On 10/5/2011 12:27 PM, Corinna Vinschen wrote:
> On Oct 5 18:04, Erwin Waterlander wrote:
>> Op 4-10-2011 20:20, Corinna Vinschen schreef:
>>> On Oct 4 20:03, Erwin Waterlander wrote:
>>>> By the way, I noticed that with the default locale C.UTF-8 the
>>>> nl_langinfo(CODESET) C function<langinfo.h> returns wrongly
>>>> "ISO-8859-1",
>>> Not for me:
>>> [...]
>>
>> My program (wcd) uses gettext/libintl. Libintl is causing the
>> effect. Libintl is not working properly with a locale C.UTF-8. That
>> is a serious problem.
>
> That's a bug in libintl8 0.18.1.1-1. It does not happen with the
> previous version 0.17-11. Hopefully this gets fixed ASAP.
There are two issues with libintl 0.18.1.1-1, and apparently both are
contentious. I don't want to update libintl/gettext until both are
resolved.
One is the discovery that libintl is always doing a very expensive
"no-op" related to relocation (from "/usr/*" to "/usr/*") even when
--disable-relocation. The discussion of potential fixes is over on the
bug-gnulib mailing list.
The other is the issue that spawned this thread, which raised questions
about how basefiles::/etc/profile.d/lang.{sh,csh} should behave, and
other related complexities. IIRC we reached an impasse with this
subthread (and replies):
http://cygwin.com/ml/cygwin/2011-09/msg00063.html
See also the various messages in this thread, during the last day or two.
So...I'm rather stuck. I can't fix anything if we don't have a plan for
what the desired behavior IS. Right now, we all (except for Bruno!)
agree that $current_behavior is bad. But how exactly to fix it -- and
whether to do so in opposition to Bruno, the actual libintl maintainer
-- is still an open question.
--
Chuck
--
Problem reports: http://cygwin.com/problems.html
FAQ: http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
