434          * @return string Raw HTML
 435          */
 436         public static function linkedStyle( $url, $media = 'all' ) {
 437                 return self::element( 'link', array(
 438                         'rel' => 'stylesheet',
 439                         'href' => $url,
 440                         'type' => 'text/css',
 441                         'media' => $media,
 442                 ) );
 443         }
 444
 445         /**
 446          * Convenience function to produce an <input> element.  This supports the
 447          * new HTML 5 input types and attributes, and will silently strip them if
 448          * $wgHtml5 is false.
 449          *
 450          * @param $name    string name attribute
 451          * @param $value   mixed  value attribute
 452          * @param $type    string type attribute
 453          * @param $attribs array  Associative array of miscellaneous extra
 454          *   attributes, passed to Html::element()
 455          * @return string Raw HTML
 456          */
 457         public static function input( $name, $value = '', $type = 'text', $attribs = array() ) {
 458                 $attribs['type'] = $type;
 459                 $attribs['value'] = $value;
 460                 $attribs['name'] = $name;
 461
 462                 return self::element( 'input', $attribs );
 463         }
 464
 465         /**
 466          * Convenience function to produce an input element with type=hidden, like
 467          * Xml::hidden.
 468          *
 469          * @param $name    string name attribute
 470          * @param $value   string value attribute
 471          * @param $attribs array  Associative array of miscellaneous extra
 472          *   attributes, passed to Html::element()
 473          * @return string Raw HTML
 474          */
 475         public static function hidden( $name, $value, $attribs = array() ) {
 476                 return self::input( $name, $value, 'hidden', $attribs );
 477         }
 478 }