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                 'svg' => 'image/svg+xml',
  57         );
  58
  59         /* Static Methods */
  60
  61         /**
  62          * Gets a list of local file paths which are referenced in a CSS style sheet
  63          *
  64          * This function will always return an empty array if the second parameter is not given or null
  65          * for backwards-compatibility.
  66          *
  67          * @param string $source CSS data to remap
  68          * @param string $path File path where the source was read from (optional)
  69          * @return array List of local file references
  70          */
  71         public static function getLocalFileReferences( $source, $path = null ) {
  72                 if ( $path === null ) {
  73                         return array();
  74                 }
  75
  76                 $path = rtrim( $path, '/' ) . '/';
  77                 $files = array();
  78
  79                 $rFlags = PREG_OFFSET_CAPTURE | PREG_SET_ORDER;
  80                 if ( preg_match_all( '/' . self::URL_REGEX . '/', $source, $matches, $rFlags ) ) {
  81                         foreach ( $matches as $match ) {
  82                                 $url = $match['file'][0];
  83
  84                                 // Skip fully-qualified and protocol-relative URLs and data URIs
  85                                 if ( substr( $url, 0, 2 ) === '//' || parse_url( $url, PHP_URL_SCHEME ) ) {
  86                                         break;
  87                                 }
  88
  89                                 $file = $path . $url;
  90                                 // Skip non-existent files
  91                                 if ( file_exists( $file ) ) {
  92                                         break;
  93                                 }
  94
  95                                 $files[] = $file;
  96                         }
  97                 }
  98                 return $files;
  99         }
 100
 101         /**
 102          * Encode an image file as a base64 data URI.
 103          * If the image file has a suitable MIME type and size, encode it as a
 104          * base64 data URI. Return false if the image type is unfamiliar or exceeds
 105          * the size limit.
 106          *
 107          * @param string $file Image file to encode.
 108          * @param string|null $type File's MIME type or null. If null, CSSMin will
 109          *     try to autodetect the type.
 110          * @param int|bool $sizeLimit If the size of the target file is greater than
 111          *     this value, decline to encode the image file and return false
 112          *     instead. If $sizeLimit is false, no limit is enforced.
 113          * @return string|bool: Image contents encoded as a data URI or false.
 114          */
 115         public static function encodeImageAsDataURI( $file, $type = null,
 116                 $sizeLimit = self::EMBED_SIZE_LIMIT
 117         ) {
 118                 if ( $sizeLimit !== false && filesize( $file ) >= $sizeLimit ) {
 119                         return false;
 120                 }
 121                 if ( $type === null ) {
 122                         $type = self::getMimeType( $file );
 123                 }
 124                 if ( !$type ) {
 125                         return false;
 126                 }
 127                 $data = base64_encode( file_get_contents( $file ) );
 128                 return 'data:' . $type . ';base64,' . $data;
 129         }
 130
 131         /**
 132          * @param $file string
 133          * @return bool|string
 134          */
 135         public static function getMimeType( $file ) {
 136                 $realpath = realpath( $file );
 137                 // Try a couple of different ways to get the mime-type of a file, in order of
 138                 // preference
 139                 if (
 140                         $realpath
 141                         && function_exists( 'finfo_file' )
 142                         && function_exists( 'finfo_open' )
 143                         && defined( 'FILEINFO_MIME_TYPE' )
 144                 ) {
 145                         // As of PHP 5.3, this is how you get the mime-type of a file; it uses the Fileinfo
 146                         // PECL extension
 147                         return finfo_file( finfo_open( FILEINFO_MIME_TYPE ), $realpath );
 148                 } elseif ( function_exists( 'mime_content_type' ) ) {
 149                         // Before this was deprecated in PHP 5.3, this was how you got the mime-type of a file
 150                         return mime_content_type( $file );
 151                 } else {
 152                         // Worst-case scenario has happened, use the file extension to infer the mime-type
 153                         $ext = strtolower( pathinfo( $file, PATHINFO_EXTENSION ) );
 154                         if ( isset( self::$mimeTypes[$ext] ) ) {
 155                                 return self::$mimeTypes[$ext];
 156                         }
 157                 }
 158                 return false;
 159         }
 160
 161         /**
 162          * Build a CSS 'url()' value for the given URL, quoting parentheses (and other funny characters)
 163          * and escaping quotes as necessary.
 164          *
 165          * @param string $url URL to process
 166          * @return string 'url()' value, usually just `"url($url)"`, quoted/escaped if necessary
 167          */
 168         public static function buildUrlValue( $url ) {
 169                 // The list below has been crafted to match URLs such as:
 170                 //   scheme://user@domain:port/~user/fi%20le.png?query=yes&really=y+s
 171                 //   data:image/png;base64,R0lGODlh/+==
 172                 if ( preg_match( '!^[\w\d:@/~.%+;,?&=-]+$!', $url ) ) {
 173                         return "url($url)";
 174                 } else {
 175                         return 'url("' . strtr( $url, array( '\\' => '\\\\', '"' => '\\"' ) ) . '")';
 176                 }
 177         }
 178
 179         /**
 180          * Remaps CSS URL paths and automatically embeds data URIs for CSS rules
 181          * or url() values preceded by an / * @embed * / comment.
 182          *
 183          * @param string $source CSS data to remap
 184          * @param string $local File path where the source was read from
 185          * @param string $remote URL path to the file
 186          * @param bool $embedData If false, never do any data URI embedding,
 187          *   even if / * @embed * / is found.
 188          * @return string Remapped CSS data
 189          */
 190         public static function remap( $source, $local, $remote, $embedData = true ) {
 191                 // High-level overview:
 192                 // * For each CSS rule in $source that includes at least one url() value:
 193                 //   * Check for an @embed comment at the start indicating that all URIs should be embedded
 194                 //   * For each url() value:
 195                 //     * Check for an @embed comment directly preceding the value
 196                 //     * If either @embed comment exists:
 197                 //       * Embedding the URL as data: URI, if it's possible / allowed
 198                 //       * Otherwise remap the URL to work in generated stylesheets
 199
 200                 // Guard against trailing slashes, because "some/remote/../foo.png"
 201                 // resolves to "some/remote/foo.png" on (some?) clients (bug 27052).
 202                 if ( substr( $remote, -1 ) == '/' ) {
 203                         $remote = substr( $remote, 0, -1 );
 204                 }
 205
 206                 // Note: This will correctly handle cases where ';', '{' or '}' appears
 207                 // in a css block comment (/* ... */). All other occurrences, e.g. in
 208                 // quoted strings, will not be handled correctly. You are advised
 209                 // not to use such characters in file names. We also match start/end of
 210                 // the string to be consistent in edge-cases ('@import url(…)').
 211                 $pattern = '/(?:^|[;{])\K(?:\/\*.*?\*\/|[^;{}])*' . CSSMin::URL_REGEX . '(?:\/\*.*?\*\/|[^;}])*(?=[;}]|$)/';
 212
 213                 return preg_replace_callback(
 214                         $pattern,
 215                         function ( $matchOuter ) use ( $local, $remote, $embedData ) {
 216                                 $rule = $matchOuter[0];
 217
 218                                 // Check for global @embed comment and remove it
 219                                 $embedAll = false;
 220                                 $rule = preg_replace( '/^(\s*)' . CSSMin::EMBED_REGEX . '\s*/', '$1', $rule, 1, $embedAll );
 221
 222                                 // Build two versions of current rule: with remapped URLs
 223                                 // and with embedded data: URIs (where possible).
 224                                 $pattern = '/(?P<embed>' . CSSMin::EMBED_REGEX . '\s*|)' . CSSMin::URL_REGEX . '/';
 225
 226                                 $ruleWithRemapped = preg_replace_callback(
 227                                         $pattern,
 228                                         function ( $match ) use ( $local, $remote ) {
 229                                                 $remapped = CSSMin::remapOne( $match['file'], $match['query'], $local, $remote, false );
 230
 231                                                 return CSSMin::buildUrlValue( $remapped );
 232                                         },
 233                                         $rule
 234                                 );
 235
 236                                 if ( $embedData ) {
 237                                         $ruleWithEmbedded = preg_replace_callback(
 238                                                 $pattern,
 239                                                 function ( $match ) use ( $embedAll, $local, $remote ) {
 240                                                         $embed = $embedAll || $match['embed'];
 241                                                         $embedded = CSSMin::remapOne(
 242                                                                 $match['file'],
 243                                                                 $match['query'],
 244                                                                 $local,
 245                                                                 $remote,
 246                                                                 $embed
 247                                                         );
 248
 249                                                         return CSSMin::buildUrlValue( $embedded );
 250                                                 },
 251                                                 $rule
 252                                         );
 253                                 }
 254
 255                                 if ( $embedData && $ruleWithEmbedded !== $ruleWithRemapped ) {
 256                                         // Build 2 CSS properties; one which uses a base64 encoded data URI in place
 257                                         // of the @embed comment to try and retain line-number integrity, and the
 258                                         // other with a remapped an versioned URL and an Internet Explorer hack
 259                                         // making it ignored in all browsers that support data URIs
 260                                         return "$ruleWithEmbedded;$ruleWithRemapped!ie";
 261                                 } else {
 262                                         // No reason to repeat twice
 263                                         return $ruleWithRemapped;
 264                                 }
 265                         }, $source );
 266         }
 267
 268         /**
 269          * Remap or embed a CSS URL path.
 270          *
 271          * @param string $file URL to remap/embed
 272          * @param string $query
 273          * @param string $local File path where the source was read from
 274          * @param string $remote URL path to the file
 275          * @param bool $embed Whether to do any data URI embedding
 276          * @return string Remapped/embedded URL data
 277          */
 278         public static function remapOne( $file, $query, $local, $remote, $embed ) {
 279                 // The full URL possibly with query, as passed to the 'url()' value in CSS
 280                 $url = $file . $query;
 281
 282                 // Skip fully-qualified and protocol-relative URLs and data URIs
 283                 if ( substr( $url, 0, 2 ) === '//' || parse_url( $url, PHP_URL_SCHEME ) ) {
 284                         return $url;
 285                 }
 286
 287                 // URLs with absolute paths like /w/index.php need to be expanded
 288                 // to absolute URLs but otherwise left alone
 289                 if ( $url !== '' && $url[0] === '/' ) {
 290                         // Replace the file path with an expanded (possibly protocol-relative) URL
 291                         // ...but only if wfExpandUrl() is even available.
 292                         // This will not be the case if we're running outside of MW
 293                         if ( function_exists( 'wfExpandUrl' ) ) {
 294                                 return wfExpandUrl( $url, PROTO_RELATIVE );
 295                         } else {
 296                                 return $url;
 297                         }
 298                 }
 299
 300                 if ( $local === false ) {
 301                         // Assume that all paths are relative to $remote, and make them absolute
 302                         return $remote . '/' . $url;
 303                 } else {
 304                         // We drop the query part here and instead make the path relative to $remote
 305                         $url = "{$remote}/{$file}";
 306                         // Path to the actual file on the filesystem
 307                         $localFile = "{$local}/{$file}";
 308                         if ( file_exists( $localFile ) ) {
 309                                 // Add version parameter as a time-stamp in ISO 8601 format,
 310                                 // using Z for the timezone, meaning GMT
 311                                 $url .= '?' . gmdate( 'Y-m-d\TH:i:s\Z', round( filemtime( $localFile ), -2 ) );
 312                                 if ( $embed ) {
 313                                         $data = self::encodeImageAsDataURI( $localFile );
 314                                         if ( $data !== false ) {
 315                                                 return $data;
 316                                         }
 317                                 }
 318                         }
 319                         // If any of these conditions failed (file missing, we don't want to embed it
 320                         // or it's not embeddable), return the URL (possibly with ?timestamp part)
 321                         return $url;
 322                 }
 323         }
 324
 325         /**
 326          * Removes whitespace from CSS data
 327          *
 328          * @param string $css CSS data to minify
 329          * @return string Minified CSS data
 330          */
 331         public static function minify( $css ) {
 332                 return trim(
 333                         str_replace(
 334                                 array( '; ', ': ', ' {', '{ ', ', ', '} ', ';}' ),
 335                                 array( ';', ':', '{', '{', ',', '}', '}' ),
 336                                 preg_replace( array( '/\s+/', '/\/\*.*?\*\//s' ), array( ' ', '' ), $css )
 337                         )
 338                 );
 339         }
 340 }