[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.
 *
 * $Id$
 */

require_once(SM_PATH . 'functions/global.php');

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