Readjusting SRU review process

[Martin Pitt]

Adam Conrad [2013-05-24 12:48 -0600]:
> And, even in the SRU case, you might have an SRU that addresses 24
> bugs: which one do you follow up on?  All 24?  The first in the list?

This is a rare case, but in that case I used any bug which was set up
with the SRU description, or if that was missing, the first one.

> At any rate, I'm having StevenK work on getting Soyuz to allow for
> a rejection comment, and from there, we can let the debate drop, since
> we'll be able to do rejection comments

That sounds nice indeed, but do you really think that this is a
blocking problem? Rejections weren't all that common in my time, maybe
that changed these days.

>  https://bugs.launchpad.net/ubuntu/+source/foo/+bug/1234/comments/5
> 
> That should address use-cases in both directions, while allowing the
> uploader to get a proper reject response with some information that
> doesn't mean they need to hunt through all 24 linked bugs for their
> excuse.

Well, if you upload an SRU for 24 bugs and are not subscribed to any
of them, then getting it rejected should not be too much of a
surprise. I think it's reasonable for that person to ask around by
herself why the package was rejected, if she also is not on IRC. But
again, this sounds like a rare corner case.

> (And, while I usually do the "ping on IRC" thing myself, this is also
> one thing that stops me from rejecting packages, as not everyone keeps
> the same IRC hours as I do and, shockingly, some people are offline
> when I try to contact them, async communication is definitely better
> in this case).

I don't feel that it should be the SRU team's responsibility to track
down uploaders who upload bad SRUs, don't subscribe to bugs, and are
not on IRC.

Martin
-- 
Martin Pitt                        | http://www.piware.de
Ubuntu Developer (www.ubuntu.com)  | Debian Developer  (www.debian.org)