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          * @param $contents string The raw HTML contents of the element: *not*
 104          *   escaped!
 105          * @return string Raw HTML
 106          */
 107         public static function rawElement( $element, $attribs = array(), $contents = '' ) {
 108                 global $wgHtml5, $wgWellFormedXml;
 109                 # This is not required in HTML 5, but let's do it anyway, for
 110                 # consistency and better compression.
 111                 $element = strtolower( $element );
 112
 113                 # Element-specific hacks to slim down output and ensure validity
 114                 if ( $element == 'input' ) {
 115                         if ( !$wgHtml5 ) {
 116                                 # With $wgHtml5 off we want to validate as XHTML 1, so we
 117                                 # strip out any fancy HTML 5-only input types for now.
 118                                 #
 119                                 # Whitelist of valid types:
 120                                 $validTypes = array(
 121                                         'hidden',
 122                                         'text',
 123                                         'password',
 124                                         'checkbox',
 125                                         'radio',
 126                                         'file',
 127                                         'submit',
 128                                         'image',
 129                                         'reset',
 130                                         'button',
 131                                 );
 132                                 if ( isset( $attribs['type'] )
 133                                 && !in_array( $attribs['type'], $validTypes ) ) {
 134                                         # Fall back to type=text, the default
 135                                         unset( $attribs['type'] );
 136                                 }
 137                                 # Here we're blacklisting some HTML5-only attributes...
 138                                 $html5attribs = array(
 139                                         'autocomplete',
 140                                         'autofocus',
 141                                         'max',
 142                                         'min',
 143                                         'multiple',
 144                                         'pattern',
 145                                         'placeholder',
 146                                         'required',
 147                                         'step',
 148                                 );
 149                                 foreach ( $html5attribs as $badAttr ) {
 150                                         unset( $attribs[$badAttr] );
 151                                 }
 152                         }
 153                 }
 154
 155                 $start = "<$element" . self::expandAttributes( $attribs );
 156                 if ( in_array( $element, self::$voidElements ) ) {
 157                         if ( $wgWellFormedXml ) {
 158                                 return "$start />";
 159                         }
 160                         return "$start>";
 161                 } else {
 162                         return "$start>$contents</$element>";
 163                 }
 164         }
 165
 166         /**
 167          * Identical to rawElement(), but HTML-escapes $contents (like
 168          * Xml::element()).
 169          */
 170         public static function element( $element, $attribs = array(), $contents = '' ) {
 171                 return self::rawElement( $element, $attribs, strtr( $contents, array(
 172                         # There's no point in escaping quotes, >, etc. in the contents of
 173                         # elements.
 174                         '&' => '&amp;',
 175                         '<' => '&lt;'
 176                 ) ) );
 177         }
 178
 179         /**
 180          * Given an associative array of element attributes, generate a string
 181          * to stick after the element name in HTML output.  Like array( 'href' =>
 182          * 'http://www.mediawiki.org/' ) becomes something like
 183          * ' href="http://www.mediawiki.org"'.  Again, this is like
 184          * Xml::expandAttributes(), but it implements some HTML-specific logic.
 185          * For instance, it will omit quotation marks if $wgWellFormedXml is false,
 186          * and will treat boolean attributes specially.
 187          *
 188          * @param $attribs array Associative array of attributes, e.g., array(
 189          *   'href' => 'http://www.mediawiki.org/' ).  Values will be HTML-escaped.
 190          * @return string HTML fragment that goes between element name and '>'
 191          *   (starting with a space if at least one attribute is output)
 192          */
 193         public static function expandAttributes( $attribs ) {
 194                 global $wgHtml5, $wgWellFormedXml;
 195
 196                 $ret = '';
 197                 foreach ( $attribs as $key => $value ) {
 198                         # See the "Attributes" section in the HTML syntax part of HTML 5,
 199                         # 9.1.2.3 as of 2009-08-10.  Most attributes can have quotation
 200                         # marks omitted, but not all.  (Although a literal " is not
 201                         # permitted, we don't check for that, since it will be escaped
 202                         # anyway.)
 203                         if ( $wgWellFormedXml || $value == ''
 204                         || preg_match( "/[ '=<>]/", $value ) ) {
 205                                 $quote = '"';
 206                         } else {
 207                                 $quote = '';
 208                         }
 209
 210                         if ( in_array( $key, self::$boolAttribs ) ) {
 211                                 # In XHTML 1.0 Transitional, the value needs to be equal to the
 212                                 # key.  In HTML 5, we can leave the value empty instead.  If we
 213                                 # don't need well-formed XML, we can omit the = entirely.
 214                                 if ( !$wgWellFormedXml ) {
 215                                         $ret .= " $key";
 216                                 } elseif ( $wgHtml5 ) {
 217                                         $ret .= " $key=\"\"";
 218                                 } else {
 219                                         $ret .= " $key=\"$key\"";
 220                                 }
 221                         } else {
 222                                 # Apparently we need to entity-encode \n, \r, \t, although the
 223                                 # spec doesn't mention that.  Since we're doing strtr() anyway,
 224                                 # and we don't need <> escaped here, we may as well not call
 225                                 # htmlspecialchars().  FIXME: verify that we actually need to
 226                                 # escape \n\r\t here, and explain why, exactly.
 227                                 $ret .= " $key=$quote" . strtr( $value, array(
 228                                         '&' => '&amp;',
 229                                         '"' => '&quot;',
 230                                         "\n" => '&#10;',
 231                                         "\r" => '&#13;',
 232                                         "\t" => '&#9;'
 233                                 ) ) . $quote;
 234                         }
 235                 }
 236                 return $ret;
 237         }
 238
 239         /**
 240          * Output a <script> tag with the given contents.  TODO: do some useful
 241          * escaping as well, like if $contents contains literal '</script>' or (for
 242          * XML) literal "]]>".
 243          *
 244          * @param $contents string JavaScript
 245          * @return string Raw HTML
 246          */
 247         public static function inlineScript( $contents ) {
 248                 global $wgHtml5, $wgJsMimeType;
 249
 250                 $attrs = array();
 251                 if ( !$wgHtml5 ) {
 252                         $attrs['type'] = $wgJsMimeType;
 253                         $contents = "/*<![CDATA[*/$contents/*]]>*/";
 254                 }
 255                 return self::rawElement( 'script', $attrs, $contents );
 256         }
 257
 258         /**
 259          * Output a <script> tag linking to the given URL, e.g.,
 260          * <script src=foo.js></script>.
 261          *
 262          * @param $url string
 263          * @return string Raw HTML
 264          */
 265         public static function linkedScript( $url ) {
 266                 global $wgHtml5, $wgJsMimeType;
 267
 268                 $attrs = array( 'src' => $url );
 269                 if ( !$wgHtml5 ) {
 270                         $attrs['type'] = $wgJsMimeType;
 271                 }
 272                 return self::element( 'script', $attrs );
 273         }
 274
 275         /**
 276          * Output a <style> tag with the given contents for the given media type
 277          * (if any).  TODO: do some useful escaping as well, like if $contents
 278          * contains literal '</style>' (admittedly unlikely).
 279          *
 280          * @param $contents string CSS
 281          * @param $media mixed A media type string, like 'screen', or null for all
 282          *   media
 283          * @return string Raw HTML
 284          */
 285         public static function inlineStyle( $contents, $media = null ) {
 286                 global $wgHtml5;
 287
 288                 $attrs = array();
 289                 if ( !$wgHtml5 ) {
 290                         # Technically we should probably add CDATA stuff here like with
 291                         # scripts, but in practice, stylesheets tend not to have
 292                         # problematic characters anyway.
 293                         $attrs['type'] = 'text/css';
 294                 }
 295                 if ( $media !== null ) {
 296                         $attrs['media'] = $media;
 297                 }
 298                 return self::rawElement( 'style', $attrs, $contents );
 299         }
 300
 301         /**
 302          * Output a <link rel=stylesheet> linking to the given URL for the given
 303          * media type (if any).
 304          *
 305          * @param $url string
 306          * @param $media mixed A media type string, like 'screen', or null for all
 307          *   media
 308          * @return string Raw HTML
 309          */
 310         public static function linkedStyle( $url, $media = null ) {
 311                 global $wgHtml5;
 312
 313                 $attrs = array( 'rel' => 'stylesheet', 'href' => $url );
 314                 if ( !$wgHtml5 ) {
 315                         $attrs['type'] = 'text/css';
 316                 }
 317                 if ( $media !== null ) {
 318                         $attrs['media'] = $media;
 319                 }
 320                 return self::element( 'link', $attrs );
 321         }
 322
 323         /**
 324          * Convenience function to produce an <input> element.  This supports the
 325          * new HTML 5 input types and attributes, and will silently strip them if
 326          * $wgHtml5 is false.
 327          *
 328          * @param $name    string name attribute
 329          * @param $value   mixed  value attribute (null = omit)
 330          * @param $type    string type attribute
 331          * @param $attribs array  Assocative array of miscellaneous extra attributes,
 332          *   passed to Html::element()
 333          * @return string Raw HTML
 334          */
 335         public static function input( $name, $value = null, $type = 'text', $attribs = array() ) {
 336                 if ( $type != 'text' ) {
 337                         $attribs['type'] = $type;
 338                 }
 339                 if ( $value !== null ) {
 340                         $attribs['value'] = $value;
 341                 }
 342                 $attribs['name'] = $name;
 343
 344                 return self::element( 'input', $attribs );
 345         }
 346 }