Add preventDefault for non-Tab, const keys in onKeyDown

[JohnC-80]

Should resolve #2055

[ljharb]

Thanks!

Let's also add a test to https://github.com/airbnb/react-dates/blob/master/test/components/DayPicker_spec.jsx#L240-L273 for this.

[JohnC-80]

Will do! Sorry I'm hitting a bit of a first-time-contributor bump... I develop on Windows so the script commands are giving me a tough time actually running tests/building locally... putting this PR up I honestly left it up to the CI to get it there... I typically would employ something like cross-env to get things working - and I might just do enough there to try and get it to spun up without making too many changes... just hasn't been automatic yet and I've had to move into some other things dayjob-wise... will keep after it, thanks @ljharb ! Any suggestions you have would be great - currently I'm hung on the build CSS step after clearing the BABEL ENV hurdle.. any input would be great. Thanks for a stellar library.

[ljharb]

No worries at all; indeed that command is not Windows-compatible. I'll fix it in master shortly, and then you can rebase and should be unblocked.

[ljharb]

Instead of pushing it to master, I prepended to this PR, so you should be set (reset your local branch to the pushed changes)

[JohnC-80]

Thanks...@ljharb sheesh, now it's stumbling on installing the peers for enzyme-adapter-react...

[ljharb]

@JohnC-80 what versions of node/npm are you using?

[JohnC-80]

@ljharb
node = v12.13.0
npm = v6.13.0

[ljharb]

hmm - that should be using ^16.0.0-0, not 16.0.0-0. if you do npm install -g npm@6, does it still fail?

[JohnC-80]

@ljharb
I'm upgraded to npm@6.14.8 and it's still bombing out with the same error installing those peerDeps for the enzyme-adapter.

[JohnC-80]

Looking at the rest of the code, I see the disableScroll functionality employed in a couple of places.. @ljharb is it present to basically do what this code does? I'm concerned that with the approach here of preventing default, there's some scrolling of some sort that's required in the picker itself or some case is missed - granted, no other tests have failed aside from the one (seems non-related?)

[ljharb]

@JohnC-80 hmm, I'm not actually sure :-/ it seems like it's worth doing a lot of manual testing here.