Provisional MicroReleaseExceptions review

[Kees Cook]

On Wed, May 07, 2014 at 03:33:14PM -0700, Brian Murray wrote:
> Steve asked that I have a look at the existing provisional micro-release
> exceptions to determine how many times the MRE was used and whether or
> not there were any issues with those micro-releases. Kees Cook and I
> put some code together that checks the package to see how many times
> a version of the package has been released to -updates for supported
> releases (including those formerly supported), whether there were any
> new crashes reported about that version of the package in the Error
> Tracker, and finally searches Launchpad for any bug reports reported by
> apport that have that specific package version in them. For any package
> the results look like:

Thank you so much for getting this finished!

> package, release updates: quantity of SRUs
>     release version: 1.2.3
>     date SRU version
>         new errors reported: count
>         launchpad bugs: #
> 
> I've posted the search results to
> https://wiki.ubuntu.com/StableReleaseUpdates/MicroReleaseExceptions/ProvisionalStatus
> 
> One thing to note is that automatic crash reporting is not enabled for
> servers and a lot of the packages with provisional MREs are server
> packages. Also, I deliberately excluded iscsitarget since its
> provisional MRE is less than a month old.

Nova, Glance, Horizon, Keystone should be full MRE (looks like a solid
history).

Cinder and quantum/neutron look similar. As do ceilometer and heat.

Mesa makes me slightly nervous (trajectory is toward more bugs per
release), but should probably get further study (are the bugs actually from
the updates?)

I think we should revoke VLC's pMRE -- it has never been used.

Ceph looks fine.

Openvswitch feels like it stopped getting updates. While it does have an
update in 2014, there's not much history here. Perhap extend its
provisional status another year?

> Finally, if the quantity of
> errors or Launchpad bugs is important then libreoffice will require some
> additional research.

Yeah, that one is interesting. I suspect it may be noise and general
problems, but as you suggest, should get more research.

Yay! Thank you again for this! :)

-Kees

-- 
Kees Cook