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         const COMMENT_REGEX = '\/\*.*?\*\/';
  44
  45         /* Protected Static Members */
  46
  47         /** @var array List of common image files extensions and MIME-types */
  48         protected static $mimeTypes = array(
  49                 'gif' => 'image/gif',
  50                 'jpe' => 'image/jpeg',
  51                 'jpeg' => 'image/jpeg',
  52                 'jpg' => 'image/jpeg',
  53                 'png' => 'image/png',
  54                 'tif' => 'image/tiff',
  55                 'tiff' => 'image/tiff',
  56                 'xbm' => 'image/x-xbitmap',
  57                 'svg' => 'image/svg+xml',
  58         );
  59
  60         /* Static Methods */
  61
  62         /**
  63          * Gets a list of local file paths which are referenced in a CSS style sheet
  64          *
  65          * This function will always return an empty array if the second parameter is not given or null
  66          * for backwards-compatibility.
  67          *
  68          * @param string $source CSS data to remap
  69          * @param string $path File path where the source was read from (optional)
  70          * @return array List of local file references
  71          */
  72         public static function getLocalFileReferences( $source, $path = null ) {
  73                 if ( $path === null ) {
  74                         return array();
  75                 }
  76
  77                 $path = rtrim( $path, '/' ) . '/';
  78                 $files = array();
  79
  80                 $rFlags = PREG_OFFSET_CAPTURE | PREG_SET_ORDER;
  81                 if ( preg_match_all( '/' . self::URL_REGEX . '/', $source, $matches, $rFlags ) ) {
  82                         foreach ( $matches as $match ) {
  83                                 $url = $match['file'][0];
  84
  85                                 // Skip fully-qualified and protocol-relative URLs and data URIs
  86                                 if ( substr( $url, 0, 2 ) === '//' || parse_url( $url, PHP_URL_SCHEME ) ) {
  87                                         break;
  88                                 }
  89
  90                                 $file = $path . $url;
  91                                 // Skip non-existent files
  92                                 if ( file_exists( $file ) ) {
  93                                         break;
  94                                 }
  95
  96                                 $files[] = $file;
  97                         }
  98                 }
  99                 return $files;
 100         }
 101
 102         /**
 103          * Encode an image file as a base64 data URI.
 104          * If the image file has a suitable MIME type and size, encode it as a
 105          * base64 data URI. Return false if the image type is unfamiliar or exceeds
 106          * the size limit.
 107          *
 108          * @param string $file Image file to encode.
 109          * @param string|null $type File's MIME type or null. If null, CSSMin will
 110          *     try to autodetect the type.
 111          * @param int|bool $sizeLimit If the size of the target file is greater than
 112          *     this value, decline to encode the image file and return false
 113          *     instead. If $sizeLimit is false, no limit is enforced.
 114          * @return string|bool: Image contents encoded as a data URI or false.
 115          */
 116         public static function encodeImageAsDataURI( $file, $type = null,
 117                 $sizeLimit = self::EMBED_SIZE_LIMIT
 118         ) {
 119                 if ( $sizeLimit !== false && filesize( $file ) >= $sizeLimit ) {
 120                         return false;
 121                 }
 122                 if ( $type === null ) {
 123                         $type = self::getMimeType( $file );
 124                 }
 125                 if ( !$type ) {
 126                         return false;
 127                 }
 128                 $data = base64_encode( file_get_contents( $file ) );
 129                 return 'data:' . $type . ';base64,' . $data;
 130         }
 131
 132         /**
 133          * @param $file string
 134          * @return bool|string
 135          */
 136         public static function getMimeType( $file ) {
 137                 $realpath = realpath( $file );
 138                 if (
 139                         $realpath
 140                         && function_exists( 'finfo_file' )
 141                         && function_exists( 'finfo_open' )
 142                         && defined( 'FILEINFO_MIME_TYPE' )
 143                 ) {
 144                         return finfo_file( finfo_open( FILEINFO_MIME_TYPE ), $realpath );
 145                 }
 146
 147                 // Infer the MIME-type from the file extension
 148                 $ext = strtolower( pathinfo( $file, PATHINFO_EXTENSION ) );
 149                 if ( isset( self::$mimeTypes[$ext] ) ) {
 150                         return self::$mimeTypes[$ext];
 151                 }
 152
 153                 return false;
 154         }
 155
 156         /**
 157          * Build a CSS 'url()' value for the given URL, quoting parentheses (and other funny characters)
 158          * and escaping quotes as necessary.
 159          *
 160          * See http://www.w3.org/TR/css-syntax-3/#consume-a-url-token
 161          *
 162          * @param string $url URL to process
 163          * @return string 'url()' value, usually just `"url($url)"`, quoted/escaped if necessary
 164          */
 165         public static function buildUrlValue( $url ) {
 166                 // The list below has been crafted to match URLs such as:
 167                 //   scheme://user@domain:port/~user/fi%20le.png?query=yes&really=y+s
 168                 //   data:image/png;base64,R0lGODlh/+==
 169                 if ( preg_match( '!^[\w\d:@/~.%+;,?&=-]+$!', $url ) ) {
 170                         return "url($url)";
 171                 } else {
 172                         return 'url("' . strtr( $url, array( '\\' => '\\\\', '"' => '\\"' ) ) . '")';
 173                 }
 174         }
 175
 176         /**
 177          * Remaps CSS URL paths and automatically embeds data URIs for CSS rules
 178          * or url() values preceded by an / * @embed * / comment.
 179          *
 180          * @param string $source CSS data to remap
 181          * @param string $local File path where the source was read from
 182          * @param string $remote URL path to the file
 183          * @param bool $embedData If false, never do any data URI embedding,
 184          *   even if / * @embed * / is found.
 185          * @return string Remapped CSS data
 186          */
 187         public static function remap( $source, $local, $remote, $embedData = true ) {
 188                 // High-level overview:
 189                 // * For each CSS rule in $source that includes at least one url() value:
 190                 //   * Check for an @embed comment at the start indicating that all URIs should be embedded
 191                 //   * For each url() value:
 192                 //     * Check for an @embed comment directly preceding the value
 193                 //     * If either @embed comment exists:
 194                 //       * Embedding the URL as data: URI, if it's possible / allowed
 195                 //       * Otherwise remap the URL to work in generated stylesheets
 196
 197                 // Guard against trailing slashes, because "some/remote/../foo.png"
 198                 // resolves to "some/remote/foo.png" on (some?) clients (bug 27052).
 199                 if ( substr( $remote, -1 ) == '/' ) {
 200                         $remote = substr( $remote, 0, -1 );
 201                 }
 202
 203                 // Replace all comments by a placeholder so they will not interfere with the remapping.
 204                 // Warning: This will also catch on anything looking like the start of a comment between
 205                 // quotation marks (e.g. "foo /* bar").
 206                 $comments = array();
 207                 $placeholder = uniqid( '', true );
 208
 209                 $pattern = '/(?!' . CSSMin::EMBED_REGEX . ')(' . CSSMin::COMMENT_REGEX . ')/s';
 210
 211                 $source = preg_replace_callback(
 212                         $pattern,
 213                         function ( $match ) use ( &$comments, $placeholder ) {
 214                                 $comments[] = $match[ 0 ];
 215                                 return $placeholder . ( count( $comments ) - 1 ) . 'x';
 216                         },
 217                         $source
 218                 );
 219
 220                 // Note: This will not correctly handle cases where ';', '{' or '}'
 221                 // appears in the rule itself, e.g. in a quoted string. You are advised
 222                 // not to use such characters in file names. We also match start/end of
 223                 // the string to be consistent in edge-cases ('@import url(…)').
 224                 $pattern = '/(?:^|[;{])\K[^;{}]*' . CSSMin::URL_REGEX . '[^;}]*(?=[;}]|$)/';
 225
 226                 $source = preg_replace_callback(
 227                         $pattern,
 228                         function ( $matchOuter ) use ( $local, $remote, $embedData, $placeholder ) {
 229                                 $rule = $matchOuter[0];
 230
 231                                 // Check for global @embed comment and remove it. Allow other comments to be present
 232                                 // before @embed (they have been replaced with placeholders at this point).
 233                                 $embedAll = false;
 234                                 $rule = preg_replace( '/^((?:\s+|' . $placeholder . '(\d+)x)*)' . CSSMin::EMBED_REGEX . '\s*/', '$1', $rule, 1, $embedAll );
 235
 236                                 // Build two versions of current rule: with remapped URLs
 237                                 // and with embedded data: URIs (where possible).
 238                                 $pattern = '/(?P<embed>' . CSSMin::EMBED_REGEX . '\s*|)' . CSSMin::URL_REGEX . '/';
 239
 240                                 $ruleWithRemapped = preg_replace_callback(
 241                                         $pattern,
 242                                         function ( $match ) use ( $local, $remote ) {
 243                                                 $remapped = CSSMin::remapOne( $match['file'], $match['query'], $local, $remote, false );
 244
 245                                                 return CSSMin::buildUrlValue( $remapped );
 246                                         },
 247                                         $rule
 248                                 );
 249
 250                                 if ( $embedData ) {
 251                                         $ruleWithEmbedded = preg_replace_callback(
 252                                                 $pattern,
 253                                                 function ( $match ) use ( $embedAll, $local, $remote ) {
 254                                                         $embed = $embedAll || $match['embed'];
 255                                                         $embedded = CSSMin::remapOne(
 256                                                                 $match['file'],
 257                                                                 $match['query'],
 258                                                                 $local,
 259                                                                 $remote,
 260                                                                 $embed
 261                                                         );
 262
 263                                                         return CSSMin::buildUrlValue( $embedded );
 264                                                 },
 265                                                 $rule
 266                                         );
 267                                 }
 268
 269                                 if ( $embedData && $ruleWithEmbedded !== $ruleWithRemapped ) {
 270                                         // Build 2 CSS properties; one which uses a base64 encoded data URI in place
 271                                         // of the @embed comment to try and retain line-number integrity, and the
 272                                         // other with a remapped an versioned URL and an Internet Explorer hack
 273                                         // making it ignored in all browsers that support data URIs
 274                                         return "$ruleWithEmbedded;$ruleWithRemapped!ie";
 275                                 } else {
 276                                         // No reason to repeat twice
 277                                         return $ruleWithRemapped;
 278                                 }
 279                         }, $source );
 280
 281                 // Re-insert comments
 282                 $pattern = '/' . $placeholder . '(\d+)x/';
 283                 $source = preg_replace_callback( $pattern, function( $match ) use ( &$comments ) {
 284                         return $comments[ $match[1] ];
 285                 }, $source );
 286
 287                 return $source;
 288
 289         }
 290
 291         /**
 292          * Remap or embed a CSS URL path.
 293          *
 294          * @param string $file URL to remap/embed
 295          * @param string $query
 296          * @param string $local File path where the source was read from
 297          * @param string $remote URL path to the file
 298          * @param bool $embed Whether to do any data URI embedding
 299          * @return string Remapped/embedded URL data
 300          */
 301         public static function remapOne( $file, $query, $local, $remote, $embed ) {
 302                 // The full URL possibly with query, as passed to the 'url()' value in CSS
 303                 $url = $file . $query;
 304
 305                 // Skip fully-qualified and protocol-relative URLs and data URIs
 306                 if ( substr( $url, 0, 2 ) === '//' || parse_url( $url, PHP_URL_SCHEME ) ) {
 307                         return $url;
 308                 }
 309
 310                 // URLs with absolute paths like /w/index.php need to be expanded
 311                 // to absolute URLs but otherwise left alone
 312                 if ( $url !== '' && $url[0] === '/' ) {
 313                         // Replace the file path with an expanded (possibly protocol-relative) URL
 314                         // ...but only if wfExpandUrl() is even available.
 315                         // This will not be the case if we're running outside of MW
 316                         if ( function_exists( 'wfExpandUrl' ) ) {
 317                                 return wfExpandUrl( $url, PROTO_RELATIVE );
 318                         } else {
 319                                 return $url;
 320                         }
 321                 }
 322
 323                 if ( $local === false ) {
 324                         // Assume that all paths are relative to $remote, and make them absolute
 325                         return $remote . '/' . $url;
 326                 } else {
 327                         // We drop the query part here and instead make the path relative to $remote
 328                         $url = "{$remote}/{$file}";
 329                         // Path to the actual file on the filesystem
 330                         $localFile = "{$local}/{$file}";
 331                         if ( file_exists( $localFile ) ) {
 332                                 // Add version parameter as a time-stamp in ISO 8601 format,
 333                                 // using Z for the timezone, meaning GMT
 334                                 $url .= '?' . gmdate( 'Y-m-d\TH:i:s\Z', round( filemtime( $localFile ), -2 ) );
 335                                 if ( $embed ) {
 336                                         $data = self::encodeImageAsDataURI( $localFile );
 337                                         if ( $data !== false ) {
 338                                                 return $data;
 339                                         }
 340                                 }
 341                         }
 342                         // If any of these conditions failed (file missing, we don't want to embed it
 343                         // or it's not embeddable), return the URL (possibly with ?timestamp part)
 344                         return $url;
 345                 }
 346         }
 347
 348         /**
 349          * Removes whitespace from CSS data
 350          *
 351          * @param string $css CSS data to minify
 352          * @return string Minified CSS data
 353          */
 354         public static function minify( $css ) {
 355                 return trim(
 356                         str_replace(
 357                                 array( '; ', ': ', ' {', '{ ', ', ', '} ', ';}' ),
 358                                 array( ';', ':', '{', '{', ',', '}', '}' ),
 359                                 preg_replace( array( '/\s+/', '/\/\*.*?\*\//s' ), array( ' ', '' ), $css )
 360                         )
 361                 );
 362         }
 363 }