Pytest evaluators2

[mmatera]

the same as #145 but without changing the name apply_N->eval_N

[rocky]

@mmatera I apologize for not having gotten this merged in before my leaves -> elements disruption.

I tried to resolve the merge conflicts, but I couldn't this is to large and complex and I just don't understand it all that well.

Would you please resolve the merge conflicts so we could merge this. It might be useful in this regard to split it up into smaller PRs.

Possibly next week (not this week), I'd like to continue with replacing "leaves" with "elements" when that is what we mean. And in particular replacing the @leaves property on Expression with @elements would probably be the next thing. (And then after that we probably could replace .get_elements() with .elements and things would be simpler and tidier.

So it would be good to fix this and any outstanding PRs that may cause conflicts.

Many thanks.

[mmatera]

OK, I think I can handle it.

[mmatera]

Do you have an example, then this rule will apply? In the best case b^a is returned, so no change?

0^a can not be evaluated in general, however, Simplify[0^a,Assumptions-> a>0] evaluates to 0.
The second rule, for a^b tries to ensure that a^0 simplifies to 1 if a!=0. However, in WMA (and in Mathics) a^0 evaluates to 1 even if a could be 0, so in the current state, that rule is irrelevant.

[mmatera]

@rocky, I merge this before merging your new PR to avoid further file conflicts.