Re: pagure -> bugzilla sync script 'tiger team'
On Tue, Apr 09, 2019 at 07:37:56PM +0200, Clement Verna wrote:
On Sun, 7 Apr 2019 at 21:04, Kevin Fenzi <kevin(a)scrye.com>
wrote:
>
> Greetings,
>
> In triaging ticket last week, we noticed there are several related to
> the pagure to bugzilla sync script.
>
> This script runs once a day (at 18utc) and syncs the point of
> contact/watchers/committers/etc to bugzilla components.
>
> Here's a meta ticket for issues:
>
> https://pagure.io/fedora-infrastructure/issue/7690
>
> That lists the other 4 tickets outstanding on it.
>
> The script is in our ansible repo:
>
https://infrastructure.fedoraproject.org/cgit/ansible.git/tree/roles/dist...
>
> In addition to the 4 tickets theree's a number of other issues with the
> script:
>
> * It fires of a query for every single package/container/module against
> fedora-scm-requests to see if there's an override for it. We should
> bundle this information up somehow, or move the overrides to a better
> place or do something to end those 25,000 queries.
>
> * Some folks wanted to add the 'packager' group to their package, with
> the idea being that they wanted anyone in packager group to be able to
> push changes. This "works", but the script then tries to mail
> devel(a)lists.fedoraproject.org as the 'address' for 'packager'.
> It would be nice to filter this out on a configurable set of groups.
>
> * We have a small number of things we don't manage in the script: the
> distribution component, the various spins components, basically anything
> thats not a package/module/container/rpm. It would be nice to manage
> those from the script too so we had a nice place to create/remove them.
>
> Anyhow, given all these issues, I thought it would be interesting to try
> and get a small team together sometime next week and attack re-doing
> this script. I'd be happpy to help from the sysadmin side of things, but
> I would love to get a few python savvy folks involved. Any takers?
I ll be happy to help working on that. It seems that we already have a
good idea of what needs to be done, do we want to spend 3-4 days on
that next week ? if this can be done during freeze ?
I am onboard for this as well :)
+1 to wait for freeze, but I'm afk next week so I'd propose we wait one more
week, this way we'll be in the middle of freeze and things should be more quiet.
It may mean, we'll have to wait for post-freeze to apply the fixed script, but I
think it's fine.
Also this can be a good opportunity for an apprentice to join the
team
and help us :-), maybe we can add this to meeting agenda what do you
think ?
The more the merrier, +1 for me :)
Pierre