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 = [
  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          * Get a list of local files referenced in a stylesheet (includes non-existent files).
  64          *
  65          * @param string $source CSS stylesheet source to process
  66          * @param string $path File path where the source was read from
  67          * @return array List of local file references
  68          */
  69         public static function getLocalFileReferences( $source, $path ) {
  70                 $stripped = preg_replace( '/' . self::COMMENT_REGEX . '/s', '', $source );
  71                 $path = rtrim( $path, '/' ) . '/';
  72                 $files = [];
  73
  74                 $rFlags = PREG_OFFSET_CAPTURE | PREG_SET_ORDER;
  75                 if ( preg_match_all( '/' . self::URL_REGEX . '/', $stripped, $matches, $rFlags ) ) {
  76                         foreach ( $matches as $match ) {
  77                                 $url = $match['file'][0];
  78
  79                                 // Skip fully-qualified and protocol-relative URLs and data URIs
  80                                 // Also skips the rare `behavior` property specifying application's default behavior
  81                                 if (
  82                                         substr( $url, 0, 2 ) === '//' ||
  83                                         parse_url( $url, PHP_URL_SCHEME ) ||
  84                                         substr( $url, 0, 9 ) === '#default#'
  85                                 ) {
  86                                         break;
  87                                 }
  88
  89                                 $files[] = $path . $url;
  90                         }
  91                 }
  92                 return $files;
  93         }
  94
  95         /**
  96          * Encode an image file as a data URI.
  97          *
  98          * If the image file has a suitable MIME type and size, encode it as a data URI, base64-encoded
  99          * for binary files or just percent-encoded otherwise. Return false if the image type is
 100          * unfamiliar or file exceeds the size limit.
 101          *
 102          * @param string $file Image file to encode.
 103          * @param string|null $type File's MIME type or null. If null, CSSMin will
 104          *     try to autodetect the type.
 105          * @param bool $ie8Compat By default, a data URI will only be produced if it can be made short
 106          *     enough to fit in Internet Explorer 8 (and earlier) URI length limit (32,768 bytes). Pass
 107          *     `false` to remove this limitation.
 108          * @return string|bool Image contents encoded as a data URI or false.
 109          */
 110         public static function encodeImageAsDataURI( $file, $type = null, $ie8Compat = true ) {
 111                 // Fast-fail for files that definitely exceed the maximum data URI length
 112                 if ( $ie8Compat && filesize( $file ) >= self::DATA_URI_SIZE_LIMIT ) {
 113                         return false;
 114                 }
 115
 116                 if ( $type === null ) {
 117                         $type = self::getMimeType( $file );
 118                 }
 119                 if ( !$type ) {
 120                         return false;
 121                 }
 122
 123                 return self::encodeStringAsDataURI( file_get_contents( $file ), $type, $ie8Compat );
 124         }
 125
 126         /**
 127          * Encode file contents as a data URI with chosen MIME type.
 128          *
 129          * The URI will be base64-encoded for binary files or just percent-encoded otherwise.
 130          *
 131          * @since 1.25
 132          *
 133          * @param string $contents File contents to encode.
 134          * @param string $type File's MIME type.
 135          * @param bool $ie8Compat See encodeImageAsDataURI().
 136          * @return string|bool Image contents encoded as a data URI or false.
 137          */
 138         public static function encodeStringAsDataURI( $contents, $type, $ie8Compat = true ) {
 139                 // Try #1: Non-encoded data URI
 140                 // The regular expression matches ASCII whitespace and printable characters.
 141                 if ( preg_match( '/^[\r\n\t\x20-\x7e]+$/', $contents ) ) {
 142                         // Do not base64-encode non-binary files (sane SVGs).
 143                         // (This often produces longer URLs, but they compress better, yielding a net smaller size.)
 144                         $uri = 'data:' . $type . ',' . rawurlencode( $contents );
 145                         if ( !$ie8Compat || strlen( $uri ) < self::DATA_URI_SIZE_LIMIT ) {
 146                                 return $uri;
 147                         }
 148                 }
 149
 150                 // Try #2: Encoded data URI
 151                 $uri = 'data:' . $type . ';base64,' . base64_encode( $contents );
 152                 if ( !$ie8Compat || strlen( $uri ) < self::DATA_URI_SIZE_LIMIT ) {
 153                         return $uri;
 154                 }
 155
 156                 // A data URI couldn't be produced
 157                 return false;
 158         }
 159
 160         /**
 161          * Serialize a string (escape and quote) for use as a CSS string value.
 162          * http://www.w3.org/TR/2013/WD-cssom-20131205/#serialize-a-string
 163          *
 164          * @param string $value
 165          * @return string
 166          * @throws Exception
 167          */
 168         public static function serializeStringValue( $value ) {
 169                 if ( strstr( $value, "\0" ) ) {
 170                         throw new Exception( "Invalid character in CSS string" );
 171                 }
 172                 $value = strtr( $value, [ '\\' => '\\\\', '"' => '\\"' ] );
 173                 $value = preg_replace_callback( '/[\x01-\x1f\x7f-\x9f]/', function ( $match ) {
 174                         return '\\' . base_convert( ord( $match[0] ), 10, 16 ) . ' ';
 175                 }, $value );
 176                 return '"' . $value . '"';
 177         }
 178
 179         /**
 180          * @param string $file
 181          * @return bool|string
 182          */
 183         public static function getMimeType( $file ) {
 184                 // Infer the MIME-type from the file extension
 185                 $ext = strtolower( pathinfo( $file, PATHINFO_EXTENSION ) );
 186                 if ( isset( self::$mimeTypes[$ext] ) ) {
 187                         return self::$mimeTypes[$ext];
 188                 }
 189
 190                 $realpath = realpath( $file );
 191                 if (
 192                         $realpath
 193                         && function_exists( 'finfo_file' )
 194                         && function_exists( 'finfo_open' )
 195                         && defined( 'FILEINFO_MIME_TYPE' )
 196                 ) {
 197                         return finfo_file( finfo_open( FILEINFO_MIME_TYPE ), $realpath );
 198                 }
 199
 200                 return false;
 201         }
 202
 203         /**
 204          * Build a CSS 'url()' value for the given URL, quoting parentheses (and other funny characters)
 205          * and escaping quotes as necessary.
 206          *
 207          * See http://www.w3.org/TR/css-syntax-3/#consume-a-url-token
 208          *
 209          * @param string $url URL to process
 210          * @return string 'url()' value, usually just `"url($url)"`, quoted/escaped if necessary
 211          */
 212         public static function buildUrlValue( $url ) {
 213                 // The list below has been crafted to match URLs such as:
 214                 //   scheme://user@domain:port/~user/fi%20le.png?query=yes&really=y+s
 215                 //   data:image/png;base64,R0lGODlh/+==
 216                 if ( preg_match( '!^[\w\d:@/~.%+;,?&=-]+$!', $url ) ) {
 217                         return "url($url)";
 218                 } else {
 219                         return 'url("' . strtr( $url, [ '\\' => '\\\\', '"' => '\\"' ] ) . '")';
 220                 }
 221         }
 222
 223         /**
 224          * Remaps CSS URL paths and automatically embeds data URIs for CSS rules
 225          * or url() values preceded by an / * @embed * / comment.
 226          *
 227          * @param string $source CSS data to remap
 228          * @param string $local File path where the source was read from
 229          * @param string $remote URL path to the file
 230          * @param bool $embedData If false, never do any data URI embedding,
 231          *   even if / * @embed * / is found.
 232          * @return string Remapped CSS data
 233          */
 234         public static function remap( $source, $local, $remote, $embedData = true ) {
 235                 // High-level overview:
 236                 // * For each CSS rule in $source that includes at least one url() value:
 237                 //   * Check for an @embed comment at the start indicating that all URIs should be embedded
 238                 //   * For each url() value:
 239                 //     * Check for an @embed comment directly preceding the value
 240                 //     * If either @embed comment exists:
 241                 //       * Embedding the URL as data: URI, if it's possible / allowed
 242                 //       * Otherwise remap the URL to work in generated stylesheets
 243
 244                 // Guard against trailing slashes, because "some/remote/../foo.png"
 245                 // resolves to "some/remote/foo.png" on (some?) clients (T29052).
 246                 if ( substr( $remote, -1 ) == '/' ) {
 247                         $remote = substr( $remote, 0, -1 );
 248                 }
 249
 250                 // Disallow U+007F DELETE, which is illegal anyway, and which
 251                 // we use for comment placeholders.
 252                 $source = str_replace( "\x7f", "?", $source );
 253
 254                 // Replace all comments by a placeholder so they will not interfere with the remapping.
 255                 // Warning: This will also catch on anything looking like the start of a comment between
 256                 // quotation marks (e.g. "foo /* bar").
 257                 $comments = [];
 258
 259                 $pattern = '/(?!' . CSSMin::EMBED_REGEX . ')(' . CSSMin::COMMENT_REGEX . ')/s';
 260
 261                 $source = preg_replace_callback(
 262                         $pattern,
 263                         function ( $match ) use ( &$comments ) {
 264                                 $comments[] = $match[ 0 ];
 265                                 return CSSMin::PLACEHOLDER . ( count( $comments ) - 1 ) . 'x';
 266                         },
 267                         $source
 268                 );
 269
 270                 // Note: This will not correctly handle cases where ';', '{' or '}'
 271                 // appears in the rule itself, e.g. in a quoted string. You are advised
 272                 // not to use such characters in file names. We also match start/end of
 273                 // the string to be consistent in edge-cases ('@import url(…)').
 274                 $pattern = '/(?:^|[;{])\K[^;{}]*' . CSSMin::URL_REGEX . '[^;}]*(?=[;}]|$)/';
 275
 276                 $source = preg_replace_callback(
 277                         $pattern,
 278                         function ( $matchOuter ) use ( $local, $remote, $embedData ) {
 279                                 $rule = $matchOuter[0];
 280
 281                                 // Check for global @embed comment and remove it. Allow other comments to be present
 282                                 // before @embed (they have been replaced with placeholders at this point).
 283                                 $embedAll = false;
 284                                 $rule = preg_replace(
 285                                         '/^((?:\s+|' .
 286                                                 CSSMin::PLACEHOLDER .
 287                                                 '(\d+)x)*)' .
 288                                                 CSSMin::EMBED_REGEX .
 289                                                 '\s*/',
 290                                         '$1',
 291                                         $rule,
 292                                         1,
 293                                         $embedAll
 294                                 );
 295
 296                                 // Build two versions of current rule: with remapped URLs
 297                                 // and with embedded data: URIs (where possible).
 298                                 $pattern = '/(?P<embed>' . CSSMin::EMBED_REGEX . '\s*|)' . CSSMin::URL_REGEX . '/';
 299
 300                                 $ruleWithRemapped = preg_replace_callback(
 301                                         $pattern,
 302                                         function ( $match ) use ( $local, $remote ) {
 303                                                 $remapped = CSSMin::remapOne( $match['file'], $match['query'], $local, $remote, false );
 304
 305                                                 return CSSMin::buildUrlValue( $remapped );
 306                                         },
 307                                         $rule
 308                                 );
 309
 310                                 if ( $embedData ) {
 311                                         // Remember the occurring MIME types to avoid fallbacks when embedding some files.
 312                                         $mimeTypes = [];
 313
 314                                         $ruleWithEmbedded = preg_replace_callback(
 315                                                 $pattern,
 316                                                 function ( $match ) use ( $embedAll, $local, $remote, &$mimeTypes ) {
 317                                                         $embed = $embedAll || $match['embed'];
 318                                                         $embedded = CSSMin::remapOne(
 319                                                                 $match['file'],
 320                                                                 $match['query'],
 321                                                                 $local,
 322                                                                 $remote,
 323                                                                 $embed
 324                                                         );
 325
 326                                                         $url = $match['file'] . $match['query'];
 327                                                         $file = "{$local}/{$match['file']}";
 328                                                         if (
 329                                                                 !self::isRemoteUrl( $url ) && !self::isLocalUrl( $url )
 330                                                                 && file_exists( $file )
 331                                                         ) {
 332                                                                 $mimeTypes[ CSSMin::getMimeType( $file ) ] = true;
 333                                                         }
 334
 335                                                         return CSSMin::buildUrlValue( $embedded );
 336                                                 },
 337                                                 $rule
 338                                         );
 339
 340                                         // Are all referenced images SVGs?
 341                                         $needsEmbedFallback = $mimeTypes !== [ 'image/svg+xml' => true ];
 342                                 }
 343
 344                                 if ( !$embedData || $ruleWithEmbedded === $ruleWithRemapped ) {
 345                                         // We're not embedding anything, or we tried to but the file is not embeddable
 346                                         return $ruleWithRemapped;
 347                                 } elseif ( $embedData && $needsEmbedFallback ) {
 348                                         // Build 2 CSS properties; one which uses a data URI in place of the @embed comment, and
 349                                         // the other with a remapped and versioned URL with an Internet Explorer 6 and 7 hack
 350                                         // making it ignored in all browsers that support data URIs
 351                                         return "$ruleWithEmbedded;$ruleWithRemapped!ie";
 352                                 } else {
 353                                         // Look ma, no fallbacks! This is for files which IE 6 and 7 don't support anyway: SVG.
 354                                         return $ruleWithEmbedded;
 355                                 }
 356                         }, $source );
 357
 358                 // Re-insert comments
 359                 $pattern = '/' . CSSMin::PLACEHOLDER . '(\d+)x/';
 360                 $source = preg_replace_callback( $pattern, function( $match ) use ( &$comments ) {
 361                         return $comments[ $match[1] ];
 362                 }, $source );
 363
 364                 return $source;
 365         }
 366
 367         /**
 368          * Is this CSS rule referencing a remote URL?
 369          *
 370          * @param string $maybeUrl
 371          * @return bool
 372          */
 373         protected static function isRemoteUrl( $maybeUrl ) {
 374                 if ( substr( $maybeUrl, 0, 2 ) === '//' || parse_url( $maybeUrl, PHP_URL_SCHEME ) ) {
 375                         return true;
 376                 }
 377                 return false;
 378         }
 379
 380         /**
 381          * Is this CSS rule referencing a local URL?
 382          *
 383          * @param string $maybeUrl
 384          * @return bool
 385          */
 386         protected static function isLocalUrl( $maybeUrl ) {
 387                 if ( $maybeUrl !== '' && $maybeUrl[0] === '/' && !self::isRemoteUrl( $maybeUrl ) ) {
 388                         return true;
 389                 }
 390                 return false;
 391         }
 392
 393         /**
 394          * Remap or embed a CSS URL path.
 395          *
 396          * @param string $file URL to remap/embed
 397          * @param string $query
 398          * @param string $local File path where the source was read from
 399          * @param string $remote URL path to the file
 400          * @param bool $embed Whether to do any data URI embedding
 401          * @return string Remapped/embedded URL data
 402          */
 403         public static function remapOne( $file, $query, $local, $remote, $embed ) {
 404                 // The full URL possibly with query, as passed to the 'url()' value in CSS
 405                 $url = $file . $query;
 406
 407                 // Expand local URLs with absolute paths like /w/index.php to possibly protocol-relative URL, if
 408                 // wfExpandUrl() is available. (This will not be the case if we're running outside of MW.)
 409                 if ( self::isLocalUrl( $url ) && function_exists( 'wfExpandUrl' ) ) {
 410                         return wfExpandUrl( $url, PROTO_RELATIVE );
 411                 }
 412
 413                 // Pass thru fully-qualified and protocol-relative URLs and data URIs, as well as local URLs if
 414                 // we can't expand them.
 415                 // Also skips the rare `behavior` property specifying application's default behavior
 416                 if (
 417                         self::isRemoteUrl( $url ) ||
 418                         self::isLocalUrl( $url ) ||
 419                         substr( $url, 0, 9 ) === '#default#'
 420                 ) {
 421                         return $url;
 422                 }
 423
 424                 if ( $local === false ) {
 425                         // Assume that all paths are relative to $remote, and make them absolute
 426                         $url = $remote . '/' . $url;
 427                 } else {
 428                         // We drop the query part here and instead make the path relative to $remote
 429                         $url = "{$remote}/{$file}";
 430                         // Path to the actual file on the filesystem
 431                         $localFile = "{$local}/{$file}";
 432                         if ( file_exists( $localFile ) ) {
 433                                 if ( $embed ) {
 434                                         $data = self::encodeImageAsDataURI( $localFile );
 435                                         if ( $data !== false ) {
 436                                                 return $data;
 437                                         }
 438                                 }
 439                                 if ( method_exists( 'OutputPage', 'transformFilePath' ) ) {
 440                                         $url = OutputPage::transformFilePath( $remote, $local, $file );
 441                                 } else {
 442                                         // Add version parameter as the first five hex digits
 443                                         // of the MD5 hash of the file's contents.
 444                                         $url .= '?' . substr( md5_file( $localFile ), 0, 5 );
 445                                 }
 446                         }
 447                         // If any of these conditions failed (file missing, we don't want to embed it
 448                         // or it's not embeddable), return the URL (possibly with ?timestamp part)
 449                 }
 450                 if ( function_exists( 'wfRemoveDotSegments' ) ) {
 451                         $url = wfRemoveDotSegments( $url );
 452                 }
 453                 return $url;
 454         }
 455
 456         /**
 457          * Removes whitespace from CSS data
 458          *
 459          * @param string $css CSS data to minify
 460          * @return string Minified CSS data
 461          */
 462         public static function minify( $css ) {
 463                 return trim(
 464                         str_replace(
 465                                 [ '; ', ': ', ' {', '{ ', ', ', '} ', ';}' ],
 466                                 [ ';', ':', '{', '{', ',', '}', '}' ],
 467                                 preg_replace( [ '/\s+/', '/\/\*.*?\*\//s' ], [ ' ', '' ], $css )
 468                         )
 469                 );
 470         }
 471 }