[spell-check] Fixes for a possible patch release

[savetheclocktower]

A couple fixes need to be applied to spell-check; this is the PR that will hold them. Opening as draft for now; I'll commentate as I add commits.

[savetheclocktower]

First we need to fix CI; there's a failing test in the spell-check specs. This test passed for me on macOS, but is failing on Linux, so I can only assume that the Linux dictionary considers foo to be a word, and macOS’s dictionary doesn't. We'll soon see if my theory is correct!

[savetheclocktower]

This commit removes the console.log I left in inadvertently… and also applies some small fixes to eliminate some ESLint gripes with the specs.

[savetheclocktower]

Finally, I'm probably going to argue for reverting the change we made to the default spell-check.grammars setting. Here's why:

• There's something a bit dumb about this package: whenever an edit is made and the editor has stopped changing for a short period of time, the entire document is re-checked. This is rather costly when the document is large (like a several-thousand-line–long source code file). The spell-checking happens in a background thread, but if you're making frequent changes to a certain file, the checking happens often, and it's CPU-intensive. It could be quite a battery drain.
• So at the very least it could be argued that adding source comment was premature, and that it would be better to revert this change and instead add the ability for spell-check to check/recheck the buffer incrementally. This is worth doing even independent of our decision with this setting.
• Still, after having used this feature for a few days, I'm completely on the fence about whether it's desirable for the average user. It has helped me spot a few typos in my code comments, but it also complains about things that aren't words — like camelCasedMethodNames — and those aren't actionable. It also complains about file extensions — so if I make reference to a highlights.scm file in a code comment, the scm has a red squiggle underneath.

For now I'd like to revert the change to spell-check.grammars setting, and then we can have the debate later on when spell-check is a bit more considerate of the user's system resources. If a user still wants it, they can add the value back in the package settings — the part that enables selective buffer checking isn't going anywhere.

[confused-Techie]

@savetheclocktower I'll take the blame for changing the default spell-check behavior. It made total sense to me, but you do make a great case for why we shouldn't have it enabled.

Maybe some time in the future, I'd love to see a way that could allow us to enable this behavior again, while having some intelligent considerations on spell checks end. Such as maybe skipping camel cased items when spell checking comments. But I cannot even begin to imagine how large of a task that might be, so it makes much more sense to disable it for now, and see if a time like that ever comes.