Split off from the stable pushes in Bodhi thread just because I'd like
to see it not get lost.
On 02/27/2010 11:35 AM, Mail Lists wrote:
> On 02/27/2010 11:27 AM, Adam Williamson wrote:
>> On Sat, 2010-02-27 at 10:57 +0100, Ralf Corsepius wrote:
>> Yeah, it's not perfect: there are cases where we have, say, a complex
>> kernel update which works fine for most people but causes a significant
>> regression for some particular bit of hardware. We wouldn't want to put
>> that update out, but it's easy for it to get five +1s before someone
>> with the specific bit of hardware comes by and gives it a -1...and even
>> then, +4 looks good if you're not reading the feedback too carefully.
One could argue that the current bodhi karma system is simply too
simplistic for real use cases. Maybe instead of just +1 -1, there
Fixes my problem
Works for me (someone testing that didn't necessarily have any of the
problem supposedly fixed by this update just noting that their system
still works ok with the update)
Doesn't fix my problem (but doesn't necessarily imply it's any worse
Causes new problems (which should, IMO, be an automatic veto of any push
to stable, requiring intervention to override)
I could see situations where you would want to push updates to stable if
say the update was supposed to solve multiple bugs, but turns out it
only solves a subset of those bugs and doesn't cause new ones, so you
would have some FMP, maybe some WFM, some DFMP, but no CNP. You'd
probably just need to leave it up to the maintainer to decided if the
bugs that are solved are important enough to push to stable before
respinning another attempt at the ones that weren't solved.
I've personally seen situations where this would be helpful with the
mdadm package when I had 5 or 6 bugs on a single update and only 4 of
them were actually solved. The update was still worth pushing while I
worked on the others again.
If you really wanted to get fancy, the FMP/DFMP options could be tied to
specific bugzillas reported against the update, and if you get a FMP for
each bugzilla listed in the update, plus at least 1 WFM, then you could
automatically push to stable. Would be much better than the 3 random
+1s you get now that don't indicate anything about test coverage or
anything like that. Also FMP/DFMP karma against a bug could be noted
both in the bodhi ticket as well as in the bug itself with a resultant
VERIFIED/FAILS_QA toggle to the bug status.