Re: [Qemu-devel] Reminder: we're now in softfreeze

[Peter Maydell]

On 26 October 2015 at 09:56, Mark Cave-Ayland
 wrote:
> Maybe we also need a [PING for-2.5] or similar? I've just sent another
> couple of pings for patches (one of which fixes the broken
> analyze-migration.py script) which have fallen through the cracks over
> the past couple of months. It would be nice to have an easy way to flag
> outstanding patches like this, especially during freeze when things get
> really busy.

That's what I'm hoping the wiki page will work for. Unlike the
mailing list, it's very short and easy to scan through :-)

-- PMM

Re: [Qemu-devel] Reminder: we're now in softfreeze

[Peter Crosthwaite]

On Thu, Oct 22, 2015 at 3:30 AM, Peter Maydell  wrote:
> [Apologies for the huge cc list, which is basically "everybody
> I have accepted a pullreq from this release cycle.]
>
> Just a reminder that we're now in softfreeze (ie "no new big
> features, start the process of cutting down towards only bugfixes
> going into master"). Hardfreeze will be on the 12th November,
> in three weeks' time. If submaintainers can aim to get pull
> requests in early rather than all on the 12th that would be
> nice, otherwise we get a big logjam on rc0 day and your
> chances of not getting new-feature code in go up sharply.
>
> For this release I'd like to try tracking known-issues on
> the wiki page: http://wiki.qemu.org/Planning/2.5
> with a list of patch series or bugs which need to be fixed
> into the release. This will hopefully avoid the problem we
> had last time around where I missed patches that should have
> gone into an rc.

Can patches fix this with the right form of query? Maybe another piece
of metadata for patches that are promoted for rc merge.

Regards,
Peter

The idea is that anybody (patch series author,
> bug submitter, submaintainer) can add to the list. Then we
> can annotate it with status (rejected as an rc bug, in a
> submaintainer tree, applied to master, etc).
>
> thanks
> -- PMM
>

Re: [Qemu-devel] Reminder: we're now in softfreeze

[Mark Cave-Ayland]

On 26/10/15 09:27, Peter Maydell wrote:

> On 26 October 2015 at 06:25, Peter Crosthwaite
>  wrote:
>> On Thu, Oct 22, 2015 at 3:30 AM, Peter Maydell  
>> wrote:
>>> For this release I'd like to try tracking known-issues on
>>> the wiki page: http://wiki.qemu.org/Planning/2.5
>>> with a list of patch series or bugs which need to be fixed
>>> into the release. This will hopefully avoid the problem we
>>> had last time around where I missed patches that should have
>>> gone into an rc.
>>
>> Can patches fix this with the right form of query? Maybe another piece
>> of metadata for patches that are promoted for rc merge.
> 
> ...only if the submitter remembers to tag it when they send
> it, and it doesn't help for issues that don't have a patch yet,
> or for tracking whether a submaintainer has taken the patch.
> 
> Tagging your patch as "[PATCH for-2.5]" is a good idea anyway
> once we're in freeze, though.

Maybe we also need a [PING for-2.5] or similar? I've just sent another
couple of pings for patches (one of which fixes the broken
analyze-migration.py script) which have fallen through the cracks over
the past couple of months. It would be nice to have an easy way to flag
outstanding patches like this, especially during freeze when things get
really busy.


ATB,

Mark.

Re: [Qemu-devel] Reminder: we're now in softfreeze

[Peter Maydell]

On 26 October 2015 at 06:25, Peter Crosthwaite
 wrote:
> On Thu, Oct 22, 2015 at 3:30 AM, Peter Maydell  
> wrote:
>> For this release I'd like to try tracking known-issues on
>> the wiki page: http://wiki.qemu.org/Planning/2.5
>> with a list of patch series or bugs which need to be fixed
>> into the release. This will hopefully avoid the problem we
>> had last time around where I missed patches that should have
>> gone into an rc.
>
> Can patches fix this with the right form of query? Maybe another piece
> of metadata for patches that are promoted for rc merge.

...only if the submitter remembers to tag it when they send
it, and it doesn't help for issues that don't have a patch yet,
or for tracking whether a submaintainer has taken the patch.

Tagging your patch as "[PATCH for-2.5]" is a good idea anyway
once we're in freeze, though.

thanks
-- PMM

Re: [Qemu-devel] Reminder: we're now in softfreeze

[Mark Cave-Ayland]

On 26/10/15 10:01, Peter Maydell wrote:

> On 26 October 2015 at 09:56, Mark Cave-Ayland
>  wrote:
>> Maybe we also need a [PING for-2.5] or similar? I've just sent another
>> couple of pings for patches (one of which fixes the broken
>> analyze-migration.py script) which have fallen through the cracks over
>> the past couple of months. It would be nice to have an easy way to flag
>> outstanding patches like this, especially during freeze when things get
>> really busy.
> 
> That's what I'm hoping the wiki page will work for. Unlike the
> mailing list, it's very short and easy to scan through :-)

Got it. I'll go add some entries later.


ATB,

Mark.

Re: [Qemu-devel] Reminder: we're now in softfreeze

[Amit Shah]

Hey Peter,

On (Thu) 22 Oct 2015 [11:30:57], Peter Maydell wrote:
> [Apologies for the huge cc list, which is basically "everybody
> I have accepted a pullreq from this release cycle.]
> 
> Just a reminder that we're now in softfreeze (ie "no new big
> features, start the process of cutting down towards only bugfixes
> going into master"). Hardfreeze will be on the 12th November,
> in three weeks' time. If submaintainers can aim to get pull
> requests in early rather than all on the 12th that would be
> nice, otherwise we get a big logjam on rc0 day and your
> chances of not getting new-feature code in go up sharply.

Juan and I hope to get the postcopy series in 2.5; it's been on the
list for a long while and it's mostly reviewed-by.  There are a few
patches that need reviews from the latest series (minor, but still
need reviewed-by).

Hope this meets the freeze rules.

Thanks,

Amit

Re: [Qemu-devel] Reminder: we're now in softfreeze

[Peter Crosthwaite]

On Mon, Oct 26, 2015 at 2:27 AM, Peter Maydell  wrote:
> On 26 October 2015 at 06:25, Peter Crosthwaite
>  wrote:
>> On Thu, Oct 22, 2015 at 3:30 AM, Peter Maydell  
>> wrote:
>>> For this release I'd like to try tracking known-issues on
>>> the wiki page: http://wiki.qemu.org/Planning/2.5
>>> with a list of patch series or bugs which need to be fixed
>>> into the release. This will hopefully avoid the problem we
>>> had last time around where I missed patches that should have
>>> gone into an rc.
>>
>> Can patches fix this with the right form of query? Maybe another piece
>> of metadata for patches that are promoted for rc merge.
>
> ...only if the submitter remembers to tag it when they send
> it,

If it is done as an annotation (e.g. like an RB) then it can be done
by reviewers after the fact.

 and it doesn't help for issues that don't have a patch yet,
> or for tracking whether a submaintainer has taken the patch.
>

Don't those issue exist regardless of freeze state? We should come up
with a more universal state tracking solution. FWIW, In my own local
system I use Applied-by: tags so I know something is sitting in
someones queue waiting for a flush.

Regards,
Peter

> Tagging your patch as "[PATCH for-2.5]" is a good idea anyway
> once we're in freeze, though.
>
> thanks
> -- PMM

[Qemu-devel] Reminder: we're now in softfreeze

[Peter Maydell]

[Apologies for the huge cc list, which is basically "everybody
I have accepted a pullreq from this release cycle.]

Just a reminder that we're now in softfreeze (ie "no new big
features, start the process of cutting down towards only bugfixes
going into master"). Hardfreeze will be on the 12th November,
in three weeks' time. If submaintainers can aim to get pull
requests in early rather than all on the 12th that would be
nice, otherwise we get a big logjam on rc0 day and your
chances of not getting new-feature code in go up sharply.

For this release I'd like to try tracking known-issues on
the wiki page: http://wiki.qemu.org/Planning/2.5
with a list of patch series or bugs which need to be fixed
into the release. This will hopefully avoid the problem we
had last time around where I missed patches that should have
gone into an rc. The idea is that anybody (patch series author,
bug submitter, submaintainer) can add to the list. Then we
can annotate it with status (rejected as an rc bug, in a
submaintainer tree, applied to master, etc).

thanks
-- PMM