Inconsistency with group knowledge notation

[twanvl]

The following two formulas should both be true, since in system K4 (also S4 and S5), K1 p -> K1 K1 p.

system K4; (K1 p & K1 K2 p) -> K1 K{1,2} p
system K4; (K1 p & K1 K2 p) -> K1 (K1 p & K2 p)

But MOLTAP incorrectly claims that the first is false. The produced counterexample is (obviously) incorrect.

This is caused by tabLocalUpIf, which only moves up in the tree if the agent set matches exactly, and 1 does not equal {1,2}, so the axiom is not applied. The correct solution would be to split into two worlds as is done in the second formula.