includes/libs/CSSJanus.php

   1 <?php
   2 /**
   3  * PHP port of CSSJanus.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 /**
  24  * This is a PHP port of CSSJanus, a utility that transforms CSS style sheets
  25  * written for LTR to RTL.
  26  *
  27  * The original Python version of CSSJanus is Copyright 2008 by Google Inc. and
  28  * is distributed under the Apache license. This PHP port is Copyright 2010 by
  29  * Roan Kattouw and is dual-licensed under the GPL (as in the comment above) and
  30  * the Apache (as in the original code) licenses.
  31  *
  32  * Original code: http://code.google.com/p/cssjanus/source/browse/trunk/cssjanus.py
  33  * License of original code: http://code.google.com/p/cssjanus/source/browse/trunk/LICENSE
  34  * @author Roan Kattouw
  35  *
  36  */
  37 class CSSJanus {
  38         // Patterns defined as null are built dynamically by buildPatterns()
  39         private static $patterns = array(
  40                 'tmpToken' => '`TMP`',
  41                 'nonAscii' => '[\200-\377]',
  42                 'unicode' => '(?:(?:\\[0-9a-f]{1,6})(?:\r\n|\s)?)',
  43                 'num' => '(?:[0-9]*\.[0-9]+|[0-9]+)',
  44                 'unit' => '(?:em|ex|px|cm|mm|in|pt|pc|deg|rad|grad|ms|s|hz|khz|%)',
  45                 'body_selector' => 'body\s*{\s*',
  46                 'direction' => 'direction\s*:\s*',
  47                 'escape' => null,
  48                 'nmstart' => null,
  49                 'nmchar' => null,
  50                 'ident' => null,
  51                 'quantity' => null,
  52                 'possibly_negative_quantity' => null,
  53                 'color' => null,
  54                 'url_special_chars' => '[!#$%&*-~]',
  55                 'valid_after_uri_chars' => '[\'\"]?\s*',
  56                 'url_chars' => null,
  57                 'lookahead_not_open_brace' => null,
  58                 'lookahead_not_closing_paren' => null,
  59                 'lookahead_for_closing_paren' => null,
  60                 'lookahead_not_letter' => '(?![a-zA-Z])',
  61                 'lookbehind_not_letter' => '(?<![a-zA-Z])',
  62                 'chars_within_selector' => '[^\}]*?',
  63                 'noflip_annotation' => '\/\*\s*@noflip\s*\*\/',
  64                 'noflip_single' => null,
  65                 'noflip_class' => null,
  66                 'comment' => '/\/\*[^*]*\*+([^\/*][^*]*\*+)*\//',
  67                 'direction_ltr' => null,
  68                 'direction_rtl' => null,
  69                 'left' => null,
  70                 'right' => null,
  71                 'left_in_url' => null,
  72                 'right_in_url' => null,
  73                 'ltr_in_url' => null,
  74                 'rtl_in_url' => null,
  75                 'cursor_east' => null,
  76                 'cursor_west' => null,
  77                 'four_notation_quantity' => null,
  78                 'four_notation_color' => null,
  79                 'border_radius' => null,
  80                 'box_shadow' => null,
  81                 'text_shadow1' => null,
  82                 'text_shadow2' => null,
  83                 'bg_horizontal_percentage' => null,
  84                 'bg_horizontal_percentage_x' => null,
  85         );
  86
  87         /**
  88          * Build patterns we can't define above because they depend on other patterns.
  89          */
  90         private static function buildPatterns() {
  91                 if ( !is_null( self::$patterns['escape'] ) ) {
  92                         // Patterns have already been built
  93                         return;
  94                 }
  95
  96                 $patterns =& self::$patterns;
  97                 $patterns['escape'] = "(?:{$patterns['unicode']}|\\[^\r\n\f0-9a-f])";
  98                 $patterns['nmstart'] = "(?:[_a-z]|{$patterns['nonAscii']}|{$patterns['escape']})";
  99                 $patterns['nmchar'] = "(?:[_a-z0-9-]|{$patterns['nonAscii']}|{$patterns['escape']})";
 100                 $patterns['ident'] = "-?{$patterns['nmstart']}{$patterns['nmchar']}*";
 101                 $patterns['quantity'] = "{$patterns['num']}(?:\s*{$patterns['unit']}|{$patterns['ident']})?";
 102                 $patterns['possibly_negative_quantity'] = "((?:-?{$patterns['quantity']})|(?:inherit|auto))";
 103                 $patterns['color'] = "(#?{$patterns['nmchar']}+|(?:rgba?|hsla?)\([ \d.,%-]+\))";
 104                 $patterns['url_chars'] = "(?:{$patterns['url_special_chars']}|{$patterns['nonAscii']}|{$patterns['escape']})*";
 105                 $patterns['lookahead_not_open_brace'] = "(?!({$patterns['nmchar']}|\r?\n|\s|#|\:|\.|\,|\+|>|\(|\))*?{)";
 106                 $patterns['lookahead_not_closing_paren'] = "(?!{$patterns['url_chars']}?{$patterns['valid_after_uri_chars']}\))";
 107                 $patterns['lookahead_for_closing_paren'] = "(?={$patterns['url_chars']}?{$patterns['valid_after_uri_chars']}\))";
 108                 $patterns['noflip_single'] = "/({$patterns['noflip_annotation']}{$patterns['lookahead_not_open_brace']}[^;}]+;?)/i";
 109                 $patterns['noflip_class'] = "/({$patterns['noflip_annotation']}{$patterns['chars_within_selector']}})/i";
 110                 $patterns['direction_ltr'] = "/({$patterns['direction']})ltr/i";
 111                 $patterns['direction_rtl'] = "/({$patterns['direction']})rtl/i";
 112                 $patterns['left'] = "/{$patterns['lookbehind_not_letter']}(left){$patterns['lookahead_not_letter']}{$patterns['lookahead_not_closing_paren']}{$patterns['lookahead_not_open_brace']}/i";
 113                 $patterns['right'] = "/{$patterns['lookbehind_not_letter']}(right){$patterns['lookahead_not_letter']}{$patterns['lookahead_not_closing_paren']}{$patterns['lookahead_not_open_brace']}/i";
 114                 $patterns['left_in_url'] = "/{$patterns['lookbehind_not_letter']}(left){$patterns['lookahead_for_closing_paren']}/i";
 115                 $patterns['right_in_url'] = "/{$patterns['lookbehind_not_letter']}(right){$patterns['lookahead_for_closing_paren']}/i";
 116                 $patterns['ltr_in_url'] = "/{$patterns['lookbehind_not_letter']}(ltr){$patterns['lookahead_for_closing_paren']}/i";
 117                 $patterns['rtl_in_url'] = "/{$patterns['lookbehind_not_letter']}(rtl){$patterns['lookahead_for_closing_paren']}/i";
 118                 $patterns['cursor_east'] = "/{$patterns['lookbehind_not_letter']}([ns]?)e-resize/";
 119                 $patterns['cursor_west'] = "/{$patterns['lookbehind_not_letter']}([ns]?)w-resize/";
 120                 $patterns['four_notation_quantity'] = "/(:\s*){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s*[;}])/i";
 121                 $patterns['four_notation_color'] = "/(-color\s*:\s*){$patterns['color']}(\s+){$patterns['color']}(\s+){$patterns['color']}(\s+){$patterns['color']}(\s*[;}])/i";
 122                 $patterns['border_radius'] = "/(border-radius\s*:\s*){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s*[;}])/i";
 123                 $patterns['box_shadow'] = "/(box-shadow\s*:\s*(?:inset\s*)?){$patterns['possibly_negative_quantity']}/i";
 124                 $patterns['text_shadow1'] = "/(text-shadow\s*:\s*){$patterns['color']}(\s*){$patterns['possibly_negative_quantity']}/i";
 125                 $patterns['text_shadow2'] = "/(text-shadow\s*:\s*){$patterns['possibly_negative_quantity']}/i";
 126                 // The two regexes below are parenthesized differently then in the original implementation to make the
 127                 // callback's job more straightforward
 128                 $patterns['bg_horizontal_percentage'] = "/(background(?:-position)?\s*:\s*[^%]*?)(-?{$patterns['num']})(%\s*(?:{$patterns['quantity']}|{$patterns['ident']}))/";
 129                 $patterns['bg_horizontal_percentage_x'] = "/(background-position-x\s*:\s*)(-?{$patterns['num']})(%)/";
 130         }
 131
 132         /**
 133          * Transform an LTR stylesheet to RTL
 134          * @param string $css stylesheet to transform
 135          * @param $swapLtrRtlInURL Boolean: If true, swap 'ltr' and 'rtl' in URLs
 136          * @param $swapLeftRightInURL Boolean: If true, swap 'left' and 'right' in URLs
 137          * @return string Transformed stylesheet
 138          */
 139         public static function transform( $css, $swapLtrRtlInURL = false, $swapLeftRightInURL = false ) {
 140                 // We wrap tokens in ` , not ~ like the original implementation does.
 141                 // This was done because ` is not a legal character in CSS and can only
 142                 // occur in URLs, where we escape it to %60 before inserting our tokens.
 143                 $css = str_replace( '`', '%60', $css );
 144
 145                 self::buildPatterns();
 146
 147                 // Tokenize single line rules with /* @noflip */
 148                 $noFlipSingle = new CSSJanus_Tokenizer( self::$patterns['noflip_single'], '`NOFLIP_SINGLE`' );
 149                 $css = $noFlipSingle->tokenize( $css );
 150
 151                 // Tokenize class rules with /* @noflip */
 152                 $noFlipClass = new CSSJanus_Tokenizer( self::$patterns['noflip_class'], '`NOFLIP_CLASS`' );
 153                 $css = $noFlipClass->tokenize( $css );
 154
 155                 // Tokenize comments
 156                 $comments = new CSSJanus_Tokenizer( self::$patterns['comment'], '`C`' );
 157                 $css = $comments->tokenize( $css );
 158
 159                 // LTR->RTL fixes start here
 160                 $css = self::fixDirection( $css );
 161                 if ( $swapLtrRtlInURL ) {
 162                         $css = self::fixLtrRtlInURL( $css );
 163                 }
 164
 165                 if ( $swapLeftRightInURL ) {
 166                         $css = self::fixLeftRightInURL( $css );
 167                 }
 168                 $css = self::fixLeftAndRight( $css );
 169                 $css = self::fixCursorProperties( $css );
 170                 $css = self::fixFourPartNotation( $css );
 171                 $css = self::fixBorderRadius( $css );
 172                 $css = self::fixBackgroundPosition( $css );
 173                 $css = self::fixShadows( $css );
 174
 175                 // Detokenize stuff we tokenized before
 176                 $css = $comments->detokenize( $css );
 177                 $css = $noFlipClass->detokenize( $css );
 178                 $css = $noFlipSingle->detokenize( $css );
 179
 180                 return $css;
 181         }
 182
 183         /**
 184          * Replace direction: ltr; with direction: rtl; and vice versa.
 185          *
 186          * The original implementation only does this inside body selectors
 187          * and misses "body\n{\ndirection:ltr;\n}". This function does not have
 188          * these problems.
 189          *
 190          * See http://code.google.com/p/cssjanus/issues/detail?id=15 and
 191          * TODO: URL
 192          * @param $css string
 193          * @return string
 194          */
 195         private static function fixDirection( $css ) {
 196                 $css = preg_replace( self::$patterns['direction_ltr'],
 197                         '$1' . self::$patterns['tmpToken'], $css );
 198                 $css = preg_replace( self::$patterns['direction_rtl'], '$1ltr', $css );
 199                 $css = str_replace( self::$patterns['tmpToken'], 'rtl', $css );
 200
 201                 return $css;
 202         }
 203
 204         /**
 205          * Replace 'ltr' with 'rtl' and vice versa in background URLs
 206          * @param $css string
 207          * @return string
 208          */
 209         private static function fixLtrRtlInURL( $css ) {
 210                 $css = preg_replace( self::$patterns['ltr_in_url'], self::$patterns['tmpToken'], $css );
 211                 $css = preg_replace( self::$patterns['rtl_in_url'], 'ltr', $css );
 212                 $css = str_replace( self::$patterns['tmpToken'], 'rtl', $css );
 213
 214                 return $css;
 215         }
 216
 217         /**
 218          * Replace 'left' with 'right' and vice versa in background URLs
 219          * @param $css string
 220          * @return string
 221          */
 222         private static function fixLeftRightInURL( $css ) {
 223                 $css = preg_replace( self::$patterns['left_in_url'], self::$patterns['tmpToken'], $css );
 224                 $css = preg_replace( self::$patterns['right_in_url'], 'left', $css );
 225                 $css = str_replace( self::$patterns['tmpToken'], 'right', $css );
 226
 227                 return $css;
 228         }
 229
 230         /**
 231          * Flip rules like left: , padding-right: , etc.
 232          * @param $css string
 233          * @return string
 234          */
 235         private static function fixLeftAndRight( $css ) {
 236                 $css = preg_replace( self::$patterns['left'], self::$patterns['tmpToken'], $css );
 237                 $css = preg_replace( self::$patterns['right'], 'left', $css );
 238                 $css = str_replace( self::$patterns['tmpToken'], 'right', $css );
 239
 240                 return $css;
 241         }
 242
 243         /**
 244          * Flip East and West in rules like cursor: nw-resize;
 245          * @param $css string
 246          * @return string
 247          */
 248         private static function fixCursorProperties( $css ) {
 249                 $css = preg_replace( self::$patterns['cursor_east'],
 250                         '$1' . self::$patterns['tmpToken'], $css );
 251                 $css = preg_replace( self::$patterns['cursor_west'], '$1e-resize', $css );
 252                 $css = str_replace( self::$patterns['tmpToken'], 'w-resize', $css );
 253
 254                 return $css;
 255         }
 256
 257         /**
 258          * Swap the second and fourth parts in four-part notation rules like
 259          * padding: 1px 2px 3px 4px;
 260          *
 261          * Unlike the original implementation, this function doesn't suffer from
 262          * the bug where whitespace is not preserved when flipping four-part rules
 263          * and four-part color rules with multiple whitespace characters between
 264          * colors are not recognized.
 265          * See http://code.google.com/p/cssjanus/issues/detail?id=16
 266          * @param $css string
 267          * @return string
 268          */
 269         private static function fixFourPartNotation( $css ) {
 270                 $css = preg_replace( self::$patterns['four_notation_quantity'], '$1$2$3$8$5$6$7$4$9', $css );
 271                 $css = preg_replace( self::$patterns['four_notation_color'], '$1$2$3$8$5$6$7$4$9', $css );
 272                 return $css;
 273         }
 274
 275         /**
 276          * Swaps appropriate corners in four-part border-radius rules.
 277          * Needs to undo the effect of fixFourPartNotation() on those rules, too.
 278          *
 279          * @param $css string
 280          * @return string
 281          */
 282         private static function fixBorderRadius( $css ) {
 283                 // Undo four_notation_quantity
 284                 $css = preg_replace( self::$patterns['border_radius'], '$1$2$3$8$5$6$7$4$9', $css );
 285                 // Do the real thing
 286                 $css = preg_replace( self::$patterns['border_radius'], '$1$4$3$2$5$8$7$6$9', $css );
 287
 288                 return $css;
 289         }
 290
 291         /**
 292          * Negates horizontal offset in box-shadow and text-shadow rules.
 293          *
 294          * @param $css string
 295          * @return string
 296          */
 297         private static function fixShadows( $css ) {
 298                 // Flips the sign of a CSS value, possibly with a unit.
 299                 // (We can't just negate the value with unary minus due to the units.)
 300                 $flipSign = function ( $cssValue ) {
 301                         // Don't mangle zeroes
 302                         if ( intval( $cssValue ) === 0 ) {
 303                                 return $cssValue;
 304                         } elseif ( $cssValue[0] === '-' ) {
 305                                 return substr( $cssValue, 1 );
 306                         } else {
 307                                 return "-" . $cssValue;
 308                         }
 309                 };
 310
 311                 $css = preg_replace_callback( self::$patterns['box_shadow'], function ( $matches ) use ( $flipSign ) {
 312                         return $matches[1] . $flipSign( $matches[2] );
 313                 }, $css );
 314
 315                 $css = preg_replace_callback( self::$patterns['text_shadow1'], function ( $matches ) use ( $flipSign ) {
 316                         return $matches[1] . $matches[2] . $matches[3] . $flipSign( $matches[4] );
 317                 }, $css );
 318
 319                 $css = preg_replace_callback( self::$patterns['text_shadow2'], function ( $matches ) use ( $flipSign ) {
 320                         return $matches[1] . $flipSign( $matches[2] );
 321                 }, $css );
 322
 323                 return $css;
 324         }
 325
 326         /**
 327          * Flip horizontal background percentages.
 328          * @param $css string
 329          * @return string
 330          */
 331         private static function fixBackgroundPosition( $css ) {
 332                 $replaced = preg_replace_callback( self::$patterns['bg_horizontal_percentage'],
 333                         array( 'self', 'calculateNewBackgroundPosition' ), $css );
 334                 if ( $replaced !== null ) {
 335                         // Check for null; sometimes preg_replace_callback() returns null here for some weird reason
 336                         $css = $replaced;
 337                 }
 338                 $replaced = preg_replace_callback( self::$patterns['bg_horizontal_percentage_x'],
 339                         array( 'self', 'calculateNewBackgroundPosition' ), $css );
 340                 if ( $replaced !== null ) {
 341                         $css = $replaced;
 342                 }
 343
 344                 return $css;
 345         }
 346
 347         /**
 348          * Callback for calculateNewBackgroundPosition()
 349          * @param $matches array
 350          * @return string
 351          */
 352         private static function calculateNewBackgroundPosition( $matches ) {
 353                 return $matches[1] . ( 100 - $matches[2] ) . $matches[3];
 354         }
 355 }
 356
 357 /**
 358  * Utility class used by CSSJanus that tokenizes and untokenizes things we want
 359  * to protect from being janused.
 360  * @author Roan Kattouw
 361  */
 362 class CSSJanus_Tokenizer {
 363         private $regex, $token;
 364         private $originals;
 365
 366         /**
 367          * Constructor
 368          * @param string $regex Regular expression whose matches to replace by a token.
 369          * @param string $token Token
 370          */
 371         public function __construct( $regex, $token ) {
 372                 $this->regex = $regex;
 373                 $this->token = $token;
 374                 $this->originals = array();
 375         }
 376
 377         /**
 378          * Replace all occurrences of $regex in $str with a token and remember
 379          * the original strings.
 380          * @param string $str to tokenize
 381          * @return string Tokenized string
 382          */
 383         public function tokenize( $str ) {
 384                 return preg_replace_callback( $this->regex, array( $this, 'tokenizeCallback' ), $str );
 385         }
 386
 387         /**
 388          * @param $matches array
 389          * @return string
 390          */
 391         private function tokenizeCallback( $matches ) {
 392                 $this->originals[] = $matches[0];
 393                 return $this->token;
 394         }
 395
 396         /**
 397          * Replace tokens with their originals. If multiple strings were tokenized, it's important they be
 398          * detokenized in exactly the SAME ORDER.
 399          * @param string $str previously run through tokenize()
 400          * @return string Original string
 401          */
 402         public function detokenize( $str ) {
 403                 // PHP has no function to replace only the first occurrence or to
 404                 // replace occurrences of the same string with different values,
 405                 // so we use preg_replace_callback() even though we don't really need a regex
 406                 return preg_replace_callback( '/' . preg_quote( $this->token, '/' ) . '/',
 407                         array( $this, 'detokenizeCallback' ), $str );
 408         }
 409
 410         /**
 411          * @param $matches
 412          * @return mixed
 413          */
 414         private function detokenizeCallback( $matches ) {
 415                 $retval = current( $this->originals );
 416                 next( $this->originals );
 417
 418                 return $retval;
 419         }
 420 }