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', 'base', 'br', 'col', 'command', 'embed', 'hr', 'img', 'input',
  46                 'keygen', 'link', 'meta', 'param', 'source'
  47         );
  48
  49         # Boolean attributes, which may have the value omitted entirely.  Manually
  50         # collected from the HTML 5 spec as of 2009-08-10.
  51         private static $boolAttribs = array( 'async', 'autobuffer', 'autofocus',
  52                 'autoplay', 'checked', 'controls', 'defer', 'disabled',
  53                 'formnovalidate', 'hidden', 'ismap', 'loop', 'multiple', 'novalidate',
  54                 'open', 'readonly', 'required', 'reversed', 'scoped', 'seamless'
  55         );
  56
  57         /**
  58          * Returns an HTML element in a string.  The major advantage here over
  59          * manually typing out the HTML is that it will escape all attribute
  60          * values.  If you're hardcoding all the attributes, or there are none, you
  61          * should probably type out the string yourself.
  62          *
  63          * This is quite similar to Xml::element(), but it implements some useful
  64          * HTML-specific logic.  For instance, there is no $allowShortTag
  65          * parameter: the closing tag is magically omitted if $element has an empty
  66          * content model.  If $wgWellFormedXml is false, then a few bytes will be
  67          * shaved off the HTML output as well.  In the future, other HTML-specific
  68          * features might be added, like allowing arrays for the values of
  69          * attributes like class= and media=.
  70          *
  71          * One notable difference to Xml::element() is that $contents is *not*
  72          * escaped.  This means that Html::element() can be usefully nested, rather
  73          * than using the rather clumsy Xml::openElement() and Xml::closeElement().
  74          *
  75          * @param $element  string The element's name, e.g., 'a'
  76          * @param $attribs  array  Associative array of attributes, e.g., array(
  77          *   'href' => 'http://www.mediawiki.org/' ).  Values will be HTML-escaped.
  78          * @param $contents string The raw HTML contents of the element: *not*
  79          *   escaped!
  80          * @return string Raw HTML
  81          */
  82         public static function rawElement( $element, $attribs = array(), $contents = '' ) {
  83                 global $wgWellFormedXml;
  84                 $element = strtolower( $element );
  85                 $start = "<$element" . self::expandAttributes( $attribs );
  86                 if ( in_array( $element, self::$voidElements ) ) {
  87                         if ( $wgWellFormedXml ) {
  88                                 return "$start />";
  89                         }
  90                         return "$start>";
  91                 } else {
  92                         return "$start>$contents</$element>";
  93                 }
  94         }
  95
  96         /**
  97          * Identical to rawElement(), but HTML-escapes $contents.
  98          */
  99         public static function element( $element, $attribs = array(), $contents = '' ) {
 100                 return self::rawElement( $element, $attribs, htmlspecialchars( $contents ) );
 101         }
 102
 103         /**
 104          * Given an associative array of element attributes, generate a string
 105          * to stick after the element name in HTML output.  Like array( 'href' =>
 106          * 'http://www.mediawiki.org/' ) becomes something like
 107          * ' href="http://www.mediawiki.org"'.  Again, this is like
 108          * Xml::expandAttributes(), but it implements some HTML-specific logic.
 109          * For instance, it will omit quotation marks if $wgWellFormedXml is false,
 110          * and will treat boolean attributes specially.
 111          *
 112          * @param $attribs array Associative array of attributes, e.g., array(
 113          *   'href' => 'http://www.mediawiki.org/' ).  Values will be HTML-escaped.
 114          * @return string HTML fragment that goes between element name and '>'
 115          *   (starting with a space if at least one attribute is output)
 116          */
 117         public static function expandAttributes( $attribs ) {
 118                 global $wgHtml5, $wgWellFormedXml;
 119
 120                 $ret = '';
 121                 foreach ( $attribs as $key => $value ) {
 122                         # See the "Attributes" section in the HTML syntax part of HTML 5,
 123                         # 9.1.2.3 as of 2009-08-10.  Most attributes can have quotation
 124                         # marks omitted, but not all.  (Although a literal " is not
 125                         # permitted, we don't check for that, since it will be escaped
 126                         # anyway.)
 127                         if ( $wgWellFormedXml || $value == ''
 128                         || preg_match( "/[ '=<>]/", $value ) ) {
 129                                 $quote = '"';
 130                         } else {
 131                                 $quote = '';
 132                         }
 133
 134                         if ( in_array( $key, self::$boolAttribs ) ) {
 135                                 # In XHTML 1.0 Transitional, the value needs to be equal to the
 136                                 # key.  In HTML 5, we can leave the value empty instead.  If we
 137                                 # don't need well-formed XML, we can omit the = entirely.
 138                                 if ( !$wgWellFormedXml ) {
 139                                         $ret .= " $key";
 140                                 } elseif ( $wgHtml5 ) {
 141                                         $ret .= " $key=\"\"";
 142                                 } else {
 143                                         $ret .= " $key=\"$key\"";
 144                                 }
 145                         } else {
 146                                 # Apparently we need to entity-encode \n, \r, \t, although the
 147                                 # spec doesn't mention that.  Since we're doing strtr() anyway,
 148                                 # and we don't need <> escaped here, we may as well not call
 149                                 # htmlspecialchars().  FIXME: verify that we actually need to
 150                                 # escape \n\r\t here, and explain why, exactly.
 151                                 $ret .= " $key=$quote" . strtr( $value, array(
 152                                         '&' => '&amp;',
 153                                         '"' => '&quot;',
 154                                         "\n" => '&#10;',
 155                                         "\r" => '&#13;',
 156                                         "\t" => '&#9;'
 157                                 ) ) . $quote;
 158                         }
 159                 }
 160                 return $ret;
 161         }
 162
 163         /**
 164          * Output a <script> tag with the given contents.  TODO: do some useful
 165          * escaping as well, like if $contents contains literal '</script>' or (for
 166          * XML) literal "]]>".
 167          *
 168          * Note that $contents will not be escaped, since JS may legitimately
 169          * contain unescaped characters like "<".  Make sure you don't output
 170          * untrusted user input here!
 171          *
 172          * @param $contents string JavaScript
 173          * @return string Raw HTML
 174          */
 175         public static function inlineScript( $contents ) {
 176                 global $wgHtml5, $wgJsMimeType;
 177
 178                 $attrs = array();
 179                 if ( !$wgHtml5 ) {
 180                         $attrs['type'] = $wgJsMimeType;
 181                         $contents = "/*<![CDATA[*/$contents/*]]>*/";
 182                 }
 183                 return self::rawElement( 'script', $attrs, $contents );
 184         }
 185
 186         /**
 187          * Output a <script> tag linking to the given URL, e.g.,
 188          * <script src=foo.js></script>.
 189          *
 190          * @param $url string
 191          * @return string Raw HTML
 192          */
 193         public static function linkedScript( $url ) {
 194                 global $wgHtml5, $wgJsMimeType;
 195
 196                 $attrs = array( 'src' => $url );
 197                 if ( !$wgHtml5 ) {
 198                         $attrs['type'] = $wgJsMimeType;
 199                 }
 200                 return self::element( 'script', $attrs );
 201         }
 202
 203         /**
 204          * Output a <style> tag with the given contents for the given media type
 205          * (if any).  TODO: do some useful escaping as well, like if $contents
 206          * contains literal '</style>' (admittedly unlikely).
 207          *
 208          * Note that $contents will not be escaped, since CSS may legitimately
 209          * contain unescaped characters like "<".  Make sure you don't output
 210          * untrusted user input here!
 211          *
 212          * @param $contents string CSS
 213          * @param $media mixed A media type string, like 'screen', or null for all
 214          *   media
 215          * @return string Raw HTML
 216          */
 217         public static function inlineStyle( $contents, $media = null ) {
 218                 global $wgHtml5;
 219
 220                 $attrs = array();
 221                 if ( !$wgHtml5 ) {
 222                         # Technically we should probably add CDATA stuff here like with
 223                         # scripts, but in practice, stylesheets tend not to have
 224                         # problematic characters anyway.
 225                         $attrs['type'] = 'text/css';
 226                 }
 227                 if ( $media !== null ) {
 228                         $attrs['media'] = $media;
 229                 }
 230                 return self::rawElement( 'style', $attrs, $contents );
 231         }
 232
 233         /**
 234          * Output a <link rel=stylesheet> linking to the given URL for the given
 235          * media type (if any).
 236          *
 237          * @param $url string
 238          * @param $media mixed A media type string, like 'screen', or null for all
 239          *   media
 240          * @return string Raw HTML
 241          */
 242         public static function linkedStyle( $url, $media = null ) {
 243                 global $wgHtml5;
 244
 245                 $attrs = array( 'rel' => 'stylesheet', 'href' => $url );
 246                 if ( !$wgHtml5 ) {
 247                         $attrs['type'] = 'text/css';
 248                 }
 249                 if ( $media !== null ) {
 250                         $attrs['media'] = $media;
 251                 }
 252                 return self::element( 'link', $attrs );
 253         }
 254
 255         /**
 256          * Convenience function to produce an <input> element.  This supports the
 257          * new HTML 5 input types and attributes, and will silently strip them if
 258          * $wgHtml5 is false.
 259          *
 260          * @param $name    string name attribute
 261          * @param $value   mixed  value attribute (null = omit)
 262          * @param $type    string type attribute
 263          * @param $attribs array  Assocative array of miscellaneous extra attributes,
 264          *   passed to Html::element()
 265          * @return string Raw HTML
 266          */
 267         public static function input( $name, $value = null, $type = 'text', $attribs = array() ) {
 268                 global $wgHtml5;
 269
 270                 if ( !$wgHtml5 ) {
 271                         if ( !in_array( $type, array( 'hidden', 'text', 'password',
 272                         'checkbox', 'radio', 'file', 'submit', 'image', 'reset', 'button'
 273                         ) ) ) {
 274                                 $type = 'text';
 275                         }
 276                         foreach ( array( 'autocomplete', 'autofocus', 'max', 'min', 'multiple',
 277                         'pattern', 'placeholder', 'required', 'step' ) as $badAttr ) {
 278                                 unset( $attribs[$badAttr] );
 279                         }
 280                 }
 281                 if ( $type != 'text' ) {
 282                         $attribs['type'] = $type;
 283                 }
 284                 if ( $value !== null ) {
 285                         $attribs['value'] = $value;
 286                 }
 287                 $attribs['name'] = $name;
 288
 289                 return self::element( 'input', $attribs );
 290         }
 291 }