Wednesday, 18 May 2016

Re: SRU never reviewed, why/how do we avoid that next time?

On Wed, May 18, 2016 at 06:08:42PM +0200, Sebastien Bacher wrote:
> Hey there,
> (sorry if that's not the perfect place but the SRU team doesn't have a
> dedicated contact point/mailing list that I know of)
> We had a libreoffice bugfix SRU in the week after the 15.10 release to
> address some of the user feedback we received from the iso version,
> that's still waiting in the queue unreviewed as far as we can tell (the
> bugs references didn't get updates, the updater/uploader didn't get
> contacted)
> That's a bit unfortunate and I'm unsure what happened there. The SRU
> team is working on best-effort basis and it's understandable that
> reviews take time sometime, but in this case the queue has been
> regularly reviewed since, it just seems that everybody ignored that item
> for some reason. I can imagine that reviewers tend to stay away from
> libreoffice updates since those can be non trivial to review, but in
> this case the diff it's pretty trivial.
> Coming to the point I would like to try to understand what has been the
> issue there (insight from the SRU team welcome ;-) and if we can find a
> solution so it doesn't happen next time?

I think a large part of the problem is that we don't have a great
process for handling things which are not accepted. An item in the
queue may not be accepted because it isn't fixed in the development
release or missing a test case among other reasons. In those situations
we comment on the bug and hope it gets addressed while leaving the item
in the queue. There isn't a way in Launchpad to record this information
about items in the queue - I personally just a browser extension to keep
notes for myself. Additionally, the SRU team doesn't have a process or
system for conveying that type of information to its members. In the
end we have long queue of items which are a mix of items that have never
been reviewed, ones that have but we are waiting on the uploader to
fix something, or ones that reviewers don't feel comfortable looking at.

One improvement to the process, from my perspective as an SRU team
member, would be rejecting items that are missing SRU information. But
it seems rude to reject an upload and throw away someones work just
because a bug is missing information, if we then give the uploader two
weeks(?) we may end up in the situation where have a long messy queue

Brian Murray