[tz] Uruguay out of DST
eggert at cs.ucla.edu
Sun Jul 12 00:49:11 UTC 2015
On 07/11/2015 01:40 PM, John Hawkinson wrote:
> This seems to me to be somewhat of a fiction, and to me an unhealthy
> one. It is not the "unofficial repository." It is the "development
> repository from which releases are made."
It's both, of course. There's no fiction here.
The experimental github repository could be moved to another hosting
site tomorrow, without affecting the official releases. In this sense
it has the same role that the old SCCS repository did. The main
difference is that the github repository is world-readable whereas the
old SCCS repository was private.
> it would help if there was a monotonic identifier that increased for
> each commit in-between releases (VERSION=2015e+20150701a perhaps,
> corresponding to 9ebb8da281487886d4110388eb139578c72d1176, the first
> commit on July 1 (UTC)). I don't know if that is something reasonable
> to require of Paul. (And, unfortunately, I don't think there's a good
> git mechanism for it.)
In the current experimental repository a monotonic identifier is the
commit timestamp in UTC.
> If a vendor could reliably say, "I can expect a tzdata release to
> follow within 14 days after a change to a rule that takes effect 14-90
> days out, within 30 days after a change to a rule that takes effect
> 90-180 days out, ..."
Many releases are special cases, and I doubt whether tzdata
redistributors could rely on such a specific formula in practice. The
best we could say would be something along the lines of "we'll try to
publish a new release a week before the change takes effect, but we
can't promise this because sometimes governments give us even less
notice than that".
If redistributors can't broadcast changes to their users within a few
days, their users will lose out on occasion. This would be true even if
the tz project generated new releases with zero delay, so there's
nothing the tz project can do to "fix" this. What needs to be fixed is
the processes some redistributors use to propagate tz updates to their
users' machines. Three months is far too long a delay. Updates clearly
can be done reasonably quickly, within a matter of days not months, as
shown by the redistributors like Ubuntu who are doing so.
Yes, Apple and Google have far more users than Ubuntu but they can scale
to those users. Yes, they have to deal with ICU but so do Ubuntu and
the other guys. Yes, Google must communicate to its own downstream
redistributors, but this problem can be addressed (see Debian and
Ubuntu). None of these problems are fundamental obstacles to getting
changes out far more quickly than what happens now, and this quickness
is needed regardless of what the tz project does.
More information about the tz