Re: -fnew-verifier

[Tom Tromey <tromey at redhat dot com>]

>> First, -findirect-dispatch actually already requires this, and
>> second, that way we don't have to document it (and then keep it
>> around for years even after it is useless).

Andrew> Why not nuke the old verifier?

Oh, also, the new verifier is not set up to do eager checking.  It
always does lazy checking and emits type assertions.  So this would be
a loss of functionality.  (I don't know whether we care, but I think
we ought to make these decisions on purpose, not as a side effect of
some other thing.)

Tom