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                 'bg_horizontal_percentage' => null,
  81                 'bg_horizontal_percentage_x' => null,
  82         );
  83
  84         /**
  85          * Build patterns we can't define above because they depend on other patterns.
  86          */
  87         private static function buildPatterns() {
  88                 if ( !is_null( self::$patterns['escape'] ) ) {
  89                         // Patterns have already been built
  90                         return;
  91                 }
  92
  93                 $patterns =& self::$patterns;
  94                 $patterns['escape'] = "(?:{$patterns['unicode']}|\\[^\r\n\f0-9a-f])";
  95                 $patterns['nmstart'] = "(?:[_a-z]|{$patterns['nonAscii']}|{$patterns['escape']})";
  96                 $patterns['nmchar'] = "(?:[_a-z0-9-]|{$patterns['nonAscii']}|{$patterns['escape']})";
  97                 $patterns['ident'] = "-?{$patterns['nmstart']}{$patterns['nmchar']}*";
  98                 $patterns['quantity'] = "{$patterns['num']}(?:\s*{$patterns['unit']}|{$patterns['ident']})?";
  99                 $patterns['possibly_negative_quantity'] = "((?:-?{$patterns['quantity']})|(?:inherit|auto))";
 100                 $patterns['color'] = "(#?{$patterns['nmchar']}+|(?:rgba?|hsla?)\([ \d.,%-]+\))";
 101                 $patterns['url_chars'] = "(?:{$patterns['url_special_chars']}|{$patterns['nonAscii']}|{$patterns['escape']})*";
 102                 $patterns['lookahead_not_open_brace'] = "(?!({$patterns['nmchar']}|\r?\n|\s|#|\:|\.|\,|\+|>|\(|\))*?{)";
 103                 $patterns['lookahead_not_closing_paren'] = "(?!{$patterns['url_chars']}?{$patterns['valid_after_uri_chars']}\))";
 104                 $patterns['lookahead_for_closing_paren'] = "(?={$patterns['url_chars']}?{$patterns['valid_after_uri_chars']}\))";
 105                 $patterns['noflip_single'] = "/({$patterns['noflip_annotation']}{$patterns['lookahead_not_open_brace']}[^;}]+;?)/i";
 106                 $patterns['noflip_class'] = "/({$patterns['noflip_annotation']}{$patterns['chars_within_selector']}})/i";
 107                 $patterns['direction_ltr'] = "/({$patterns['direction']})ltr/i";
 108                 $patterns['direction_rtl'] = "/({$patterns['direction']})rtl/i";
 109                 $patterns['left'] = "/{$patterns['lookbehind_not_letter']}(left){$patterns['lookahead_not_letter']}{$patterns['lookahead_not_closing_paren']}{$patterns['lookahead_not_open_brace']}/i";
 110                 $patterns['right'] = "/{$patterns['lookbehind_not_letter']}(right){$patterns['lookahead_not_letter']}{$patterns['lookahead_not_closing_paren']}{$patterns['lookahead_not_open_brace']}/i";
 111                 $patterns['left_in_url'] = "/{$patterns['lookbehind_not_letter']}(left){$patterns['lookahead_for_closing_paren']}/i";
 112                 $patterns['right_in_url'] = "/{$patterns['lookbehind_not_letter']}(right){$patterns['lookahead_for_closing_paren']}/i";
 113                 $patterns['ltr_in_url'] = "/{$patterns['lookbehind_not_letter']}(ltr){$patterns['lookahead_for_closing_paren']}/i";
 114                 $patterns['rtl_in_url'] = "/{$patterns['lookbehind_not_letter']}(rtl){$patterns['lookahead_for_closing_paren']}/i";
 115                 $patterns['cursor_east'] = "/{$patterns['lookbehind_not_letter']}([ns]?)e-resize/";
 116                 $patterns['cursor_west'] = "/{$patterns['lookbehind_not_letter']}([ns]?)w-resize/";
 117                 $patterns['four_notation_quantity'] = "/(:\s*){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s*[;}])/i";
 118                 $patterns['four_notation_color'] = "/(-color\s*:\s*){$patterns['color']}(\s+){$patterns['color']}(\s+){$patterns['color']}(\s+){$patterns['color']}(\s*[;}])/i";
 119                 $patterns['border_radius'] = "/(border-radius\s*:\s*){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s*[;}])/i";
 120                 // The two regexes below are parenthesized differently then in the original implementation to make the
 121                 // callback's job more straightforward
 122                 $patterns['bg_horizontal_percentage'] = "/(background(?:-position)?\s*:\s*[^%]*?)(-?{$patterns['num']})(%\s*(?:{$patterns['quantity']}|{$patterns['ident']}))/";
 123                 $patterns['bg_horizontal_percentage_x'] = "/(background-position-x\s*:\s*)(-?{$patterns['num']})(%)/";
 124         }
 125
 126         /**
 127          * Transform an LTR stylesheet to RTL
 128          * @param string $css stylesheet to transform
 129          * @param $swapLtrRtlInURL Boolean: If true, swap 'ltr' and 'rtl' in URLs
 130          * @param $swapLeftRightInURL Boolean: If true, swap 'left' and 'right' in URLs
 131          * @return string Transformed stylesheet
 132          */
 133         public static function transform( $css, $swapLtrRtlInURL = false, $swapLeftRightInURL = false ) {
 134                 // We wrap tokens in ` , not ~ like the original implementation does.
 135                 // This was done because ` is not a legal character in CSS and can only
 136                 // occur in URLs, where we escape it to %60 before inserting our tokens.
 137                 $css = str_replace( '`', '%60', $css );
 138
 139                 self::buildPatterns();
 140
 141                 // Tokenize single line rules with /* @noflip */
 142                 $noFlipSingle = new CSSJanus_Tokenizer( self::$patterns['noflip_single'], '`NOFLIP_SINGLE`' );
 143                 $css = $noFlipSingle->tokenize( $css );
 144
 145                 // Tokenize class rules with /* @noflip */
 146                 $noFlipClass = new CSSJanus_Tokenizer( self::$patterns['noflip_class'], '`NOFLIP_CLASS`' );
 147                 $css = $noFlipClass->tokenize( $css );
 148
 149                 // Tokenize comments
 150                 $comments = new CSSJanus_Tokenizer( self::$patterns['comment'], '`C`' );
 151                 $css = $comments->tokenize( $css );
 152
 153                 // LTR->RTL fixes start here
 154                 $css = self::fixDirection( $css );
 155                 if ( $swapLtrRtlInURL ) {
 156                         $css = self::fixLtrRtlInURL( $css );
 157                 }
 158
 159                 if ( $swapLeftRightInURL ) {
 160                         $css = self::fixLeftRightInURL( $css );
 161                 }
 162                 $css = self::fixLeftAndRight( $css );
 163                 $css = self::fixCursorProperties( $css );
 164                 $css = self::fixFourPartNotation( $css );
 165                 $css = self::fixBorderRadius( $css );
 166                 $css = self::fixBackgroundPosition( $css );
 167
 168                 // Detokenize stuff we tokenized before
 169                 $css = $comments->detokenize( $css );
 170                 $css = $noFlipClass->detokenize( $css );
 171                 $css = $noFlipSingle->detokenize( $css );
 172
 173                 return $css;
 174         }
 175
 176         /**
 177          * Replace direction: ltr; with direction: rtl; and vice versa.
 178          *
 179          * The original implementation only does this inside body selectors
 180          * and misses "body\n{\ndirection:ltr;\n}". This function does not have
 181          * these problems.
 182          *
 183          * See http://code.google.com/p/cssjanus/issues/detail?id=15 and
 184          * TODO: URL
 185          * @param $css string
 186          * @return string
 187          */
 188         private static function fixDirection( $css ) {
 189                 $css = preg_replace( self::$patterns['direction_ltr'],
 190                         '$1' . self::$patterns['tmpToken'], $css );
 191                 $css = preg_replace( self::$patterns['direction_rtl'], '$1ltr', $css );
 192                 $css = str_replace( self::$patterns['tmpToken'], 'rtl', $css );
 193
 194                 return $css;
 195         }
 196
 197         /**
 198          * Replace 'ltr' with 'rtl' and vice versa in background URLs
 199          * @param $css string
 200          * @return string
 201          */
 202         private static function fixLtrRtlInURL( $css ) {
 203                 $css = preg_replace( self::$patterns['ltr_in_url'], self::$patterns['tmpToken'], $css );
 204                 $css = preg_replace( self::$patterns['rtl_in_url'], 'ltr', $css );
 205                 $css = str_replace( self::$patterns['tmpToken'], 'rtl', $css );
 206
 207                 return $css;
 208         }
 209
 210         /**
 211          * Replace 'left' with 'right' and vice versa in background URLs
 212          * @param $css string
 213          * @return string
 214          */
 215         private static function fixLeftRightInURL( $css ) {
 216                 $css = preg_replace( self::$patterns['left_in_url'], self::$patterns['tmpToken'], $css );
 217                 $css = preg_replace( self::$patterns['right_in_url'], 'left', $css );
 218                 $css = str_replace( self::$patterns['tmpToken'], 'right', $css );
 219
 220                 return $css;
 221         }
 222
 223         /**
 224          * Flip rules like left: , padding-right: , etc.
 225          * @param $css string
 226          * @return string
 227          */
 228         private static function fixLeftAndRight( $css ) {
 229                 $css = preg_replace( self::$patterns['left'], self::$patterns['tmpToken'], $css );
 230                 $css = preg_replace( self::$patterns['right'], 'left', $css );
 231                 $css = str_replace( self::$patterns['tmpToken'], 'right', $css );
 232
 233                 return $css;
 234         }
 235
 236         /**
 237          * Flip East and West in rules like cursor: nw-resize;
 238          * @param $css string
 239          * @return string
 240          */
 241         private static function fixCursorProperties( $css ) {
 242                 $css = preg_replace( self::$patterns['cursor_east'],
 243                         '$1' . self::$patterns['tmpToken'], $css );
 244                 $css = preg_replace( self::$patterns['cursor_west'], '$1e-resize', $css );
 245                 $css = str_replace( self::$patterns['tmpToken'], 'w-resize', $css );
 246
 247                 return $css;
 248         }
 249
 250         /**
 251          * Swap the second and fourth parts in four-part notation rules like
 252          * padding: 1px 2px 3px 4px;
 253          *
 254          * Unlike the original implementation, this function doesn't suffer from
 255          * the bug where whitespace is not preserved when flipping four-part rules
 256          * and four-part color rules with multiple whitespace characters between
 257          * colors are not recognized.
 258          * See http://code.google.com/p/cssjanus/issues/detail?id=16
 259          * @param $css string
 260          * @return string
 261          */
 262         private static function fixFourPartNotation( $css ) {
 263                 $css = preg_replace( self::$patterns['four_notation_quantity'], '$1$2$3$8$5$6$7$4$9', $css );
 264                 $css = preg_replace( self::$patterns['four_notation_color'], '$1$2$3$8$5$6$7$4$9', $css );
 265
 266                 return $css;
 267         }
 268
 269         /**
 270          * Swaps appropriate corners in four-part border-radius rules.
 271          * Needs to undo the effect of fixFourPartNotation() on those rules, too.
 272          *
 273          * @param $css string
 274          * @return string
 275          */
 276         private static function fixBorderRadius( $css ) {
 277                 // Undo four_notation_quantity
 278                 $css = preg_replace( self::$patterns['border_radius'], '$1$2$3$8$5$6$7$4$9', $css );
 279                 // Do the real thing
 280                 $css = preg_replace( self::$patterns['border_radius'], '$1$4$3$2$5$8$7$6$9', $css );
 281
 282                 return $css;
 283         }
 284
 285         /**
 286          * Flip horizontal background percentages.
 287          * @param $css string
 288          * @return string
 289          */
 290         private static function fixBackgroundPosition( $css ) {
 291                 $replaced = preg_replace_callback( self::$patterns['bg_horizontal_percentage'],
 292                         array( 'self', 'calculateNewBackgroundPosition' ), $css );
 293                 if ( $replaced !== null ) {
 294                         // Check for null; sometimes preg_replace_callback() returns null here for some weird reason
 295                         $css = $replaced;
 296                 }
 297                 $replaced = preg_replace_callback( self::$patterns['bg_horizontal_percentage_x'],
 298                         array( 'self', 'calculateNewBackgroundPosition' ), $css );
 299                 if ( $replaced !== null ) {
 300                         $css = $replaced;
 301                 }
 302
 303                 return $css;
 304         }
 305
 306         /**
 307          * Callback for calculateNewBackgroundPosition()
 308          * @param $matches array
 309          * @return string
 310          */
 311         private static function calculateNewBackgroundPosition( $matches ) {
 312                 return $matches[1] . ( 100 - $matches[2] ) . $matches[3];
 313         }
 314 }
 315
 316 /**
 317  * Utility class used by CSSJanus that tokenizes and untokenizes things we want
 318  * to protect from being janused.
 319  * @author Roan Kattouw
 320  */
 321 class CSSJanus_Tokenizer {
 322         private $regex, $token;
 323         private $originals;
 324
 325         /**
 326          * Constructor
 327          * @param string $regex Regular expression whose matches to replace by a token.
 328          * @param string $token Token
 329          */
 330         public function __construct( $regex, $token ) {
 331                 $this->regex = $regex;
 332                 $this->token = $token;
 333                 $this->originals = array();
 334         }
 335
 336         /**
 337          * Replace all occurrences of $regex in $str with a token and remember
 338          * the original strings.
 339          * @param string $str to tokenize
 340          * @return string Tokenized string
 341          */
 342         public function tokenize( $str ) {
 343                 return preg_replace_callback( $this->regex, array( $this, 'tokenizeCallback' ), $str );
 344         }
 345
 346         /**
 347          * @param $matches array
 348          * @return string
 349          */
 350         private function tokenizeCallback( $matches ) {
 351                 $this->originals[] = $matches[0];
 352                 return $this->token;
 353         }
 354
 355         /**
 356          * Replace tokens with their originals. If multiple strings were tokenized, it's important they be
 357          * detokenized in exactly the SAME ORDER.
 358          * @param string $str previously run through tokenize()
 359          * @return string Original string
 360          */
 361         public function detokenize( $str ) {
 362                 // PHP has no function to replace only the first occurrence or to
 363                 // replace occurrences of the same string with different values,
 364                 // so we use preg_replace_callback() even though we don't really need a regex
 365                 return preg_replace_callback( '/' . preg_quote( $this->token, '/' ) . '/',
 366                         array( $this, 'detokenizeCallback' ), $str );
 367         }
 368
 369         /**
 370          * @param $matches
 371          * @return mixed
 372          */
 373         private function detokenizeCallback( $matches ) {
 374                 $retval = current( $this->originals );
 375                 next( $this->originals );
 376
 377                 return $retval;
 378         }
 379 }