Re: Replacement of the locales package
On Mon, Oct 25, 2004 at 10:52:51PM +0100, Edmund GRIMLEY EVANS wrote:
> > Ok. BTW did you file a bugreport to upstream bugzilla?
>
> Er, no. What's the URL for that, please?
http://sources.redhat.com/bugzilla
You need an account, then file a bug against the localedata component.
You can browse current bugs by clicking on "Query existing bug reports",
select "component" and hit "Search".
> > Some remarks about your locale:
> > * In LC_IDENTIFICATION, language is the 2-letter ISO 639 language code.
> > * In LC_NAME, %d is not defined, it should certainly be %p instead,
> > * In LC_ADDRESS, lang_lib is the 3-letter abbreviation from ISO 639-2.
One more point, lang_term should be defined here instead of lang_lib.
> > Any objection against this patch?
>
> No objection, but I'll have to trust you about changing %d to %p as I
> don't understand LC_NAME.
Nobody understands those locale files, but
http://www.student.uit.no/~pere/linux/glibc/
has nice pointers, You may also read
http://sources.redhat.com/ml/libc-alpha/2003-06/msg00184.html
this message explains what category is for in LC_IDENTIFICATION.
(I did not want to raise this issue in the previous post, but you
may fix these category lines too)
At this point, if your locale file follows ISO 14652, LC_NAME definition
can be found in
http://www.open-std.org/jtc1/sc22/wg20/docs/fcd14652.txt
section 4.9 (hence my remark about %d/%p).
As ISO 14652 is supposed to be backward compatible with POSIX, I suppose
that POSIX does not define %d in LC_NAME, but I have no reference.
> I currently have rather a lot of locales with %d in name_fmt:
>
> $ grep name_fmt /usr/share/i18n/locales/* | grep '<U0025><U0064>' | wc -l
> 110
That's funny, it looks like old locale files are wrong whereas newer
ones are right.
Denis
Reply to: