Heikki Linnakangas <heikki(at)enterprisedb(dot)com> writes:
> I wonder if this would help to clean up the equivalence class hacks in
> Greg's ordered append patch?
Pretty hard to say at this early stage, but the more I think about it
the more I like the idea of never using the same Var to represent two
values that aren't guaranteed equal. The maybe-its-null business has
been a thorn in the side of any sort of close semantic analysis for
a long time, but somehow I never perceived what a bogus idea that was.
I'm almost more excited about that than about fixing the original
problem ...
regards, tom lane