tim | TMI: Release week

Today I thought I would focus on closing remaining release blockers; we're trying to punt most of the bugs marked 0.6 to some point in the future, as I mentioned last week. I didn't have a super-productive time, but we are down to 72 open issues milestoned for 0.6. I think I isolated the mysterious bit of code -- in trans::base::copy_arg_to_allocas -- that's causing by-value self not to work (as manifested in segfaults or bad frees), for example in #4850 and #4439. But I don't know exactly why the code is there or how to fix it.

Otherwise, I tried fixing a few bugs only to realize the work had already been done and the bug just needed to be closed.