Closed OlivierBBB closed 1 month ago
but there should be a binary constraint for BIN
Because it is marked binary@prove
... right?
With Franklin's patch we get the following constraints from the example above:
(1 - BIN) * Z
BIN * N
But, as I understand it, we should also have a binary
constraint as well.
See #92 for further discussion around the binary@prove
issue.
Currently Corset doesn't take advantage of binary-ness of the condition when compiling constraints à la
produces
There are several issues here:
BIN
doesn't require the computation ofC/INV[BIN]
NORM[BIN]
constraintsBIN
Franklin suggested this fix