The dispatching mechanism was more trouble than it was worth but we did
lose some flexibility. Do you really need to unify Sequential or is
unifying with a concrete type like PersistentVector work well enough for
your use case?
David
On Tue, Jan 8, 2013 at 2:01 AM, Austin Haas
Ok, thanks, I'll try to get it to work with PersistentVector.
I just started looking at core.logic's internals, and I'm still fairly new to
Clojure, so everything is a little fuzzy. I took Sequential for granted, since
that is referenced in the existing code.
Thanks for the help.
-austin
I've updated the experimental core.logic Datomic support so that you can
unify PersistentVector and Datoms again. In a real system I'd probably
recommend providing your own tuple type that does not implement Sequential
for doing unification with Datoms.
David
On Tue, Jan 8, 2013 at 11:55 AM,
Thanks! I had just finished making similar changes. It was a good exercise and
I'm glad to be able to compare code.
A couple of questions:
1. Why is there no case for unifying a Datom with another Datom?
2. Why aren't multimethods used for dispatching to the appropriate unification
On Tue, Jan 8, 2013 at 3:29 PM, Austin Haas aus...@pettomato.com wrote:
Thanks! I had just finished making similar changes. It was a good exercise
and I'm glad to be able to compare code.
A couple of questions:
1. Why is there no case for unifying a Datom with another Datom?
Oversight. I
The datomic unification code in core.logic has bit-rotted. It depends on
IUnifyWithSequential, which was removed in this commit:
https://github.com/clojure/core.logic/commit/bbc4e820128d5a0745ce3d79cd3bbd9401a1bf55
I'm trying to understand how to update the code, but I don't get how