Handle correct unicodedata method exception

[wesinator]

unicodedata methods return ValueError, not TypeError. https://docs.python.org/3/library/unicodedata.html
fix #6

@yamatt

[yamatt]

Amazing. Thanks @wesinator this is my first PR since taking over the project. So let me check a few things and I hope to get back to you soon.

[yamatt]

So I've had a look around @wesinator.

So I agree with you that exception type should be ValueError according to the docs and as in your PR.

Although from testing, what happens is that unicode.name does in fact raise a TypeError if you enter an empty unicode string.

Then what happens is the TypeError is caught, but nothing is done with it except that there's a comment about Python 2 -- which is long out of support. Then the code continues on to the section below the error block, that then throws another own TypeError.

Personally I find the flow in this method a bit strange.

I'm not certain of the intention of this bit of code. Possibly one option is to ensure we have a valid char parameter in this function. But I'm also thinking this might mess up some deployments depending on how they defensively they wrote their code.

Could I ask, if you wouldn't mind, providing a bit more detail on the issue you encountered here? Was it just to align it to the docs? Or were you encountering an error?

[wesinator]

Amazing. Thanks @wesinator this is my first PR since taking over the project. So let me check a few things and I hope to get back to you soon.

#5 was the first?
I think you have to manually follow your own forks to get notified of activity - unfortunately Github doesn't have the fork author automatically follow their own repo yet.

[wesinator]

That's interesting that it throws a TypeError, the Python docs don't seem complete.

This was to fix someone else's issue, #6

It looks like the purpose of to_ascii is to convert glyphs to ascii representations -

homoglyphs/README.md

Line 124 in fe302e5

### Converting glyphs to ASCII chars

I'll look at updating the PR. Also I'll probably open a python issue for the undocumented TypeError, since that also seems odd because other non-empty string types work in that method.