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