includes/libs/CSSMin.php

   1 <?php
   2 /**
   3  * Minification of CSS stylesheets.
   4  *
   5  * Copyright 2010 Wikimedia Foundation
   6  *
   7  * Licensed under the Apache License, Version 2.0 (the "License"); you may
   8  * not use this file except in compliance with the License.
   9  * You may obtain a copy of the License at
  10  *
  11  *              http://www.apache.org/licenses/LICENSE-2.0
  12  *
  13  * Unless required by applicable law or agreed to in writing, software distributed
  14  * under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS
  15  * OF ANY KIND, either express or implied. See the License for the
  16  * specific language governing permissions and limitations under the License.
  17  *
  18  * @file
  19  * @version 0.1.1 -- 2010-09-11
  20  * @author Trevor Parscal <tparscal@wikimedia.org>
  21  * @copyright Copyright 2010 Wikimedia Foundation
  22  * @license http://www.apache.org/licenses/LICENSE-2.0
  23  */
  24
  25 /**
  26  * Transforms CSS data
  27  *
  28  * This class provides minification, URL remapping, URL extracting, and data-URL embedding.
  29  */
  30 class CSSMin {
  31
  32         /* Constants */
  33
  34         /**
  35          * Maximum file size to still qualify for in-line embedding as a data-URI
  36          *
  37          * 24,576 is used because Internet Explorer has a 32,768 byte limit for data URIs,
  38          * which when base64 encoded will result in a 1/3 increase in size.
  39          */
  40         const EMBED_SIZE_LIMIT = 24576;
  41         const URL_REGEX = 'url\(\s*[\'"]?(?P<file>[^\?\)\'"]*)(?P<query>\??[^\)\'"]*)[\'"]?\s*\)';
  42
  43         /* Protected Static Members */
  44
  45         /** @var array List of common image files extensions and mime-types */
  46         protected static $mimeTypes = array(
  47                 'gif' => 'image/gif',
  48                 'jpe' => 'image/jpeg',
  49                 'jpeg' => 'image/jpeg',
  50                 'jpg' => 'image/jpeg',
  51                 'png' => 'image/png',
  52                 'tif' => 'image/tiff',
  53                 'tiff' => 'image/tiff',
  54                 'xbm' => 'image/x-xbitmap',
  55         );
  56
  57         /* Static Methods */
  58
  59         /**
  60          * Gets a list of local file paths which are referenced in a CSS style sheet
  61          *
  62          * @param string $source CSS data to remap
  63          * @param string $path File path where the source was read from (optional)
  64          * @return array List of local file references
  65          */
  66         public static function getLocalFileReferences( $source, $path = null ) {
  67                 $files = array();
  68                 $rFlags = PREG_OFFSET_CAPTURE | PREG_SET_ORDER;
  69                 if ( preg_match_all( '/' . self::URL_REGEX . '/', $source, $matches, $rFlags ) ) {
  70                         foreach ( $matches as $match ) {
  71                                 $file = ( isset( $path )
  72                                         ? rtrim( $path, '/' ) . '/'
  73                                         : '' ) . "{$match['file'][0]}";
  74
  75                                 // Only proceed if we can access the file
  76                                 if ( !is_null( $path ) && file_exists( $file ) ) {
  77                                         $files[] = $file;
  78                                 }
  79                         }
  80                 }
  81                 return $files;
  82         }
  83
  84         /**
  85          * @param $file string
  86          * @return bool|string
  87          */
  88         protected static function getMimeType( $file ) {
  89                 $realpath = realpath( $file );
  90                 // Try a couple of different ways to get the mime-type of a file, in order of
  91                 // preference
  92                 if (
  93                         $realpath
  94                         && function_exists( 'finfo_file' )
  95                         && function_exists( 'finfo_open' )
  96                         && defined( 'FILEINFO_MIME_TYPE' )
  97                 ) {
  98                         // As of PHP 5.3, this is how you get the mime-type of a file; it uses the Fileinfo
  99                         // PECL extension
 100                         return finfo_file( finfo_open( FILEINFO_MIME_TYPE ), $realpath );
 101                 } elseif ( function_exists( 'mime_content_type' ) ) {
 102                         // Before this was deprecated in PHP 5.3, this was how you got the mime-type of a file
 103                         return mime_content_type( $file );
 104                 } else {
 105                         // Worst-case scenario has happened, use the file extension to infer the mime-type
 106                         $ext = strtolower( pathinfo( $file, PATHINFO_EXTENSION ) );
 107                         if ( isset( self::$mimeTypes[$ext] ) ) {
 108                                 return self::$mimeTypes[$ext];
 109                         }
 110                 }
 111                 return false;
 112         }
 113
 114         /**
 115          * Remaps CSS URL paths and automatically embeds data URIs for URL rules
 116          * preceded by an /* @embed * / comment
 117          *
 118          * @param string $source CSS data to remap
 119          * @param string $local File path where the source was read from
 120          * @param string $remote URL path to the file
 121          * @param bool $embedData If false, never do any data URI embedding, even if / * @embed * / is found
 122          * @return string Remapped CSS data
 123          */
 124         public static function remap( $source, $local, $remote, $embedData = true ) {
 125                 $pattern = '/((?P<embed>\s*\/\*\s*\@embed\s*\*\/)(?P<pre>[^\;\}]*))?' .
 126                         self::URL_REGEX . '(?P<post>[^;]*)[\;]?/';
 127                 $offset = 0;
 128                 while ( preg_match( $pattern, $source, $match, PREG_OFFSET_CAPTURE, $offset ) ) {
 129                         // Skip fully-qualified URLs and data URIs
 130                         $urlScheme = parse_url( $match['file'][0], PHP_URL_SCHEME );
 131                         if ( $urlScheme ) {
 132                                 // Move the offset to the end of the match, leaving it alone
 133                                 $offset = $match[0][1] + strlen( $match[0][0] );
 134                                 continue;
 135                         }
 136                         // URLs with absolute paths like /w/index.php need to be expanded
 137                         // to absolute URLs but otherwise left alone
 138                         if ( $match['file'][0] !== '' && $match['file'][0][0] === '/' ) {
 139                                 // Replace the file path with an expanded (possibly protocol-relative) URL
 140                                 // ...but only if wfExpandUrl() is even available.
 141                                 // This will not be the case if we're running outside of MW
 142                                 $lengthIncrease = 0;
 143                                 if ( function_exists( 'wfExpandUrl' ) ) {
 144                                         $expanded = wfExpandUrl( $match['file'][0], PROTO_RELATIVE );
 145                                         $origLength = strlen( $match['file'][0] );
 146                                         $lengthIncrease = strlen( $expanded ) - $origLength;
 147                                         $source = substr_replace( $source, $expanded,
 148                                                 $match['file'][1], $origLength
 149                                         );
 150                                 }
 151                                 // Move the offset to the end of the match, leaving it alone
 152                                 $offset = $match[0][1] + strlen( $match[0][0] ) + $lengthIncrease;
 153                                 continue;
 154                         }
 155
 156                         // Guard against double slashes, because "some/remote/../foo.png"
 157                         // resolves to "some/remote/foo.png" on (some?) clients (bug 27052).
 158                         if ( substr( $remote, -1 ) == '/' ) {
 159                                 $remote = substr( $remote, 0, -1 );
 160                         }
 161
 162                         // Shortcuts
 163                         $embed = $match['embed'][0];
 164                         $pre = $match['pre'][0];
 165                         $post = $match['post'][0];
 166                         $query = $match['query'][0];
 167                         $url = "{$remote}/{$match['file'][0]}";
 168                         $file = "{$local}/{$match['file'][0]}";
 169
 170                         $replacement = false;
 171
 172                         if ( $local !== false && file_exists( $file ) ) {
 173                                 // Add version parameter as a time-stamp in ISO 8601 format,
 174                                 // using Z for the timezone, meaning GMT
 175                                 $url .= '?' . gmdate( 'Y-m-d\TH:i:s\Z', round( filemtime( $file ), -2 ) );
 176                                 // Embedding requires a bit of extra processing, so let's skip that if we can
 177                                 if ( $embedData && $embed ) {
 178                                         $type = self::getMimeType( $file );
 179                                         // Detect when URLs were preceeded with embed tags, and also verify file size is
 180                                         // below the limit
 181                                         if (
 182                                                 $type
 183                                                 && $match['embed'][1] > 0
 184                                                 && filesize( $file ) < self::EMBED_SIZE_LIMIT
 185                                         ) {
 186                                                 // Strip off any trailing = symbols (makes browsers freak out)
 187                                                 $data = base64_encode( file_get_contents( $file ) );
 188                                                 // Build 2 CSS properties; one which uses a base64 encoded data URI in place
 189                                                 // of the @embed comment to try and retain line-number integrity, and the
 190                                                 // other with a remapped an versioned URL and an Internet Explorer hack
 191                                                 // making it ignored in all browsers that support data URIs
 192                                                 $replacement = "{$pre}url(data:{$type};base64,{$data}){$post};";
 193                                                 $replacement .= "{$pre}url({$url}){$post}!ie;";
 194                                         }
 195                                 }
 196                                 if ( $replacement === false ) {
 197                                         // Assume that all paths are relative to $remote, and make them absolute
 198                                         $replacement = "{$embed}{$pre}url({$url}){$post};";
 199                                 }
 200                         } elseif ( $local === false ) {
 201                                 // Assume that all paths are relative to $remote, and make them absolute
 202                                 $replacement = "{$embed}{$pre}url({$url}{$query}){$post};";
 203                         }
 204                         if ( $replacement !== false ) {
 205                                 // Perform replacement on the source
 206                                 $source = substr_replace(
 207                                         $source, $replacement, $match[0][1], strlen( $match[0][0] )
 208                                 );
 209                                 // Move the offset to the end of the replacement in the source
 210                                 $offset = $match[0][1] + strlen( $replacement );
 211                                 continue;
 212                         }
 213                         // Move the offset to the end of the match, leaving it alone
 214                         $offset = $match[0][1] + strlen( $match[0][0] );
 215                 }
 216                 return $source;
 217         }
 218
 219         /**
 220          * Removes whitespace from CSS data
 221          *
 222          * @param string $css CSS data to minify
 223          * @return string Minified CSS data
 224          */
 225         public static function minify( $css ) {
 226                 return trim(
 227                         str_replace(
 228                                 array( '; ', ': ', ' {', '{ ', ', ', '} ', ';}' ),
 229                                 array( ';', ':', '{', '{', ',', '}', '}' ),
 230                                 preg_replace( array( '/\s+/', '/\/\*.*?\*\//s' ), array( ' ', '' ), $css )
 231                         )
 232                 );
 233         }
 234 }