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