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         /** @var string Strip marker for comments. **/
  35         const PLACEHOLDER = "\x7fPLACEHOLDER\x7f";
  36
  37         /**
  38          * Internet Explorer data URI length limit. See encodeImageAsDataURI().
  39          */
  40         const DATA_URI_SIZE_LIMIT = 32768;
  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 data URI.
 104          *
 105          * If the image file has a suitable MIME type and size, encode it as a data URI, base64-encoded
 106          * for binary files or just percent-encoded otherwise. Return false if the image type is
 107          * unfamiliar or file exceeds the size limit.
 108          *
 109          * @param string $file Image file to encode.
 110          * @param string|null $type File's MIME type or null. If null, CSSMin will
 111          *     try to autodetect the type.
 112          * @param bool $ie8Compat By default, a data URI will only be produced if it can be made short
 113          *     enough to fit in Internet Explorer 8 (and earlier) URI length limit (32,768 bytes). Pass
 114          *     `false` to remove this limitation.
 115          * @return string|bool Image contents encoded as a data URI or false.
 116          */
 117         public static function encodeImageAsDataURI( $file, $type = null, $ie8Compat = true ) {
 118                 // Fast-fail for files that definitely exceed the maximum data URI length
 119                 if ( $ie8Compat && filesize( $file ) >= self::DATA_URI_SIZE_LIMIT ) {
 120                         return false;
 121                 }
 122
 123                 if ( $type === null ) {
 124                         $type = self::getMimeType( $file );
 125                 }
 126                 if ( !$type ) {
 127                         return false;
 128                 }
 129
 130                 return self::encodeStringAsDataURI( file_get_contents( $file ), $type, $ie8Compat );
 131         }
 132
 133         /**
 134          * Encode file contents as a data URI with chosen MIME type.
 135          *
 136          * The URI will be base64-encoded for binary files or just percent-encoded otherwise.
 137          *
 138          * @since 1.25
 139          *
 140          * @param string $contents File contents to encode.
 141          * @param string $type File's MIME type.
 142          * @param bool $ie8Compat See encodeImageAsDataURI().
 143          * @return string|bool Image contents encoded as a data URI or false.
 144          */
 145         public static function encodeStringAsDataURI( $contents, $type, $ie8Compat = true ) {
 146                 // Try #1: Non-encoded data URI
 147                 // The regular expression matches ASCII whitespace and printable characters.
 148                 if ( preg_match( '/^[\r\n\t\x20-\x7e]+$/', $contents ) ) {
 149                         // Do not base64-encode non-binary files (sane SVGs).
 150                         // (This often produces longer URLs, but they compress better, yielding a net smaller size.)
 151                         $uri = 'data:' . $type . ',' . rawurlencode( $contents );
 152                         if ( !$ie8Compat || strlen( $uri ) < self::DATA_URI_SIZE_LIMIT ) {
 153                                 return $uri;
 154                         }
 155                 }
 156
 157                 // Try #2: Encoded data URI
 158                 $uri = 'data:' . $type . ';base64,' . base64_encode( $contents );
 159                 if ( !$ie8Compat || strlen( $uri ) < self::DATA_URI_SIZE_LIMIT ) {
 160                         return $uri;
 161                 }
 162
 163                 // A data URI couldn't be produced
 164                 return false;
 165         }
 166
 167         /**
 168          * @param $file string
 169          * @return bool|string
 170          */
 171         public static function getMimeType( $file ) {
 172                 $realpath = realpath( $file );
 173                 if (
 174                         $realpath
 175                         && function_exists( 'finfo_file' )
 176                         && function_exists( 'finfo_open' )
 177                         && defined( 'FILEINFO_MIME_TYPE' )
 178                 ) {
 179                         return finfo_file( finfo_open( FILEINFO_MIME_TYPE ), $realpath );
 180                 }
 181
 182                 // Infer the MIME-type from the file extension
 183                 $ext = strtolower( pathinfo( $file, PATHINFO_EXTENSION ) );
 184                 if ( isset( self::$mimeTypes[$ext] ) ) {
 185                         return self::$mimeTypes[$ext];
 186                 }
 187
 188                 return false;
 189         }
 190
 191         /**
 192          * Build a CSS 'url()' value for the given URL, quoting parentheses (and other funny characters)
 193          * and escaping quotes as necessary.
 194          *
 195          * See http://www.w3.org/TR/css-syntax-3/#consume-a-url-token
 196          *
 197          * @param string $url URL to process
 198          * @return string 'url()' value, usually just `"url($url)"`, quoted/escaped if necessary
 199          */
 200         public static function buildUrlValue( $url ) {
 201                 // The list below has been crafted to match URLs such as:
 202                 //   scheme://user@domain:port/~user/fi%20le.png?query=yes&really=y+s
 203                 //   data:image/png;base64,R0lGODlh/+==
 204                 if ( preg_match( '!^[\w\d:@/~.%+;,?&=-]+$!', $url ) ) {
 205                         return "url($url)";
 206                 } else {
 207                         return 'url("' . strtr( $url, array( '\\' => '\\\\', '"' => '\\"' ) ) . '")';
 208                 }
 209         }
 210
 211         /**
 212          * Remaps CSS URL paths and automatically embeds data URIs for CSS rules
 213          * or url() values preceded by an / * @embed * / comment.
 214          *
 215          * @param string $source CSS data to remap
 216          * @param string $local File path where the source was read from
 217          * @param string $remote URL path to the file
 218          * @param bool $embedData If false, never do any data URI embedding,
 219          *   even if / * @embed * / is found.
 220          * @return string Remapped CSS data
 221          */
 222         public static function remap( $source, $local, $remote, $embedData = true ) {
 223                 // High-level overview:
 224                 // * For each CSS rule in $source that includes at least one url() value:
 225                 //   * Check for an @embed comment at the start indicating that all URIs should be embedded
 226                 //   * For each url() value:
 227                 //     * Check for an @embed comment directly preceding the value
 228                 //     * If either @embed comment exists:
 229                 //       * Embedding the URL as data: URI, if it's possible / allowed
 230                 //       * Otherwise remap the URL to work in generated stylesheets
 231
 232                 // Guard against trailing slashes, because "some/remote/../foo.png"
 233                 // resolves to "some/remote/foo.png" on (some?) clients (bug 27052).
 234                 if ( substr( $remote, -1 ) == '/' ) {
 235                         $remote = substr( $remote, 0, -1 );
 236                 }
 237
 238                 // Disallow U+007F DELETE, which is illegal anyway, and which
 239                 // we use for comment placeholders.
 240                 $source = str_replace( "\x7f", "?", $source );
 241
 242                 // Replace all comments by a placeholder so they will not interfere with the remapping.
 243                 // Warning: This will also catch on anything looking like the start of a comment between
 244                 // quotation marks (e.g. "foo /* bar").
 245                 $comments = array();
 246
 247                 $pattern = '/(?!' . CSSMin::EMBED_REGEX . ')(' . CSSMin::COMMENT_REGEX . ')/s';
 248
 249                 $source = preg_replace_callback(
 250                         $pattern,
 251                         function ( $match ) use ( &$comments ) {
 252                                 $comments[] = $match[ 0 ];
 253                                 return CSSMin::PLACEHOLDER . ( count( $comments ) - 1 ) . 'x';
 254                         },
 255                         $source
 256                 );
 257
 258                 // Note: This will not correctly handle cases where ';', '{' or '}'
 259                 // appears in the rule itself, e.g. in a quoted string. You are advised
 260                 // not to use such characters in file names. We also match start/end of
 261                 // the string to be consistent in edge-cases ('@import url(…)').
 262                 $pattern = '/(?:^|[;{])\K[^;{}]*' . CSSMin::URL_REGEX . '[^;}]*(?=[;}]|$)/';
 263
 264                 $source = preg_replace_callback(
 265                         $pattern,
 266                         function ( $matchOuter ) use ( $local, $remote, $embedData ) {
 267                                 $rule = $matchOuter[0];
 268
 269                                 // Check for global @embed comment and remove it. Allow other comments to be present
 270                                 // before @embed (they have been replaced with placeholders at this point).
 271                                 $embedAll = false;
 272                                 $rule = preg_replace( '/^((?:\s+|' . CSSMin::PLACEHOLDER . '(\d+)x)*)' . CSSMin::EMBED_REGEX . '\s*/', '$1', $rule, 1, $embedAll );
 273
 274                                 // Build two versions of current rule: with remapped URLs
 275                                 // and with embedded data: URIs (where possible).
 276                                 $pattern = '/(?P<embed>' . CSSMin::EMBED_REGEX . '\s*|)' . CSSMin::URL_REGEX . '/';
 277
 278                                 $ruleWithRemapped = preg_replace_callback(
 279                                         $pattern,
 280                                         function ( $match ) use ( $local, $remote ) {
 281                                                 $remapped = CSSMin::remapOne( $match['file'], $match['query'], $local, $remote, false );
 282
 283                                                 return CSSMin::buildUrlValue( $remapped );
 284                                         },
 285                                         $rule
 286                                 );
 287
 288                                 if ( $embedData ) {
 289                                         // Remember the occurring MIME types to avoid fallbacks when embedding some files.
 290                                         $mimeTypes = array();
 291
 292                                         $ruleWithEmbedded = preg_replace_callback(
 293                                                 $pattern,
 294                                                 function ( $match ) use ( $embedAll, $local, $remote, &$mimeTypes ) {
 295                                                         $embed = $embedAll || $match['embed'];
 296                                                         $embedded = CSSMin::remapOne(
 297                                                                 $match['file'],
 298                                                                 $match['query'],
 299                                                                 $local,
 300                                                                 $remote,
 301                                                                 $embed
 302                                                         );
 303
 304                                                         $url = $match['file'] . $match['query'];
 305                                                         $file = $local . $match['file'];
 306                                                         if (
 307                                                                 !CSSMin::isRemoteUrl( $url ) && !CSSMin::isLocalUrl( $url )
 308                                                                 && file_exists( $file )
 309                                                         ) {
 310                                                                 $mimeTypes[ CSSMin::getMimeType( $file ) ] = true;
 311                                                         }
 312
 313                                                         return CSSMin::buildUrlValue( $embedded );
 314                                                 },
 315                                                 $rule
 316                                         );
 317
 318                                         // Are all referenced images SVGs?
 319                                         $needsEmbedFallback = $mimeTypes !== array( 'image/svg+xml' => true );
 320                                 }
 321
 322                                 if ( !$embedData || $ruleWithEmbedded === $ruleWithRemapped ) {
 323                                         // We're not embedding anything, or we tried to but the file is not embeddable
 324                                         return $ruleWithRemapped;
 325                                 } elseif ( $embedData && $needsEmbedFallback ) {
 326                                         // Build 2 CSS properties; one which uses a data URI in place of the @embed comment, and
 327                                         // the other with a remapped and versioned URL with an Internet Explorer 6 and 7 hack
 328                                         // making it ignored in all browsers that support data URIs
 329                                         return "$ruleWithEmbedded;$ruleWithRemapped!ie";
 330                                 } else {
 331                                         // Look ma, no fallbacks! This is for files which IE 6 and 7 don't support anyway: SVG.
 332                                         return $ruleWithEmbedded;
 333                                 }
 334                         }, $source );
 335
 336                 // Re-insert comments
 337                 $pattern = '/' . CSSMin::PLACEHOLDER . '(\d+)x/';
 338                 $source = preg_replace_callback( $pattern, function( $match ) use ( &$comments ) {
 339                         return $comments[ $match[1] ];
 340                 }, $source );
 341
 342                 return $source;
 343
 344         }
 345
 346         /**
 347          * Is this CSS rule referencing a remote URL?
 348          *
 349          * @private Until we require PHP 5.5 and we can access self:: from closures.
 350          * @param string $maybeUrl
 351          * @return bool
 352          */
 353         public static function isRemoteUrl( $maybeUrl ) {
 354                 if ( substr( $maybeUrl, 0, 2 ) === '//' || parse_url( $maybeUrl, PHP_URL_SCHEME ) ) {
 355                         return true;
 356                 }
 357                 return false;
 358         }
 359
 360         /**
 361          * Is this CSS rule referencing a local URL?
 362          *
 363          * @private Until we require PHP 5.5 and we can access self:: from closures.
 364          * @param string $maybeUrl
 365          * @return bool
 366          */
 367         public static function isLocalUrl( $maybeUrl ) {
 368                 if ( $maybeUrl !== '' && $maybeUrl[0] === '/' && !self::isRemoteUrl( $maybeUrl ) ) {
 369                         return true;
 370                 }
 371                 return false;
 372         }
 373
 374         /**
 375          * Remap or embed a CSS URL path.
 376          *
 377          * @param string $file URL to remap/embed
 378          * @param string $query
 379          * @param string $local File path where the source was read from
 380          * @param string $remote URL path to the file
 381          * @param bool $embed Whether to do any data URI embedding
 382          * @return string Remapped/embedded URL data
 383          */
 384         public static function remapOne( $file, $query, $local, $remote, $embed ) {
 385                 // The full URL possibly with query, as passed to the 'url()' value in CSS
 386                 $url = $file . $query;
 387
 388                 // Expand local URLs with absolute paths like /w/index.php to possibly protocol-relative URL, if
 389                 // wfExpandUrl() is available. (This will not be the case if we're running outside of MW.)
 390                 if ( self::isLocalUrl( $url ) && function_exists( 'wfExpandUrl' ) ) {
 391                         return wfExpandUrl( $url, PROTO_RELATIVE );
 392                 }
 393
 394                 // Pass thru fully-qualified and protocol-relative URLs and data URIs, as well as local URLs if
 395                 // we can't expand them.
 396                 if ( self::isRemoteUrl( $url ) || self::isLocalUrl( $url ) ) {
 397                         return $url;
 398                 }
 399
 400                 if ( $local === false ) {
 401                         // Assume that all paths are relative to $remote, and make them absolute
 402                         $url = $remote . '/' . $url;
 403                 } else {
 404                         // We drop the query part here and instead make the path relative to $remote
 405                         $url = "{$remote}/{$file}";
 406                         // Path to the actual file on the filesystem
 407                         $localFile = "{$local}/{$file}";
 408                         if ( file_exists( $localFile ) ) {
 409                                 // Add version parameter as the first five hex digits
 410                                 // of the MD5 hash of the file's contents.
 411                                 $url .= '?' . substr( md5_file( $localFile ), 0, 5 );
 412                                 if ( $embed ) {
 413                                         $data = self::encodeImageAsDataURI( $localFile );
 414                                         if ( $data !== false ) {
 415                                                 return $data;
 416                                         }
 417                                 }
 418                         }
 419                         // If any of these conditions failed (file missing, we don't want to embed it
 420                         // or it's not embeddable), return the URL (possibly with ?timestamp part)
 421                 }
 422                 if ( function_exists( 'wfRemoveDotSegments' ) ) {
 423                         $url = wfRemoveDotSegments( $url );
 424                 }
 425                 return $url;
 426         }
 427
 428         /**
 429          * Removes whitespace from CSS data
 430          *
 431          * @param string $css CSS data to minify
 432          * @return string Minified CSS data
 433          */
 434         public static function minify( $css ) {
 435                 return trim(
 436                         str_replace(
 437                                 array( '; ', ': ', ' {', '{ ', ', ', '} ', ';}' ),
 438                                 array( ';', ':', '{', '{', ',', '}', '}' ),
 439                                 preg_replace( array( '/\s+/', '/\/\*.*?\*\//s' ), array( ' ', '' ), $css )
 440                         )
 441                 );
 442         }
 443 }