Le 03/12/2013 20:26, Oliver Grawert a écrit :
hi,
On Di, 2013-12-03 at 14:06 -0500, Rick Spencer wrote:
Seems like with everything moving we are having a hard time getting a
good image out. Any thought to stopping changes that don't address the
blockers until we get a good image?
with everything stopped it means that changes start to pile up and we
end up getting gigantic change sets which then makes it hard to identify
a single breakage in a landing ...
we should put our focus indeed on the main blockers, but should still go
on landing stuff that is not actually related to the current issues and
check it so we can make sure it does not interfere or introduce new
breakage. the blockers usually only have one or two specialists that can
fix them which produces an artificial bottleneck for all of us if we
stop all unrelated landings because of them.
lets go on as we do and simply see that we can put as much focus onto
the blockers as possible, but lets not block unrelated landings for
this.
getting a better automated testing infrastructure and more dogfooders
that also find the corner cases appears more intresting than just
blocking the whole world for something that can only be fixed by a small
fraction of the team imho ...
Knowing that it seems we'll only get the 2 Unity8 fixes next week as
the specialist isn't available this week. But those 2 are not
blockers, just high priority bugs worth mentioning IMHO as regressing
(but already in promoted image).
However, we ensure that we won't get any new release without a fix in
components that have a blocker bug (so no Unity8 release before the
bugs are fixed for instance).
Didier
