Seems reasonable, I can't think of a good reason to update pending
either. Maybe if you are cloning a repo with unresolved conflicts it
could leave effects in the working directory which sometimes need
to be in pending? Anyway, I'm not personally too worried if that
scenario gets a bit worse.
Hmm. I hadn't thought what happens when you clone unresolved
conflicts. I think we should do the equivalent of mark-conflicts in
that case, but I just checked and it looks as if we don't. This is
unacceptable. An invitation to desaster! I can't believe we have no
test for this. I'll open a new issue.
(To come back to the issue at hand: pending should be updated when we
fix that, of course, but ATM should be unaffected.)