include/languages.php

   1 <?php
   2
   3 /**
   4  * SquirrelMail internationalization functions
   5  *
   6  * This file contains variuos functions that are needed to do
   7  * internationalization of SquirrelMail.
   8  *
   9  * Internally the output character set is used. Other characters are
  10  * encoded using Unicode entities according to HTML 4.0.
  11  *
  12  * Before 1.5.2 functions were stored in functions/i18n.php. Script is moved
  13  * because it executes some code in order to detect functions supported by
  14  * existing PHP installation and implements fallback functions when required
  15  * functions are not available. Scripts in functions/ directory should not
  16  * setup anything when they are loaded.
  17  * @copyright &copy; 1999-2007 The SquirrelMail Project Team
  18  * @license http://opensource.org/licenses/gpl-license.php GNU Public License
  19  * @version $Id$
  20  * @package squirrelmail
  21  * @subpackage i18n
  22  */
  23
  24
  25 /**
  26  * Wrapper for textdomain(), bindtextdomain() and
  27  * bind_textdomain_codeset() primarily intended for
  28  * plugins when changing into their own text domain
  29  * and back again.
  30  *
  31  * Note that if plugins using this function have
  32  * their translation files located in the SquirrelMail
  33  * locale directory, the second argument is optional.
  34  *
  35  * @param string $domain_name The name of the text domain
  36  *                            (usually the plugin name, or
  37  *                            "squirrelmail") being switched to.
  38  * @param string $directory   The directory that contains
  39  *                            all translations for the domain
  40  *                            (OPTIONAL; default is SquirrelMail
  41  *                            locale directory).
  42  *
  43  * @return void
  44  *
  45  * @since 1.5.2 and 1.4.10
  46  */
  47 function sq_change_text_domain($domain_name, $directory='') {
  48
  49     if (empty($directory)) $directory = SM_PATH . 'locale/';
  50
  51     static $domains_already_seen = array();
  52
  53     // only need to call bindtextdomain() once
  54     //
  55     if (in_array($domain_name, $domains_already_seen)) {
  56         sq_textdomain($domain_name);
  57         return;
  58     }
  59
  60     $domains_already_seen[] = $domain_name;
  61
  62     sq_bindtextdomain($domain_name, $directory);
  63     sq_textdomain($domain_name);
  64
  65 }
  66
  67 /**
  68  * Gettext bindtextdomain wrapper.
  69  *
  70  * Wrapper solves differences between php versions in order to provide
  71  * ngettext support. Should be used if translation uses ngettext
  72  * functions.
  73  *
  74  * This also provides a bind_textdomain_codeset call to make sure the
  75  * domain's encoding will not be overridden.
  76  *
  77  * @since 1.4.10 and 1.5.1
  78  * @param string $domain gettext domain name
  79  * @param string $dir directory that contains all translations
  80  * @return string path to translation directory
  81  */
  82 function sq_bindtextdomain($domain,$dir) {
  83     global $l10n, $gettext_flags, $sm_notAlias;
  84
  85     if ($gettext_flags==7) {
  86         // gettext extension without ngettext
  87         if (substr($dir, -1) != '/') $dir .= '/';
  88         $mofile=$dir . $sm_notAlias . '/LC_MESSAGES/' . $domain . '.mo';
  89         $input = new FileReader($mofile);
  90         $l10n[$domain] = new gettext_reader($input);
  91     }
  92
  93     $dir=bindtextdomain($domain,$dir);
  94
  95     // set codeset in order to avoid gettext charset conversions
  96     if (function_exists('bind_textdomain_codeset')
  97      && isset($languages[$sm_notAlias]['CHARSET'])) {
  98
  99         // Japanese translation uses different internal charset
 100         if ($sm_notAlias == 'ja_JP') {
 101             bind_textdomain_codeset ($domain_name, 'EUC-JP');
 102         } else {
 103             bind_textdomain_codeset ($domain_name, $languages[$sm_notAlias]['CHARSET']);
 104         }
 105
 106     }
 107
 108     return $dir;
 109 }
 110
 111 /**
 112  * Gettext textdomain wrapper.
 113  * Makes sure that gettext_domain global is modified.
 114  * @since 1.5.1
 115  * @param string $name gettext domain name
 116  * @return string gettext domain name
 117  */
 118 function sq_textdomain($domain) {
 119     global $gettext_domain;
 120     $gettext_domain=textdomain($domain);
 121     return $gettext_domain;
 122 }
 123
 124 /**
 125  * php setlocale function wrapper
 126  *
 127  * From php 4.3.0 it is possible to use arrays in order to set locale.
 128  * php gettext extension works only when locale is set. This wrapper
 129  * function allows to use more than one locale name.
 130  *
 131  * @param int $category locale category name. Use php named constants
 132  *     (LC_ALL, LC_COLLATE, LC_CTYPE, LC_MONETARY, LC_NUMERIC, LC_TIME)
 133  * @param mixed $locale option contains array with possible locales or string with one locale
 134  * @return string name of set locale or false, if all locales fail.
 135  * @since 1.5.1 and 1.4.5
 136  * @see http://www.php.net/setlocale
 137  */
 138 function sq_setlocale($category,$locale) {
 139     if (is_string($locale)) {
 140         // string with only one locale
 141         $ret = setlocale($category,$locale);
 142     } elseif (! check_php_version(4,3)) {
 143         // older php version (second setlocale argument must be string)
 144         $ret=false;
 145         $index=0;
 146         while ( ! $ret && $index<count($locale)) {
 147             $ret=setlocale($category,$locale[$index]);
 148             $index++;
 149         }
 150     } else {
 151         // php 4.3.0 or better, use entire array
 152         $ret=setlocale($category,$locale);
 153     }
 154
 155     /* safety checks */
 156     if (preg_match("/^.*\/.*\/.*\/.*\/.*\/.*$/",$ret)) {
 157         /**
 158          * Welcome to We-Don't-Follow-Own-Fine-Manual department
 159          * OpenBSD 3.8, 3.9-current and maybe later versions
 160          * return invalid response to setlocale command.
 161          * SM bug report #1427512.
 162          */
 163         $ret = false;
 164     }
 165     return $ret;
 166 }
 167
 168 /**
 169  * Converts string from given charset to charset, that can be displayed by user translation.
 170  *
 171  * Function by default returns html encoded strings, if translation uses different encoding.
 172  * If Japanese translation is used - function returns string converted to euc-jp
 173  * If iconv or recode functions are enabled and translation uses utf-8 - function returns utf-8 encoded string.
 174  * If $charset is not supported - function returns unconverted string.
 175  *
 176  * sanitizing of html tags is also done by this function.
 177  *
 178  * @param string $charset
 179  * @param string $string Text to be decoded
 180  * @param boolean $force_decode converts string to html without $charset!=$default_charset check.
 181  * Argument is available since 1.5.1 and 1.4.5.
 182  * @param boolean $save_html disables htmlspecialchars() in order to preserve
 183  *  html formating. Use with care. Available since 1.5.1
 184  * @return string decoded string
 185  */
 186 function charset_decode ($charset, $string, $force_decode=false, $save_html=false) {
 187     global $languages, $squirrelmail_language, $default_charset;
 188     global $use_php_recode, $use_php_iconv, $aggressive_decoding;
 189
 190     if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
 191         function_exists($languages[$squirrelmail_language]['XTRA_CODE'] . '_decode')) {
 192         $string = call_user_func($languages[$squirrelmail_language]['XTRA_CODE'] . '_decode', $string);
 193     }
 194
 195     $charset = strtolower($charset);
 196
 197     set_my_charset();
 198
 199     // Variables that allow to use functions without function_exist() calls
 200     if (! isset($use_php_recode) || $use_php_recode=="" ) {
 201         $use_php_recode=false; }
 202     if (! isset($use_php_iconv) || $use_php_iconv=="" ) {
 203         $use_php_iconv=false; }
 204
 205     // Don't do conversion if charset is the same.
 206     if ( ! $force_decode && $charset == strtolower($default_charset) )
 207         return ($save_html ? $string : htmlspecialchars($string));
 208
 209     // catch iso-8859-8-i thing
 210     if ( $charset == "iso-8859-8-i" )
 211         $charset = "iso-8859-8";
 212
 213     /*
 214      * Recode converts html special characters automatically if you use
 215      * 'charset..html' decoding. There is no documented way to put -d option
 216      * into php recode function call.
 217      */
 218     if ( $use_php_recode ) {
 219         if ( $default_charset == "utf-8" ) {
 220             // other charsets can be converted to utf-8 without loss.
 221             // and output string is smaller
 222             $string = recode_string($charset . "..utf-8",$string);
 223             return ($save_html ? $string : htmlspecialchars($string));
 224         } else {
 225             $string = recode_string($charset . "..html",$string);
 226             // recode does not convert single quote, htmlspecialchars does.
 227             $string = str_replace("'", '&#039;', $string);
 228             // undo html specialchars
 229             if ($save_html)
 230                 $string=str_replace(array('&amp;','&quot;','&lt;','&gt;'),
 231                                     array('&','"','<','>'),$string);
 232             return $string;
 233         }
 234     }
 235
 236     // iconv functions does not have html target and can be used only with utf-8
 237     if ( $use_php_iconv && $default_charset=='utf-8') {
 238         $string = iconv($charset,$default_charset,$string);
 239         return ($save_html ? $string : htmlspecialchars($string));
 240     }
 241
 242     // If we don't use recode and iconv, we'll do it old way.
 243
 244     /* All HTML special characters are 7 bit and can be replaced first */
 245     if (! $save_html) $string = htmlspecialchars ($string);
 246
 247     /* controls cpu and memory intensive decoding cycles */
 248     if (! isset($aggressive_decoding) || $aggressive_decoding=="" ) {
 249         $aggressive_decoding=false; }
 250
 251     $decode=fixcharset($charset);
 252     $decodefile=SM_PATH . 'functions/decode/' . $decode . '.php';
 253     if ($decode != 'index' && file_exists($decodefile)) {
 254         include_once($decodefile);
 255         // send $save_html argument to decoding function. needed for iso-2022-xx decoding.
 256         $ret = call_user_func('charset_decode_'.$decode, $string, $save_html);
 257     } else {
 258         $ret = $string;
 259     }
 260     return( $ret );
 261 }
 262
 263 /**
 264  * Converts html string to given charset
 265  * @since 1.5.1 and 1.4.4
 266  * @param string $string
 267  * @param string $charset
 268  * @param boolean $htmlencode keep htmlspecialchars encoding
 269  * @return string
 270  */
 271 function charset_encode($string,$charset,$htmlencode=true) {
 272     global $default_charset;
 273
 274     $encode=fixcharset($charset);
 275     $encodefile=SM_PATH . 'functions/encode/' . $encode . '.php';
 276     if ($encode != 'index' && file_exists($encodefile)) {
 277         include_once($encodefile);
 278         $ret = call_user_func('charset_encode_'.$encode, $string);
 279     } elseif(file_exists(SM_PATH . 'functions/encode/us_ascii.php')) {
 280         // function replaces all 8bit html entities with question marks.
 281         // it is used when other encoding functions are unavailable
 282         include_once(SM_PATH . 'functions/encode/us_ascii.php');
 283         $ret = charset_encode_us_ascii($string);
 284     } else {
 285         /**
 286          * fix for yahoo users that remove all us-ascii related things
 287          */
 288         $ret = $string;
 289     }
 290
 291     /**
 292      * Undo html special chars, some places (like compose form) have
 293      * own sanitizing functions and don't need html symbols.
 294      * Undo chars only after encoding in order to prevent conversion of
 295      * html entities in plain text emails.
 296      */
 297     if (! $htmlencode ) {
 298         $ret = str_replace(array('&amp;','&gt;','&lt;','&quot;'),array('&','>','<','"'),$ret);
 299     }
 300     return( $ret );
 301 }
 302
 303 /**
 304  * Combined decoding and encoding functions
 305  *
 306  * If conversion is done to charset different that utf-8, unsupported symbols
 307  * will be replaced with question marks.
 308  * @since 1.5.1 and 1.4.4
 309  * @param string $in_charset initial charset
 310  * @param string $string string that has to be converted
 311  * @param string $out_charset final charset
 312  * @param boolean $htmlencode keep htmlspecialchars encoding
 313  * @return string converted string
 314  */
 315 function charset_convert($in_charset,$string,$out_charset,$htmlencode=true) {
 316     $string=charset_decode($in_charset,$string,true);
 317     $string=sqi18n_convert_entities($string);
 318     $string=charset_encode($string,$out_charset,$htmlencode);
 319     return $string;
 320 }
 321
 322 /**
 323  * Makes charset name suitable for decoding cycles
 324  *
 325  * @since 1.5.0 and 1.4.4
 326  * @param string $charset Name of charset
 327  * @return string $charset Adjusted name of charset
 328  */
 329 function fixcharset($charset) {
 330     /* remove minus and characters that might be used in paths from charset
 331      * name in order to be able to use it in function names and include calls.
 332      */
 333     $charset=preg_replace("/[-:.\/\\\]/",'_',$charset);
 334
 335     // OE ks_c_5601_1987 > cp949
 336     $charset=str_replace('ks_c_5601_1987','cp949',$charset);
 337     // Moz x-euc-tw > euc-tw
 338     $charset=str_replace('x_euc','euc',$charset);
 339     // Moz x-windows-949 > cp949
 340     $charset=str_replace('x_windows_','cp',$charset);
 341
 342     // windows-125x and cp125x charsets
 343     $charset=str_replace('windows_','cp',$charset);
 344
 345     // ibm > cp
 346     $charset=str_replace('ibm','cp',$charset);
 347
 348     // iso-8859-8-i -> iso-8859-8
 349     // use same cycle until I'll find differences
 350     $charset=str_replace('iso_8859_8_i','iso_8859_8',$charset);
 351
 352     return $charset;
 353 }
 354
 355 /**
 356  * Set up the language to be output
 357  * if $do_search is true, then scan the browser information
 358  * for a possible language that we know
 359  *
 360  * Function sets system locale environment (LC_ALL, LANG, LANGUAGE),
 361  * gettext translation bindings and html header information.
 362  *
 363  * Function returns error codes, if there is some fatal error.
 364  *  0 = no error,
 365  *  1 = mbstring support is not present,
 366  *  2 = mbstring support is not present, user's translation reverted to en_US.
 367  *
 368  * @param string $sm_language translation used by user's interface
 369  * @param bool $do_search use browser's preferred language detection functions. Defaults to false.
 370  * @param bool $default set $sm_language to $squirrelmail_default_language if language detection fails or language is not set. Defaults to false.
 371  * @return int function execution error codes.
 372  */
 373 function set_up_language($sm_language, $do_search = false, $default = false) {
 374
 375     static $SetupAlready = 0;
 376     global $use_gettext, $languages,
 377            $squirrelmail_language, $squirrelmail_default_language, $default_charset,
 378            $sm_notAlias, $username, $data_dir;
 379
 380     if ($SetupAlready) {
 381         return;
 382     }
 383
 384     $SetupAlready = TRUE;
 385     sqgetGlobalVar('HTTP_ACCEPT_LANGUAGE',  $accept_lang, SQ_SERVER);
 386
 387     /**
 388      * If function is asked to detect preferred language
 389      *  OR squirrelmail default language is set to empty string
 390      *    AND
 391      * squirrelmail language ($sm_language) is empty string
 392      * (not set in user's prefs and no cookie with language info)
 393      *    AND
 394      * browser provides list of preferred languages
 395      *  THEN
 396      * get preferred language from HTTP_ACCEPT_LANGUAGE header
 397      */
 398     if (($do_search || empty($squirrelmail_default_language)) &&
 399         ! $sm_language &&
 400         isset($accept_lang)) {
 401         // TODO: use more than one language, if first language is not available
 402         // FIXME: function assumes that string contains two or more characters.
 403         // FIXME: some languages use 5 chars
 404         $sm_language = substr($accept_lang, 0, 2);
 405     }
 406
 407     /**
 408      * If language preference is not set OR script asks to use default language
 409      *  AND
 410      * default squirrelmail language is not set to empty string
 411      *  THEN
 412      * use default squirrelmail language value from configuration.
 413      */
 414     if ((!$sm_language||$default) &&
 415         ! empty($squirrelmail_default_language)) {
 416         $squirrelmail_language = $squirrelmail_default_language;
 417         $sm_language = $squirrelmail_default_language;
 418     }
 419
 420     /** provide failsafe language when detection fails */
 421     if (! $sm_language) $sm_language='en_US';
 422
 423     $sm_notAlias = $sm_language;
 424
 425     // Catching removed translation
 426     // System reverts to English translation if user prefs contain translation
 427     // that is not available in $languages array
 428     if (!isset($languages[$sm_notAlias])) {
 429         $sm_notAlias="en_US";
 430     }
 431
 432     while (isset($languages[$sm_notAlias]['ALIAS'])) {
 433         $sm_notAlias = $languages[$sm_notAlias]['ALIAS'];
 434     }
 435
 436     if ( isset($sm_language) &&
 437          $use_gettext &&
 438          $sm_language != '' &&
 439          isset($languages[$sm_notAlias]['CHARSET']) ) {
 440         sq_bindtextdomain( 'squirrelmail', SM_PATH . 'locale/' );
 441         sq_textdomain( 'squirrelmail' );
 442
 443         // Use LOCALE key, if it is set.
 444         if (isset($languages[$sm_notAlias]['LOCALE'])){
 445             $longlocale=$languages[$sm_notAlias]['LOCALE'];
 446         } else {
 447             $longlocale=$sm_notAlias;
 448         }
 449
 450         // try setting locale
 451         $retlocale=sq_setlocale(LC_ALL, $longlocale);
 452
 453         // check if locale is set and assign that locale to $longlocale
 454         // in order to use it in putenv calls.
 455         if (! is_bool($retlocale)) {
 456             $longlocale=$retlocale;
 457         } elseif (is_array($longlocale)) {
 458             // setting of all locales failed.
 459             // we need string instead of array used in LOCALE key.
 460             $longlocale=$sm_notAlias;
 461         }
 462
 463         if ( !((bool)ini_get('safe_mode')) &&
 464              getenv( 'LC_ALL' ) != $longlocale ) {
 465             putenv( "LC_ALL=$longlocale" );
 466             putenv( "LANG=$longlocale" );
 467             putenv( "LANGUAGE=$longlocale" );
 468             putenv( "LC_NUMERIC=C" );
 469             if ($sm_notAlias=='tr_TR') putenv( "LC_CTYPE=C" );
 470         }
 471         // Workaround for plugins that use numbers with floating point
 472         // It might be removed if plugins use correct decimal delimiters
 473         // according to locale settings.
 474         setlocale(LC_NUMERIC, 'C');
 475         // Workaround for specific Turkish strtolower/strtoupper rules.
 476         // Many functions expect English conversion rules.
 477         if ($sm_notAlias=='tr_TR') setlocale(LC_CTYPE,'C');
 478
 479         /**
 480          * Set text direction/alignment variables
 481          * When language environment is setup, scripts can use these globals
 482          * without accessing $languages directly and making checks for optional
 483          * array key.
 484          */
 485         global $text_direction, $left_align, $right_align;
 486         if (isset($languages[$sm_notAlias]['DIR']) &&
 487             $languages[$sm_notAlias]['DIR'] == 'rtl') {
 488             /**
 489              * Text direction
 490              * @global string $text_direction
 491              */
 492             $text_direction='rtl';
 493             /**
 494              * Left alignment
 495              * @global string $left_align
 496              */
 497             $left_align='right';
 498             /**
 499              * Right alignment
 500              * @global string $right_align
 501              */
 502             $right_align='left';
 503         } else {
 504             $text_direction='ltr';
 505             $left_align='left';
 506             $right_align='right';
 507         }
 508
 509         $squirrelmail_language = $sm_notAlias;
 510         if ($squirrelmail_language == 'ja_JP') {
 511             header ('Content-Type: text/html; charset=EUC-JP');
 512             if (!function_exists('mb_internal_encoding')) {
 513                 // Error messages can't be displayed here
 514                 $error = 1;
 515                 // Revert to English if possible.
 516                 if (function_exists('setPref')  && $username!='' && $data_dir!="") {
 517                     setPref($data_dir, $username, 'language', "en_US");
 518                     $error = 2;
 519                 }
 520                 // stop further execution in order not to get php errors on mb_internal_encoding().
 521                 return $error;
 522             }
 523             if (function_exists('mb_language')) {
 524                 mb_language('Japanese');
 525             }
 526             mb_internal_encoding('EUC-JP');
 527             mb_http_output('pass');
 528         } elseif ($squirrelmail_language == 'en_US') {
 529             header( 'Content-Type: text/html; charset=' . $default_charset );
 530         } else {
 531             header( 'Content-Type: text/html; charset=' . $languages[$sm_notAlias]['CHARSET'] );
 532         }
 533         /**
 534          * mbstring.func_overload fix (#929644).
 535          *
 536          * php mbstring extension can replace standard string functions with their multibyte
 537          * equivalents. See http://www.php.net/ref.mbstring#mbstring.overload. This feature
 538          * was added in php v.4.2.0
 539          *
 540          * Some SquirrelMail functions work with 8bit strings in bytes. If interface is forced
 541          * to use mbstring functions and mbstring internal encoding is set to multibyte charset,
 542          * interface can't trust regular string functions. Due to mbstring overloading design
 543          * limits php scripts can't control this setting.
 544          *
 545          * This hack should fix some issues related to 8bit strings in passwords. Correct fix is
 546          * to disable mbstring overloading. Japanese translation uses different internal encoding.
 547          */
 548         if ($squirrelmail_language != 'ja_JP' &&
 549             function_exists('mb_internal_encoding') &&
 550             check_php_version(4,2,0) &&
 551             (int)ini_get('mbstring.func_overload')!=0) {
 552             mb_internal_encoding('pass');
 553         }
 554     }
 555     return 0;
 556 }
 557
 558 /**
 559  * Sets default_charset variable according to the one that is used by user's translations.
 560  *
 561  * Function changes global $default_charset variable in order to be sure, that it
 562  * contains charset used by user's translation. Sanity of $squirrelmail_language
 563  * and $default_charset combination is also tested.
 564  *
 565  * There can be a $default_charset setting in the
 566  * config.php file, but the user may have a different language
 567  * selected for a user interface. This function checks the
 568  * language selected by the user and tags the outgoing messages
 569  * with the appropriate charset corresponding to the language
 570  * selection. This is "more right" (tm), than just stamping the
 571  * message blindly with the system-wide $default_charset.
 572  */
 573 function set_my_charset(){
 574     global $data_dir, $username, $default_charset, $languages, $squirrelmail_language;
 575
 576     $my_language = getPref($data_dir, $username, 'language');
 577     if (!$my_language) {
 578         $my_language = $squirrelmail_language ;
 579     }
 580     // Catch removed translation
 581     if (!isset($languages[$my_language])) {
 582         $my_language="en_US";
 583     }
 584     while (isset($languages[$my_language]['ALIAS'])) {
 585         $my_language = $languages[$my_language]['ALIAS'];
 586     }
 587     $my_charset = $languages[$my_language]['CHARSET'];
 588     if ($my_language!='en_US') {
 589         $default_charset = $my_charset;
 590     }
 591 }
 592
 593 /**
 594  * Replaces non-braking spaces inserted by some browsers with regular space
 595  *
 596  * This function can be used to replace non-braking space symbols
 597  * that are inserted in forms by some browsers instead of normal
 598  * space symbol.
 599  *
 600  * @param string $string Text that needs to be cleaned
 601  * @param string $charset Charset used in text
 602  * @return string Cleaned text
 603  */
 604 function cleanup_nbsp($string,$charset) {
 605
 606   // reduce number of case statements
 607   if (stristr('iso-8859-',substr($charset,0,9))){
 608     $output_charset="iso-8859-x";
 609   }
 610   if (stristr('windows-125',substr($charset,0,11))){
 611     $output_charset="cp125x";
 612   }
 613   if (stristr('koi8',substr($charset,0,4))){
 614     $output_charset="koi8-x";
 615   }
 616   if (! isset($output_charset)){
 617     $output_charset=strtolower($charset);
 618   }
 619
 620 // where is non-braking space symbol
 621 switch($output_charset):
 622  case "iso-8859-x":
 623  case "cp125x":
 624  case "iso-2022-jp":
 625   $nbsp="\xA0";
 626   break;
 627  case "koi8-x":
 628    $nbsp="\x9A";
 629    break;
 630  case "utf-8":
 631    $nbsp="\xC2\xA0";
 632    break;
 633  default:
 634    // don't change string if charset is unmatched
 635    return $string;
 636 endswitch;
 637
 638 // return space instead of non-braking space.
 639  return str_replace($nbsp,' ',$string);
 640 }
 641
 642 /**
 643  * Function informs if it is safe to convert given charset to the one that is used by user.
 644  *
 645  * It is safe to use conversion only if user uses utf-8 encoding and when
 646  * converted charset is similar to the one that is used by user.
 647  *
 648  * @param string $input_charset Charset of text that needs to be converted
 649  * @return bool is it possible to convert to user's charset
 650  */
 651 function is_conversion_safe($input_charset) {
 652     global $languages, $sm_notAlias, $default_charset, $lossy_encoding;
 653
 654     if (isset($lossy_encoding) && $lossy_encoding )
 655         return true;
 656
 657     // convert to lower case
 658     $input_charset = strtolower($input_charset);
 659
 660     // Is user's locale Unicode based ?
 661     if ( $default_charset == "utf-8" ) {
 662         return true;
 663     }
 664
 665     // Charsets that are similar
 666     switch ($default_charset) {
 667     case "windows-1251":
 668         if ( $input_charset == "iso-8859-5" ||
 669              $input_charset == "koi8-r" ||
 670              $input_charset == "koi8-u" ) {
 671             return true;
 672         } else {
 673             return false;
 674         }
 675     case "windows-1257":
 676         if ( $input_charset == "iso-8859-13" ||
 677              $input_charset == "iso-8859-4" ) {
 678             return true;
 679         } else {
 680             return false;
 681         }
 682     case "iso-8859-4":
 683         if ( $input_charset == "iso-8859-13" ||
 684              $input_charset == "windows-1257" ) {
 685             return true;
 686         } else {
 687             return false;
 688         }
 689     case "iso-8859-5":
 690         if ( $input_charset == "windows-1251" ||
 691              $input_charset == "koi8-r" ||
 692              $input_charset == "koi8-u" ) {
 693             return true;
 694         } else {
 695             return false;
 696         }
 697     case "iso-8859-13":
 698         if ( $input_charset == "iso-8859-4" ||
 699              $input_charset == "windows-1257" ) {
 700             return true;
 701         } else {
 702             return false;
 703         }
 704     case "koi8-r":
 705         if ( $input_charset == "windows-1251" ||
 706              $input_charset == "iso-8859-5" ||
 707              $input_charset == "koi8-u" ) {
 708             return true;
 709         } else {
 710             return false;
 711         }
 712     case "koi8-u":
 713         if ( $input_charset == "windows-1251" ||
 714              $input_charset == "iso-8859-5" ||
 715              $input_charset == "koi8-r" ) {
 716             return true;
 717         } else {
 718             return false;
 719         }
 720     default:
 721         return false;
 722     }
 723 }
 724
 725 /**
 726  * Converts html character entities to numeric entities
 727  *
 728  * SquirrelMail encoding functions work only with numeric entities.
 729  * This function fixes issues with decoding functions that might convert
 730  * some symbols to character entities. Issue is specific to PHP recode
 731  * extension decoding. Function is used internally in charset_convert()
 732  * function.
 733  * @param string $str string that might contain html character entities
 734  * @return string string with character entities converted to decimals.
 735  * @since 1.5.2
 736  */
 737 function sqi18n_convert_entities($str) {
 738
 739     $entities = array(
 740         // Latin 1
 741         '&nbsp;'   => '&#160;',
 742         '&iexcl;'  => '&#161;',
 743         '&cent;'   => '&#162;',
 744         '&pound;'  => '&#163;',
 745         '&curren;' => '&#164;',
 746         '&yen;'    => '&#165;',
 747         '&brvbar;' => '&#166;',
 748         '&sect;'   => '&#167;',
 749         '&uml;'    => '&#168;',
 750         '&copy;'   => '&#169;',
 751         '&ordf;'   => '&#170;',
 752         '&laquo;'  => '&#171;',
 753         '&not;'    => '&#172;',
 754         '&shy;'    => '&#173;',
 755         '&reg;'    => '&#174;',
 756         '&macr;'   => '&#175;',
 757         '&deg;'    => '&#176;',
 758         '&plusmn;' => '&#177;',
 759         '&sup2;'   => '&#178;',
 760         '&sup3;'   => '&#179;',
 761         '&acute;'  => '&#180;',
 762         '&micro;'  => '&#181;',
 763         '&para;'   => '&#182;',
 764         '&middot;' => '&#183;',
 765         '&cedil;'  => '&#184;',
 766         '&sup1;'   => '&#185;',
 767         '&ordm;'   => '&#186;',
 768         '&raquo;'  => '&#187;',
 769         '&frac14;' => '&#188;',
 770         '&frac12;' => '&#189;',
 771         '&frac34;' => '&#190;',
 772         '&iquest;' => '&#191;',
 773         '&Agrave;' => '&#192;',
 774         '&Aacute;' => '&#193;',
 775         '&Acirc;'  => '&#194;',
 776         '&Atilde;' => '&#195;',
 777         '&Auml;'   => '&#196;',
 778         '&Aring;'  => '&#197;',
 779         '&AElig;'  => '&#198;',
 780         '&Ccedil;' => '&#199;',
 781         '&Egrave;' => '&#200;',
 782         '&Eacute;' => '&#201;',
 783         '&Ecirc;'  => '&#202;',
 784         '&Euml;'   => '&#203;',
 785         '&Igrave;' => '&#204;',
 786         '&Iacute;' => '&#205;',
 787         '&Icirc;'  => '&#206;',
 788         '&Iuml;'   => '&#207;',
 789         '&ETH;'    => '&#208;',
 790         '&Ntilde;' => '&#209;',
 791         '&Ograve;' => '&#210;',
 792         '&Oacute;' => '&#211;',
 793         '&Ocirc;'  => '&#212;',
 794         '&Otilde;' => '&#213;',
 795         '&Ouml;'   => '&#214;',
 796         '&times;'  => '&#215;',
 797         '&Oslash;' => '&#216;',
 798         '&Ugrave;' => '&#217;',
 799         '&Uacute;' => '&#218;',
 800         '&Ucirc;'  => '&#219;',
 801         '&Uuml;'   => '&#220;',
 802         '&Yacute;' => '&#221;',
 803         '&THORN;'  => '&#222;',
 804         '&szlig;'  => '&#223;',
 805         '&agrave;' => '&#224;',
 806         '&aacute;' => '&#225;',
 807         '&acirc;'  => '&#226;',
 808         '&atilde;' => '&#227;',
 809         '&auml;'   => '&#228;',
 810         '&aring;'  => '&#229;',
 811         '&aelig;'  => '&#230;',
 812         '&ccedil;' => '&#231;',
 813         '&egrave;' => '&#232;',
 814         '&eacute;' => '&#233;',
 815         '&ecirc;'  => '&#234;',
 816         '&euml;'   => '&#235;',
 817         '&igrave;' => '&#236;',
 818         '&iacute;' => '&#237;',
 819         '&icirc;'  => '&#238;',
 820         '&iuml;'   => '&#239;',
 821         '&eth;'    => '&#240;',
 822         '&ntilde;' => '&#241;',
 823         '&ograve;' => '&#242;',
 824         '&oacute;' => '&#243;',
 825         '&ocirc;'  => '&#244;',
 826         '&otilde;' => '&#245;',
 827         '&ouml;'   => '&#246;',
 828         '&divide;' => '&#247;',
 829         '&oslash;' => '&#248;',
 830         '&ugrave;' => '&#249;',
 831         '&uacute;' => '&#250;',
 832         '&ucirc;'  => '&#251;',
 833         '&uuml;'   => '&#252;',
 834         '&yacute;' => '&#253;',
 835         '&thorn;'  => '&#254;',
 836         '&yuml;'   => '&#255;',
 837         // Latin Extended-A
 838         '&OElig;'  => '&#338;',
 839         '&oelig;'  => '&#339;',
 840         '&Scaron;' => '&#352;',
 841         '&scaron;' => '&#353;',
 842         '&Yuml;'   => '&#376;',
 843         // Spacing Modifier Letters
 844         '&circ;'   => '&#710;',
 845         '&tilde;'  => '&#732;',
 846         // General Punctuation
 847         '&ensp;'   => '&#8194;',
 848         '&emsp;'   => '&#8195;',
 849         '&thinsp;' => '&#8201;',
 850         '&zwnj;'   => '&#8204;',
 851         '&zwj;'    => '&#8205;',
 852         '&lrm;'    => '&#8206;',
 853         '&rlm;'    => '&#8207;',
 854         '&ndash;'  => '&#8211;',
 855         '&mdash;'  => '&#8212;',
 856         '&lsquo;'  => '&#8216;',
 857         '&rsquo;'  => '&#8217;',
 858         '&sbquo;'  => '&#8218;',
 859         '&ldquo;'  => '&#8220;',
 860         '&rdquo;'  => '&#8221;',
 861         '&bdquo;'  => '&#8222;',
 862         '&dagger;' => '&#8224;',
 863         '&Dagger;' => '&#8225;',
 864         '&permil;' => '&#8240;',
 865         '&lsaquo;' => '&#8249;',
 866         '&rsaquo;' => '&#8250;',
 867         '&euro;'   => '&#8364;',
 868         // Latin Extended-B
 869         '&fnof;' => '&#402;',
 870         // Greek
 871         '&Alpha;'  => '&#913;',
 872         '&Beta;'   => '&#914;',
 873         '&Gamma;'  => '&#915;',
 874         '&Delta;'  => '&#916;',
 875         '&Epsilon;' => '&#917;',
 876         '&Zeta;'   => '&#918;',
 877         '&Eta;'    => '&#919;',
 878         '&Theta;'  => '&#920;',
 879         '&Iota;'   => '&#921;',
 880         '&Kappa;'  => '&#922;',
 881         '&Lambda;' => '&#923;',
 882         '&Mu;'     => '&#924;',
 883         '&Nu;'     => '&#925;',
 884         '&Xi;'     => '&#926;',
 885         '&Omicron;' => '&#927;',
 886         '&Pi;'     => '&#928;',
 887         '&Rho;'    => '&#929;',
 888         '&Sigma;'  => '&#931;',
 889         '&Tau;'    => '&#932;',
 890         '&Upsilon;' => '&#933;',
 891         '&Phi;'    => '&#934;',
 892         '&Chi;'    => '&#935;',
 893         '&Psi;'    => '&#936;',
 894         '&Omega;'  => '&#937;',
 895         '&alpha;'  => '&#945;',
 896         '&beta;'   => '&#946;',
 897         '&gamma;'  => '&#947;',
 898         '&delta;'  => '&#948;',
 899         '&epsilon;' => '&#949;',
 900         '&zeta;'   => '&#950;',
 901         '&eta;'    => '&#951;',
 902         '&theta;'  => '&#952;',
 903         '&iota;'   => '&#953;',
 904         '&kappa;'  => '&#954;',
 905         '&lambda;' => '&#955;',
 906         '&mu;'     => '&#956;',
 907         '&nu;'     => '&#957;',
 908         '&xi;'     => '&#958;',
 909         '&omicron;' => '&#959;',
 910         '&pi;'     => '&#960;',
 911         '&rho;'    => '&#961;',
 912         '&sigmaf;' => '&#962;',
 913         '&sigma;'  => '&#963;',
 914         '&tau;'    => '&#964;',
 915         '&upsilon;' => '&#965;',
 916         '&phi;'    => '&#966;',
 917         '&chi;'    => '&#967;',
 918         '&psi;'    => '&#968;',
 919         '&omega;'  => '&#969;',
 920         '&thetasym;' => '&#977;',
 921         '&upsih;'  => '&#978;',
 922         '&piv;'    => '&#982;',
 923         // General Punctuation
 924         '&bull;'   => '&#8226;',
 925         '&hellip;' => '&#8230;',
 926         '&prime;'  => '&#8242;',
 927         '&Prime;'  => '&#8243;',
 928         '&oline;'  => '&#8254;',
 929         '&frasl;'  => '&#8260;',
 930         // Letterlike Symbols
 931         '&weierp;' => '&#8472;',
 932         '&image;'  => '&#8465;',
 933         '&real;'   => '&#8476;',
 934         '&trade;'  => '&#8482;',
 935         '&alefsym;' => '&#8501;',
 936         // Arrows
 937         '&larr;'   => '&#8592;',
 938         '&uarr;'   => '&#8593;',
 939         '&rarr;'   => '&#8594;',
 940         '&darr;'   => '&#8595;',
 941         '&harr;'   => '&#8596;',
 942         '&crarr;'  => '&#8629;',
 943         '&lArr;'   => '&#8656;',
 944         '&uArr;'   => '&#8657;',
 945         '&rArr;'   => '&#8658;',
 946         '&dArr;'   => '&#8659;',
 947         '&hArr;'   => '&#8660;',
 948         // Mathematical Operators
 949         '&forall;' => '&#8704;',
 950         '&part;'   => '&#8706;',
 951         '&exist;'  => '&#8707;',
 952         '&empty;'  => '&#8709;',
 953         '&nabla;'  => '&#8711;',
 954         '&isin;'   => '&#8712;',
 955         '&notin;'  => '&#8713;',
 956         '&ni;'     => '&#8715;',
 957         '&prod;'   => '&#8719;',
 958         '&sum;'    => '&#8721;',
 959         '&minus;'  => '&#8722;',
 960         '&lowast;' => '&#8727;',
 961         '&radic;'  => '&#8730;',
 962         '&prop;'   => '&#8733;',
 963         '&infin;'  => '&#8734;',
 964         '&ang;'    => '&#8736;',
 965         '&and;'    => '&#8743;',
 966         '&or;'     => '&#8744;',
 967         '&cap;'    => '&#8745;',
 968         '&cup;'    => '&#8746;',
 969         '&int;'    => '&#8747;',
 970         '&there4;' => '&#8756;',
 971         '&sim;'    => '&#8764;',
 972         '&cong;'   => '&#8773;',
 973         '&asymp;'  => '&#8776;',
 974         '&ne;'     => '&#8800;',
 975         '&equiv;'  => '&#8801;',
 976         '&le;'     => '&#8804;',
 977         '&ge;'     => '&#8805;',
 978         '&sub;'    => '&#8834;',
 979         '&sup;'    => '&#8835;',
 980         '&nsub;'   => '&#8836;',
 981         '&sube;'   => '&#8838;',
 982         '&supe;'   => '&#8839;',
 983         '&oplus;'  => '&#8853;',
 984         '&otimes;' => '&#8855;',
 985         '&perp;'   => '&#8869;',
 986         '&sdot;'   => '&#8901;',
 987         // Miscellaneous Technical
 988         '&lceil;'  => '&#8968;',
 989         '&rceil;'  => '&#8969;',
 990         '&lfloor;' => '&#8970;',
 991         '&rfloor;' => '&#8971;',
 992         '&lang;'   => '&#9001;',
 993         '&rang;'   => '&#9002;',
 994         // Geometric Shapes
 995         '&loz;'    => '&#9674;',
 996         // Miscellaneous Symbols
 997         '&spades;' => '&#9824;',
 998         '&clubs;'  => '&#9827;',
 999         '&hearts;' => '&#9829;',
1000         '&diams;'  => '&#9830;');
1001
1002     $str = str_replace(array_keys($entities), array_values($entities), $str);
1003
1004     return $str;
1005 }
1006
1007 /* ------------------------------ main --------------------------- */
1008
1009 global $squirrelmail_language, $languages, $use_gettext;
1010
1011 if (! sqgetGlobalVar('squirrelmail_language',$squirrelmail_language,SQ_COOKIE)) {
1012     $squirrelmail_language = '';
1013 }
1014
1015 /**
1016  * Array specifies the available translations.
1017  *
1018  * Structure of array:
1019  * $languages['language']['variable'] = 'value'
1020  *
1021  * Possible 'variable' names:
1022  *  NAME      - Translation name in English
1023  *  CHARSET   - Encoding used by translation
1024  *  ALIAS     - used when 'language' is only short name and 'value' should provide long language name
1025  *  ALTNAME   - Native translation name. Any 8bit symbols must be html encoded.
1026  *  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
1027  *  DIR       - Text direction. Used to define Right-to-Left languages. Possible values 'rtl' or 'ltr'. If undefined - defaults to 'ltr'
1028  *  XTRA_CODE - translation uses special functions. See doc/i18n.txt
1029  *
1030  * Each 'language' definition requires NAME+CHARSET or ALIAS variables.
1031  *
1032  * @name $languages
1033  * @global array $languages
1034  */
1035 $languages['en_US']['NAME']    = 'English';
1036 $languages['en_US']['CHARSET'] = 'iso-8859-1';
1037 $languages['en_US']['LOCALE']  = 'en_US.ISO8859-1';
1038 $languages['en']['ALIAS'] = 'en_US';
1039
1040 /**
1041  * Automatic translation loading from setup.php files.
1042  * Solution for bug. 1240889.
1043  * setup.php file can contain $languages array entries and XTRA_CODE functions.
1044  */
1045 if (is_dir(SM_PATH . 'locale') &&
1046     is_readable(SM_PATH . 'locale')) {
1047     $localedir = dir(SM_PATH . 'locale');
1048     while($lang_dir=$localedir->read()) {
1049         // remove trailing slash, if present
1050         if (substr($lang_dir,-1)=='/') {
1051             $lang_dir = substr($lang_dir,0,-1);
1052         }
1053         if ($lang_dir != '..' && $lang_dir != '.' && $lang_dir != 'CVS' &&
1054             $lang_dir != '.svn' && is_dir(SM_PATH.'locale/'.$lang_dir) &&
1055             file_exists(SM_PATH.'locale/'.$lang_dir.'/setup.php')) {
1056             include_once(SM_PATH.'locale/'.$lang_dir.'/setup.php');
1057         }
1058     }
1059     $localedir->close();
1060 }
1061
1062 /* Detect whether gettext is installed. */
1063 $gettext_flags = 0;
1064 if (function_exists('_')) {
1065     $gettext_flags += 1;
1066 }
1067 if (function_exists('bindtextdomain')) {
1068     $gettext_flags += 2;
1069 }
1070 if (function_exists('textdomain')) {
1071     $gettext_flags += 4;
1072 }
1073 if (function_exists('ngettext')) {
1074     $gettext_flags += 8;
1075 }
1076
1077 /* If gettext is fully loaded, cool */
1078 if ($gettext_flags == 15) {
1079     $use_gettext = true;
1080 }
1081
1082 /* If ngettext support is missing, load it */
1083 elseif ($gettext_flags == 7) {
1084     $use_gettext = true;
1085     // load internal ngettext functions
1086     include_once(SM_PATH . 'class/l10n.class.php');
1087     include_once(SM_PATH . 'functions/ngettext.php');
1088 }
1089
1090 /* If we can fake gettext, try that */
1091 elseif ($gettext_flags == 0) {
1092     $use_gettext = true;
1093     include_once(SM_PATH . 'functions/gettext.php');
1094 } else {
1095     /* Uh-ho.  A weird install */
1096     if (! $gettext_flags & 1) {
1097       /**
1098        * Function is used as replacement in broken installs
1099        * @ignore
1100        */
1101         function _($str) {
1102             return $str;
1103         }
1104     }
1105     if (! $gettext_flags & 2) {
1106       /**
1107        * Function is used as replacement in broken installs
1108        * @ignore
1109        */
1110         function bindtextdomain() {
1111             return;
1112         }
1113     }
1114     if (! $gettext_flags & 4) {
1115       /**
1116        * Function is used as replacemet in broken installs
1117        * @ignore
1118        */
1119         function textdomain() {
1120             return;
1121         }
1122     }
1123     if (! $gettext_flags & 8) {
1124         /**
1125          * Function is used as replacemet in broken installs
1126          * @ignore
1127          */
1128         function ngettext($str,$str2,$number) {
1129             if ($number>1) {
1130                 return $str2;
1131             } else {
1132                 return $str;
1133             }
1134         }
1135     }
1136     if (! function_exists('dgettext')) {
1137         /**
1138          * Replacement for broken setups.
1139          * @ignore
1140          */
1141         function dgettext($domain,$str) {
1142             return $str;
1143         }
1144     }
1145     if (! function_exists('dngettext')) {
1146         /**
1147          * Replacement for broken setups
1148          * @ignore
1149          */
1150         function dngettext($domain,$str1,$strn,$number) {
1151             return ($number==1 ? $str1 : $strn);
1152         }
1153     }
1154 }