[bug] Suspected wrong graph for isfull? (Interpreter wrong)

[acl-cqc]

In examples/infer2.brat:

isfull(n :: #) -> Bool
isfull(succ(doub(n))) = isfull(n)
isfull(0) = true
isfull(_) = false

Interpreter wrongly produces false for 3 and 15, despite interpreting correctly

hasfulllen(n :: #, Vec(Bool, n)) -> Bool
hasfulllen(n, x ,- (xs =,= ys)) = hasfulllen(!, xs)
hasfulllen(_, []) = true
hasfulllen(_, _) = false

suggests some difference in the graph computing the values vs the lengths, tho it's amusing that the lengths are correct and the values wrong not the other way around! Might this possibly relate to #30 ??