[libvirt] Availability of libvirt-1.3.3-rc2

[Laine Stump]

On 04/01/2016 11:39 AM, Martin Kletzander wrote:
> On Fri, Apr 01, 2016 at 11:54:30AM +0800, Daniel Veillard wrote:
>>  I pushed the rc2 tag in git head and signed tarball and rpms to the
>> usual place:
>>
>>   ftp://libvrt.org/libvirt/
>>
>>  I didn't really had time to test it TBH, seems we don't have too 
>> much red
>> in the CI https://ci.centos.org/view/libvirt-project/ , mostly perl 
>> tests,
>> and virt manager, better than a few days ago !
>>
>>  Please give it more testing and hopefully I can make a release this 
>> w.e.
>> or early monday,
>>
>>  thanks !
>>
>
> There are two things that worry me and should be fixed before the
> release.  It's nothing big and I should have it worked out during the
> weekend, so I would just like to ask you to not do the release on
> weekend and rather leave it for Monday.  I'll post any updates here.
>
> I don't want to go to the details very much as I still have errands to
> do, but one of them Laine encountered when trying to nodedev-detach an
> SRIOV card and it's just a matter of nodedev XML parsing.  The other one
> was found by Guido and that is that we don't regenerate the per-domain
> directory name (which has ID in it) on subsequent domain restarts.
> Again, just a matter of resetting it properly, I believe.  I'm just
> writing this because I won't get to that till Saturday evening from now,
> sorry for that.

There is a different regression that I also found which I've just posted 
patches for:

https://www.redhat.com/archives/libvir-list/2016-April/msg00067.html

Because the fix involves adding in a new function that may or may not be 
compiled in depending on platform and options, I would appreciate if 
people on odd systems would try applying the patches and doing a build 
(since there won't be another release candidate). (of course it 
shouldn't be too bad, as it's only called from qemu, so it shouldn't 
cause headaches for bhve, for example, but I've been burned by this in 
the past so I'm extra cautious).

I should be around as late as Monday AM (US EST) to push, but if it's 
not ACKed by then, I'll be on a plane for the rest of Monday - if that 
happens I would appreciate if the reviewer pushed it for me.