>>>>> "Georg" == Georg Baum <[EMAIL PROTECTED]> writes:
Georg> Mainly that not all has been converted yet. Some strings go Georg> unchanged from the file to memory. I tried to describe that Georg> here: http://wiki.lyx.org/devel/pmwiki.php/Devel/Unicode Something that is not very clear to me is whether we are going to force gettext to produce utf-8 always. What if the user's locale is iso-8859-1? In this case, it seems to me that the debug info output to console should be in latin1, shouldn't they? Also, I just discovered the existence of bind_textdomain_codeset; isn't it something that would be useful (can we set codeset to ucs-4?) The bind_textdomain_codeset function sets the output codeset for mes- sage catalogs for domain domainname. A message domain is a set of translatable msgid messages. Usually, every software package has its own message domain. By default, the gettext family of functions returns translated messages in the locale's character encoding, which can be retrieved as nl_lang- info(CODESET). The need for calling bind_textdomain_codeset arises for programs which store strings in a locale independent way (e.g. UTF-8) and want to avoid an extra character set conversion on the returned translated messages. JMarc
