Makeboxes overhaul

[mmatera]

In this PR, I propose some changes in the way that the underlining MakeBoxes mechanism works. In Mathics 1.1 (and the current master branch) there are BoxConstruct classes, hard-linked to the interpreter, that defines how (the FullForm of ) an expression is translated to a set of rules to build the output, that can be processed by the front-end.
In WMA, the interpreter is designed in a way that MakeBoxes can be redefined to adapt the output to different front-ends.
With the proposed changes, the behavior of MakeBoxes in Mathics would be closer to the WMA, as described in Wolfram Reference: Textual Input and Output
This would allow for example, to overload the MakeBoxes for different classes of expressions (Graphics, Strings, etc) by the Pymathics mechanism, in a way that different front-ends build and render boxes in the most suitable way, according to the capabilities of the interface.

This will help also to make a transparent interface for the Matplotlib pymathics package.

[mmatera]

Overall I really like this. Thanks so much for doing it.
Your welcome!
The main question I have is, how can I see the new wonderful behavior? Would you describe something please?
At this point, the idea is that you cannot see it from mathics-core: everything should work as before. To see the differences, you would need to see in the code, or trying to write a pymathics module that uses this new behaviour.

Is there any test I can run to see the difference?
In principle, if you load (my version of) pymathics-matplotlib, you would see the difference (in one case it works, otherwise no).

[rocky]

Should we be able to write a test that alters Makebox[] in a valid way and see the change in behavior there?

[mmatera]

Overall I really like this. Thanks so much for doing it.

The main question I have is, how can I see the new wonderful behavior? Would you describe something please?

Is there any test I can run to see the difference?
I added some basic tests in te

Overall I really like this. Thanks so much for doing it.

The main question I have is, how can I see the new wonderful behavior? Would you describe something please?

Is there any test I can run to see the difference?

@rocky, please look at test/test_custom_boxconstruct.py to see some very basic tests.

If there are no extra thoughts, observations, or complaints, I would like to merge this during the weekend to finish implementing the pymathics-matplotlib and pymathics-asy.

[rocky]

This name feels a bit weird to me. In a PR off of this I might have changed the name. I don't remember. Right now I can't come up with another name.

[rocky]

Ok. I can't come up with another name and it problably isn't that important either.

Sorry for the delay - thanks for your patience.

[rocky]

Overall I really like this. Thanks so much for doing it.
The main question I have is, how can I see the new wonderful behavior? Would you describe something please?
Is there any test I can run to see the difference?
I added some basic tests in te

Overall I really like this. Thanks so much for doing it.
The main question I have is, how can I see the new wonderful behavior? Would you describe something please?
Is there any test I can run to see the difference?

@rocky, please look at test/test_custom_boxconstruct.py to see some very basic tests.

If there are no extra thoughts, observations, or complaints, I would like to merge this during the weekend to finish implementing the pymathics-matplotlib and pymathics-asy.

Will do with pleasure, but I don't I'll have time to look at before the weekend.