bind_textdomain_codeset

jon at rejon dot org (2006-08-31 00:58:04)

So, by using this function and by setting the LANGUAGE variable, you should be able to have a functioning i18n system that is not dependent upon a system's locale installation. Here is a sample from a method I created for a language class. Look for it on http://wiki.creativecommons.org/Developer soon:
function Init ()
{
// set the LANGUAGE environmental variable
// This one for some reason makes a difference FU@#$%^&*!CK
// and when combined with bind_textdomain_codeset allows one
// to set locale independent of server locale setup!!!
if ( false == putenv("LANGUAGE=" . $this->_language ) )
CCDebug::Log(sprintf("Could not set the ENV variable LANGUAGE = %s",
$this->_language));
// set the LANG environmental variable
if ( false == putenv("LANG=" . $this->_language ) )
CCDebug::Log(sprintf("Could not set the ENV variable LANG = %s",
$this->_language));
// if locales are not installed in locale folder, they will not
// get set! This is usually in /usr/lib/locale
// Also, the backup language should always be the default language
// because of this...see the NOTE in the class description
// Try first what we want but with the .utf8, which is what the locale
// setting on most systems want (and is most compatible
// Then just try the standard lang encoding asked for, and then if
// all else fails, just try the default language
// LC_ALL is said to be used, but it has nasty usage in some languages
// in swapping commas and periods! Thus try LC_MESSAGE if on one of
// those systems.
// It is supposedly not defined on WINDOWS, so am including it here
// for possible uncommenting if a problem is shown
//
// if (!defined('LC_MESSAGES')) define('LC_MESSAGES', 6);
// yes, setlocale is case-sensitive...arg
$locale_set = setlocale(LC_ALL, $this->_language . ".utf8",
$this->_language . ".UTF8",
$this->_language . ".utf-8",
$this->_language . ".UTF-8",
$this->_language,
CC_LANG);
// if we don't get the setting we want, make sure to complain!
if ( ( $locale_set != $this->_language && CC_LANG == $locale_set) ||
empty($locale_set) )
{
CCDebug::Log(
sprintf("Tried: setlocale to '%s', but could only set to '%s'.", $this->_language, $locale_set) );
}
$bindtextdomain_set = bindtextdomain($this->_domain,
CC_LANG_LOCALE . "/" . $this->_locale_pref );
if ( empty($bindtextdomain_set) )
CCDebug::Log(
sprintf("Tried: bindtextdomain, '%s', to directory, '%s', " .
"but received '%s'",
$this->_domain, CC_LANG_LOCALE . "/" . $this->_locale_pref,
$bindtextdomain_set) );
bind_textdomain_codeset($this->_domain, "UTF-8");
$textdomain_set = textdomain($this->_domain);
if ( empty($textdomain_set) )
CCDebug::Log(sprintf("Tried: set textdomain to '%s', but got '%s'",
$this->_domain, $textdomain_set));
} // end of method Init ()

przemekkus at interia dot pl (2006-06-05 11:54:05)

I had problems with German "umlauts" when using gettext. So, this is how it can be resolved:
I've put these lines itom my PHP code:
$domain = "messages";
bindtextdomain($domain, "path_to_messages_dir");
bind_textdomain_codeset($domain, 'ISO-8859-15');
textdomain($domain);
It works!

duckx at mezimail dot com (2004-05-24 11:33:24)

First the the url of the gettext manual changed:
http://www.gnu.org/software/gettext/manual/

Secondly, lets explain a little bit what this fonction does.
By default, gettext will use the LC_CTYPE of the language you choose (for example fr_FR).
This LC_CTYPE is extracted from your locales.alias file in your configuration dir (Should be /etc/locales.alias).
By default, the encoding is frequently iso-8859-1.

So if you want to make your site utf-8 aware, you need to bind your domain with the right encoding.
Here is a sample:
<?php
   $locale="fr_FR.UTF-8"
   setlocale(LC_MESSAGES,  $locale);
   $domain = 'your_text_domain';
   bindtextdomain($domain, './translations_path');
   textdomain($domain);
   bind_textdomain_codeset($domain, 'UTF-8');

?>

As quoted in other notes, the translations path should be like
/translations_path
   /de_DE/
       /LC_MESSAGES
   /fr_FR/
       /LC_MESSAGES
   ...

Your translation goes in the LC_MESSAGES dirs ... Hopes this helps :)