[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#782960: closed by Jonathan Wiltshire <jmw@debian.org> (Re: Bug#782960: unblock: mariadb-10.0/10.0.17-1)



On 2015-04-21 6:41, Otto Kekäläinen wrote:
Hello!

2015-04-21 0:15 GMT+03:00 Debian Bug Tracking System <owner@bugs.debian.org>:
..
I don't see anything that's appropriate for a point release either, so I'm
going to close this bug.
[...]
Could you please reconsider?

For the MySQL and MariaDB packages point releases are very stable and
the normal way to upgrade packages in production environments (and
distributions). For this reason the MySQL and MariaDB packages have
been granted the micro release exception in Ubuntu
(https://wiki.ubuntu.com/StableReleaseUpdates/MicroReleaseExceptions)
as an example of distribution policy.

It's got very little to do with how good upstream are at stable releases and very much more to do with the fact that, as you're no doubt aware, we're now less than a week away from release and in deep freeze.

A point release will anyway be pushed via the security channel sooner
or later. I think it would make sense to push unblock mariadb-10.017-1
now directly.

A little earlier in the freeze, quite possibly.

Personally, I'm unconvinced that the GCC5 fix would have been appropriate at /any/ stage of this freeze. Fixes for failures with the default compiler, sure. Fixes for non-default compilers, possibly with a good rationale. A compiler that's not even in unstable, much less in the release, otoh...

I have CC'd the security team if they want to chip in and say that
letting 10.0.17-1 in is OK.

I'm going to assume there was something lost in translation in that statement. While the security team's input is always valuable and will often form part of our decision process, whether or not a package is suitable for unblock is ultimately the release team's decision.

Salvatore's confirmed that this will most likely get resolved via a release through the security archive, and that's of course fine. Personally at least the GCC5 change would be even less appropriate under those circumstances, but that's not my decision.

Regards,

Adam


Reply to: