On Sat, 09 Jan 2016 19:45:43 -0000
"Andre Robatino" <robatino(a)fedoraproject.org> wrote:
I think you're looking at the wrong time period.
I don't think so, but it's possible. :)
went to stable on 2015-11-23 (see
this time, 1.6.17-3 had already been pushed to testing and was just
sitting there. Then it was submitted for stable on 2016-01-06 and
went to stable on 2016-01-07 (see
downgrading the newer version.
The time it went stable doesn't matter.
The time(s) that matter were when the newer one was submitted as an
update and what time the previous older one was being pushed.
My contention is that the older one was in a push (and thus locked)
when the newer one was submitted (so bodhi couldn't obsolete it). Then
both updates were still live after that. (one in testing and one
pending testing in the next push).
Ideally here bodhi would queue up the obsoletion event and after the
push check to see if any updates went to testing that should be
obsoleted (then it would be unpushed and the new one would go out in
the next updates-testing push).