Problem 44382. Parse me a Lisp

it seems look-up table solutions are becoming a bit of an issue, could you please add a few test cases to discourage this? +1 for random test cases that are a bit harder to trick, for example something like n=randi(100,1,randi(10)); expr = "(+" + string(num2str(n))+ ")"; assert(isequal(eval_lisp(expr), sum(n)));

I think another test case with multiple expressions would be equally useful. Even something as simple as (+ 1 (* 2 3) 4) would prevent some of the look-up solutions. Make some of those values random, and it will work even better.

Good problem, and better than 44374. Precedence is clear.