includes/libs/CSSMin.php

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