Evaluator for closed expressions

[valis]

It is possible to implement a more efficient evaluator for closed expressions. For example, a function defined by recursion over Nat can be evaluated using a for-loop instead of recursion. This will help with some stack overflow errors.

[ice1000]

Is this fixed? I recall @valis has implemented tail-call optimization.

Can we have some benchmark tests?

[valis]

No. It is still easy to get stack overflow. I'm going to fix this, but it still makes sense to implement a separate evaluator (or even compiler) for closed expressions.