Re: Back-branch releases upcoming this week

As has been mentioned a couple times, we're well overdue for updates of
the back branches. Seems like time to get that done, so we'll be
wrapping 8.2.x and up this Thursday for release Monday the 26th.

regards, tom lane

On Tue, 2011-09-20 at 01:37 -0400, Tom Lane wrote:
> As has been mentioned a couple times, we're well overdue for updates
> of the back branches. Seems like time to get that done, so we'll be
> wrapping 8.2.x and up this Thursday for release Monday the 26th.

Can we also specify a final release version for 8.2? This set will be
8.2.21, and I propose to EOL 8.2 as of 8.2.22.

Regards,
--
Devrim GÜNDÜZ
Principal Systems Engineer @ EnterpriseDB: http://www.enterprisedb.com
PostgreSQL Danışmanı/Consultant, Red Hat Certified Engineer
Community: devrim~PostgreSQL.org, devrim.gunduz~linux.org.tr
http://www.gunduz.org Twitter: http://twitter.com/devrimgunduz

On Tue, Sep 20, 2011 at 12:37 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> As has been mentioned a couple times, we're well overdue for updates of
> the back branches.  Seems like time to get that done, so we'll be
> wrapping 8.2.x and up this Thursday for release Monday the 26th.

8.2 up, including 9.1.1? I'm not sure our QA guys will be able to cope
with verification of so many individual installers in that timeframe -
8.2 - 9.0 is hard enough to do in one go.

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On 09/20/2011 02:46 AM, Devrim GÜNDÜZ wrote:
> On Tue, 2011-09-20 at 01:37 -0400, Tom Lane wrote:
>> As has been mentioned a couple times, we're well overdue for updates
>> of the back branches. Seems like time to get that done, so we'll be
>> wrapping 8.2.x and up this Thursday for release Monday the 26th.
> Can we also specify a final release version for 8.2? This set will be
> 8.2.21, and I propose to EOL 8.2 as of 8.2.22.
>
>

I don't see why we should deviate from the policy at
<http://wiki.postgresql.org/wiki/PostgreSQL_Release_Support_Policy>. In
all probability, that means there will be one more release for 8.2 after
this, but I don't think we need to determine that now.

cheers

andrew

Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
> On 09/20/2011 02:46 AM, Devrim GNDZ wrote:
>> Can we also specify a final release version for 8.2? This set will be
>> 8.2.21, and I propose to EOL 8.2 as of 8.2.22.

> I don't see why we should deviate from the policy at
> <http://wiki.postgresql.org/wiki/PostgreSQL_Release_Support_Policy>. In
> all probability, that means there will be one more release for 8.2 after
> this, but I don't think we need to determine that now.

Per that policy, there will certainly be at least one more 8.2.x
release, namely the first one "after December 2011". There could be
more than one, if we are pressed into making a set of releases between
now and December.

I don't think we've yet decided what the policy means if a release
happens during the stated calendar month, which seems rather likely
this time around in view of our historical record of doing updates
roughly quarterly. Should we settle that detail now? That is,
does "after December" really mean "in or after December", or did we
really mean "after"?

regards, tom lane

On 09/20/2011 10:28 AM, Tom Lane wrote:
>
> I don't think we've yet decided what the policy means if a release
> happens during the stated calendar month, which seems rather likely
> this time around in view of our historical record of doing updates
> roughly quarterly. Should we settle that detail now? That is,
> does "after December" really mean "in or after December", or did we
> really mean "after"?
>
>

If we really want to get that specific, let's just say that the EOL date
is at the end of the designated month.

cheers

andrew

2011/9/20 Andrew Dunstan <andrew(at)dunslane(dot)net>:
>
>
> On 09/20/2011 10:28 AM, Tom Lane wrote:
>>
>> I don't think we've yet decided what the policy means if a release
>> happens during the stated calendar month, which seems rather likely
>> this time around in view of our historical record of doing updates
>> roughly quarterly.  Should we settle that detail now?  That is,
>> does "after December" really mean "in or after December", or did we
>> really mean "after"?
>>
>>
>
> If we really want to get that specific, let's just say that the EOL date is
> at the end of the designated month.

+1

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Dave Page <dpage(at)pgadmin(dot)org> writes:
> 2011/9/20 Andrew Dunstan <andrew(at)dunslane(dot)net>:
>> On 09/20/2011 10:28 AM, Tom Lane wrote:
>>> does "after December" really mean "in or after December", or did we
>>> really mean "after"?

>> If we really want to get that specific, let's just say that the EOL date is
>> at the end of the designated month.

> +1

OK, so "after" really means "after", ie, the last 8.2.x release will be
timestamped 2012-something. Fine with me.

regards, tom lane

2011/9/20 Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>:
> Dave Page <dpage(at)pgadmin(dot)org> writes:
>> 2011/9/20 Andrew Dunstan <andrew(at)dunslane(dot)net>:
>>> On 09/20/2011 10:28 AM, Tom Lane wrote:
>>>> does "after December" really mean "in or after December", or did we
>>>> really mean "after"?
>
>>> If we really want to get that specific, let's just say that the EOL date is
>>> at the end of the designated month.
>
>> +1
>
> OK, so "after" really means "after", ie, the last 8.2.x release will be
> timestamped 2012-something.  Fine with me.

It's unfortunate, but it seems to me it's the only interpretation that
doesn't risk taking someone by surprise.

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Dave Page <dpage(at)pgadmin(dot)org> writes:
> On Tue, Sep 20, 2011 at 12:37 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> As has been mentioned a couple times, we're well overdue for updates of
>> the back branches. Seems like time to get that done, so we'll be
>> wrapping 8.2.x and up this Thursday for release Monday the 26th.

> 8.2 up, including 9.1.1? I'm not sure our QA guys will be able to cope
> with verification of so many individual installers in that timeframe -
> 8.2 - 9.0 is hard enough to do in one go.

Well, all the pre-9.1 branches are definitely badly in need of updates.
9.1 maybe could go without at this point, but we do have one crasher bug
and one serious memory leak fixed there, neither new in 9.1. I'd just
as soon not establish a precedent for not releasing the same fixes at
the same time in all branches.

How about we wrap them all, but you could let your team slip the 9.1
update for a day or so if they need more time? It's certainly less
critical than the older branches.

regards, tom lane

On Tue, Sep 20, 2011 at 3:57 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Dave Page <dpage(at)pgadmin(dot)org> writes:
>> On Tue, Sep 20, 2011 at 12:37 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>>> As has been mentioned a couple times, we're well overdue for updates of
>>> the back branches.  Seems like time to get that done, so we'll be
>>> wrapping 8.2.x and up this Thursday for release Monday the 26th.
>
>> 8.2 up, including 9.1.1? I'm not sure our QA guys will be able to cope
>> with verification of so many individual installers in that timeframe -
>> 8.2 - 9.0 is hard enough to do in one go.
>
> Well, all the pre-9.1 branches are definitely badly in need of updates.
> 9.1 maybe could go without at this point, but we do have one crasher bug
> and one serious memory leak fixed there, neither new in 9.1.  I'd just
> as soon not establish a precedent for not releasing the same fixes at
> the same time in all branches.
>
> How about we wrap them all, but you could let your team slip the 9.1
> update for a day or so if they need more time?  It's certainly less
> critical than the older branches.

OK, well we can push the installers much more quickly over the CDN anyway.

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Tom Lane wrote:
> Dave Page <dpage(at)pgadmin(dot)org> writes:
>> On Tue, Sep 20, 2011 at 12:37 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>>> As has been mentioned a couple times, we're well overdue for updates of
>>> the back branches. �Seems like time to get that done, so we'll be
>>> wrapping 8.2.x and up this Thursday for release Monday the 26th.
>
>> 8.2 up, including 9.1.1? I'm not sure our QA guys will be able to cope
>> with verification of so many individual installers in that timeframe -
>> 8.2 - 9.0 is hard enough to do in one go.
>
> Well, all the pre-9.1 branches are definitely badly in need of updates.
> 9.1 maybe could go without at this point, but we do have one crasher bug
> and one serious memory leak fixed there, neither new in 9.1. I'd just
> as soon not establish a precedent for not releasing the same fixes at
> the same time in all branches.
>
> How about we wrap them all, but you could let your team slip the 9.1
> update for a day or so if they need more time? It's certainly less
> critical than the older branches.

I would prefer that all branches have synchronized patch releases as they seem
to have had in the past, and that the latest production is included, 9.1.1 in
this case, even if its change set is more minor. -- Darren Duncan

On tis, 2011-09-20 at 10:28 -0400, Tom Lane wrote:
> I don't think we've yet decided what the policy means if a release
> happens during the stated calendar month, which seems rather likely
> this time around in view of our historical record of doing updates
> roughly quarterly. Should we settle that detail now? That is,
> does "after December" really mean "in or after December", or did we
> really mean "after"?

The policy states that we will support branches until that time, which
should reasonably include the whole time period stated. That is, 8.2 is
supported until December 31.

But it does not say that we are obliged to make another release after
the EOL with all the patches that have accumulated between the previous
release and the EOL. And it certainly does not say that we are obliged
to keep patching after EOL until that next release happens. It does say
that that would "normally" happen, but it doesn't have to. Previously,
we have argued that we should make another release because the previous
patching effort would otherwise have been wasted. Maybe so. But let's
keep this in perspective. If we made another release on December 13, we
shouldn't have to keep patching after that, unless there is an
emergency.