Some heuristics (I think at least merge-and-shrink and LM-cut) currently
special-case handling of certain "unnecessary" conditional effects of the type
"if A = 0, set A to 1", where the domain of A is {0, 1} (so this could be
replaced by an unconditional "set A to 1").
Once issue7 is resolved, this special-casing should no longer be necessary and
should be removed.
|