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 HTML5, 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 HTML5, 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 HTML5 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/' ).  See expandAttributes() for
 103          *   further documentation.
 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 $wgWellFormedXml;
 110                 $start = self::openElement( $element, $attribs );
 111                 if ( in_array( $element, self::$voidElements ) ) {
 112                         if ( $wgWellFormedXml ) {
 113                                 # Silly XML.
 114                                 return substr( $start, 0, -1 ) . ' />';
 115                         }
 116                         return $start;
 117                 } else {
 118                         return "$start$contents</$element>";
 119                 }
 120         }
 121
 122         /**
 123          * Identical to rawElement(), but HTML-escapes $contents (like
 124          * Xml::element()).
 125          */
 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
 129                         # elements.
 130                         '&' => '&amp;',
 131                         '<' => '&lt;'
 132                 ) ) );
 133         }
 134
 135         /**
 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).
 138          */
 139         public static function openElement( $element, $attribs = array() ) {
 140                 global $wgHtml5, $wgWellFormedXml;
 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( $element );
 145
 146                 # Remove HTML5-only attributes if we aren't doing HTML5
 147                 if ( !$wgHtml5 ) {
 148                         if ( $element == 'input' ) {
 149                                 # Whitelist of valid XHTML1 types
 150                                 $validTypes = array(
 151                                         'hidden',
 152                                         'text',
 153                                         'password',
 154                                         'checkbox',
 155                                         'radio',
 156                                         'file',
 157                                         'submit',
 158                                         'image',
 159                                         'reset',
 160                                         'button',
 161                                 );
 162                                 if ( isset( $attribs['type'] )
 163                                 && !in_array( $attribs['type'], $validTypes ) ) {
 164                                         # Fall back to type=text, the default
 165                                         unset( $attribs['type'] );
 166                                 }
 167                         }
 168                         if ( $element == 'textarea' && isset( $attribs['maxlength'] ) ) {
 169                                 unset( $attribs['maxlength'] );
 170                         }
 171                         # Here we're blacklisting some HTML5-only attributes...
 172                         $html5attribs = array(
 173                                 'autocomplete',
 174                                 'autofocus',
 175                                 'max',
 176                                 'min',
 177                                 'multiple',
 178                                 'pattern',
 179                                 'placeholder',
 180                                 'required',
 181                                 'step',
 182                                 'spellcheck',
 183                         );
 184                         foreach ( $html5attribs as $badAttr ) {
 185                                 unset( $attribs[$badAttr] );
 186                         }
 187                 }
 188
 189                 return "<$element" . self::expandAttributes(
 190                         self::dropDefaults( $element, $attribs ) ) . '>';
 191         }
 192
 193         /**
 194          * Given an element name and an associative array of element attributes,
 195          * return an array that is functionally identical to the input array, but
 196          * possibly smaller.  In particular, attributes might be stripped if they
 197          * are given their default values.
 198          *
 199          * This method is not guaranteed to remove all redundant attributes, only
 200          * some common ones and some others selected arbitrarily at random.  It
 201          * only guarantees that the output array should be functionally identical
 202          * to the input array (currently per the HTML 5 draft as of 2009-09-06).
 203          *
 204          * @param $element string Name of the element, e.g., 'a'
 205          * @param $attribs array  Associative array of attributes, e.g., array(
 206          *   'href' => 'http://www.mediawiki.org/' ).  See expandAttributes() for
 207          *   further documentation.
 208          * @return array An array of attributes functionally identical to $attribs
 209          */
 210         private static function dropDefaults( $element, $attribs ) {
 211                 # Don't bother doing anything if we aren't outputting HTML5; it's too
 212                 # much of a pain to maintain two sets of defaults.
 213                 global $wgHtml5;
 214                 if ( !$wgHtml5 ) {
 215                         return $attribs;
 216                 }
 217
 218                 static $attribDefaults = array(
 219                         'area' => array( 'shape' => 'rect' ),
 220                         'button' => array(
 221                                 'formaction' => 'GET',
 222                                 'formenctype' => 'application/x-www-form-urlencoded',
 223                                 'type' => 'submit',
 224                         ),
 225                         'canvas' => array(
 226                                 'height' => '150',
 227                                 'width' => '300',
 228                         ),
 229                         'command' => array( 'type' => 'command' ),
 230                         'form' => array(
 231                                 'action' => 'GET',
 232                                 'autocomplete' => 'on',
 233                                 'enctype' => 'application/x-www-form-urlencoded',
 234                         ),
 235                         'input' => array(
 236                                 'formaction' => 'GET',
 237                                 'type' => 'text',
 238                                 'value' => '',
 239                         ),
 240                         'keygen' => array( 'keytype' => 'rsa' ),
 241                         'link' => array( 'media' => 'all' ),
 242                         'menu' => array( 'type' => 'list' ),
 243                         # Note: the use of text/javascript here instead of other JavaScript
 244                         # MIME types follows the HTML5 spec.
 245                         'script' => array( 'type' => 'text/javascript' ),
 246                         'style' => array(
 247                                 'media' => 'all',
 248                                 'type' => 'text/css',
 249                         ),
 250                         'textarea' => array( 'wrap' => 'soft' ),
 251                 );
 252
 253                 $element = strtolower( $element );
 254
 255                 foreach ( $attribs as $attrib => $value ) {
 256                         $lcattrib = strtolower( $attrib );
 257                         $value = strval( $value );
 258
 259                         # Simple checks using $attribDefaults
 260                         if ( isset( $attribDefaults[$element][$lcattrib] ) &&
 261                         $attribDefaults[$element][$lcattrib] == $value ) {
 262                                 unset( $attribs[$attrib] );
 263                         }
 264
 265                         if ( $lcattrib == 'class' && $value == '' ) {
 266                                 unset( $attribs[$attrib] );
 267                         }
 268                 }
 269
 270                 # More subtle checks
 271                 if ( $element === 'link' && isset( $attribs['type'] )
 272                 && strval( $attribs['type'] ) == 'text/css' ) {
 273                         unset( $attribs['type'] );
 274                 }
 275                 if ( $element === 'select' && isset( $attribs['size'] ) ) {
 276                         if ( in_array( 'multiple', $attribs )
 277                                 || ( isset( $attribs['multiple'] ) && $attribs['multiple'] !== false )
 278                         ) {
 279                                 # A multi-select
 280                                 if ( strval( $attribs['size'] ) == '4' ) {
 281                                         unset( $attribs['size'] );
 282                                 }
 283                         } else {
 284                                 # Single select
 285                                 if ( strval( $attribs['size'] ) == '1' ) {
 286                                         unset( $attribs['size'] );
 287                                 }
 288                         }
 289                 }
 290
 291                 return $attribs;
 292         }
 293
 294         /**
 295          * Given an associative array of element attributes, generate a string
 296          * to stick after the element name in HTML output.  Like array( 'href' =>
 297          * 'http://www.mediawiki.org/' ) becomes something like
 298          * ' href="http://www.mediawiki.org"'.  Again, this is like
 299          * Xml::expandAttributes(), but it implements some HTML-specific logic.
 300          * For instance, it will omit quotation marks if $wgWellFormedXml is false,
 301          * and will treat boolean attributes specially.
 302          *
 303          * @param $attribs array Associative array of attributes, e.g., array(
 304          *   'href' => 'http://www.mediawiki.org/' ).  Values will be HTML-escaped.
 305          *   A value of false means to omit the attribute.  For boolean attributes,
 306          *   you can omit the key, e.g., array( 'checked' ) instead of
 307          *   array( 'checked' => 'checked' ) or such.
 308          * @return string HTML fragment that goes between element name and '>'
 309          *   (starting with a space if at least one attribute is output)
 310          */
 311         public static function expandAttributes( $attribs ) {
 312                 global $wgHtml5, $wgWellFormedXml;
 313
 314                 $ret = '';
 315                 $attribs = (array)$attribs;
 316                 foreach ( $attribs as $key => $value ) {
 317                         if ( $value === false ) {
 318                                 continue;
 319                         }
 320
 321                         # For boolean attributes, support array( 'foo' ) instead of
 322                         # requiring array( 'foo' => 'meaningless' ).
 323                         if ( is_int( $key )
 324                         && in_array( strtolower( $value ), self::$boolAttribs ) ) {
 325                                 $key = $value;
 326                         }
 327
 328                         # Not technically required in HTML5, but required in XHTML 1.0,
 329                         # and we'd like consistency and better compression anyway.
 330                         $key = strtolower( $key );
 331
 332                         # See the "Attributes" section in the HTML syntax part of HTML5,
 333                         # 9.1.2.3 as of 2009-08-10.  Most attributes can have quotation
 334                         # marks omitted, but not all.  (Although a literal " is not
 335                         # permitted, we don't check for that, since it will be escaped
 336                         # anyway.)
 337                         #
 338                         # See also research done on further characters that need to be
 339                         # escaped: http://code.google.com/p/html5lib/issues/detail?id=93
 340                         $badChars = "\\x00- '=<>`/\x{00a0}\x{1680}\x{180e}\x{180F}\x{2000}\x{2001}"
 341                                 . "\x{2002}\x{2003}\x{2004}\x{2005}\x{2006}\x{2007}\x{2008}\x{2009}"
 342                                 . "\x{200A}\x{2028}\x{2029}\x{202F}\x{205F}\x{3000}";
 343                         if ( $wgWellFormedXml || $value === ''
 344                         || preg_match( "![$badChars]!u", $value ) ) {
 345                                 $quote = '"';
 346                         } else {
 347                                 $quote = '';
 348                         }
 349
 350                         if ( in_array( $key, self::$boolAttribs ) ) {
 351                                 # In XHTML 1.0 Transitional, the value needs to be equal to the
 352                                 # key.  In HTML5, we can leave the value empty instead.  If we
 353                                 # don't need well-formed XML, we can omit the = entirely.
 354                                 if ( !$wgWellFormedXml ) {
 355                                         $ret .= " $key";
 356                                 } elseif ( $wgHtml5 ) {
 357                                         $ret .= " $key=\"\"";
 358                                 } else {
 359                                         $ret .= " $key=\"$key\"";
 360                                 }
 361                         } else {
 362                                 # Apparently we need to entity-encode \n, \r, \t, although the
 363                                 # spec doesn't mention that.  Since we're doing strtr() anyway,
 364                                 # and we don't need <> escaped here, we may as well not call
 365                                 # htmlspecialchars().  FIXME: verify that we actually need to
 366                                 # escape \n\r\t here, and explain why, exactly.
 367                                 #
 368                                 # We could call Sanitizer::encodeAttribute() for this, but we
 369                                 # don't because we're stubborn and like our marginal savings on
 370                                 # byte size from not having to encode unnecessary quotes.
 371                                 $map = array(
 372                                         '&' => '&amp;',
 373                                         '"' => '&quot;',
 374                                         "\n" => '&#10;',
 375                                         "\r" => '&#13;',
 376                                         "\t" => '&#9;'
 377                                 );
 378                                 if ( $wgWellFormedXml ) {
 379                                         # '<' must be escaped in attributes for XML for some
 380                                         # reason, per spec: http://www.w3.org/TR/xml/#NT-AttValue
 381                                         $map['<'] = '&lt;';
 382                                 }
 383                                 $ret .= " $key=$quote" . strtr( $value, $map ) . $quote;
 384                         }
 385                 }
 386                 return $ret;
 387         }
 388
 389         /**
 390          * Output a <script> tag with the given contents.  TODO: do some useful
 391          * escaping as well, like if $contents contains literal '</script>' or (for
 392          * XML) literal "]]>".
 393          *
 394          * @param $contents string JavaScript
 395          * @return string Raw HTML
 396          */
 397         public static function inlineScript( $contents ) {
 398                 global $wgHtml5, $wgJsMimeType, $wgWellFormedXml;
 399
 400                 $attrs = array();
 401                 if ( !$wgHtml5 ) {
 402                         $attrs['type'] = $wgJsMimeType;
 403                 }
 404                 if ( $wgWellFormedXml && preg_match( '/[<&]/', $contents ) ) {
 405                         $contents = "/*<![CDATA[*/$contents/*]]>*/";
 406                 }
 407                 return self::rawElement( 'script', $attrs, $contents );
 408         }
 409
 410         /**
 411          * Output a <script> tag linking to the given URL, e.g.,
 412          * <script src=foo.js></script>.
 413          *
 414          * @param $url string
 415          * @return string Raw HTML
 416          */
 417         public static function linkedScript( $url ) {
 418                 global $wgHtml5, $wgJsMimeType;
 419
 420                 $attrs = array( 'src' => $url );
 421                 if ( !$wgHtml5 ) {
 422                         $attrs['type'] = $wgJsMimeType;
 423                 }
 424                 return self::element( 'script', $attrs );
 425         }
 426
 427         /**
 428          * Output a <style> tag with the given contents for the given media type
 429          * (if any).  TODO: do some useful escaping as well, like if $contents
 430          * contains literal '</style>' (admittedly unlikely).
 431          *
 432          * @param $contents string CSS
 433          * @param $media mixed A media type string, like 'screen'
 434          * @return string Raw HTML
 435          */
 436         public static function inlineStyle( $contents, $media = 'all' ) {
 437                 global $wgWellFormedXml;
 438
 439                 if ( $wgWellFormedXml && preg_match( '/[<&]/', $contents ) ) {
 440                         $contents = "/*<![CDATA[*/$contents/*]]>*/";
 441                 }
 442                 return self::rawElement( 'style', array(
 443                         'type' => 'text/css',
 444                         'media' => $media,
 445                 ), $contents );
 446         }
 447
 448         /**
 449          * Output a <link rel=stylesheet> linking to the given URL for the given
 450          * media type (if any).
 451          *
 452          * @param $url string
 453          * @param $media mixed A media type string, like 'screen'
 454          * @return string Raw HTML
 455          */
 456         public static function linkedStyle( $url, $media = 'all' ) {
 457                 return self::element( 'link', array(
 458                         'rel' => 'stylesheet',
 459                         'href' => $url,
 460                         'type' => 'text/css',
 461                         'media' => $media,
 462                 ) );
 463         }
 464
 465         /**
 466          * Convenience function to produce an <input> element.  This supports the
 467          * new HTML5 input types and attributes, and will silently strip them if
 468          * $wgHtml5 is false.
 469          *
 470          * @param $name    string name attribute
 471          * @param $value   mixed  value attribute
 472          * @param $type    string type attribute
 473          * @param $attribs array  Associative array of miscellaneous extra
 474          *   attributes, passed to Html::element()
 475          * @return string Raw HTML
 476          */
 477         public static function input( $name, $value = '', $type = 'text', $attribs = array() ) {
 478                 $attribs['type'] = $type;
 479                 $attribs['value'] = $value;
 480                 $attribs['name'] = $name;
 481
 482                 return self::element( 'input', $attribs );
 483         }
 484
 485         /**
 486          * Convenience function to produce an input element with type=hidden, like
 487          * Xml::hidden.
 488          *
 489          * @param $name    string name attribute
 490          * @param $value   string value attribute
 491          * @param $attribs array  Associative array of miscellaneous extra
 492          *   attributes, passed to Html::element()
 493          * @return string Raw HTML
 494          */
 495         public static function hidden( $name, $value, $attribs = array() ) {
 496                 return self::input( $name, $value, 'hidden', $attribs );
 497         }
 498
 499         /**
 500          * Convenience function to produce an <input> element.  This supports leaving
 501          * out the cols= and rows= which Xml requires and are required by HTML4/XHTML
 502          * but not required by HTML5 and will silently set cols="" and rows="" if
 503          * $wgHtml5 is false and cols and rows are omitted (HTML4 validates present
 504          * but empty cols="" and rows="" as valid).
 505          *
 506          * @param $name    string name attribute
 507          * @param $value   string value attribute
 508          * @param $attribs array  Associative array of miscellaneous extra
 509          *   attributes, passed to Html::element()
 510          * @return string Raw HTML
 511          */
 512         public static function textarea( $name, $value = '', $attribs = array() ) {
 513                 global $wgHtml5;
 514                 $attribs['name'] = $name;
 515                 if ( !$wgHtml5 ) {
 516                         if ( !isset( $attribs['cols'] ) )
 517                                 $attribs['cols'] = "";
 518                         if ( !isset( $attribs['rows'] ) )
 519                                 $attribs['rows'] = "";
 520                 }
 521                 return self::element( 'textarea', $attribs, $value );
 522         }
 523 }