This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH] Canonical types (1/3)
> This is completely orthogonal to canonical types. I see no reason to
> hold up Doug's patch, which does not change the behavior of the
> compiler, but makes it a good bit quicker. Doug's patch isn't going to
> make it any harder to change this behavior later.
I agree. My comment was based on the confusion as to what was being changed.
However, I'm wondering whether "narrowing" what we consider a "variant"
really is a good idea given this change. In some sense, TYPE_CANONICAL
and TYPE_MAIN_VARIANT represent the same concept. Do we really need them
both or is there some way that we can unify both concepts? We'd certainly
save considerable space were we able to do that.