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 DATA_URI_SIZE_LIMIT = 32768;
  42         const URL_REGEX = 'url\(\s*[\'"]?(?P<file>[^\?\)\'"]*?)(?P<query>\?[^\)\'"]*?|)[\'"]?\s*\)';
  43         const EMBED_REGEX = '\/\*\s*\@embed\s*\*\/';
  44         const COMMENT_REGEX = '\/\*.*?\*\/';
  45
  46         /* Protected Static Members */
  47
  48         /** @var array List of common image files extensions and MIME-types */
  49         protected static $mimeTypes = array(
  50                 'gif' => 'image/gif',
  51                 'jpe' => 'image/jpeg',
  52                 'jpeg' => 'image/jpeg',
  53                 'jpg' => 'image/jpeg',
  54                 'png' => 'image/png',
  55                 'tif' => 'image/tiff',
  56                 'tiff' => 'image/tiff',
  57                 'xbm' => 'image/x-xbitmap',
  58                 'svg' => 'image/svg+xml',
  59         );
  60
  61         /* Static Methods */
  62
  63         /**
  64          * Gets a list of local file paths which are referenced in a CSS style sheet
  65          *
  66          * This function will always return an empty array if the second parameter is not given or null
  67          * for backwards-compatibility.
  68          *
  69          * @param string $source CSS data to remap
  70          * @param string $path File path where the source was read from (optional)
  71          * @return array List of local file references
  72          */
  73         public static function getLocalFileReferences( $source, $path = null ) {
  74                 if ( $path === null ) {
  75                         return array();
  76                 }
  77
  78                 $path = rtrim( $path, '/' ) . '/';
  79                 $files = array();
  80
  81                 $rFlags = PREG_OFFSET_CAPTURE | PREG_SET_ORDER;
  82                 if ( preg_match_all( '/' . self::URL_REGEX . '/', $source, $matches, $rFlags ) ) {
  83                         foreach ( $matches as $match ) {
  84                                 $url = $match['file'][0];
  85
  86                                 // Skip fully-qualified and protocol-relative URLs and data URIs
  87                                 if ( substr( $url, 0, 2 ) === '//' || parse_url( $url, PHP_URL_SCHEME ) ) {
  88                                         break;
  89                                 }
  90
  91                                 $file = $path . $url;
  92                                 // Skip non-existent files
  93                                 if ( file_exists( $file ) ) {
  94                                         break;
  95                                 }
  96
  97                                 $files[] = $file;
  98                         }
  99                 }
 100                 return $files;
 101         }
 102
 103         /**
 104          * Encode an image file as a data URI.
 105          *
 106          * If the image file has a suitable MIME type and size, encode it as a data URI, base64-encoded
 107          * for binary files or just percent-encoded otherwise. Return false if the image type is
 108          * unfamiliar or file exceeds the size limit.
 109          *
 110          * @param string $file Image file to encode.
 111          * @param string|null $type File's MIME type or null. If null, CSSMin will
 112          *     try to autodetect the type.
 113          * @param int|bool $sizeLimit If the size of the target file is greater than
 114          *     this value, decline to encode the image file and return false
 115          *     instead. If $sizeLimit is false, no limit is enforced.
 116          * @return string|bool Image contents encoded as a data URI or false.
 117          */
 118         public static function encodeImageAsDataURI( $file, $type = null,
 119                 $sizeLimit = self::EMBED_SIZE_LIMIT
 120         ) {
 121                 if ( $sizeLimit !== false && filesize( $file ) >= $sizeLimit ) {
 122                         return false;
 123                 }
 124                 if ( $type === null ) {
 125                         $type = self::getMimeType( $file );
 126                 }
 127                 if ( !$type ) {
 128                         return false;
 129                 }
 130
 131                 $contents = file_get_contents( $file );
 132                 // Only whitespace and printable ASCII characters
 133                 $isText = (bool)preg_match( '/^[\r\n\t\x20-\x7e]+$/', $contents );
 134
 135                 if ( $isText ) {
 136                         // Do not base64-encode non-binary files (sane SVGs), unless that'd exceed URI length limit.
 137                         // (This often produces longer URLs, but they compress better, yielding a net smaller size.)
 138                         $uri = 'data:' . $type . ',' . rawurlencode( $contents );
 139                         if ( strlen( $uri ) >= self::DATA_URI_SIZE_LIMIT ) {
 140                                 $uri = 'data:' . $type . ';base64,' . base64_encode( $contents );
 141                         }
 142                 } else {
 143                         $uri = 'data:' . $type . ';base64,' . base64_encode( $contents );
 144                 }
 145
 146                 return $uri;
 147         }
 148
 149         /**
 150          * @param $file string
 151          * @return bool|string
 152          */
 153         public static function getMimeType( $file ) {
 154                 $realpath = realpath( $file );
 155                 if (
 156                         $realpath
 157                         && function_exists( 'finfo_file' )
 158                         && function_exists( 'finfo_open' )
 159                         && defined( 'FILEINFO_MIME_TYPE' )
 160                 ) {
 161                         return finfo_file( finfo_open( FILEINFO_MIME_TYPE ), $realpath );
 162                 }
 163
 164                 // Infer the MIME-type from the file extension
 165                 $ext = strtolower( pathinfo( $file, PATHINFO_EXTENSION ) );
 166                 if ( isset( self::$mimeTypes[$ext] ) ) {
 167                         return self::$mimeTypes[$ext];
 168                 }
 169
 170                 return false;
 171         }
 172
 173         /**
 174          * Build a CSS 'url()' value for the given URL, quoting parentheses (and other funny characters)
 175          * and escaping quotes as necessary.
 176          *
 177          * See http://www.w3.org/TR/css-syntax-3/#consume-a-url-token
 178          *
 179          * @param string $url URL to process
 180          * @return string 'url()' value, usually just `"url($url)"`, quoted/escaped if necessary
 181          */
 182         public static function buildUrlValue( $url ) {
 183                 // The list below has been crafted to match URLs such as:
 184                 //   scheme://user@domain:port/~user/fi%20le.png?query=yes&really=y+s
 185                 //   data:image/png;base64,R0lGODlh/+==
 186                 if ( preg_match( '!^[\w\d:@/~.%+;,?&=-]+$!', $url ) ) {
 187                         return "url($url)";
 188                 } else {
 189                         return 'url("' . strtr( $url, array( '\\' => '\\\\', '"' => '\\"' ) ) . '")';
 190                 }
 191         }
 192
 193         /**
 194          * Remaps CSS URL paths and automatically embeds data URIs for CSS rules
 195          * or url() values preceded by an / * @embed * / comment.
 196          *
 197          * @param string $source CSS data to remap
 198          * @param string $local File path where the source was read from
 199          * @param string $remote URL path to the file
 200          * @param bool $embedData If false, never do any data URI embedding,
 201          *   even if / * @embed * / is found.
 202          * @return string Remapped CSS data
 203          */
 204         public static function remap( $source, $local, $remote, $embedData = true ) {
 205                 // High-level overview:
 206                 // * For each CSS rule in $source that includes at least one url() value:
 207                 //   * Check for an @embed comment at the start indicating that all URIs should be embedded
 208                 //   * For each url() value:
 209                 //     * Check for an @embed comment directly preceding the value
 210                 //     * If either @embed comment exists:
 211                 //       * Embedding the URL as data: URI, if it's possible / allowed
 212                 //       * Otherwise remap the URL to work in generated stylesheets
 213
 214                 // Guard against trailing slashes, because "some/remote/../foo.png"
 215                 // resolves to "some/remote/foo.png" on (some?) clients (bug 27052).
 216                 if ( substr( $remote, -1 ) == '/' ) {
 217                         $remote = substr( $remote, 0, -1 );
 218                 }
 219
 220                 // Replace all comments by a placeholder so they will not interfere with the remapping.
 221                 // Warning: This will also catch on anything looking like the start of a comment between
 222                 // quotation marks (e.g. "foo /* bar").
 223                 $comments = array();
 224                 $placeholder = uniqid( '', true );
 225
 226                 $pattern = '/(?!' . CSSMin::EMBED_REGEX . ')(' . CSSMin::COMMENT_REGEX . ')/s';
 227
 228                 $source = preg_replace_callback(
 229                         $pattern,
 230                         function ( $match ) use ( &$comments, $placeholder ) {
 231                                 $comments[] = $match[ 0 ];
 232                                 return $placeholder . ( count( $comments ) - 1 ) . 'x';
 233                         },
 234                         $source
 235                 );
 236
 237                 // Note: This will not correctly handle cases where ';', '{' or '}'
 238                 // appears in the rule itself, e.g. in a quoted string. You are advised
 239                 // not to use such characters in file names. We also match start/end of
 240                 // the string to be consistent in edge-cases ('@import url(…)').
 241                 $pattern = '/(?:^|[;{])\K[^;{}]*' . CSSMin::URL_REGEX . '[^;}]*(?=[;}]|$)/';
 242
 243                 $source = preg_replace_callback(
 244                         $pattern,
 245                         function ( $matchOuter ) use ( $local, $remote, $embedData, $placeholder ) {
 246                                 $rule = $matchOuter[0];
 247
 248                                 // Check for global @embed comment and remove it. Allow other comments to be present
 249                                 // before @embed (they have been replaced with placeholders at this point).
 250                                 $embedAll = false;
 251                                 $rule = preg_replace( '/^((?:\s+|' . $placeholder . '(\d+)x)*)' . CSSMin::EMBED_REGEX . '\s*/', '$1', $rule, 1, $embedAll );
 252
 253                                 // Build two versions of current rule: with remapped URLs
 254                                 // and with embedded data: URIs (where possible).
 255                                 $pattern = '/(?P<embed>' . CSSMin::EMBED_REGEX . '\s*|)' . CSSMin::URL_REGEX . '/';
 256
 257                                 $ruleWithRemapped = preg_replace_callback(
 258                                         $pattern,
 259                                         function ( $match ) use ( $local, $remote ) {
 260                                                 $remapped = CSSMin::remapOne( $match['file'], $match['query'], $local, $remote, false );
 261
 262                                                 return CSSMin::buildUrlValue( $remapped );
 263                                         },
 264                                         $rule
 265                                 );
 266
 267                                 if ( $embedData ) {
 268                                         $ruleWithEmbedded = preg_replace_callback(
 269                                                 $pattern,
 270                                                 function ( $match ) use ( $embedAll, $local, $remote ) {
 271                                                         $embed = $embedAll || $match['embed'];
 272                                                         $embedded = CSSMin::remapOne(
 273                                                                 $match['file'],
 274                                                                 $match['query'],
 275                                                                 $local,
 276                                                                 $remote,
 277                                                                 $embed
 278                                                         );
 279
 280                                                         return CSSMin::buildUrlValue( $embedded );
 281                                                 },
 282                                                 $rule
 283                                         );
 284                                 }
 285
 286                                 if ( $embedData && $ruleWithEmbedded !== $ruleWithRemapped ) {
 287                                         // Build 2 CSS properties; one which uses a base64 encoded data URI in place
 288                                         // of the @embed comment to try and retain line-number integrity, and the
 289                                         // other with a remapped an versioned URL and an Internet Explorer hack
 290                                         // making it ignored in all browsers that support data URIs
 291                                         return "$ruleWithEmbedded;$ruleWithRemapped!ie";
 292                                 } else {
 293                                         // No reason to repeat twice
 294                                         return $ruleWithRemapped;
 295                                 }
 296                         }, $source );
 297
 298                 // Re-insert comments
 299                 $pattern = '/' . $placeholder . '(\d+)x/';
 300                 $source = preg_replace_callback( $pattern, function( $match ) use ( &$comments ) {
 301                         return $comments[ $match[1] ];
 302                 }, $source );
 303
 304                 return $source;
 305
 306         }
 307
 308         /**
 309          * Remap or embed a CSS URL path.
 310          *
 311          * @param string $file URL to remap/embed
 312          * @param string $query
 313          * @param string $local File path where the source was read from
 314          * @param string $remote URL path to the file
 315          * @param bool $embed Whether to do any data URI embedding
 316          * @return string Remapped/embedded URL data
 317          */
 318         public static function remapOne( $file, $query, $local, $remote, $embed ) {
 319                 // The full URL possibly with query, as passed to the 'url()' value in CSS
 320                 $url = $file . $query;
 321
 322                 // Skip fully-qualified and protocol-relative URLs and data URIs
 323                 if ( substr( $url, 0, 2 ) === '//' || parse_url( $url, PHP_URL_SCHEME ) ) {
 324                         return $url;
 325                 }
 326
 327                 // URLs with absolute paths like /w/index.php need to be expanded
 328                 // to absolute URLs but otherwise left alone
 329                 if ( $url !== '' && $url[0] === '/' ) {
 330                         // Replace the file path with an expanded (possibly protocol-relative) URL
 331                         // ...but only if wfExpandUrl() is even available.
 332                         // This will not be the case if we're running outside of MW
 333                         if ( function_exists( 'wfExpandUrl' ) ) {
 334                                 return wfExpandUrl( $url, PROTO_RELATIVE );
 335                         } else {
 336                                 return $url;
 337                         }
 338                 }
 339
 340                 if ( $local === false ) {
 341                         // Assume that all paths are relative to $remote, and make them absolute
 342                         return $remote . '/' . $url;
 343                 } else {
 344                         // We drop the query part here and instead make the path relative to $remote
 345                         $url = "{$remote}/{$file}";
 346                         // Path to the actual file on the filesystem
 347                         $localFile = "{$local}/{$file}";
 348                         if ( file_exists( $localFile ) ) {
 349                                 // Add version parameter as a time-stamp in ISO 8601 format,
 350                                 // using Z for the timezone, meaning GMT
 351                                 $url .= '?' . gmdate( 'Y-m-d\TH:i:s\Z', round( filemtime( $localFile ), -2 ) );
 352                                 if ( $embed ) {
 353                                         $data = self::encodeImageAsDataURI( $localFile );
 354                                         if ( $data !== false ) {
 355                                                 return $data;
 356                                         }
 357                                 }
 358                         }
 359                         // If any of these conditions failed (file missing, we don't want to embed it
 360                         // or it's not embeddable), return the URL (possibly with ?timestamp part)
 361                         return $url;
 362                 }
 363         }
 364
 365         /**
 366          * Removes whitespace from CSS data
 367          *
 368          * @param string $css CSS data to minify
 369          * @return string Minified CSS data
 370          */
 371         public static function minify( $css ) {
 372                 return trim(
 373                         str_replace(
 374                                 array( '; ', ': ', ' {', '{ ', ', ', '} ', ';}' ),
 375                                 array( ';', ':', '{', '{', ',', '}', '}' ),
 376                                 preg_replace( array( '/\s+/', '/\/\*.*?\*\//s' ), array( ' ', '' ), $css )
 377                         )
 378                 );
 379         }
 380 }