<?php
/**
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License along
 * with this program; if not, write to the Free Software Foundation, Inc.,
 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
 * http://www.gnu.org/copyleft/gpl.html
 *
 */

/**
 * This is a PHP port of CSSJanus, a utility that transforms CSS style sheets
 * written for LTR to RTL.
 *
 * The original Python version of CSSJanus is Copyright 2008 by Google Inc. and
 * is distributed under the Apache license. This PHP port is Copyright 2010 by
 * Roan Kattouw and is dual-licensed under the GPL (as in the comment above) and
 * the Apache (as in the original code) licenses.
 *
 * Original code: http://code.google.com/p/cssjanus/source/browse/trunk/cssjanus.py
 * License of original code: http://code.google.com/p/cssjanus/source/browse/trunk/LICENSE
 * @author Roan Kattouw
 *
 */
class CSSJanus {
	// Patterns defined as null are built dynamically by buildPatterns()
	private static $patterns = array(
		'tmpToken' => '`TMP`',
		'nonAscii' => '[\200-\377]',
		'unicode' => '(?:(?:\\[0-9a-f]{1,6})(?:\r\n|\s)?)',
		'num' => '(?:[0-9]*\.[0-9]+|[0-9]+)',
		'unit' => '(?:em|ex|px|cm|mm|in|pt|pc|deg|rad|grad|ms|s|hz|khz|%)',
		'body_selector' => 'body\s*{\s*',
		'direction' => 'direction\s*:\s*',
		'escape' => null,
		'nmstart' => null,
		'nmchar' => null,
		'ident' => null,
		'quantity' => null,
		'possibly_negative_quantity' => null,
		'color' => null,
		'url_special_chars' => '[!#$%&*-~]',
		'valid_after_uri_chars' => '[\'\"]?\s*',
		'url_chars' => null,
		'lookahead_not_open_brace' => null,
		'lookahead_not_closing_paren' => null,
		'lookahead_for_closing_paren' => null,
		'lookbehind_not_letter' => '(?<![a-zA-Z])',
		'chars_within_selector' => '[^\}]*?',
		'noflip_annotation' => '\/\*\s*@noflip\s*\*\/',
		'noflip_single' => null,
		'noflip_class' => null,
		'comment' => '/\/\*[^*]*\*+([^\/*][^*]*\*+)*\//',
		'direction_ltr' => null,
		'direction_rtl' => null,
		'left' => null,
		'right' => null,
		'left_in_url' => null,
		'right_in_url' => null,
		'ltr_in_url' => null,
		'rtl_in_url' => null,
		'cursor_east' => null,
		'cursor_west' => null,
		'four_notation_quantity' => null,
		'four_notation_color' => null,
		'bg_horizontal_percentage' => null,
		'bg_horizontal_percentage_x' => null,
	);

	/**
	 * Build patterns we can't define above because they depend on other patterns.
	 */
	private static function buildPatterns() {
		if ( !is_null( self::$patterns['escape'] ) ) {
			// Patterns have already been built
			return;
		}

		$patterns =& self::$patterns;
		$patterns['escape'] = "(?:{$patterns['unicode']}|\\[^\r\n\f0-9a-f])";
		$patterns['nmstart'] = "(?:[_a-z]|{$patterns['nonAscii']}|{$patterns['escape']})";
		$patterns['nmchar'] = "(?:[_a-z0-9-]|{$patterns['nonAscii']}|{$patterns['escape']})";
		$patterns['ident'] = "-?{$patterns['nmstart']}{$patterns['nmchar']}*";
		$patterns['quantity'] = "{$patterns['num']}(?:\s*{$patterns['unit']}|{$patterns['ident']})?";
		$patterns['possibly_negative_quantity'] = "((?:-?{$patterns['quantity']})|(?:inherit|auto))";
		$patterns['color'] = "(#?{$patterns['nmchar']}+)";
		$patterns['url_chars'] = "(?:{$patterns['url_special_chars']}|{$patterns['nonAscii']}|{$patterns['escape']})*";
		$patterns['lookahead_not_open_brace'] = "(?!({$patterns['nmchar']}|\r?\n|\s|#|\:|\.|\,|\+|>)*?{)";
		$patterns['lookahead_not_closing_paren'] = "(?!{$patterns['url_chars']}?{$patterns['valid_after_uri_chars']}\))";
		$patterns['lookahead_for_closing_paren'] = "(?={$patterns['url_chars']}?{$patterns['valid_after_uri_chars']}\))";
		$patterns['noflip_single'] = "/({$patterns['noflip_annotation']}{$patterns['lookahead_not_open_brace']}[^;}]+;?)/i";
		$patterns['noflip_class'] = "/({$patterns['noflip_annotation']}{$patterns['chars_within_selector']}})/i";
		$patterns['direction_ltr'] = "/({$patterns['direction']})ltr/i";
		$patterns['direction_rtl'] = "/({$patterns['direction']})rtl/i";
		$patterns['left'] = "/{$patterns['lookbehind_not_letter']}(left){$patterns['lookahead_not_closing_paren']}{$patterns['lookahead_not_open_brace']}/i";
		$patterns['right'] = "/{$patterns['lookbehind_not_letter']}(right){$patterns['lookahead_not_closing_paren']}{$patterns['lookahead_not_open_brace']}/i";
		$patterns['left_in_url'] = "/{$patterns['lookbehind_not_letter']}(left){$patterns['lookahead_for_closing_paren']}/i";
		$patterns['right_in_url'] = "/{$patterns['lookbehind_not_letter']}(right){$patterns['lookahead_for_closing_paren']}/i";
		$patterns['ltr_in_url'] = "/{$patterns['lookbehind_not_letter']}(ltr){$patterns['lookahead_for_closing_paren']}/i";
		$patterns['rtl_in_url'] = "/{$patterns['lookbehind_not_letter']}(rtl){$patterns['lookahead_for_closing_paren']}/i";
		$patterns['cursor_east'] = "/{$patterns['lookbehind_not_letter']}([ns]?)e-resize/";
		$patterns['cursor_west'] = "/{$patterns['lookbehind_not_letter']}([ns]?)w-resize/";
		$patterns['four_notation_quantity'] = "/{$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}(\s+){$patterns['possibly_negative_quantity']}/i";
		$patterns['four_notation_color'] = "/(-color\s*:\s*){$patterns['color']}(\s+){$patterns['color']}(\s+){$patterns['color']}(\s+){$patterns['color']}/i";
		// The two regexes below are parenthesized differently then in the original implementation to make the
		// callback's job more straightforward
		$patterns['bg_horizontal_percentage'] = "/(background(?:-position)?\s*:\s*[^%]*?)(-?{$patterns['num']})(%\s*(?:{$patterns['quantity']}|{$patterns['ident']}))/";
		$patterns['bg_horizontal_percentage_x'] = "/(background-position-x\s*:\s*)(-?{$patterns['num']})(%)/";
	}

	/**
	 * Transform an LTR stylesheet to RTL
	 * @param $css String: stylesheet to transform
	 * @param $swapLtrRtlInURL Boolean: If true, swap 'ltr' and 'rtl' in URLs
	 * @param $swapLeftRightInURL Boolean: If true, swap 'left' and 'right' in URLs
	 * @return Transformed stylesheet
	 */
	public static function transform( $css, $swapLtrRtlInURL = false, $swapLeftRightInURL = false ) {
		// We wrap tokens in ` , not ~ like the original implementation does.
		// This was done because ` is not a legal character in CSS and can only
		// occur in URLs, where we escape it to %60 before inserting our tokens.
		$css = str_replace( '`', '%60', $css );

		self::buildPatterns();

		// Tokenize single line rules with /* @noflip */
		$noFlipSingle = new CSSJanus_Tokenizer( self::$patterns['noflip_single'], '`NOFLIP_SINGLE`' );
		$css = $noFlipSingle->tokenize( $css );

		// Tokenize class rules with /* @noflip */
		$noFlipClass = new CSSJanus_Tokenizer( self::$patterns['noflip_class'], '`NOFLIP_CLASS`' );
		$css = $noFlipClass->tokenize( $css );

		// Tokenize comments
		$comments = new CSSJanus_Tokenizer( self::$patterns['comment'], '`C`' );
		$css = $comments->tokenize( $css );

		// LTR->RTL fixes start here
		$css = self::fixDirection( $css );
		if ( $swapLtrRtlInURL ) {
			$css = self::fixLtrRtlInURL( $css );
		}

		if ( $swapLeftRightInURL ) {
			$css = self::fixLeftRightInURL( $css );
		}
		$css = self::fixLeftAndRight( $css );
		$css = self::fixCursorProperties( $css );
		$css = self::fixFourPartNotation( $css );
		$css = self::fixBackgroundPosition( $css );

		// Detokenize stuff we tokenized before
		$css = $comments->detokenize( $css );
		$css = $noFlipClass->detokenize( $css );
		$css = $noFlipSingle->detokenize( $css );

		return $css;
	}

	/**
	 * Replace direction: ltr; with direction: rtl; and vice versa.
	 *
	 * The original implementation only does this inside body selectors
	 * and misses "body\n{\ndirection:ltr;\n}". This function does not have
	 * these problems.
	 *
	 * See http://code.google.com/p/cssjanus/issues/detail?id=15 and
	 * TODO: URL
	 * @param $css string
	 * @return string
	 */
	private static function fixDirection( $css ) {
		$css = preg_replace( self::$patterns['direction_ltr'],
			'$1' . self::$patterns['tmpToken'], $css );
		$css = preg_replace( self::$patterns['direction_rtl'], '$1ltr', $css );
		$css = str_replace( self::$patterns['tmpToken'], 'rtl', $css );

		return $css;
	}

	/**
	 * Replace 'ltr' with 'rtl' and vice versa in background URLs
	 * @param $css string
	 * @return string
	 */
	private static function fixLtrRtlInURL( $css ) {
		$css = preg_replace( self::$patterns['ltr_in_url'], self::$patterns['tmpToken'], $css );
		$css = preg_replace( self::$patterns['rtl_in_url'], 'ltr', $css );
		$css = str_replace( self::$patterns['tmpToken'], 'rtl', $css );

		return $css;
	}

	/**
	 * Replace 'left' with 'right' and vice versa in background URLs
	 * @param $css string
	 * @return string
	 */
	private static function fixLeftRightInURL( $css ) {
		$css = preg_replace( self::$patterns['left_in_url'], self::$patterns['tmpToken'], $css );
		$css = preg_replace( self::$patterns['right_in_url'], 'left', $css );
		$css = str_replace( self::$patterns['tmpToken'], 'right', $css );

		return $css;
	}

	/**
	 * Flip rules like left: , padding-right: , etc.
	 * @param $css string
	 * @return string
	 */
	private static function fixLeftAndRight( $css ) {
		$css = preg_replace( self::$patterns['left'], self::$patterns['tmpToken'], $css );
		$css = preg_replace( self::$patterns['right'], 'left', $css );
		$css = str_replace( self::$patterns['tmpToken'], 'right', $css );

		return $css;
	}

	/**
	 * Flip East and West in rules like cursor: nw-resize;
	 * @param $css string
	 * @return string
	 */
	private static function fixCursorProperties( $css ) {
		$css = preg_replace( self::$patterns['cursor_east'],
			'$1' . self::$patterns['tmpToken'], $css );
		$css = preg_replace( self::$patterns['cursor_west'], '$1e-resize', $css );
		$css = str_replace( self::$patterns['tmpToken'], 'w-resize', $css );

		return $css;
	}

	/**
	 * Swap the second and fourth parts in four-part notation rules like
	 * padding: 1px 2px 3px 4px;
	 *
	 * Unlike the original implementation, this function doesn't suffer from
	 * the bug where whitespace is not preserved when flipping four-part rules
	 * and four-part color rules with multiple whitespace characters between
	 * colors are not recognized.
	 * See http://code.google.com/p/cssjanus/issues/detail?id=16
	 * @param $css string
	 * @return string
	 */
	private static function fixFourPartNotation( $css ) {
		$css = preg_replace( self::$patterns['four_notation_quantity'], '$1$2$7$4$5$6$3', $css );
		$css = preg_replace( self::$patterns['four_notation_color'], '$1$2$3$8$5$6$7$4', $css );

		return $css;
	}

	/**
	 * Flip horizontal background percentages.
	 * @param $css string
	 * @return string
	 */
	private static function fixBackgroundPosition( $css ) {
		$css = preg_replace_callback( self::$patterns['bg_horizontal_percentage'],
			array( 'self', 'calculateNewBackgroundPosition' ), $css );
		$css = preg_replace_callback( self::$patterns['bg_horizontal_percentage_x'],
			array( 'self', 'calculateNewBackgroundPosition' ), $css );

		return $css;
	}

	/**
	 * Callback for calculateNewBackgroundPosition()
	 * @param $matches array
	 * @return string
	 */
	private static function calculateNewBackgroundPosition( $matches ) {
		return $matches[1] . ( 100 - $matches[2] ) . $matches[3];
	}
}

/**
 * Utility class used by CSSJanus that tokenizes and untokenizes things we want
 * to protect from being janused.
 * @author Roan Kattouw
 */
class CSSJanus_Tokenizer {
	private $regex, $token;
	private $originals;

	/**
	 * Constructor
	 * @param $regex string Regular expression whose matches to replace by a token.
	 * @param $token string Token
	 */
	public function __construct( $regex, $token ) {
		$this->regex = $regex;
		$this->token = $token;
		$this->originals = array();
	}

	/**
	 * Replace all occurrences of $regex in $str with a token and remember
	 * the original strings.
	 * @param $str String to tokenize
	 * @return string Tokenized string
	 */
	public function tokenize( $str ) {
		return preg_replace_callback( $this->regex, array( $this, 'tokenizeCallback' ), $str );
	}

	/**
	 * @param $matches array
	 * @return string
	 */
	private function tokenizeCallback( $matches ) {
		$this->originals[] = $matches[0];
		return $this->token;
	}

	/**
	 * Replace tokens with their originals. If multiple strings were tokenized, it's important they be
	 * detokenized in exactly the SAME ORDER.
	 * @param $str String: previously run through tokenize()
	 * @return string Original string
	 */
	public function detokenize( $str ) {
		// PHP has no function to replace only the first occurrence or to
		// replace occurrences of the same string with different values,
		// so we use preg_replace_callback() even though we don't really need a regex
		return preg_replace_callback( '/' . preg_quote( $this->token, '/' ) . '/',
			array( $this, 'detokenizeCallback' ), $str );
	}

	/**
	 * @param $matches
	 * @return mixed
	 */
	private function detokenizeCallback( $matches ) {
		$retval = current( $this->originals );
		next( $this->originals );

		return $retval;
	}
}