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         # A nested associative array of element => content attribute => default
  87         # value.  Attributes that have the default value will be omitted, since
  88         # they're pointless.  Currently the list hasn't been systematically
  89         # populated.
  90         private static $attribDefaults = array(
  91                 'input' => array(
  92                         'value' => '',
  93                         'type' => 'text',
  94                 ),
  95         );
  96
  97         /**
  98          * Returns an HTML element in a string.  The major advantage here over
  99          * manually typing out the HTML is that it will escape all attribute
 100          * values.  If you're hardcoding all the attributes, or there are none, you
 101          * should probably type out the string yourself.
 102          *
 103          * This is quite similar to Xml::tags(), but it implements some useful
 104          * HTML-specific logic.  For instance, there is no $allowShortTag
 105          * parameter: the closing tag is magically omitted if $element has an empty
 106          * content model.  If $wgWellFormedXml is false, then a few bytes will be
 107          * shaved off the HTML output as well.  In the future, other HTML-specific
 108          * features might be added, like allowing arrays for the values of
 109          * attributes like class= and media=.
 110          *
 111          * @param $element  string The element's name, e.g., 'a'
 112          * @param $attribs  array  Associative array of attributes, e.g., array(
 113          *   'href' => 'http://www.mediawiki.org/' ).  Values will be HTML-escaped.
 114          *   A value of false means to omit the attribute.
 115          * @param $contents string The raw HTML contents of the element: *not*
 116          *   escaped!
 117          * @return string Raw HTML
 118          */
 119         public static function rawElement( $element, $attribs = array(), $contents = '' ) {
 120                 global $wgHtml5, $wgWellFormedXml;
 121                 # This is not required in HTML 5, but let's do it anyway, for
 122                 # consistency and better compression.
 123                 $element = strtolower( $element );
 124
 125                 # Element-specific hacks to slim down output and ensure validity
 126                 if ( $element == 'input' ) {
 127                         if ( !$wgHtml5 ) {
 128                                 # With $wgHtml5 off we want to validate as XHTML 1, so we
 129                                 # strip out any fancy HTML 5-only input types for now.
 130                                 #
 131                                 # Whitelist of valid types:
 132                                 $validTypes = array(
 133                                         'hidden',
 134                                         'text',
 135                                         'password',
 136                                         'checkbox',
 137                                         'radio',
 138                                         'file',
 139                                         'submit',
 140                                         'image',
 141                                         'reset',
 142                                         'button',
 143                                 );
 144                                 if ( isset( $attribs['type'] )
 145                                 && !in_array( $attribs['type'], $validTypes ) ) {
 146                                         # Fall back to type=text, the default
 147                                         unset( $attribs['type'] );
 148                                 }
 149                                 # Here we're blacklisting some HTML5-only attributes...
 150                                 $html5attribs = array(
 151                                         'autocomplete',
 152                                         'autofocus',
 153                                         'max',
 154                                         'min',
 155                                         'multiple',
 156                                         'pattern',
 157                                         'placeholder',
 158                                         'required',
 159                                         'step',
 160                                 );
 161                                 foreach ( $html5attribs as $badAttr ) {
 162                                         unset( $attribs[$badAttr] );
 163                                 }
 164                         }
 165                 }
 166
 167                 # Don't bother outputting the default values for attributes
 168                 foreach ( $attribs as $attrib => $value ) {
 169                         $lcattrib = strtolower( $attrib );
 170                         if ( isset( self::$attribDefaults[$element][$lcattrib] ) &&
 171                         self::$attribDefaults[$element][$lcattrib] === $value ) {
 172                                 unset( $attribs[$attrib] );
 173                         }
 174                 }
 175
 176                 $start = "<$element" . self::expandAttributes( $attribs );
 177                 if ( in_array( $element, self::$voidElements ) ) {
 178                         if ( $wgWellFormedXml ) {
 179                                 return "$start />";
 180                         }
 181                         return "$start>";
 182                 } else {
 183                         return "$start>$contents</$element>";
 184                 }
 185         }
 186
 187         /**
 188          * Identical to rawElement(), but HTML-escapes $contents (like
 189          * Xml::element()).
 190          */
 191         public static function element( $element, $attribs = array(), $contents = '' ) {
 192                 return self::rawElement( $element, $attribs, strtr( $contents, array(
 193                         # There's no point in escaping quotes, >, etc. in the contents of
 194                         # elements.
 195                         '&' => '&amp;',
 196                         '<' => '&lt;'
 197                 ) ) );
 198         }
 199
 200         /**
 201          * Given an associative array of element attributes, generate a string
 202          * to stick after the element name in HTML output.  Like array( 'href' =>
 203          * 'http://www.mediawiki.org/' ) becomes something like
 204          * ' href="http://www.mediawiki.org"'.  Again, this is like
 205          * Xml::expandAttributes(), but it implements some HTML-specific logic.
 206          * For instance, it will omit quotation marks if $wgWellFormedXml is false,
 207          * and will treat boolean attributes specially.
 208          *
 209          * @param $attribs array Associative array of attributes, e.g., array(
 210          *   'href' => 'http://www.mediawiki.org/' ).  Values will be HTML-escaped.
 211          *   A value of false means to omit the attribute.
 212          * @return string HTML fragment that goes between element name and '>'
 213          *   (starting with a space if at least one attribute is output)
 214          */
 215         public static function expandAttributes( $attribs ) {
 216                 global $wgHtml5, $wgWellFormedXml;
 217
 218                 $ret = '';
 219                 foreach ( $attribs as $key => $value ) {
 220                         if ( $value === false ) {
 221                                 continue;
 222                         }
 223
 224                         # For boolean attributes, support array( 'foo' ) instead of
 225                         # requiring array( 'foo' => 'meaningless' ).
 226                         if ( is_int( $key )
 227                         && in_array( strtolower( $value ), self::$boolAttribs ) ) {
 228                                 $key = $value;
 229                         }
 230
 231                         # Not technically required in HTML 5, but required in XHTML 1.0,
 232                         # and we'd like consistency and better compression anyway.
 233                         $key = strtolower( $key );
 234
 235                         # See the "Attributes" section in the HTML syntax part of HTML 5,
 236                         # 9.1.2.3 as of 2009-08-10.  Most attributes can have quotation
 237                         # marks omitted, but not all.  (Although a literal " is not
 238                         # permitted, we don't check for that, since it will be escaped
 239                         # anyway.)
 240                         if ( $wgWellFormedXml || $value == ''
 241                         || preg_match( "/[ '=<>]/", $value ) ) {
 242                                 $quote = '"';
 243                         } else {
 244                                 $quote = '';
 245                         }
 246
 247                         if ( in_array( $key, self::$boolAttribs ) ) {
 248                                 # In XHTML 1.0 Transitional, the value needs to be equal to the
 249                                 # key.  In HTML 5, we can leave the value empty instead.  If we
 250                                 # don't need well-formed XML, we can omit the = entirely.
 251                                 if ( !$wgWellFormedXml ) {
 252                                         $ret .= " $key";
 253                                 } elseif ( $wgHtml5 ) {
 254                                         $ret .= " $key=\"\"";
 255                                 } else {
 256                                         $ret .= " $key=\"$key\"";
 257                                 }
 258                         } else {
 259                                 # Apparently we need to entity-encode \n, \r, \t, although the
 260                                 # spec doesn't mention that.  Since we're doing strtr() anyway,
 261                                 # and we don't need <> escaped here, we may as well not call
 262                                 # htmlspecialchars().  FIXME: verify that we actually need to
 263                                 # escape \n\r\t here, and explain why, exactly.
 264                                 $ret .= " $key=$quote" . strtr( $value, array(
 265                                         '&' => '&amp;',
 266                                         '"' => '&quot;',
 267                                         "\n" => '&#10;',
 268                                         "\r" => '&#13;',
 269                                         "\t" => '&#9;'
 270                                 ) ) . $quote;
 271                         }
 272                 }
 273                 return $ret;
 274         }
 275
 276         /**
 277          * Output a <script> tag with the given contents.  TODO: do some useful
 278          * escaping as well, like if $contents contains literal '</script>' or (for
 279          * XML) literal "]]>".
 280          *
 281          * @param $contents string JavaScript
 282          * @return string Raw HTML
 283          */
 284         public static function inlineScript( $contents ) {
 285                 global $wgHtml5, $wgJsMimeType, $wgWellFormedXml;
 286
 287                 $attrs = array();
 288                 if ( !$wgHtml5 ) {
 289                         $attrs['type'] = $wgJsMimeType;
 290                 }
 291                 if ( $wgWellFormedXml && preg_match( '/[<&]/', $contents ) ) {
 292                         $contents = "/*<![CDATA[*/$contents/*]]>*/";
 293                 }
 294                 return self::rawElement( 'script', $attrs, $contents );
 295         }
 296
 297         /**
 298          * Output a <script> tag linking to the given URL, e.g.,
 299          * <script src=foo.js></script>.
 300          *
 301          * @param $url string
 302          * @return string Raw HTML
 303          */
 304         public static function linkedScript( $url ) {
 305                 global $wgHtml5, $wgJsMimeType;
 306
 307                 $attrs = array( 'src' => $url );
 308                 if ( !$wgHtml5 ) {
 309                         $attrs['type'] = $wgJsMimeType;
 310                 }
 311                 return self::element( 'script', $attrs );
 312         }
 313
 314         /**
 315          * Output a <style> tag with the given contents for the given media type
 316          * (if any).  TODO: do some useful escaping as well, like if $contents
 317          * contains literal '</style>' (admittedly unlikely).
 318          *
 319          * @param $contents string CSS
 320          * @param $media mixed A media type string, like 'screen', or null for all
 321          *   media
 322          * @return string Raw HTML
 323          */
 324         public static function inlineStyle( $contents, $media = null ) {
 325                 global $wgHtml5, $wgWellFormedXml;
 326
 327                 $attrs = array();
 328                 if ( !$wgHtml5 ) {
 329                         $attrs['type'] = 'text/css';
 330                 }
 331                 if ( $wgWellFormedXml && preg_match( '/[<&]/', $contents ) ) {
 332                         $contents = "/*<![CDATA[*/$contents/*]]>*/";
 333                 }
 334                 if ( $media !== null ) {
 335                         $attrs['media'] = $media;
 336                 }
 337                 return self::rawElement( 'style', $attrs, $contents );
 338         }
 339
 340         /**
 341          * Output a <link rel=stylesheet> linking to the given URL for the given
 342          * media type (if any).
 343          *
 344          * @param $url string
 345          * @param $media mixed A media type string, like 'screen', or null for all
 346          *   media
 347          * @return string Raw HTML
 348          */
 349         public static function linkedStyle( $url, $media = null ) {
 350                 global $wgHtml5;
 351
 352                 $attrs = array( 'rel' => 'stylesheet', 'href' => $url );
 353                 if ( !$wgHtml5 ) {
 354                         $attrs['type'] = 'text/css';
 355                 }
 356                 if ( $media !== null ) {
 357                         $attrs['media'] = $media;
 358                 }
 359                 return self::element( 'link', $attrs );
 360         }
 361
 362         /**
 363          * Convenience function to produce an <input> element.  This supports the
 364          * new HTML 5 input types and attributes, and will silently strip them if
 365          * $wgHtml5 is false.
 366          *
 367          * @param $name    string name attribute
 368          * @param $value   mixed  value attribute (null = omit)
 369          * @param $type    string type attribute
 370          * @param $attribs array  Associative array of miscellaneous extra
 371          *   attributes, passed to Html::element()
 372          * @return string Raw HTML
 373          */
 374         public static function input( $name, $value = null, $type = 'text', $attribs = array() ) {
 375                 $attribs['type'] = $type;
 376                 if ( $value !== null ) {
 377                         $attribs['value'] = $value;
 378                 }
 379                 $attribs['name'] = $name;
 380
 381                 return self::element( 'input', $attribs );
 382         }
 383
 384         /**
 385          * Convenience function to produce an input element with type=hidden, like
 386          * Xml::hidden.
 387          *
 388          * @param $name    string name attribute
 389          * @param $value   string value attribute
 390          * @param $attribs array  Associative array of miscellaneous extra
 391          *   attributes, passed to Html::element()
 392          * @return string Raw HTML
 393          */
 394         public static function hidden( $name, $value, $attribs = array() ) {
 395                 return self::input( $name, $value, 'hidden', $attribs );
 396         }
 397 }