[Qemu-devel] qdev merge plans?

[Qemu-devel] qdev merge plans?

[Gerd Hoffmann]

   Hi Anthony,

Your queue continues bitrotting.  BlueSwirl did more qdev conversions 
with the result that more fixups are needed to make the tree build again 
once your queue is pushed ...

For now I did incremental fixups and didn't change the patches you have 
queued up already.  At least for the build failure this isn't ideal 
though as it leaves a few revisions which don't build which is bad for 
bisecting.

So what is your plan to handle the mess^H^H^H^Hmerge?

I plan to respin my patches tomorrow and can adjust the patches the way 
you want to have it.

cheers,
   Gerd

PS: complete patch queue (including work-in-progress bits) is at
http://git.et.redhat.com/?p=qemu-kraxel.git;a=shortlog;h=refs/heads/qdev.v11

Re: [Qemu-devel] qdev merge plans?

[Anthony Liguori]

Gerd Hoffmann wrote:
>   Hi Anthony,
>
> Your queue continues bitrotting.  BlueSwirl did more qdev conversions 
> with the result that more fixups are needed to make the tree build 
> again once your queue is pushed ...
>
> For now I did incremental fixups and didn't change the patches you 
> have queued up already.  At least for the build failure this isn't 
> ideal though as it leaves a few revisions which don't build which is 
> bad for bisecting.
>
> So what is your plan to handle the mess^H^H^H^Hmerge?

Early in your series, you introduce -device.  But -device syntax doesn't 
seem to be resolved yet.

I don't think it's appropriate to pull in -device this close to the 
feature freeze when the command line syntax is still being debated.  You 
only countered Paul's suggestions this morning so I don't think there's 
any indication that this discussion has converged yet.

Quite a lot of your queue consists of things you've only posted as 
Proof-of-concept series (like qdev/isa).

So let me ask you, what is the set of patches that you think is 
non-contended, posted as a non-RFC, but not yet merged?

Regards,

Anthony Liguori

Re: [Qemu-devel] qdev merge plans?

[Gerd Hoffmann]

On 07/15/09 00:25, Anthony Liguori wrote:
> Gerd Hoffmann wrote:
>> Hi Anthony,
>>
>> Your queue continues bitrotting. BlueSwirl did more qdev conversions
>> with the result that more fixups are needed to make the tree build
>> again once your queue is pushed ...
>>
>> For now I did incremental fixups and didn't change the patches you
>> have queued up already. At least for the build failure this isn't
>> ideal though as it leaves a few revisions which don't build which is
>> bad for bisecting.
>>
>> So what is your plan to handle the mess^H^H^H^Hmerge?
>
> Early in your series, you introduce -device. But -device syntax doesn't
> seem to be resolved yet.

The patch sits in your queue nevertheless, so I'm not sure what your 
plan is ...

> I don't think it's appropriate to pull in -device this close to the
> feature freeze when the command line syntax is still being debated. You
> only countered Paul's suggestions this morning so I don't think there's
> any indication that this discussion has converged yet.

Ok, so it sounds like you'll plan to drop the patch(es) from the queue. 
  Ok.  I'll go post fresh patch series based on upstream/master then, so 
you can replace the qdev bits in your queue.

> Quite a lot of your queue consists of things you've only posted as
> Proof-of-concept series (like qdev/isa).

Yes, as mentioned there are quite a few work-in-progress patches.

> So let me ask you, what is the set of patches that you think is
> non-contended, posted as a non-RFC, but not yet merged?

I'll sort and repost the patches today.  Rough plan is:

(1) Split the patch series in your queue into two:
     - One with the properties bits (should be ready to merge).
     - One with the -device stuff (for further discussion
       and maybe merge).
(2) Repost qdev/compat bits, slightly updated according to
     review comments (should be ready to merge too).

cheers,
   Gerd