[squirrelmail.git] / functions / i18n.php

<?php

/**
 * i18n.php
 *
 * Copyright (c) 1999-2003 The SquirrelMail Project Team
 * Licensed under the GNU GPL. For full terms see the file COPYING.
 *
 * This file contains variuos functions that are needed to do
 * internationalization of SquirrelMail.
 *
 * Internally the output character set is used. Other characters are
 * encoded using Unicode entities according to HTML 4.0.
 *
 * $Id$
 * @package squirrelmail
 */

/** Everything uses global.php... */
require_once(SM_PATH . 'functions/global.php');

/**
 * Decodes a string to the internal encoding from the given charset
 * 
 * @param string $charset
 * @param string $string Text to be decoded
 * @return string Decoded text
 */
function charset_decode ($charset, $string) {
    global $languages, $squirrelmail_language, $default_charset;
    global $use_php_recode, $use_php_iconv, $agresive_decoding;

    if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
        function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
        $string = $languages[$squirrelmail_language]['XTRA_CODE']('decode', $string);
    }

    $charset = strtolower($charset);

    set_my_charset();

    // Variables that allow to use functions without function_exist() calls
    if (! isset($use_php_recode) || $use_php_recode=="" ) {
	     $use_php_recode=false; }
    if (! isset($use_php_iconv) || $use_php_iconv=="" ) {
         $use_php_iconv=false; }

    // Don't do conversion if charset is the same.
    if ( $charset == strtolower($default_charset) )
          return htmlspecialchars($string);

    // catch iso-8859-8-i thing
    if ( $charset == "iso-8859-8-i" )
              $charset = "iso-8859-8";

    /*
     * Recode converts html special characters automatically if you use 
     * 'charset..html' decoding. There is no documented way to put -d option 
     * into php recode function call.
     */
    if ( $use_php_recode ) {
      if ( $default_charset == "utf-8" ) {
	// other charsets can be converted to utf-8 without loss.
	// and output string is smaller
	$string = recode_string($charset . "..utf-8",$string);
	return htmlspecialchars($string);
      } else {
	$string = recode_string($charset . "..html",$string);
	// recode does not convert single quote, htmlspecialchars does.
	$string = str_replace("'", '&#039;', $string);
	return $string;
      }
    }

    // iconv functions does not have html target and can be used only with utf-8
    if ( $use_php_iconv && $default_charset=='utf-8') {
      $string = iconv($charset,$default_charset,$string);
      return htmlspecialchars($string);
    }

    // If we don't use recode and iconv, we'll do it old way.

    /* All HTML special characters are 7 bit and can be replaced first */
    
    $string = htmlspecialchars ($string);

    /* controls cpu and memory intensive decoding cycles */
    if (! isset($agresive_decoding) || $agresive_decoding=="" ) {
         $agresive_decoding=false; }

    if (ereg('iso-8859-([[:digit:]]+)', $charset, $res)) {
        if ($res[1] == '1') {
	    include_once(SM_PATH . 'functions/decode/iso8859-1.php');
            $ret = charset_decode_iso8859_1 ($string);
        } else if ($res[1] == '2') {
	    include_once(SM_PATH . 'functions/decode/iso8859-2.php');
            $ret = charset_decode_iso8859_2 ($string);
        } else if ($res[1] == '3') {
	    include_once(SM_PATH . 'functions/decode/iso8859-3.php');
            $ret = charset_decode_iso8859_3 ($string);
        } else if ($res[1] == '4') {
	    include_once(SM_PATH . 'functions/decode/iso8859-4.php');
            $ret = charset_decode_iso8859_4 ($string);
        } else if ($res[1] == '5') {
	    include_once(SM_PATH . 'functions/decode/iso8859-5.php');
            $ret = charset_decode_iso8859_5 ($string);
        } else if ($res[1] == '6') {
	    include_once(SM_PATH . 'functions/decode/iso8859-6.php');
            $ret = charset_decode_iso8859_6 ($string);
        } else if ($res[1] == '7') {
	    include_once(SM_PATH . 'functions/decode/iso8859-7.php');
            $ret = charset_decode_iso8859_7 ($string);
        } else if ($res[1] == '8') {
	    include_once(SM_PATH . 'functions/decode/iso8859-8.php');
            $ret = charset_decode_iso8859_8 ($string);
        } else if ($res[1] == '9') {
	    include_once(SM_PATH . 'functions/decode/iso8859-9.php');
            $ret = charset_decode_iso8859_9 ($string);
        } else if ($res[1] == '10') {
	    include_once(SM_PATH . 'functions/decode/iso8859-10.php');
            $ret = charset_decode_iso8859_10 ($string);
        } else if ($res[1] == '11') {
	    include_once(SM_PATH . 'functions/decode/iso8859-11.php');
            $ret = charset_decode_iso8859_11 ($string);
        } else if ($res[1] == '13') {
	    include_once(SM_PATH . 'functions/decode/iso8859-13.php');
            $ret = charset_decode_iso8859_13 ($string);
        } else if ($res[1] == '14') {
	    include_once(SM_PATH . 'functions/decode/iso8859-14.php');
            $ret = charset_decode_iso8859_14 ($string);
        } else if ($res[1] == '15') {
	    include_once(SM_PATH . 'functions/decode/iso8859-15.php');
            $ret = charset_decode_iso8859_15 ($string);
        } else if ($res[1] == '16') {
	    include_once(SM_PATH . 'functions/decode/iso8859-16.php');
            $ret = charset_decode_iso8859_16 ($string);
        } else {
            $ret = charset_decode_iso_8859_default ($string);
        }
    } else if ($charset == 'ns_4551-1') {
        $ret = charset_decode_ns_4551_1 ($string);
    } else if ($charset == 'koi8-r') {
        include_once(SM_PATH . 'functions/decode/koi8-r.php');
        $ret = charset_decode_koi8r ($string);
    } else if ($charset == 'koi8-u') {
        include_once(SM_PATH . 'functions/decode/koi8-u.php');
        $ret = charset_decode_koi8u ($string);
    } else if ($charset == 'windows-1250') {
        include_once(SM_PATH . 'functions/decode/cp1250.php');
        $ret = charset_decode_cp1250 ($string);
    } else if ($charset == 'windows-1251') {
        include_once(SM_PATH . 'functions/decode/cp1251.php');
        $ret = charset_decode_cp1251 ($string);
    } else if ($charset == 'windows-1252') {
        include_once(SM_PATH . 'functions/decode/cp1252.php');
        $ret = charset_decode_cp1252 ($string);
    } else if ($charset == 'windows-1253') {
        include_once(SM_PATH . 'functions/decode/cp1253.php');
	$ret = charset_decode_cp1253 ($string);
    } else if ($charset == 'windows-1254') {
        include_once(SM_PATH . 'functions/decode/cp1254.php');
	$ret = charset_decode_cp1254 ($string);
    } else if ($charset == 'windows-1255') {
        include_once(SM_PATH . 'functions/decode/cp1255.php');
	$ret = charset_decode_cp1255 ($string);
    } else if ($charset == 'windows-1256') {
        include_once(SM_PATH . 'functions/decode/cp1256.php');
	$ret = charset_decode_cp1256 ($string);
    } else if ($charset == 'windows-1257') {
        include_once(SM_PATH . 'functions/decode/cp1257.php');
        $ret = charset_decode_cp1257 ($string);
    } else if ($charset == 'windows-1258') {
        include_once(SM_PATH . 'functions/decode/cp1258.php');
        $ret = charset_decode_cp1258 ($string);
    } else if ($charset == 'x-mac-roman') {
        include_once(SM_PATH . 'functions/decode/cp10000.php');
        $ret = charset_decode_cp10000 ($string);
    } else if ($charset == 'x-mac-greek') {
        include_once(SM_PATH . 'functions/decode/cp10006.php');
        $ret = charset_decode_cp10006 ($string);
    } else if ($charset == 'x-mac-cyrillic') {
        include_once(SM_PATH . 'functions/decode/cp10007.php');
        $ret = charset_decode_cp10007 ($string);
    } else if ($charset == 'x-mac-ukrainian') {
        include_once(SM_PATH . 'functions/decode/cp10017.php');
        $ret = charset_decode_cp10017 ($string);
    } else if ($charset == 'x-mac-centraleurroman') {
        include_once(SM_PATH . 'functions/decode/cp10029.php');
        $ret = charset_decode_cp10029 ($string);
    } else if ($charset == 'x-mac-icelandic') {
        include_once(SM_PATH . 'functions/decode/cp10079.php');
        $ret = charset_decode_cp10079 ($string);
    } else if ($charset == 'x-mac-turkish') {
        include_once(SM_PATH . 'functions/decode/cp10081.php');
        $ret = charset_decode_cp10081 ($string);
    } else if ($charset == 'ibm855') {
        include_once(SM_PATH . 'functions/decode/cp855.php');
        $ret = charset_decode_cp855 ($string);
    } else if ($charset == 'ibm866') {
        include_once(SM_PATH . 'functions/decode/cp866.php');
        $ret = charset_decode_cp866 ($string);
    } else if ($charset == 'iso-ir-111') {
        include_once(SM_PATH . 'functions/decode/iso-ir-111.php');
        $ret = charset_decode_iso_ir_111 ($string);
    } else if ($charset == 'tis-620') {
        include_once(SM_PATH . 'functions/decode/tis620.php');
        $ret = charset_decode_tis620 ($string);
    } else if ($charset == 'big5' and $agresive_decoding ) {
        include_once(SM_PATH . 'functions/decode/big5.php');
        $ret = charset_decode_big5 ($string);
    } else if ($charset == 'gb2312' and $agresive_decoding ) {
        include_once(SM_PATH . 'functions/decode/gb2312.php');
        $ret = charset_decode_gb2312 ($string);
    } else if ($charset == 'utf-8' && $squirrelmail_language != 'ja_JP') {
        include_once(SM_PATH . 'functions/decode/utf-8.php');
	$ret = charset_decode_utf8 ($string);
    } else {
        $ret = $string;
    }
    return( $ret );
}


/* Remove all 8 bit characters from all other ISO-8859 character sets */
function charset_decode_iso_8859_default ($string) {
    return (strtr($string, "\240\241\242\243\244\245\246\247".
                    "\250\251\252\253\254\255\256\257".
                    "\260\261\262\263\264\265\266\267".
                    "\270\271\272\273\274\275\276\277".
                    "\300\301\302\303\304\305\306\307".
                    "\310\311\312\313\314\315\316\317".
                    "\320\321\322\323\324\325\326\327".
                    "\330\331\332\333\334\335\336\337".
                    "\340\341\342\343\344\345\346\347".
                    "\350\351\352\353\354\355\356\357".
                    "\360\361\362\363\364\365\366\367".
                    "\370\371\372\373\374\375\376\377",
                    "????????????????????????????????????????".
                    "????????????????????????????????????????".
                    "????????????????????????????????????????".
                    "????????"));

}

/*
 * This is the same as ISO-646-NO and is used by some
 * Microsoft programs when sending Norwegian characters
 */
function charset_decode_ns_4551_1 ($string) {
    /*
     * These characters are:
     * Latin capital letter AE
     * Latin capital letter O with stroke
     * Latin capital letter A with ring above
     * and the same as small letters
     */
Commit	Line	Data
59177427	1	<?php
1fd97780	2
35586184	3	/**
	4	* i18n.php
	5	*
76911253	6	* Copyright (c) 1999-2003 The SquirrelMail Project Team
35586184	7	* Licensed under the GNU GPL. For full terms see the file COPYING.
	8	*
	9	* This file contains variuos functions that are needed to do
	10	* internationalization of SquirrelMail.
	11	*
	12	* Internally the output character set is used. Other characters are
	13	* encoded using Unicode entities according to HTML 4.0.
	14	*
	15	* $Id$
d6c32258	16	* @package squirrelmail
35586184	17	*/
35586184	18
d6c32258	19	/** Everything uses global.php... */
961ca3d8	20	require_once(SM_PATH . 'functions/global.php');
961ca3d8	21
d6c32258	22	/**
	23	* Decodes a string to the internal encoding from the given charset
	24	*
	25	* @param string $charset
	26	* @param string $string Text to be decoded
	27	* @return string Decoded text
	28	*/
a2a7852b	29	function charset_decode ($charset, $string) {
3ec81e63	30	global $languages, $squirrelmail_language, $default_charset;
edf2c0ba	31	global $use_php_recode, $use_php_iconv, $agresive_decoding;
a2a7852b	32
3714db45	33	if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
3714db45	34	function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
6fbd125b	35	$string = $languages[$squirrelmail_language]['XTRA_CODE']('decode', $string);
6fbd125b	36	}
b05c8961	37
3ec81e63	38	$charset = strtolower($charset);
	39
	40	set_my_charset();
	41
	42	// Variables that allow to use functions without function_exist() calls
edf2c0ba	43	if (! isset($use_php_recode) \|\| $use_php_recode=="" ) {
	44	$use_php_recode=false; }
	45	if (! isset($use_php_iconv) \|\| $use_php_iconv=="" ) {
	46	$use_php_iconv=false; }
3ec81e63	47
	48	// Don't do conversion if charset is the same.
	49	if ( $charset == strtolower($default_charset) )
	50	return htmlspecialchars($string);
	51
	52	// catch iso-8859-8-i thing
	53	if ( $charset == "iso-8859-8-i" )
	54	$charset = "iso-8859-8";
	55
	56	/*
	57	* Recode converts html special characters automatically if you use
	58	* 'charset..html' decoding. There is no documented way to put -d option
	59	* into php recode function call.
	60	*/
	61	if ( $use_php_recode ) {
	62	if ( $default_charset == "utf-8" ) {
	63	// other charsets can be converted to utf-8 without loss.
	64	// and output string is smaller
	65	$string = recode_string($charset . "..utf-8",$string);
	66	return htmlspecialchars($string);
	67	} else {
	68	$string = recode_string($charset . "..html",$string);
	69	// recode does not convert single quote, htmlspecialchars does.
	70	$string = str_replace("'", ''', $string);
	71	return $string;
	72	}
	73	}
	74
	75	// iconv functions does not have html target and can be used only with utf-8
	76	if ( $use_php_iconv && $default_charset=='utf-8') {
	77	$string = iconv($charset,$default_charset,$string);
	78	return htmlspecialchars($string);
	79	}
	80
	81	// If we don't use recode and iconv, we'll do it old way.
	82
a2a7852b	83	/* All HTML special characters are 7 bit and can be replaced first */
cef054e4	84
098ea084	85	$string = htmlspecialchars ($string);
a2a7852b	86
5dd23dac	87	/* controls cpu and memory intensive decoding cycles */
edf2c0ba	88	if (! isset($agresive_decoding) \|\| $agresive_decoding=="" ) {
edf2c0ba	89	$agresive_decoding=false; }
5dd23dac	90
a2a7852b	91	if (ereg('iso-8859-([[:digit:]]+)', $charset, $res)) {
a2a7852b	92	if ($res[1] == '1') {
5dd23dac	93	include_once(SM_PATH . 'functions/decode/iso8859-1.php');
5dd23dac	94	$ret = charset_decode_iso8859_1 ($string);
a2a7852b	95	} else if ($res[1] == '2') {
5dd23dac	96	include_once(SM_PATH . 'functions/decode/iso8859-2.php');
5dd23dac	97	$ret = charset_decode_iso8859_2 ($string);
3a66bed2	98	} else if ($res[1] == '3') {
	99	include_once(SM_PATH . 'functions/decode/iso8859-3.php');
	100	$ret = charset_decode_iso8859_3 ($string);
9be313d5	101	} else if ($res[1] == '4') {
3a66bed2	102	include_once(SM_PATH . 'functions/decode/iso8859-4.php');
3a66bed2	103	$ret = charset_decode_iso8859_4 ($string);
94965562	104	} else if ($res[1] == '5') {
3a66bed2	105	include_once(SM_PATH . 'functions/decode/iso8859-5.php');
3a66bed2	106	$ret = charset_decode_iso8859_5 ($string);
ef82d2d5	107	} else if ($res[1] == '6') {
5dd23dac	108	include_once(SM_PATH . 'functions/decode/iso8859-6.php');
5dd23dac	109	$ret = charset_decode_iso8859_6 ($string);
a2a7852b	110	} else if ($res[1] == '7') {
5dd23dac	111	include_once(SM_PATH . 'functions/decode/iso8859-7.php');
5dd23dac	112	$ret = charset_decode_iso8859_7 ($string);
3a66bed2	113	} else if ($res[1] == '8') {
	114	include_once(SM_PATH . 'functions/decode/iso8859-8.php');
	115	$ret = charset_decode_iso8859_8 ($string);
3ab35042	116	} else if ($res[1] == '9') {
5dd23dac	117	include_once(SM_PATH . 'functions/decode/iso8859-9.php');
5dd23dac	118	$ret = charset_decode_iso8859_9 ($string);
3a66bed2	119	} else if ($res[1] == '10') {
	120	include_once(SM_PATH . 'functions/decode/iso8859-10.php');
	121	$ret = charset_decode_iso8859_10 ($string);
	122	} else if ($res[1] == '11') {
	123	include_once(SM_PATH . 'functions/decode/iso8859-11.php');
	124	$ret = charset_decode_iso8859_11 ($string);
9be313d5	125	} else if ($res[1] == '13') {
3a66bed2	126	include_once(SM_PATH . 'functions/decode/iso8859-13.php');
	127	$ret = charset_decode_iso8859_13 ($string);
	128	} else if ($res[1] == '14') {
	129	include_once(SM_PATH . 'functions/decode/iso8859-14.php');
	130	$ret = charset_decode_iso8859_14 ($string);
a2a7852b	131	} else if ($res[1] == '15') {
5dd23dac	132	include_once(SM_PATH . 'functions/decode/iso8859-15.php');
5dd23dac	133	$ret = charset_decode_iso8859_15 ($string);
3a66bed2	134	} else if ($res[1] == '16') {
	135	include_once(SM_PATH . 'functions/decode/iso8859-16.php');
	136	$ret = charset_decode_iso8859_16 ($string);
a2a7852b	137	} else {
	138	$ret = charset_decode_iso_8859_default ($string);
	139	}
	140	} else if ($charset == 'ns_4551-1') {
	141	$ret = charset_decode_ns_4551_1 ($string);
	142	} else if ($charset == 'koi8-r') {
5dd23dac	143	include_once(SM_PATH . 'functions/decode/koi8-r.php');
a2a7852b	144	$ret = charset_decode_koi8r ($string);
1c0e847f	145	} else if ($charset == 'koi8-u') {
5dd23dac	146	include_once(SM_PATH . 'functions/decode/koi8-u.php');
1c0e847f	147	$ret = charset_decode_koi8u ($string);
5dd23dac	148	} else if ($charset == 'windows-1250') {
	149	include_once(SM_PATH . 'functions/decode/cp1250.php');
	150	$ret = charset_decode_cp1250 ($string);
a2a7852b	151	} else if ($charset == 'windows-1251') {
5dd23dac	152	include_once(SM_PATH . 'functions/decode/cp1251.php');
	153	$ret = charset_decode_cp1251 ($string);
	154	} else if ($charset == 'windows-1252') {
	155	include_once(SM_PATH . 'functions/decode/cp1252.php');
	156	$ret = charset_decode_cp1252 ($string);
3ab35042	157	} else if ($charset == 'windows-1253') {
5dd23dac	158	include_once(SM_PATH . 'functions/decode/cp1253.php');
5dd23dac	159	$ret = charset_decode_cp1253 ($string);
3ab35042	160	} else if ($charset == 'windows-1254') {
5dd23dac	161	include_once(SM_PATH . 'functions/decode/cp1254.php');
5dd23dac	162	$ret = charset_decode_cp1254 ($string);
c48a8ca7	163	} else if ($charset == 'windows-1255') {
5dd23dac	164	include_once(SM_PATH . 'functions/decode/cp1255.php');
5dd23dac	165	$ret = charset_decode_cp1255 ($string);
c48a8ca7	166	} else if ($charset == 'windows-1256') {
5dd23dac	167	include_once(SM_PATH . 'functions/decode/cp1256.php');
5dd23dac	168	$ret = charset_decode_cp1256 ($string);
c37a12f8	169	} else if ($charset == 'windows-1257') {
3a66bed2	170	include_once(SM_PATH . 'functions/decode/cp1257.php');
3a66bed2	171	$ret = charset_decode_cp1257 ($string);
5dd23dac	172	} else if ($charset == 'windows-1258') {
	173	include_once(SM_PATH . 'functions/decode/cp1258.php');
	174	$ret = charset_decode_cp1258 ($string);
e9a71964	175	} else if ($charset == 'x-mac-roman') {
	176	include_once(SM_PATH . 'functions/decode/cp10000.php');
	177	$ret = charset_decode_cp10000 ($string);
	178	} else if ($charset == 'x-mac-greek') {
	179	include_once(SM_PATH . 'functions/decode/cp10006.php');
	180	$ret = charset_decode_cp10006 ($string);
	181	} else if ($charset == 'x-mac-cyrillic') {
	182	include_once(SM_PATH . 'functions/decode/cp10007.php');
	183	$ret = charset_decode_cp10007 ($string);
	184	} else if ($charset == 'x-mac-ukrainian') {
	185	include_once(SM_PATH . 'functions/decode/cp10017.php');
	186	$ret = charset_decode_cp10017 ($string);
	187	} else if ($charset == 'x-mac-centraleurroman') {
	188	include_once(SM_PATH . 'functions/decode/cp10029.php');
	189	$ret = charset_decode_cp10029 ($string);
	190	} else if ($charset == 'x-mac-icelandic') {
	191	include_once(SM_PATH . 'functions/decode/cp10079.php');
	192	$ret = charset_decode_cp10079 ($string);
	193	} else if ($charset == 'x-mac-turkish') {
	194	include_once(SM_PATH . 'functions/decode/cp10081.php');
	195	$ret = charset_decode_cp10081 ($string);
	196	} else if ($charset == 'ibm855') {
	197	include_once(SM_PATH . 'functions/decode/cp855.php');
	198	$ret = charset_decode_cp855 ($string);
	199	} else if ($charset == 'ibm866') {
	200	include_once(SM_PATH . 'functions/decode/cp866.php');
	201	$ret = charset_decode_cp866 ($string);
0e4337e7	202	} else if ($charset == 'iso-ir-111') {
	203	include_once(SM_PATH . 'functions/decode/iso-ir-111.php');
	204	$ret = charset_decode_iso_ir_111 ($string);
7af26ef8	205	} else if ($charset == 'tis-620') {
	206	include_once(SM_PATH . 'functions/decode/tis620.php');
	207	$ret = charset_decode_tis620 ($string);
5dd23dac	208	} else if ($charset == 'big5' and $agresive_decoding ) {
	209	include_once(SM_PATH . 'functions/decode/big5.php');
	210	$ret = charset_decode_big5 ($string);
	211	} else if ($charset == 'gb2312' and $agresive_decoding ) {
	212	include_once(SM_PATH . 'functions/decode/gb2312.php');
	213	$ret = charset_decode_gb2312 ($string);
96d985e5	214	} else if ($charset == 'utf-8' && $squirrelmail_language != 'ja_JP') {
5dd23dac	215	include_once(SM_PATH . 'functions/decode/utf-8.php');
3ab35042	216	$ret = charset_decode_utf8 ($string);
a2a7852b	217	} else {
	218	$ret = $string;
	219	}
	220	return( $ret );
	221	}
	222
a2a7852b	223
	224	/* Remove all 8 bit characters from all other ISO-8859 character sets */
	225	function charset_decode_iso_8859_default ($string) {
	226	return (strtr($string, "\240\241\242\243\244\245\246\247".
1fd97780	227	"\250\251\252\253\254\255\256\257".
	228	"\260\261\262\263\264\265\266\267".
	229	"\270\271\272\273\274\275\276\277".
	230	"\300\301\302\303\304\305\306\307".
	231	"\310\311\312\313\314\315\316\317".
	232	"\320\321\322\323\324\325\326\327".
	233	"\330\331\332\333\334\335\336\337".
	234	"\340\341\342\343\344\345\346\347".
	235	"\350\351\352\353\354\355\356\357".
	236	"\360\361\362\363\364\365\366\367".
a2a7852b	237	"\370\371\372\373\374\375\376\377",
1fd97780	238	"????????????????????????????????????????".
	239	"????????????????????????????????????????".
	240	"????????????????????????????????????????".
	241	"????????"));
a2a7852b	242
	243	}
	244
	245	/*
	246	* This is the same as ISO-646-NO and is used by some
	247	* Microsoft programs when sending Norwegian characters
	248	*/
	249	function charset_decode_ns_4551_1 ($string) {
	250	/*
	251	* These characters are:
	252	* Latin capital letter AE
	253	* Latin capital letter O with stroke
	254	* Latin capital letter A with ring above
	255	* and the same as small letters
	256	*/
	257