Red Hat Bugzilla – Bug 999535
glibc should build an uz_UZ.UTF-8 locale, not only the Latin1 locales uz_UZ uz_UZ.iso88591
Last modified: 2016-11-24 10:57:04 EST
Version-Release number of selected component (if applicable):
Description of problem:
We have no UTF-8 locale for Uzbek written in Latin script:
$ locale -a | grep ^uz
$ LC_ALL=uz_UZ locale charmap
I.e., although we have an UTF-8 locale for Uzbek written in Cyrillic
script, the Uzbek locales for Latin script are available only in
But Latin script is more common nowadays for Uzbek, see:
CLDR also made Latin script the default for Uzbek recently, see
http://unicode.org/cldr/trac/ticket/6548 (Update country/language/pop for Uzbek)
http://unicode.org/cldr/trac/changeset/9155 (cldrbug 6548: Updated Uzbekistan Latin/Cyrillic? usage to better reflect modern education / usage.)
Therefore, the default locale chosen by Anaconda for Uzbek should
also use Latin script now, but it is not nice to use a non-UTF-8 locale
Therefore, we should also build an uz_UZ.UTF-8 locale.
Created attachment 788886 [details]
Patch for our glibc package in f19 and rawhide to build
an uz_UZ.UTF-8 locale.
I did a test build of glibc for f19 with the patch from comment#1 and
$ locale -a | grep ^uz
$ LANG=uz_UZ.UTF-8 locale charmap
$ LANG=uz_UZ.UTF-8 date
21 Avgust, 2013 yil, Chorshanba
The problem still exists in Fedora-20-Beta-TC6-x86_64-netinst.iso
Changes to SUPPORTED should happen upstream with consent from the entire community since they represent a firm commitment to provide support for that language in that specific format.
For Fedora to add entries to SUPPORTED would put us under a lot of pressure to support the language in that format without much help from upstream. What we really want to do is remain aligned with upstream. The support should come through upstream into Fedora.
I'm marking this as CLOSED / WONTFIX precisely because the change needs to happen upstream. Please file an upstream bug in sourceware.org tracker for glibc and we can get the ball rolling upstream. Once upstream has a patch you can reopen issues to backport the SUPPORTED changes.
(In reply to Mike FABIAN from comment #6)
Excellent! Thanks Mike. I've taken this upstream and asked Chris Leonard to comment. I hope I wasn't too blunt marking this CLOSED/WONTFIX, but some things need to be handled upstream before Fedora can easily commit to them.
I've adjusted the closed state to CLOSED/UPSTREAM since that's more accurate depiction of the issue.