(WIP) forth: implement

[cmccandless]

Resolves #739

TODO:

• add README
• add forth to config.json
• create solution template forth.py
• write test cases forth_test.py
• write example solution example.py

[N-Parsons]

It would be best to avoid using input as an argument in your example.py since input is a built-in function and this can cause confusion.

It would also be preferable for your example.py to use named exception types in your try: ... except: ... blocks rather than just catching all possible exceptions.

I think that the tests would be better implemented using exceptions rather than returning None. In gforth, most of these errors would raise a "stack underflow", so we could define a custom StackUnderflow exception. Division by zero could raise a ZeroDivisionError and other errors would probably be fine with ValueError.

[ilya-khadykin]

We've switched to snake_case naming convention for topics to increase consistency among tracks, so could you please choose them from the following list - https://github.com/exercism/problem-specifications/blob/master/TOPICS.txt

[cmccandless]

Topics renamed to conform to naming convention.

[ilya-khadykin]

Overall it looks good 👍
I've left some comments

@N-Parsons pointed to some issues, I think they should be addressed as well

[ilya-khadykin]

Are these test cases adopted from https://github.com/exercism/problem-specifications/blob/master/exercises/forth/canonical-data.json?
If so, please state that in the comment denoting version as well

[cmccandless]

Yes, they were adapted from the canonical data. Comment with test data version has been added.

[N-Parsons]

Don't forget that from example import ... needs to be changed back to from forth import ....

[ilya-khadykin]

Not sure why difficulty is set to maximum value, it should be lower

[cmccandless]

Difficulty lowered to 5.

[N-Parsons]

Minor point, but I think this might be better as a StackUnderflowError to make its meaning a bit more clear and to differentiate it from arithmetic underflow.

[cmccandless]

When you put it that way, I agree. Original intent was to match convention, since there is an existing OverflowError.

[N-Parsons]

Looks good

[ilya-khadykin]

👍

[ilya-khadykin]

Thanks for your work and congrats on implementing this one 👍, @cmccandless

[ilya-khadykin]

@cmccandless, one more thing: will you please remove WIP tag when you are done with the PR to avoid confusion in future PRs?