Tagging upstream stable tracking bugs and mass requesting feedback
On Mon, May 31, 2010 at 12:03:51PM +0200, Stefan Bader wrote:
> This is one of the ideas that sort of were discussed in smaller groups at UDS
> and I think we should bring into a bigger group and hopefully end up in
> The challenge we face with upstream stable releases is that these often consist
> of a quite big amount of patches. And this makes it nearly impossible to always
> make connections to existing bugs. Sometimes we don't even know that a
> particular patch would fix a bug (if its not finished triaging or even sometimes
> its less that obvious from descriptions in the patch what effect the bug fixed
> in that patch has).
> So basically the idea was to tag the tracking bugs for the upstream stable
> releases with the series (e.g. lucid) and something like 'upstream-stable' and
> then, when we put a kernel into proposed which contains at least one of those,
> let the arsenal scripts update all bugs which are tagged with the same series to
> carefully ask to look at the proposed kernel.
> The message for the update needs to be carefully phrased to make it clear that
> this does not necessarily means that particular bug is fixed but that it just
> would be worth checking as there are quite a lot of fixes in general.
I like this in that it gives us a bit more of a specific venue for
requesting a test from the original reporter other than the normal "Hey,
this bug is stale please retest." that we currently have. I think it
would be relatively trivial to create something script wise to have
kernel-janitor ask for a test of the pre-proposed or proposed to have
bug reporters try.
> Another plus on this would be that it could help to get more attention on that
> proposed kernel and thus a better regression testing coverage.
> How does this sound?
> kernel-team mailing list
kernel-team mailing list