launchpad-dev team mailing list archive

[Bryce Harrington <bryce@xxxxxxxxxxxxx>]

On Fri, Apr 15, 2011 at 09:55:02AM +0100, Jonathan Lange wrote:
> On Fri, Apr 15, 2011 at 5:59 AM, Robert Collins
> <robertc@xxxxxxxxxxxxxxxxx> wrote:
> > This isn't deeply thought out, but I wonder if we can get rid of
> > 'product', 'distro' and 'distrosourcepackage' bugtasks (and thus
> > conjoined masters). If this mail gets a reasonable level of interest,
> > I will LEP this up and socialise it more widely.
> >
> Sounds like a good idea to me.
> 
> There's already some confusion in the UI about bugtasks that are
> assigned to the current development focus of a distro. It would be
> nice if this change alleviated some of that.
> 
> Ubuntu would be the biggest hit by this change, so it's worth talking
> directly to some key people there. I think Kate Stewart (the RM) and
> key people in QA (e.g. Brian Murray, Jeremy Foshee) would be a good
> start.

Having been very confused by the distrosourcepackage code myself, this
improvement sounds very worthwhile, particularly if it improves
performance, too.

It's true this would significantly change how Ubuntu bugs are handled,
but I don't see this as a bad thing.  If the change is done well it
could alleviate several problems, including what Martin referenced:

Martin wrote:
> It seems like the carrying-over behaviour for Ubuntu is not quite
> perfect at the moment anyhow: fairly often bugs that aren't fixed in a
> particular series don't want to be automatically carried over, because
> there is a good chance they are fixed upstream.  Many of them are bulk
> set to incomplete.

It's definitely true that this isn't perfect.  We end up needing to
request people continually re-test bugs, which gets irritating for all
involved, and is time consuming triager work.

But without doing that, we end up with an overwhelming backlog of bugs.
Having all bugs targeted to an affected series and not carrying them
over automatically would alleviate this, and permit developers to start
with a clean slate each release which would make it easier to focus on
regressions and release-specific problems.

Some package maintainers may dislike the proposed change due to being
accustomed to having bugs auto-carried over.  It might be worthwhile
considering adding UI or tools to help address these needs.

Another benefit of this is that often people will simply close a bug
once it's confirmed fixed in the current development series, even if it
is still an issue in earlier series.  This change would enable us to
more easily keep accurate track of these types of bugs.

Presently in Ubuntu series nominations are often used for marking
"bugs with a fix worth backporting", and some package maintainers may
dislike the proposed change since this would be harder.  But use of
milestones should address this need, I think.

Currently we tag bugs by release name.  If the proposed change is
implemented, it would be helpful to conduct a mass tag-to-nomination
change, to ensure bugs are properly targeted in the new system.

Anyway, this change could be a bit bumpy for Ubuntu but IMHO it ends up
providing a lot of benefit.

Bryce