3 * SquirrelMail internationalization functions
5 * Copyright (c) 1999-2005 The SquirrelMail Project Team
6 * Licensed under the GNU GPL. For full terms see the file COPYING.
8 * This file contains variuos functions that are needed to do
9 * internationalization of SquirrelMail.
11 * Internally the output character set is used. Other characters are
12 * encoded using Unicode entities according to HTML 4.0.
15 * @package squirrelmail
20 if (! defined('SM_PATH')) define('SM_PATH','../');
22 /** Everything uses global.php... */
23 require_once(SM_PATH
. 'functions/global.php');
26 * Gettext bindtextdomain wrapper.
28 * Wrapper solves differences between php versions in order to provide
29 * ngettext support. Should be used if translation uses ngettext
32 * @param string $domain gettext domain name
33 * @param string $dir directory that contains all translations
34 * @return string path to translation directory
36 function sq_bindtextdomain($domain,$dir) {
37 global $l10n, $gettext_flags, $sm_notAlias;
39 if ($gettext_flags==7) {
40 // gettext extension without ngettext
41 if (substr($dir, -1) != '/') $dir .= '/';
42 $mofile=$dir . $sm_notAlias . '/LC_MESSAGES/' . $domain . '.mo';
43 $input = new FileReader($mofile);
44 $l10n[$domain] = new gettext_reader($input);
47 $dir=bindtextdomain($domain,$dir);
53 * Gettext textdomain wrapper.
54 * Makes sure that gettext_domain global is modified.
56 * @param string $name gettext domain name
57 * @return string gettext domain name
59 function sq_textdomain($domain) {
60 global $gettext_domain;
61 $gettext_domain=textdomain($domain);
62 return $gettext_domain;
66 * php setlocale function wrapper
68 * From php 4.3.0 it is possible to use arrays in order to set locale.
69 * php gettext extension works only when locale is set. This wrapper
70 * function allows to use more than one locale name.
72 * @param int $category locale category name. Use php named constants
73 * (LC_ALL, LC_COLLATE, LC_CTYPE, LC_MONETARY, LC_NUMERIC, LC_TIME)
74 * @param mixed $locale option contains array with possible locales or string with one locale
75 * @return string name of set locale or false, if all locales fail.
76 * @since 1.5.1 and 1.4.5
77 * @see http://www.php.net/setlocale
79 function sq_setlocale($category,$locale) {
80 // string with only one locale
81 if (is_string($locale))
82 return setlocale($category,$locale);
84 if (! check_php_version(4,3)) {
87 while ( ! $ret && $index<count($locale)) {
88 $ret=setlocale($category,$locale[$index]);
92 // php 4.3.0 or better, use entire array
93 $ret=setlocale($category,$locale);
99 * Converts string from given charset to charset, that can be displayed by user translation.
101 * Function by default returns html encoded strings, if translation uses different encoding.
102 * If Japanese translation is used - function returns string converted to euc-jp
103 * If iconv or recode functions are enabled and translation uses utf-8 - function returns utf-8 encoded string.
104 * If $charset is not supported - function returns unconverted string.
106 * sanitizing of html tags is also done by this function.
108 * @param string $charset
109 * @param string $string Text to be decoded
110 * @param boolean $force_decode converts string to html without $charset!=$default_charset check.
111 * Argument is available since 1.5.1 and 1.4.5.
112 * @return string decoded string
114 function charset_decode ($charset, $string, $force_decode=false) {
115 global $languages, $squirrelmail_language, $default_charset;
116 global $use_php_recode, $use_php_iconv, $aggressive_decoding;
118 if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
119 function_exists($languages[$squirrelmail_language]['XTRA_CODE'] . '_decode')) {
120 $string = call_user_func($languages[$squirrelmail_language]['XTRA_CODE'] . '_decode', $string);
123 $charset = strtolower($charset);
127 // Variables that allow to use functions without function_exist() calls
128 if (! isset($use_php_recode) ||
$use_php_recode=="" ) {
129 $use_php_recode=false; }
130 if (! isset($use_php_iconv) ||
$use_php_iconv=="" ) {
131 $use_php_iconv=false; }
133 // Don't do conversion if charset is the same.
134 if ( ! $force_decode && $charset == strtolower($default_charset) )
135 return htmlspecialchars($string);
137 // catch iso-8859-8-i thing
138 if ( $charset == "iso-8859-8-i" )
139 $charset = "iso-8859-8";
142 * Recode converts html special characters automatically if you use
143 * 'charset..html' decoding. There is no documented way to put -d option
144 * into php recode function call.
146 if ( $use_php_recode ) {
147 if ( $default_charset == "utf-8" ) {
148 // other charsets can be converted to utf-8 without loss.
149 // and output string is smaller
150 $string = recode_string($charset . "..utf-8",$string);
151 return htmlspecialchars($string);
153 $string = recode_string($charset . "..html",$string);
154 // recode does not convert single quote, htmlspecialchars does.
155 $string = str_replace("'", ''', $string);
160 // iconv functions does not have html target and can be used only with utf-8
161 if ( $use_php_iconv && $default_charset=='utf-8') {
162 $string = iconv($charset,$default_charset,$string);
163 return htmlspecialchars($string);
166 // If we don't use recode and iconv, we'll do it old way.
168 /* All HTML special characters are 7 bit and can be replaced first */
170 $string = htmlspecialchars ($string);
172 /* controls cpu and memory intensive decoding cycles */
173 if (! isset($aggressive_decoding) ||
$aggressive_decoding=="" ) {
174 $aggressive_decoding=false; }
176 $decode=fixcharset($charset);
177 $decodefile=SM_PATH
. 'functions/decode/' . $decode . '.php';
178 if (file_exists($decodefile)) {
179 include_once($decodefile);
180 $ret = call_user_func('charset_decode_'.$decode, $string);
188 * Converts html string to given charset
189 * @since 1.5.1 and 1.4.4
190 * @param string $string
191 * @param string $charset
192 * @param boolean $htmlencode keep htmlspecialchars encoding
195 function charset_encode($string,$charset,$htmlencode=true) {
196 global $default_charset;
198 $encode=fixcharset($charset);
199 $encodefile=SM_PATH
. 'functions/encode/' . $encode . '.php';
200 if (file_exists($encodefile)) {
201 include_once($encodefile);
202 $ret = call_user_func('charset_encode_'.$encode, $string);
203 } elseif(file_exists(SM_PATH
. 'functions/encode/us_ascii.php')) {
204 // function replaces all 8bit html entities with question marks.
205 // it is used when other encoding functions are unavailable
206 include_once(SM_PATH
. 'functions/encode/us_ascii.php');
207 $ret = charset_encode_us_ascii($string);
210 * fix for yahoo users that remove all us-ascii related things
216 * Undo html special chars, some places (like compose form) have
217 * own sanitizing functions and don't need html symbols.
218 * Undo chars only after encoding in order to prevent conversion of
219 * html entities in plain text emails.
221 if (! $htmlencode ) {
222 $ret = str_replace(array('&','>','<','"'),array('&','>','<','"'),$ret);
228 * Combined decoding and encoding functions
230 * If conversion is done to charset different that utf-8, unsupported symbols
231 * will be replaced with question marks.
232 * @since 1.5.1 and 1.4.4
233 * @param string $in_charset initial charset
234 * @param string $string string that has to be converted
235 * @param string $out_charset final charset
236 * @param boolean $htmlencode keep htmlspecialchars encoding
237 * @return string converted string
239 function charset_convert($in_charset,$string,$out_charset,$htmlencode=true) {
240 $string=charset_decode($in_charset,$string,true);
241 $string=charset_encode($string,$out_charset,$htmlencode);
246 * Makes charset name suitable for decoding cycles
248 * @since 1.5.0 and 1.4.4
249 * @param string $charset Name of charset
250 * @return string $charset Adjusted name of charset
252 function fixcharset($charset) {
253 /* remove minus and characters that might be used in paths from charset
254 * name in order to be able to use it in function names and include calls.
256 $charset=preg_replace("/[-:.\/\\\]/",'_',$charset);
258 // windows-125x and cp125x charsets
259 $charset=str_replace('windows_','cp',$charset);
262 $charset=str_replace('ibm','cp',$charset);
264 // iso-8859-8-i -> iso-8859-8
265 // use same cycle until I'll find differences
266 $charset=str_replace('iso_8859_8_i','iso_8859_8',$charset);
272 * Set up the language to be output
273 * if $do_search is true, then scan the browser information
274 * for a possible language that we know
276 * Function sets system locale environment (LC_ALL, LANG, LANGUAGE),
277 * gettext translation bindings and html header information.
279 * Function returns error codes, if there is some fatal error.
281 * 1 = mbstring support is not present,
282 * 2 = mbstring support is not present, user's translation reverted to en_US.
284 * @param string $sm_language translation used by user's interface
285 * @param bool $do_search use browser's preferred language detection functions. Defaults to false.
286 * @param bool $default set $sm_language to $squirrelmail_default_language if language detection fails or language is not set. Defaults to false.
287 * @return int function execution error codes.
289 function set_up_language($sm_language, $do_search = false, $default = false) {
291 static $SetupAlready = 0;
292 global $use_gettext, $languages,
293 $squirrelmail_language, $squirrelmail_default_language, $default_charset,
294 $sm_notAlias, $username, $data_dir;
300 $SetupAlready = TRUE;
301 sqgetGlobalVar('HTTP_ACCEPT_LANGUAGE', $accept_lang, SQ_SERVER
);
304 * If function is asked to detect preferred language
305 * OR squirrelmail default language is set to empty string
307 * squirrelmail language ($sm_language) is empty string
308 * (not set in user's prefs and no cookie with language info)
310 * browser provides list of preferred languages
312 * get preferred language from HTTP_ACCEPT_LANGUAGE header
314 if (($do_search ||
empty($squirrelmail_default_language)) &&
316 isset($accept_lang)) {
317 // TODO: use more than one language, if first language is not available
318 // FIXME: function assumes that string contains two or more characters.
319 // FIXME: some languages use 5 chars
320 $sm_language = substr($accept_lang, 0, 2);
324 * If language preference is not set OR script asks to use default language
326 * default squirrelmail language is not set to empty string
328 * use default squirrelmail language value from configuration.
330 if ((!$sm_language||
$default) &&
331 ! empty($squirrelmail_default_language)) {
332 $squirrelmail_language = $squirrelmail_default_language;
333 $sm_language = $squirrelmail_default_language;
336 /** provide failsafe language when detection fails */
337 if (! $sm_language) $sm_language='en_US';
339 $sm_notAlias = $sm_language;
341 // Catching removed translation
342 // System reverts to English translation if user prefs contain translation
343 // that is not available in $languages array
344 if (!isset($languages[$sm_notAlias])) {
345 $sm_notAlias="en_US";
348 while (isset($languages[$sm_notAlias]['ALIAS'])) {
349 $sm_notAlias = $languages[$sm_notAlias]['ALIAS'];
352 if ( isset($sm_language) &&
354 $sm_language != '' &&
355 isset($languages[$sm_notAlias]['CHARSET']) ) {
356 sq_bindtextdomain( 'squirrelmail', SM_PATH
. 'locale/' );
357 sq_textdomain( 'squirrelmail' );
359 // set codeset in order to avoid gettext charset conversions
360 if (function_exists('bind_textdomain_codeset')) {
361 // Japanese translation uses different internal charset
362 if ($sm_notAlias == 'ja_JP') {
363 bind_textdomain_codeset ('squirrelmail', 'EUC-JP');
365 bind_textdomain_codeset ('squirrelmail', $languages[$sm_notAlias]['CHARSET'] );
369 // Use LOCALE key, if it is set.
370 if (isset($languages[$sm_notAlias]['LOCALE'])){
371 $longlocale=$languages[$sm_notAlias]['LOCALE'];
373 $longlocale=$sm_notAlias;
376 // try setting locale
377 $retlocale=sq_setlocale(LC_ALL
, $longlocale);
379 // check if locale is set and assign that locale to $longlocale
380 // in order to use it in putenv calls.
381 if (! is_bool($retlocale)) {
382 $longlocale=$retlocale;
383 } elseif (is_array($longlocale)) {
384 // setting of all locales failed.
385 // we need string instead of array used in LOCALE key.
386 $longlocale=$sm_notAlias;
389 if ( !((bool)ini_get('safe_mode')) &&
390 getenv( 'LC_ALL' ) != $longlocale ) {
391 putenv( "LC_ALL=$longlocale" );
392 putenv( "LANG=$longlocale" );
393 putenv( "LANGUAGE=$longlocale" );
394 putenv( "LC_NUMERIC=C" );
395 if ($sm_notAlias=='tr_TR') putenv( "LC_CTYPE=C" );
397 // Workaround for plugins that use numbers with floating point
398 // It might be removed if plugins use correct decimal delimiters
399 // according to locale settings.
400 setlocale(LC_NUMERIC
, 'C');
401 // Workaround for specific Turkish strtolower/strtoupper rules.
402 // Many functions expect English conversion rules.
403 if ($sm_notAlias=='tr_TR') setlocale(LC_CTYPE
,'C');
405 // Set text direction/alignment variables
406 // These don't appear to be used... are they safe to remove?
407 if (isset($languages[$sm_notAlias]['DIR']) &&
408 $languages[$sm_notAlias]['DIR'] == 'rtl') {
411 * @global string $text_direction
413 $text_direction='rtl';
416 * @global string $left_align
421 * @global string $right_align
425 $text_direction='ltr';
427 $right_align='right';
430 $squirrelmail_language = $sm_notAlias;
431 if ($squirrelmail_language == 'ja_JP') {
432 header ('Content-Type: text/html; charset=EUC-JP');
433 if (!function_exists('mb_internal_encoding')) {
434 // Error messages can't be displayed here
436 // Revert to English if possible.
437 if (function_exists('setPref') && $username!='' && $data_dir!="") {
438 setPref($data_dir, $username, 'language', "en_US");
441 // stop further execution in order not to get php errors on mb_internal_encoding().
444 if (function_exists('mb_language')) {
445 mb_language('Japanese');
447 mb_internal_encoding('EUC-JP');
448 mb_http_output('pass');
449 } elseif ($squirrelmail_language == 'en_US') {
450 header( 'Content-Type: text/html; charset=' . $default_charset );
452 header( 'Content-Type: text/html; charset=' . $languages[$sm_notAlias]['CHARSET'] );
455 * mbstring.func_overload fix (#929644).
457 * php mbstring extension can replace standard string functions with their multibyte
458 * equivalents. See http://www.php.net/ref.mbstring#mbstring.overload. This feature
459 * was added in php v.4.2.0
461 * Some SquirrelMail functions work with 8bit strings in bytes. If interface is forced
462 * to use mbstring functions and mbstring internal encoding is set to multibyte charset,
463 * interface can't trust regular string functions. Due to mbstring overloading design
464 * limits php scripts can't control this setting.
466 * This hack should fix some issues related to 8bit strings in passwords. Correct fix is
467 * to disable mbstring overloading. Japanese translation uses different internal encoding.
469 if ($squirrelmail_language != 'ja_JP' &&
470 function_exists('mb_internal_encoding') &&
471 check_php_version(4,2,0) &&
472 (int)ini_get('mbstring.func_overload')!=0) {
473 mb_internal_encoding('pass');
480 * Sets default_charset variable according to the one that is used by user's translations.
482 * Function changes global $default_charset variable in order to be sure, that it
483 * contains charset used by user's translation. Sanity of $squirrelmail_language
484 * and $default_charset combination is also tested.
486 * There can be a $default_charset setting in the
487 * config.php file, but the user may have a different language
488 * selected for a user interface. This function checks the
489 * language selected by the user and tags the outgoing messages
490 * with the appropriate charset corresponding to the language
491 * selection. This is "more right" (tm), than just stamping the
492 * message blindly with the system-wide $default_charset.
494 function set_my_charset(){
495 global $data_dir, $username, $default_charset, $languages, $squirrelmail_language;
497 $my_language = getPref($data_dir, $username, 'language');
499 $my_language = $squirrelmail_language ;
501 // Catch removed translation
502 if (!isset($languages[$my_language])) {
503 $my_language="en_US";
505 while (isset($languages[$my_language]['ALIAS'])) {
506 $my_language = $languages[$my_language]['ALIAS'];
508 $my_charset = $languages[$my_language]['CHARSET'];
509 if ($my_language!='en_US') {
510 $default_charset = $my_charset;
515 * Replaces non-braking spaces inserted by some browsers with regular space
517 * This function can be used to replace non-braking space symbols
518 * that are inserted in forms by some browsers instead of normal
521 * @param string $string Text that needs to be cleaned
522 * @param string $charset Charset used in text
523 * @return string Cleaned text
525 function cleanup_nbsp($string,$charset) {
527 // reduce number of case statements
528 if (stristr('iso-8859-',substr($charset,0,9))){
529 $output_charset="iso-8859-x";
531 if (stristr('windows-125',substr($charset,0,11))){
532 $output_charset="cp125x";
534 if (stristr('koi8',substr($charset,0,4))){
535 $output_charset="koi8-x";
537 if (! isset($output_charset)){
538 $output_charset=strtolower($charset);
541 // where is non-braking space symbol
542 switch($output_charset):
555 // don't change string if charset is unmatched
559 // return space instead of non-braking space.
560 return str_replace($nbsp,' ',$string);
564 * Function informs if it is safe to convert given charset to the one that is used by user.
566 * It is safe to use conversion only if user uses utf-8 encoding and when
567 * converted charset is similar to the one that is used by user.
569 * @param string $input_charset Charset of text that needs to be converted
570 * @return bool is it possible to convert to user's charset
572 function is_conversion_safe($input_charset) {
573 global $languages, $sm_notAlias, $default_charset, $lossy_encoding;
575 if (isset($lossy_encoding) && $lossy_encoding )
578 // convert to lower case
579 $input_charset = strtolower($input_charset);
581 // Is user's locale Unicode based ?
582 if ( $default_charset == "utf-8" ) {
586 // Charsets that are similar
587 switch ($default_charset):
589 if ( $input_charset == "iso-8859-5" ||
590 $input_charset == "koi8-r" ||
591 $input_charset == "koi8-u" ) {
597 if ( $input_charset == "iso-8859-13" ||
598 $input_charset == "iso-8859-4" ) {
604 if ( $input_charset == "iso-8859-13" ||
605 $input_charset == "windows-1257" ) {
611 if ( $input_charset == "windows-1251" ||
612 $input_charset == "koi8-r" ||
613 $input_charset == "koi8-u" ) {
619 if ( $input_charset == "iso-8859-4" ||
620 $input_charset == "windows-1257" ) {
626 if ( $input_charset == "windows-1251" ||
627 $input_charset == "iso-8859-5" ||
628 $input_charset == "koi8-u" ) {
634 if ( $input_charset == "windows-1251" ||
635 $input_charset == "iso-8859-5" ||
636 $input_charset == "koi8-r" ) {
647 /* ------------------------------ main --------------------------- */
649 global $squirrelmail_language, $languages, $use_gettext;
651 if (! sqgetGlobalVar('squirrelmail_language',$squirrelmail_language,SQ_COOKIE
)) {
652 $squirrelmail_language = '';
656 * Array specifies the available translations.
658 * Structure of array:
659 * $languages['language']['variable'] = 'value'
661 * Possible 'variable' names:
662 * NAME - Translation name in English
663 * CHARSET - Encoding used by translation
664 * ALIAS - used when 'language' is only short name and 'value' should provide long language name
665 * ALTNAME - Native translation name. Any 8bit symbols must be html encoded.
666 * LOCALE - Full locale name (in xx_XX.charset format). It can use array with more than one locale name since 1.4.5 and 1.5.1
667 * DIR - Text direction. Used to define Right-to-Left languages. Possible values 'rtl' or 'ltr'. If undefined - defaults to 'ltr'
668 * XTRA_CODE - translation uses special functions. See doc/i18n.txt
670 * Each 'language' definition requires NAME+CHARSET or ALIAS variables.
673 * @global array $languages
675 $languages['en_US']['NAME'] = 'English';
676 $languages['en_US']['CHARSET'] = 'iso-8859-1';
677 $languages['en_US']['LOCALE'] = 'en_US.ISO8859-1';
678 $languages['en']['ALIAS'] = 'en_US';
681 * Automatic translation loading from setup.php files.
682 * Solution for bug. 1240889.
683 * setup.php file can contain $languages array entries and XTRA_CODE functions.
685 if (is_dir(SM_PATH
. 'locale') &&
686 is_readable(SM_PATH
. 'locale')) {
687 $localedir = dir(SM_PATH
. 'locale');
688 while($lang_dir=$localedir->read()) {
689 // remove trailing slash, if present
690 if (substr($lang_dir,-1)=='/') {
691 $lang_dir = substr($lang_dir,0,-1);
693 if ($lang_dir != '..' && $lang_dir != '.' && $lang_dir != 'CVS' &&
694 is_dir(SM_PATH
.'locale/'.$lang_dir) &&
695 file_exists(SM_PATH
.'locale/'.$lang_dir.'/setup.php')) {
696 include_once(SM_PATH
.'locale/'.$lang_dir.'/setup.php');
702 /* Detect whether gettext is installed. */
704 if (function_exists('_')) {
707 if (function_exists('bindtextdomain')) {
710 if (function_exists('textdomain')) {
713 if (function_exists('ngettext')) {
717 /* If gettext is fully loaded, cool */
718 if ($gettext_flags == 15) {
722 /* If ngettext support is missing, load it */
723 elseif ($gettext_flags == 7) {
725 // load internal ngettext functions
726 include_once(SM_PATH
. 'class/l10n.class.php');
727 include_once(SM_PATH
. 'functions/ngettext.php');
730 /* If we can fake gettext, try that */
731 elseif ($gettext_flags == 0) {
733 include_once(SM_PATH
. 'functions/gettext.php');
735 /* Uh-ho. A weird install */
736 if (! $gettext_flags & 1) {
738 * Function is used as replacement in broken installs
745 if (! $gettext_flags & 2) {
747 * Function is used as replacement in broken installs
750 function bindtextdomain() {
754 if (! $gettext_flags & 4) {
756 * Function is used as replacemet in broken installs
759 function textdomain() {
763 if (! $gettext_flags & 8) {
765 * Function is used as replacemet in broken installs
768 function ngettext($str,$str2,$number) {
776 if (! function_exists('dgettext')) {
778 * Replacement for broken setups.
781 function dgettext($domain,$str) {
785 if (! function_exists('dngettext')) {
787 * Replacement for broken setups
790 function dngettext($domain,$str1,$strn,$number) {
791 return ($number==1 ?
$str1 : $strn);