Thank you, that was quite interesting to read through - if there are other similar cases, I'd love to hear about them.
There are some but Mozilla is one of the biggest one. Actually, I just studied FLOSS licenses evolution and I really think it is a feasible thing.
One significant obstacle, of course, is that we have a lot of anonymous editors where it's effectively impossible to trace the person who holds the copyright (as opposed to the computer from which they made the contribution). I'm guessing that Mozilla didn't have this problem. We probably also have a much larger volume of people who are not contactable via email, since we don't require an email address in order to sign up for an account.
The good solution is a well done tagging job... And time.
And while I don't know how many people have actually contributed code to Mozilla, I would guess that we're on a different level in terms of sheer numbers. I have this sneaking suspicion that the relicensing process would not scale very well, shall we say.
I think it would, but is it a priority today ?
The possibility of rewriting content we're unable to relicense is interesting to consider. It strikes me that one potential use for Magnus Manske's article validation tool would be to flag revisions when an article has been rewritten so as to remove the content that we can't secure permission to relicense. But anyway, if people are serious about actually relicensing, the longer they wait, the harder it will be.
Then we should prepare something around it.