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