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