vendor/symfony/src/Symfony/Component/CssSelector/Tokenizer.php
author ymh <ymh.work@gmail.com>
Sat, 24 Sep 2011 15:40:41 +0200
changeset 0 7f95f8617b0b
permissions -rwxr-xr-x
first commit

<?php

/*
 * This file is part of the Symfony package.
 *
 * (c) Fabien Potencier <fabien@symfony.com>
 *
 * For the full copyright and license information, please view the LICENSE
 * file that was distributed with this source code.
 */

namespace Symfony\Component\CssSelector;

use Symfony\Component\CssSelector\Exception\ParseException;

/**
 * Tokenizer lexes a CSS Selector to tokens.
 *
 * This component is a port of the Python lxml library,
 * which is copyright Infrae and distributed under the BSD license.
 *
 * @author Fabien Potencier <fabien@symfony.com>
 */
class Tokenizer
{
    /**
     * Takes a CSS selector and returns an array holding the Tokens
     * it contains.
     *
     * @param  string $s The selector to lex.
     *
     * @return array Token[]
     */
    public function tokenize($s)
    {
        if (function_exists('mb_internal_encoding') && ((int) ini_get('mbstring.func_overload')) & 2) {
            $mbEncoding = mb_internal_encoding();
            mb_internal_encoding('ASCII');
        }

        $tokens = array();
        $pos = 0;
        $s = preg_replace('#/\*.*?\*/#s', '', $s);

        while (true) {
            if (preg_match('#\s+#A', $s, $match, 0, $pos)) {
                $precedingWhitespacePos = $pos;
                $pos += strlen($match[0]);
            } else {
                $precedingWhitespacePos = 0;
            }

            if ($pos >= strlen($s)) {
                if (isset($mbEncoding)) {
                    mb_internal_encoding($mbEncoding);
                }

                return $tokens;
            }

            if (preg_match('#[+-]?\d*n(?:[+-]\d+)?#A', $s, $match, 0, $pos) && 'n' !== $match[0]) {
                $sym = substr($s, $pos, strlen($match[0]));
                $tokens[] = new Token('Symbol', $sym, $pos);
                $pos += strlen($match[0]);

                continue;
            }

            $c = $s[$pos];
            $c2 = substr($s, $pos, 2);
            if (in_array($c2, array('~=', '|=', '^=', '$=', '*=', '::', '!='))) {
                $tokens[] = new Token('Token', $c2, $pos);
                $pos += 2;

                continue;
            }

            if (in_array($c, array('>', '+', '~', ',', '.', '*', '=', '[', ']', '(', ')', '|', ':', '#'))) {
                if (in_array($c, array('.', '#', '[')) && $precedingWhitespacePos > 0) {
                    $tokens[] = new Token('Token', ' ', $precedingWhitespacePos);
                }
                $tokens[] = new Token('Token', $c, $pos);
                ++$pos;

                continue;
            }

            if ('"' === $c || "'" === $c) {
                // Quoted string
                $oldPos = $pos;
                list($sym, $pos) = $this->tokenizeEscapedString($s, $pos);

                $tokens[] = new Token('String', $sym, $oldPos);

                continue;
            }

            $oldPos = $pos;
            list($sym, $pos) = $this->tokenizeSymbol($s, $pos);

            $tokens[] = new Token('Symbol', $sym, $oldPos);

            continue;
        }
    }

    /**
     * Tokenizes a quoted string (i.e. 'A string quoted with \' characters'),
     * and returns an array holding the unquoted string contained by $s and
     * the new position from which tokenizing should take over.
     *
     * @throws ParseException When expected closing is not found
     *
     * @param  string  $s   The selector string containing the quoted string.
     * @param  integer $pos The starting position for the quoted string.
     *
     * @return array
     */
    private function tokenizeEscapedString($s, $pos)
    {
        $quote = $s[$pos];

        $pos = $pos + 1;
        $start = $pos;
        while (true) {
            $next = strpos($s, $quote, $pos);
            if (false === $next) {
                throw new ParseException(sprintf('Expected closing %s for string in: %s', $quote, substr($s, $start)));
            }

            $result = substr($s, $start, $next - $start);
            if ('\\' === $result[strlen($result) - 1]) {
                // next quote character is escaped
                $pos = $next + 1;
                continue;
            }

            if (false !== strpos($result, '\\')) {
                $result = $this->unescapeStringLiteral($result);
            }

            return array($result, $next + 1);
        }
    }

    /**
     * Unescapes a string literal and returns the unescaped string.
     *
     * @throws ParseException When invalid escape sequence is found
     *
     * @param  string $literal The string literal to unescape.
     *
     * @return string
     */
    private function unescapeStringLiteral($literal)
    {
        return preg_replace_callback('#(\\\\(?:[A-Fa-f0-9]{1,6}(?:\r\n|\s)?|[^A-Fa-f0-9]))#', function ($matches) use ($literal)
        {
            if ($matches[0][0] == '\\' && strlen($matches[0]) > 1) {
                $matches[0] = substr($matches[0], 1);
                if (in_array($matches[0][0], array('0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'A', 'B', 'C', 'D', 'E', 'F', 'a', 'b', 'c', 'd', 'e', 'f'))) {
                    return chr(trim($matches[0]));
                }
            } else {
                throw new ParseException(sprintf('Invalid escape sequence %s in string %s', $matches[0], $literal));
            }
        }, $literal);
    }

    /**
     * Lexes selector $s and returns an array holding the name of the symbol
     * contained in it and the new position from which tokenizing should take
     * over.
     *
     * @throws ParseException When Unexpected symbol is found
     *
     * @param  string  $s   The selector string.
     * @param  integer $pos The position in $s at which the symbol starts.
     *
     * @return array
     */
    private function tokenizeSymbol($s, $pos)
    {
        $start = $pos;

        if (!preg_match('#[^\w\-]#', $s, $match, PREG_OFFSET_CAPTURE, $pos)) {
            // Goes to end of s
            return array(substr($s, $start), strlen($s));
        }

        $matchStart = $match[0][1];

        if ($matchStart == $pos) {
            throw new ParseException(sprintf('Unexpected symbol: %s at %s', $s[$pos], $pos));
        }

        $result = substr($s, $start, $matchStart - $start);
        $pos = $matchStart;

        return array($result, $pos);
    }
}