On Fri, 2010-06-25 at 22:50 -0700, Adam Williamson wrote:
I talked to notting &c about this earlier, and we've hit this
situation
before. The 'scenario' is simply that there's really no screening
between 'submit' and 'push' for stable updates, and this one was
submitted to stable before any negative karma came in. There's no
reliable process whereby whoever's doing the push to stable actually
looks at feedback before doing it; unless they happen to have been made
aware that a particular package shouldn't be pushed, they just push
everything that's been submitted.
There actually is a way to get feedback on updates before pushing.
Bodhi's admin request list interface gives everything in the list headed
to stable a color based on some heuristics. So it's quite easy to scan
the list and look at the red updates, which have negative karma, or
haven't been in testing long. However, I think all of our bodhi admins
who kick off the pushes use the command-line instead.
I think that in this case, the command-line bodhi client could be
improved to make it obvious that they're about to push an update with
negative karma.
[...]
The requirement for proventester feedback for critpath updates, when
we
turn it on, should also catch problems like this in the critpath. Evo
isn't critpath, though, I believe.
evolution-data-server is in the critpath, and having the proventester
feedback policy implemented would have definitely caught this issue.
luke