itojun@iijlab.net wrote about Re: -current setlocale() problems:
:>	yes. there's some weird sideeffect ni setlocale(LC_ALL).  it is
:>	a bug.  i'll look into it soon.  thanks.
:	lib/libc/locale/setlocale.c 1.31 should fix it.
:	please test and report.

Works fine.  Failed properly.  I had a bit of a problem trying to test
with a success until I noticed I was doing it on Big5 encoding (oops -
forgot that multibyte wasn't ready yet :-) )
Noticed this reverses the error checks from v1.29 but I presume that's
a side effect of your return value changes from the other routines(?).

Another question someone may be able to answer:  why are directories in
/usr/share/locale without encoding methods (eg. en_AU) being created?
Ah, may have thought of the answer.  They're sitting there for some
potential LC_MESSAGES files, right?  Though shouldn't we be creating
those directories when the parents are created?

:#	we need some regression test suite.

Yes.  What's the general rule about those who suggest such things become
responsible for them?  :-)

Regards,
-- 
Geoff Wing : <gcw@pobox.com>
Rxvt Stuff : <gcw@rxvt.org>
Zsh Stuff  : <gcw@zsh.org>