On Wed, Oct 31, 2018 at 2:48 PM Lukas Ruzicka <lruzicka(a)redhat.com> wrote:
>
> Hello,
> I have been assigned to organize a discussion about this issue
https://pagure.io/fedora-qa/issue/567.
>
> I have thought about some possibilities (see lower or the issue) how the behaviour
should be defined. I would like you to comment on it and add your ideas to the material. I
hope this discussion will result in one transparent, documented and testable scenario in
the end.
>
> Thanks a lot.
> Lukas
I agree with Stephen at
https://pagure.io/fedora-qa/issue/567#comment-538601 , this might
be a misunderstanding. I don't think we need to change when updates-testing is
enabled/disabled in the installed system. That seems to be working quite fine. But we want
to define how anaconda should behave when installing from online repos, whether it should
use updates-testing for installation or not, or when.
So the possible options probably are:
* always disabled during installation
* always enabled during installation up to the final RC
* respect the default values in fedora-repos (enabled only as long as it is enabled in
the installed system). Note that this is likely decided at compose time.
And then we need to define what happens when the user enables/disables additional updates
using the checkbox in "installation source" spoke. Or, in case we keep
updates-testing enabled at least sometime, whether the GUI should change somehow to
reflect that.
I want to cast a vote that anaconda should never enable the
updates-testing repository absent an explicit user request (such as
adding them as supplementary install sources). The advantage here is
that we are far more likely to succeed with installations and then
testing will take place once they have a viable starting point.