fixes in other contexts are not indicated on buglistings

[Christian Reis]

For bugs reported on multiple contexts, and in particular for Ubuntu tasks with additional upstream tasks, it is really useful to be able to see what bugs have been fixed upstream. This is what the fixed-upstream advanced search option does, but we should step ahead with this, and actually display a special symbol next to the bug if the bug has been fixed elsewhere already.

This will probably require a column cache in bugtask, which the advanced search report could use to make the query faster.

[Stuart Bishop (stub)]

Why is a flag proposed for the BugTask table rather than the Bug table? From the use case, putting it here is denormalized.

The logic for when the flag is being set and unset needs to be specified, particularly when a Bug has multiple upstream tasks. At the moment I have an implementation in my review queue but no use case or logic to verify it against.

[Eleanor Berger (intellectronica)]

The logic is identical to the logic of the equivalent feature in searches.

[Björn Tillenius (bjornt)]

This feature needs a bit more discussion before it's read to code. The same logic used in the search doesn't quite work here.

[Christian Reis (kiko)]

Why don't we fix this by simply caching a fixed_bugtask_count counter on the bug, and using that to decide whether to display something special or not?

[Eleanor Berger (intellectronica)]

I think the problem that bothered Björn is that it's not enough to say that the bug was fixed _elsewhere_ - we care about the relation between the fixed bugtask and the unfixed on (is one an upstream of the other, etc). The solution you (and stub) are proposing makes sense, but I take the point that it might be of little use to users.

[Christian Reis (kiko)]

On Thu, Jan 17, 2008 at 02:56:22PM -0000, Tom Berger wrote:
> etc). The solution you (and stub) are proposing makes sense, but I take
> the point that it might be of little use to users.

In practice doesn't it give the right indication in most cases where we
have multiple bugtasks?

[Björn Tillenius (bjornt)]

On Thu, Jan 17, 2008 at 02:56:22PM -0000, Tom Berger wrote:
> I think the problem that bothered Björn is that it's not enough to say
> that the bug was fixed _elsewhere_

Actually, no. My concern was the the current advanced search logic is
bascially, 'has this bug been fixed in a Product'. That means that if
you looked at a closed bug for a Product, this flag would always be
True, even if that's the only place where the bug has been fixed. I
think having a 'fixed elsewhere' flag is more useful. What we have to
think of is how we define 'elsewhere'. For example, is it useful, when
you look at Ubuntu bugs, to see that the bug has been fixed in dapper?

I think it is quite useful, but once upon a time I did add a search
option for 'fixed elsewhere'. This wasn't deemed useful, and was
replaced with the current logic. I can't remember the reasons, though.

[Björn Tillenius (bjornt)]

Untargeting for now to leave room for 2.0 tasks.