On Tue, 2017-12-19 at 12:16 +0100, Tomáš Kopeček wrote:
> Dne 18.12.2017 v 20:18 Kevin Fenzi napsal(a):
>> On 12/18/2017 03:19 AM, Neal Gompa wrote:
>>> On Mon, Dec 18, 2017 at 2:35 AM, Miroslav Suchý <msuchy(a)redhat.co
>>> m> wrote:
>>>> Dne 15.12.2017 v 17:00 Dennis Gregorovic napsal(a):
>>>>> Hi Miroslav. Thank you for the heads up. Do you have a list
>>>>> of
>>>>> changes that will included in the release?
>>>>
>>>> Just a list of commits since last tag:
>>>>
https://github.com/rpm-software-management/mock/commits/devel
>>>> Of course, once released there will be release notes as usual.
>>>>
>>>>> Are there plans to test
>>>>> the new release with Koji or other tools that depend on mock
>>>>> before
>>>>> enabling karma?
>>>>
>>>> I do not operate Koji. So I do not know. I *hope* that someone
>>>> tests with Koji.
>>>> My team is testing it with Copr, I am doing running regular
>>>> mock tests (included in mock itself), providing release
>>>> notes and these heads up.
>>>>
>>>> But AFAIK fedora infra decided to stay with mock 1.3.4:
>>>>
https://koji.fedoraproject.org/koji/buildinfo?buildID=100763
>>>> 8
>>>> I am not really sure why.
>>>>
>>>
>>> It was because the private tmpfs /tmp thing from nspawn broke
>>> Koji builds.
>>>
>>> This is now fixed[1] in Koji 1.15, so Fedora Infrastructure has
>>> no
>>> reason to remain on 1.3.4.
>>>
>>> [1]:
https://pagure.io/koji/pull-request/602
>>
>> no.
>>
>> nspawn still does not work for a bunch of things koji does, some
>> known,
>> some unknown.
>
> I've fixed only very basic issue, but we're not able to build e.g.
> livemedia images at all as these depend on loopback devices, and yes,
> I
> also sometimes hit some random weird issues, which are not tracked
> yet,
> as I wasn't able to reproduce them.
>
I am successfully building livemedia images with koji-1.14 and only
needed to `koji edit-tag f25-build -x mock.new_chroot=False`. I'll
admit I haven't tried rebasing my custom spin on F27 yet, but I
wouldn't think that'd make a difference.
Yes, it shouldn't change for new_chroot=False - problems are with
nspawn-based approach.
_______________________________________________
buildsys mailing list -- buildsys(a)lists.fedoraproject.org
To unsubscribe send an email to buildsys-leave(a)lists.fedoraproject.org
--
Tomas Kopecek <tkopecek(a)redhat.com>
Release Engineering Development, RedHat