Re: RFC: target FreeBSD version for Jessie - FreeBSD 10.1 Release schedule
Hi,
On 13/07/14 08:34, Petr Salinger wrote:
> the FreeBSD 10.1 Release schedule have been announced [1], key dates:
Thanks for mentioning it, I've been waiting to hear this.
> Code freeze begins 5 September 2014
> releng/10.1 branch 3 October 2014
> RELEASE announcement 29 October 2014
A little more detail at:
http://www.freebsd.org/releases/10.1R/schedule.html
(By the way, the 9.3 release stayed perfectly on schedule this time.)
The timing looks a bit tighter than I hoped, but here's what I think:
We could put stable-10 snapshots (ALPHA builds) into experimental any
time. (Maybe we could try to do this more regularly, like a
monthly/weekly thing?) In September the BETA builds start, and we'd
want as many people as possible testing those.
RC builds from the releng/10.1 branch could go directly to sid in
October (set for 10 days before migration, to give more time to test?).
Whether it reaches testing, depends on any new problems we hit and how
quickly we can fix them.
Upstream shouldn't change the kernel much between the RC builds and
final release, unless there is some major problem to fix. If it affects
us, release-critical bugs can still be fixed for a while during the freeze:
https://release.debian.org/jessie/freeze_policy.html
> * pre-approved fixes (until the 5th of January 2015);
> * targeted fixes for release critical bugs (i.e., bugs of severity critical, grave, and serious) in all packages (applies during the entire freeze (ok for TPU));
I estimate the freeze might be quite long, unfortunately - remember that
GNU/Linux is doing its systemd integration.
Nearer the time, if we still only have a BETA or RC in sid and serious
bugs, we may decide to just stay with 10.0. If a BETA or RC is
*already* in testing, we may want to ask the release time about getting
a pre-approved unblock when the final release comes.
> a) Jessie with 10.0 codebase
> b) Jessie close to 10.1 codebase as much as possible
(In retrospect, some of the 10.x features most interesting to us have
been backported into 9.x to some extent... but let's move on).
> pros of 10.1 over 10.0:
> - better security support during Jessie lifetime:
> 10.0 eol [3] January 2015
> 10.1 will be probably Extended release, estimate is therefore
> November 2016
This is the most compelling reason for me to prefer 10.1. The jessie
release won't happen until 2015, and we'd like upstream security support
for as long as possible; backporting fixes ourselves is not always easy.
Regards,
--
Steven Chamberlain
steven@pyro.eu.org
Reply to: