Native speaker of English (not Italian :))

Che stupido sono io...Please disregard my post. Jeez what an idiot I can be sometimes. Sorry Giacomo...

I agree with an earlier poster. If you have cases where signature is somthing like [1,1], it is FALSE to claim that "signature is guaranteed to have at least 3 numbers." If you ask me to choose between believing that "guarantee" and my own lying eyes regarding the case of [1,1], I'm going to believe my own lying eyes and assert that [1, 1] has (at most!) two numbers. PLEASE accurately describe the rules on the input.

But this has nothing to do with "it is GUARANTEED that the signature contains 3 values". It's distressing and unfair that the solver, counting on this guarantee, is assaulted with INCORRECT results because some of thie inputs DO NOT HAVE 3 values.

Same result for me as for pindio58

There's just a typo in the unreachable cases sample test where start = (0,0), dest = (7, 7), and obstacles are the squares one knight's move away from dest. It shouldn't be calling func(s, e, o) -- it should be calling attack(s, e, o).

I used BigInteger instead of BigDecimal. DOH! But it worked! "I could have done it in a much more complicated way."