RE: proposal - minibuffer completion functions

[Drew Adams]

> >> For what it's worth: in 99% of the cases, advising is superior
> >> to redefining.  The above case is no exception.
> >
> > It's irrelevant to the proposal.  Please read the proposal.
> 
> Yes, it's a separate issue.  I was just pointing it out so that in the
> future you'll hopefully use advice rather than redefinition.

I will not, for that application. I don't agree that advice is superior in that
case, but the details are not worth exploring here. I do sometimes use advice,
FWIW.