Another RC?, was: Re: Akonadi-Nepomuk Feeder Improvements

[Scott Kitterman]

Jos Poortvliet <jos at opensuse.org> wrote:

>On Sunday 30 December 2012 20:55:22 Scott Kitterman wrote:
>> On Monday, December 31, 2012 12:24:18 AM Sebastian Kügler wrote:
>> > On Sunday, December 30, 2012 02:41:43 Scott Kitterman wrote:
>> > > Assuming 4.10.1 and 4.10.2 slip similarly, that would result in
>the
>> > > next
>> > > Kubuntu release having 4.10.1 instead of 4.10.2.  We might also
>have
>> > > to
>> > > make  some adjustments to our internal testing milestone
>schedule.
>> > > 
>> > > 4.10.2 would come out a day or two before Kubuntu 13.04 (far to
>late
>> > > for
>> > > pre- release updates) so we'd get to release without the current
>KDE
>> > > SC.  We do ship the point releases as post-release updates, so
>they
>> > > would
>> > > get to users eventually, but post-release QA is a lot more work
>for
>> > > us.
>> > > 
>> > > This isn't precisely a problem, but changing the release cycle
>now is
>> > > not
>> > > idea  for us.  As long as 4.11 drops back into the usual time
>slot
>> > > (and
>> > > doesn't also slip), then the impact would not be major.
>> > 
>> > Would adding in two instead of three weeks work for you?
>> 
>> No.  Looking at the schedule, it would need to be only a week later
>than
>> the current schedule to make it.  If the 4.10.0 -> 4.10.1 and 4.10.1
>->
>> 4.10.2 intervals were slightly reduced, maybe we could get there?
>> 
>> Scott K
>
>Just out of curiosity - how big a deal is it to ship with 4.10.1
>instead of 
>4.10.2? Does it matter, other than the number? As a user, I'm rather
>fond of 
>having a much more usable KDE PIM in 4.10... I can't judge the
>technical 
>side of this argument but the impact of crappy Nepomuk<->KDEPIM
>interaction 
>is still very noticeable in the RC...

It's difficult to say as it mostly depends on what fixes are in 4.10.2 and that's not known yet.  We did release with a .1 before (4.7, IIRC) and we expended significant effort cherry picking before the release immediately followed by a QA cycle for 4.7.2 as a post release update.   

So, in my one experience,  releasing with a .1 resulted in a noticeably poorer experience for users and significant work for the Kubuntu team as packagers/testers.  I'd rather avoid it.

If we can move 4.10.0 by three weeks and .1/.2 by two and one respectively,  then (as Martin suggests in a later mail) Kubuntu should be fine.  

It's difficult for me to make a firm recommendation on this as I'm not clear on the benefits of the proposed changes.  I do generally think it's better to adjust the schedule if these late adds are allowed in.

Scott K