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         const EMBED_REGEX = '\/\*\s*\@embed\s*\*\/';
  43
  44         /* Protected Static Members */
  45
  46         /** @var array List of common image files extensions and mime-types */
  47         protected static $mimeTypes = array(
  48                 'gif' => 'image/gif',
  49                 'jpe' => 'image/jpeg',
  50                 'jpeg' => 'image/jpeg',
  51                 'jpg' => 'image/jpeg',
  52                 'png' => 'image/png',
  53                 'tif' => 'image/tiff',
  54                 'tiff' => 'image/tiff',
  55                 'xbm' => 'image/x-xbitmap',
  56         );
  57
  58         /* Static Methods */
  59
  60         /**
  61          * Gets a list of local file paths which are referenced in a CSS style sheet
  62          *
  63          * @param string $source CSS data to remap
  64          * @param string $path File path where the source was read from (optional)
  65          * @return array List of local file references
  66          */
  67         public static function getLocalFileReferences( $source, $path = null ) {
  68                 $files = array();
  69                 $rFlags = PREG_OFFSET_CAPTURE | PREG_SET_ORDER;
  70                 if ( preg_match_all( '/' . self::URL_REGEX . '/', $source, $matches, $rFlags ) ) {
  71                         foreach ( $matches as $match ) {
  72                                 $file = ( isset( $path )
  73                                         ? rtrim( $path, '/' ) . '/'
  74                                         : '' ) . "{$match['file'][0]}";
  75
  76                                 // Only proceed if we can access the file
  77                                 if ( !is_null( $path ) && file_exists( $file ) ) {
  78                                         $files[] = $file;
  79                                 }
  80                         }
  81                 }
  82                 return $files;
  83         }
  84
  85         /**
  86          * Encode an image file as a base64 data URI.
  87          * If the image file has a suitable MIME type and size, encode it as a
  88          * base64 data URI. Return false if the image type is unfamiliar or exceeds
  89          * the size limit.
  90          *
  91          * @param string $file Image file to encode.
  92          * @param string|null $type File's MIME type or null. If null, CSSMin will
  93          *     try to autodetect the type.
  94          * @param int|bool $sizeLimit If the size of the target file is greater than
  95          *     this value, decline to encode the image file and return false
  96          *     instead. If $sizeLimit is false, no limit is enforced.
  97          * @return string|bool: Image contents encoded as a data URI or false.
  98          */
  99         public static function encodeImageAsDataURI( $file, $type = null, $sizeLimit = self::EMBED_SIZE_LIMIT ) {
 100                 if ( $sizeLimit !== false && filesize( $file ) >= $sizeLimit ) {
 101                         return false;
 102                 }
 103                 if ( $type === null ) {
 104                         $type = self::getMimeType( $file );
 105                 }
 106                 if ( !$type ) {
 107                         return false;
 108                 }
 109                 $data = base64_encode( file_get_contents( $file ) );
 110                 return 'data:' . $type . ';base64,' . $data;
 111         }
 112
 113         /**
 114          * @param $file string
 115          * @return bool|string
 116          */
 117         public static function getMimeType( $file ) {
 118                 $realpath = realpath( $file );
 119                 // Try a couple of different ways to get the mime-type of a file, in order of
 120                 // preference
 121                 if (
 122                         $realpath
 123                         && function_exists( 'finfo_file' )
 124                         && function_exists( 'finfo_open' )
 125                         && defined( 'FILEINFO_MIME_TYPE' )
 126                 ) {
 127                         // As of PHP 5.3, this is how you get the mime-type of a file; it uses the Fileinfo
 128                         // PECL extension
 129                         return finfo_file( finfo_open( FILEINFO_MIME_TYPE ), $realpath );
 130                 } elseif ( function_exists( 'mime_content_type' ) ) {
 131                         // Before this was deprecated in PHP 5.3, this was how you got the mime-type of a file
 132                         return mime_content_type( $file );
 133                 } else {
 134                         // Worst-case scenario has happened, use the file extension to infer the mime-type
 135                         $ext = strtolower( pathinfo( $file, PATHINFO_EXTENSION ) );
 136                         if ( isset( self::$mimeTypes[$ext] ) ) {
 137                                 return self::$mimeTypes[$ext];
 138                         }
 139                 }
 140                 return false;
 141         }
 142
 143         /**
 144          * Build a CSS 'url()' value for the given URL, quoting parentheses (and other funny characters)
 145          * and escaping quotes as necessary.
 146          *
 147          * @param string $url URL to process
 148          * @return string 'url()' value, usually just `"url($url)"`, quoted/escaped if necessary
 149          */
 150         public static function buildUrlValue( $url ) {
 151                 // The list below has been crafted to match URLs such as:
 152                 //   scheme://user@domain:port/~user/fi%20le.png?query=yes&really=y+s
 153                 //   data:image/png;base64,R0lGODlh/+==
 154                 if ( preg_match( '!^[\w\d:@/~.%+;,?&=-]+$!', $url ) ) {
 155                         return "url($url)";
 156                 } else {
 157                         return 'url("' . strtr( $url, array( '\\' => '\\\\', '"' => '\\"' ) ) . '")';
 158                 }
 159         }
 160
 161         /**
 162          * Remaps CSS URL paths and automatically embeds data URIs for CSS rules or url() values
 163          * preceded by an / * @embed * / comment.
 164          *
 165          * @param string $source CSS data to remap
 166          * @param string $local File path where the source was read from
 167          * @param string $remote URL path to the file
 168          * @param bool $embedData If false, never do any data URI embedding, even if / * @embed * / is found
 169          * @return string Remapped CSS data
 170          */
 171         public static function remap( $source, $local, $remote, $embedData = true ) {
 172                 // High-level overview:
 173                 // * For each CSS rule in $source that includes at least one url() value:
 174                 //   * Check for an @embed comment at the start indicating that all URIs should be embedded
 175                 //   * For each url() value:
 176                 //     * Check for an @embed comment directly preceding the value
 177                 //     * If either @embed comment exists:
 178                 //       * Embedding the URL as data: URI, if it's possible / allowed
 179                 //       * Otherwise remap the URL to work in generated stylesheets
 180
 181                 // Guard against trailing slashes, because "some/remote/../foo.png"
 182                 // resolves to "some/remote/foo.png" on (some?) clients (bug 27052).
 183                 if ( substr( $remote, -1 ) == '/' ) {
 184                         $remote = substr( $remote, 0, -1 );
 185                 }
 186
 187                 // Note: This will not correctly handle cases where ';', '{' or '}' appears in the rule itself,
 188                 // e.g. in a quoted string. You are advised not to use such characters in file names.
 189                 // We also match start/end of the string to be consistent in edge-cases ('@import url(…)').
 190                 $pattern = '/(?:^|[;{])\K[^;{}]*' . CSSMin::URL_REGEX . '[^;}]*(?=[;}]|$)/';
 191                 return preg_replace_callback( $pattern, function ( $matchOuter ) use ( $local, $remote, $embedData ) {
 192                         $rule = $matchOuter[0];
 193
 194                         // Check for global @embed comment and remove it
 195                         $embedAll = false;
 196                         $rule = preg_replace( '/^(\s*)' . CSSMin::EMBED_REGEX . '\s*/', '$1', $rule, 1, $embedAll );
 197
 198                         // Build two versions of current rule: with remapped URLs and with embedded data: URIs (where possible)
 199                         $pattern = '/(?P<embed>' . CSSMin::EMBED_REGEX . '\s*|)' . CSSMin::URL_REGEX . '/';
 200
 201                         $ruleWithRemapped = preg_replace_callback( $pattern, function ( $match ) use ( $local, $remote ) {
 202                                 $remapped = CSSMin::remapOne( $match['file'], $match['query'], $local, $remote, false );
 203                                 return CSSMin::buildUrlValue( $remapped );
 204                         }, $rule );
 205
 206                         if ( $embedData ) {
 207                                 $ruleWithEmbedded = preg_replace_callback( $pattern, function ( $match ) use ( $embedAll, $local, $remote ) {
 208                                         $embed = $embedAll || $match['embed'];
 209                                         $embedded = CSSMin::remapOne( $match['file'], $match['query'], $local, $remote, $embed );
 210                                         return CSSMin::buildUrlValue( $embedded );
 211                                 }, $rule );
 212                         }
 213
 214                         if ( $embedData && $ruleWithEmbedded !== $ruleWithRemapped ) {
 215                                 // Build 2 CSS properties; one which uses a base64 encoded data URI in place
 216                                 // of the @embed comment to try and retain line-number integrity, and the
 217                                 // other with a remapped an versioned URL and an Internet Explorer hack
 218                                 // making it ignored in all browsers that support data URIs
 219                                 return "$ruleWithEmbedded;$ruleWithRemapped!ie";
 220                         } else {
 221                                 // No reason to repeat twice
 222                                 return $ruleWithRemapped;
 223                         }
 224                 }, $source );
 225
 226                 return $source;
 227         }
 228
 229         /**
 230          * Remap or embed a CSS URL path.
 231          *
 232          * @param string $file URL to remap/embed
 233          * @param string $query
 234          * @param string $local File path where the source was read from
 235          * @param string $remote URL path to the file
 236          * @param bool $embed Whether to do any data URI embedding
 237          * @return string Remapped/embedded URL data
 238          */
 239         public static function remapOne( $file, $query, $local, $remote, $embed ) {
 240                 // The full URL possibly with query, as passed to the 'url()' value in CSS
 241                 $url = $file . $query;
 242
 243                 // Skip fully-qualified and protocol-relative URLs and data URIs
 244                 $urlScheme = substr( $url, 0, 2 ) === '//' || parse_url( $url, PHP_URL_SCHEME );
 245                 if ( $urlScheme ) {
 246                         return $url;
 247                 }
 248
 249                 // URLs with absolute paths like /w/index.php need to be expanded
 250                 // to absolute URLs but otherwise left alone
 251                 if ( $url !== '' && $url[0] === '/' ) {
 252                         // Replace the file path with an expanded (possibly protocol-relative) URL
 253                         // ...but only if wfExpandUrl() is even available.
 254                         // This will not be the case if we're running outside of MW
 255                         if ( function_exists( 'wfExpandUrl' ) ) {
 256                                 return wfExpandUrl( $url, PROTO_RELATIVE );
 257                         } else {
 258                                 return $url;
 259                         }
 260                 }
 261
 262                 if ( $local === false ) {
 263                         // Assume that all paths are relative to $remote, and make them absolute
 264                         return $remote . '/' . $url;
 265                 } else {
 266                         // We drop the query part here and instead make the path relative to $remote
 267                         $url = "{$remote}/{$file}";
 268                         // Path to the actual file on the filesystem
 269                         $localFile = "{$local}/{$file}";
 270                         if ( file_exists( $localFile ) ) {
 271                                 // Add version parameter as a time-stamp in ISO 8601 format,
 272                                 // using Z for the timezone, meaning GMT
 273                                 $url .= '?' . gmdate( 'Y-m-d\TH:i:s\Z', round( filemtime( $localFile ), -2 ) );
 274                                 if ( $embed ) {
 275                                         $data = self::encodeImageAsDataURI( $localFile );
 276                                         if ( $data !== false ) {
 277                                                 return $data;
 278                                         }
 279                                 }
 280                         }
 281                         // If any of these conditions failed (file missing, we don't want to embed it
 282                         // or it's not embeddable), return the URL (possibly with ?timestamp part)
 283                         return $url;
 284                 }
 285         }
 286
 287         /**
 288          * Removes whitespace from CSS data
 289          *
 290          * @param string $css CSS data to minify
 291          * @return string Minified CSS data
 292          */
 293         public static function minify( $css ) {
 294                 return trim(
 295                         str_replace(
 296                                 array( '; ', ': ', ' {', '{ ', ', ', '} ', ';}' ),
 297                                 array( ';', ':', '{', '{', ',', '}', '}' ),
 298                                 preg_replace( array( '/\s+/', '/\/\*.*?\*\//s' ), array( ' ', '' ), $css )
 299                         )
 300                 );
 301         }
 302 }