[dev] Question marks and problems with gettext reencoding...

[Romans Malinovskis]

-- Monday, November 19, 2001 -- 4:00:59 PM -- Rudi Benkoviè ----------------
>> not yet, but my own modified code works very funny... it randomly
>> changes charset sometimes...
>> sometimes  it  change it in the middle of the page.. could this be
>> because of multiple paralel httpd
>> processes?
> I wrote that little patch for PHP. Needless to say, it doesn't work for me.
> Google found a little tip:
> http://mail.nl.linux.org/linux-utf8/2000-10/msg00002.html ... And yes, my devel
> server is running glibc 2.1.92. Fun. :)
> I'll try to do a foo translation in UTF-8 and/or upgrade box running RH7.2 to
> PHP from CVS.
> Anyone else care to test this instead of me? If only I could buy some time...
> 1kg will do. ;)
i use distribution i assembled myself. I didn't install glibc's "locales" since i never predicted i
will need one, but now i do. (I use glibc 2.2.3, i suspect latter versions will work similary)

localedef -v -c -i ru_RU /usr/share/locale/ru_win
helps  to  create  /usr/share/locale/ru_RU/LC_*,  which  is  needed  in order for gettext to work.

imp/po/README describes how to test that.

debian  &  earler glibc's use /etc/locale.gen and /usr/sbin/locale-gen for that, but i couldn't test
that.

BIND_TEXTDOMAIN_CODESET  is needed only for broken locale configuration, setlocale(LC_ALL should set
em all (as gnu-guys say in docs..)


I would be happy if there would be a workaraund for bind_textdomain_codeset...

romans



> --

> Rudi