Yeah, well, you can never trust users to be accurate in bug reports. :-( Sorry about that.Well, ok, but your initial report began: Make an RCS file writable with chmod +w without locking it. If I instead use M-x toggle-read-only, make changes, and use C-x v v, then I see the problem.
I'm obviously not in any position of authority here, but as for my personal opinion, I would have no objection whatsoever to making vc-mistrust-permissions default to true for safety's sake. Might also want to put a warning in the documentation of the variable about what might happen if you set it to false and then use vc-next-action on a file with non-checked-out changes.OK. It looks like that variable only affects RCS and SCCS? In which case I am guessing that few people will care if the default changes, so maybe we should just do that...