Merge lp://staging/~racb/ubuntu/oneiric/samba/862129 into lp://staging/ubuntu/oneiric/samba

Outstanding questions I posted from that bug that I would like reviewed before merging:

1) Is this the right approach?
2) Is dpkg-query the right way to get the status of the update-inetd package?
3) What about triggers-awaiting and triggers-pending? What should the behaviour be in these cases? Is it OK to treat update-inetd as not available in these cases?

Thanks.

The effect of this change is to make the system non-deterministic. If
we want update-inetd to work from postrm scripts, don't we instead need
to make sure that it always works even if unconfigured, as if it were an
Essential package? This would be much safer and the resulting system
would be easier to understand.

I'm also unconvinced about the reliability of the data returned by
dpkg-query from a maintainer script. It may be better now, but
historically this didn't reliably reflect the state of a running dpkg
instance.

To be honest, I think there are too many questions about this kind of
thing for it to be safe to attempt a fix for this for oneiric. I'd
rather we got it right in precise, and then maybe backported the fix.

It might be worth pointing out that calling update-inetd in a postrm is vaguely crack anyway, as we're removing the service entries AFTER we've removed the binaries from the disk.

Moving the update-inetd call to prerm (where I'd argue it belongs) solves everything we're trying to solve with this complex branch, since the dependencies are still there during prerm.

Thanks Colin and Adam.

I was trying to do the minimal change possible to get the fix into oneiric - as it's causing upgrade failures at the moment. If that's not going to happen, then I'll file a Debian bug as the postrm hasn't been modified by Ubuntu in any way.