Comment 122 for bug 294712

(In reply to comment #116)
> Why are you assuming that content policies necessarily check the TYPE_* type
> before the MIME guess?

I'm not. If they're checking the mime guess and are set to do something for
this mime ("policy-user-override"), then I'd have to guess that a) they're
doing the right thing :). Or b) they'll have to update. I still don't
understand what the danger is in this, the mime is for objects only as of now,
and are supposed to be helpful when it's something that can be deciphered. This
is a new notification, why should I suspect that someone will get tripped on

> What happened to just getting dwitte to ok the patch you'd been working on, or
> tell you exactly what he wants changed about it?

Well, comment 103 happened. I thought he pretty much gave us his opinion that
he didn't like the approach of overriding content policies, and his alternative choice wasn't entirely possible. Note: this mime guess override thing is very similar to the approach he suggested.

Either way, I left them both up, and I'm deferring to whomever will make the
final decision on how exactly to tackle this bug. I thought I would share this
alternative approach, being that it was extremely simple to put together and
provided an alternative to what we had, which was deemed unsatisfactory.