Improved rcube_charset::detect() with BOM checks from rcube_vcard::detect_e...

Aleksander Machniak

2012-11-15 0679b22150dfded4383d0729940a888fcc7a1b99

 program/include/rcube_charset.php

@@ -655,10 +655,19 @@
     */
    public static function detect($string, $failover='')
    {
        if (!function_exists('mb_detect_encoding')) {
            return $failover;
        }
        if (substr($string, 0, 4) == "\0\0\xFE\xFF") return 'UTF-32BE';  // Big Endian
        if (substr($string, 0, 4) == "\xFF\xFE\0\0") return 'UTF-32LE';  // Little Endian
        if (substr($string, 0, 2) == "\xFE\xFF")     return 'UTF-16BE';  // Big Endian
        if (substr($string, 0, 2) == "\xFF\xFE")     return 'UTF-16LE';  // Little Endian
        if (substr($string, 0, 3) == "\xEF\xBB\xBF") return 'UTF-8';

        // heuristics
        if ($string[0] == "\0" && $string[1] == "\0" && $string[2] == "\0" && $string[3] != "\0") return 'UTF-32BE';
        if ($string[0] != "\0" && $string[1] == "\0" && $string[2] == "\0" && $string[3] == "\0") return 'UTF-32LE';
        if ($string[0] == "\0" && $string[1] != "\0" && $string[2] == "\0" && $string[3] != "\0") return 'UTF-16BE';
        if ($string[0] != "\0" && $string[1] == "\0" && $string[2] != "\0" && $string[3] == "\0") return 'UTF-16LE';

        if (function_exists('mb_detect_encoding')) {
        // FIXME: the order is important, because sometimes
        // iso string is detected as euc-jp and etc.
        $enc = array(
@@ -671,6 +680,24 @@
        );

        $result = mb_detect_encoding($string, join(',', $enc));
        }
        else {
            // No match, check for UTF-8
            // from http://w3.org/International/questions/qa-forms-utf-8.html
            if (preg_match('/\A(
                [\x09\x0A\x0D\x20-\x7E]
                | [\xC2-\xDF][\x80-\xBF]
                | \xE0[\xA0-\xBF][\x80-\xBF]
                | [\xE1-\xEC\xEE\xEF][\x80-\xBF]{2}
                | \xED[\x80-\x9F][\x80-\xBF]
                | \xF0[\x90-\xBF][\x80-\xBF]{2}
                | [\xF1-\xF3][\x80-\xBF]{3}
                | \xF4[\x80-\x8F][\x80-\xBF]{2}
                )*\z/xs', substr($string, 0, 2048))
            ) {
            return 'UTF-8';
            }
        }

        return $result ? $result : $failover;
    }