[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[bug#60788] Policy for system tests?
From: |
Ludovic Courtès |
Subject: |
[bug#60788] Policy for system tests? |
Date: |
Thu, 20 Apr 2023 12:03:04 +0200 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/28.2 (gnu/linux) |
Hi,
Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:
> Ludovic Courtès <ludo@gnu.org> writes:
[...]
>> I think a system test would be nice, we generally require it upfront,
>> but since Maxim wrote it can come later, let’s not let it block this
>> patch any longer.
>
> I didn't mean to lower our standards; I wasn't sure if that policy of
> ours was strict, since a few system services do not have corresponding
> tests, IIRC. I wary a bit that demanding a system test for each added
> service may cause scalability problems in the long run, as each demand a
> disk-heavy image to be generated and the test to run in a VM, which
> makes it expensive/slow. On the other hand, it's nice to know about any
> regressions when they happen rather than on a reboot...
Yeah, it’s an unwritten policy; I think we’ve consistently required it
for some time now. It’s useful because otherwise it’s hard to tell what
the status is for a service.
(Speaking of which, we do *not* have that policy for Home services,
because we don’t even have a test strategy, and that’s something we
should fix before it’s too late.)
> If we have such a policy, perhaps we should explicit it in
> our documented contribution guidelines?
Yes!
Also, we should split the submission guidelines into different
categories: packages, services, doc, core, etc.
Ludo’.