Unable to invalidate linux-source-2.6.24 bug

[Leann Ogasawara]

I am trying to invalidate the following bug:

https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.24/+bug/133666

The reason is that beginning with the Hardy release, the kernel source package naming convention changed from 'linux-source-2.6.xx' to just 'linux'. Therefore, I want to mark the linux-source-2.6.24 task as Invalid. When trying to do so I get an error page saying this bug has already been reported on linux. I've attached a screenshot. Let me know if you need anything else. Thanks!

[James Henstridge (jamesh)]

This looks like a regression of bug 81014.

[Björn Tillenius (bjornt)]

No, this is not a regression of bug 81014. The issue here is that linux-source-2.6.24 has never been published as a source package, but it has as a binary package.

The fix is to make sure that we always look up the corresponding source package (linux in this case) and target the bug to that one instead.

[Christian Reis (kiko)]

Wow, this is really weird. Isn't this just broken data in this specific case that needs to be fixed? What is that actual bugtask open against? A source package name that doesn't actually exist in Ubuntu?

[Christian Reis (kiko)]

Well, now I seem to have f***** it up completely. I changed the existing tasks over to firefox and evolution (yay for random picks) and have marked them invalid. In the process I've declined the nominations for Hardy and this ended up declining the linux nomination as well. I guess I owe apologies and ice cream for trying to fix this in my usual sneaky way..

[Christian Reis (kiko)]

Is this actually being caused by packages with specific names no longer being published in the current Ubuntu release?

[Christian Reis (kiko)]

Oh. It's because we shouldn't even /have/ these bugtasks. Gar!

[Christian Reis (kiko)]

OK, so here's a plan.

First, we fix bug 201547 to actually stop getting these bugs filed. I think we could add an assertion somewhere to ensure this never happens again, but Bjorn can decide that.

Once we are sure we're not getting any new bugs filed on unpublished sources, we go through all bugs filed against unpublished sources, and a) if they contain more than one task, nuke (DELETE FROM BugTask!) the task on the unpublished source b) if they contain only one task, report them so they can be fixed manually.

How's that?

[Björn Tillenius (bjornt)]

Actually, bug 201547 has nothing to do with this bug. I think it's only +distrotask that allows bugs to be targeted to linux-source-2.6.24. This is reported as bug 232498. After that bug is fixed (a fix is on its way), we can clean up the data, as kiko suggested.

[Björn Tillenius (bjornt)]

It's time to clean up the bad data.