Re: GNU libc
On Mar 6, Fernando wrote
> In article <email@example.com>, David Engel wrote:
> >The same holds true for the timezone. However, in this case, I do
> >know that several data files in the timezone package are currently
> >wrong, but are correct in the glibc timezones package.
> The data files are not wrong. According to the upstream documentation, there
> are two ways to generate them: one is POSIX-conformant but wrong and one is
> "right" (meaning it gives the actual time as defined by the International
> Earth Rotation Service:
> # The International Earth Rotation Service periodically uses leap seconds
> # to keep UTC to within 0.9 s of TAI (atomic time); see
> # Terry J Quinn, The BIPM and the accurate measure of time,
> # Proc IEEE 79, 7 (July 1991), 894-905.
> POSIX ignores the leap second issue and therefore the POSIX-conformant time is
OK, this at least explains the difference. I'm copying this to Ulrich
Drepper. I think you two should get together and decide which way is
the "best" way then.
> I had generated POSIX-conformant data files previously but I switched to the
> "right" ones in the latest release. I did not realize the consequences this
> change had for mirrors. As I am unsure of whether the "right" thing is the
> good one, I can revert to the old behavior if that is preferred.
> In any case, it seems that the timezone package could be obsoleted by the new
> timezones one which comes with glibc. Both use exactly the same upstream
> source, so it makes not much sense to keep both (out of sync...)
> I think David Engel and I should agree on which package to keep.
It probably doesn't make a big difference either way. There is a
slight benefit to using the version provided by glibc in that there is
one less upstream source to track.
David Engel ODS Networks
firstname.lastname@example.org 1001 E. Arapaho Road
(972) 234-6400 Richardson, TX 75081