Temporarily using `best=False' for EPEL-8 builds in Copr

[Jakub Kadlcik]

Hello,

currently, there is a problem with building EPEL-8 packages because of
DNF bugs regarding modularity (see RHBZ 1758459).

The only known workaround is to use DNF with `best=False'. Even though
it is something you don't really want to use longterm, we are patching
mock configs epel-8-* chroots, so there can be at least some EPEL-8
building in Copr.

(I am doing the change right now, but it won't affect already
spined-up builders. It will take a while until they get recycled and
use new configuration)

Once the DNF issues get resolved, we are going to drop those patches
and use `best=True` again.

For additional information, please see RHBZ 1756681 and RHBZ 1758467.


Thank you for understanding,
Jakub

[Nico Kadel-Garcia]

>  On Oct 10, 2019, at 7:34 PM, Jakub Kadlcik
> <jkadlcik(a)redhat.com&gt; wrote:
>  
>  Hello,
>  
>  currently, there is a problem with building EPEL-8 packages because of
>  DNF bugs regarding modularity (see RHBZ 1758459).
>  
>  The only known workaround is to use DNF with `best=False'. Even though
>  it is something you don't really want to use longterm, we are patching
>  mock configs epel-8-* chroots, so there can be at least some EPEL-8
>  building in Copr. 
These bugs exist in RHEL 8 as well as CentOS 8 based environments. Why do you think that
modularity will *ever* work this way?

[Leigh Scott]

Shame on Redhat for using an untested feature
https://bugzilla.redhat.com/show_bug.cgi?id=1671683

[Jakub Kadlcik]

>  These bugs exist in RHEL 8 as well as CentOS 8 based environments.
> Why do you think that modularity will *ever* work this way? 
I don't know whether this question was aimed directly to me, but if so
... I am sorry, I can't help you here. You will probably need to
talk directly to the modularity guys or start a discussion on
devel(a)lists.fedoraproject.org . I was just trying to apply a known
workaround,
so at least some EPEL-8 building can be done in Copr.


Jakub

On Fri, Oct 11, 2019 at 6:17 AM Leigh Scott <leigh123linux(a)gmail.com&gt; wrote:
>  Shame on Redhat for using an untested feature
> https://bugzilla.redhat.com/show_bug.cgi?id=1671683
>  _______________________________________________
>  devel mailing list -- devel(a)lists.fedoraproject.org
>  To unsubscribe send an email to devel-leave(a)lists.fedoraproject.org
>  Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>  List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>  List Archives:
> https://lists.fedoraproject.org/archives/list/[e-mail]

[Nico Kadel-Garcia]

On Sun, Oct 13, 2019 at 5:58 PM Jakub Kadlcik <jkadlcik(a)redhat.com&gt; wrote:
>  > These bugs exist in RHEL 8 as well as CentOS 8 based environments. Why do you think
> that modularity will *ever* work this way?
>
>  I don't know whether this question was aimed directly to me, but if so
>  ... I am sorry, I can't help you here. You will probably need to
>  talk directly to the modularity guys or start a discussion on
>  devel(a)lists.fedoraproject.org . I was just trying to apply a known
>  workaround,
>  so at least some EPEL-8 building can be done in Copr. 
I agree that the work around works. You have no evidence or even
suggestion that the workaround will ever be unnecessary? I'd agree.