possible validator framework

[jverzani]

I had a window of time, so thought I'd sketch this out. It does move a few things around, but I think separating the things in runner and question_types is going to make extensions come more easily.

There are other ways this could be done, but I think this approach makes writing different validators really quite easy. I'm not in love with the names used.

[atantos]

I appreciate your time and effort! I will be back in a few days and, then, I will take the appropriate time to look at it carefully.

[jverzani]

Okay, it is a matter of opinion, but I think this second approach is easier for a user to write their own validator. There are basically two types InputValidator and OutputValidator. These take a predicate function (either f(input, question_answer) or f(result, question_answer). This should make it relatively easy for users to write specific validators. In addition, there are counterpart function for the R ones. The one possibly confusing part for users is that these functions return validators that gets passed to the validator argument. (Though it doesn't seem that confusing to me, but is a bit more expected when a struct is used, like DefaultCodeValidator.)

[atantos]

Dear @jverzani .
Thanks for the work you put into this PR! I’m sorry for the delay. The past couple of months have been a bit rough, and I haven’t been able to give this the attention it deserves yet.
I haven’t forgotten about it, and I do intend to come back to it once a few things settle. Thanks a lot for your patience!