[tz] [tz-announce] 2021b release of tz code and data available

Paul Eggert eggert at cs.ucla.edu
Tue Sep 28 18:38:13 UTC 2021


On 9/28/21 10:01, Paul Ganssle via tz wrote:
> For those on the list who re-package this, what did / do you plan to 
> release? A version using `PACKRATDATA=backzone`, or no?

I don't know of major distributions using PACKRATDATA=backzone, and as I 
mentioned earlier I wouldn't recommend switching to that option if 
you're worried about pre-1970 churn, as it would change many more 
pre-1970 timestamps than 2021b does. PACKRATDATA=backzone is intended 
mostly for astrologers and suchlike who want access to pre-1970 data 
even if it's out-of-scope and/or woefully incomplete and/or wrong.

For what it's worth, Red Hat rawhide rebased to 2021b on Saturday, and 
also included my followup patch to fix the Jan Mayen typo. See:

https://src.fedoraproject.org/rpms/tzdata/c/c1c7aed6464460ab060ed09cd87ad8723cd5d296

It's their normal practice to test this sort of thing extensively before 
distributing to Fedora and eventually to RHEL users, and they're quite 
good about letting us know of any issues that come up.

As Matthew noted, Debian cherry-picked the Samoa and Fiji changes, as 
well as the update to leap second expiration. Debian has cherry-picked 
in the past, and I'm not surprised they did it this time too given all 
the uproar on the mailing list. Consideration of packaging 2021b is one 
of Debian's high-priority items in their tracker list; see 
<https://tracker.debian.org/pkg/tzdata>.


More information about the tz mailing list