launchpad-dev team mailing list archive

[Robert Collins <robertc@xxxxxxxxxxxxxxxxx>]

On Thu, Mar 22, 2012 at 2:49 AM, Guilherme Salgado
<guilherme.salgado@xxxxxxxxxx> wrote:
> Except it crashes if you pass any(*milestones) *and*
> exclude_conjoined_tasks=True because _build_exclude_conjoined_clause()
> assumes params.milestone is a single IMilestone instance.  And fixing
> _build_exclude_conjoined_clause() seems to involve doing up to 3 joins
> for every milestone passed in, which sounds scary to me.

I've thought about this a little bit, and here is what I think is up:
 - bug search is crufty
 - but less crufty than having a different search out there

->

I *really* don't like the idea of another bug search implementation
existing. What we have is fraught enough. I think in terms of LoC,
maintenance and overall sustainability it is much better to fix what
we have (it doesn't have to issue one query, its allow to issue N -
and does today, but N is constant for a given search irrespective of
the number of found results).

*but*

can you just search on productseries (and not product) or do you
really have both product and series targeted bugs targeted to the same
milestones? That would let you defer diving into bugSearch while still
meeting your goals...
 -  milestones are meant to have the same toplevel|series split tasks
have, so probably this is more of that tech debt we raised during the
discussion about milestones a few months back.
 - OnlySeriesTasks would remove the non-series bugtasks and conjoined
masters anyway, and that is in the list of things Ubuntu wants done,..

I'll try and find you on IRC, to get a better understanding of the use
case involved.

-Rob