feat: enable uninstalling PatchedMediaKeysApple

[martinstark]

Close #4469

The idea is to only enable uninstalling when it is desired, in order to not hold things in memory unnecessarily.

[github-actions]

Incremental code coverage: 97.83%

[martinstark]

Will re-open once I run some more tests.

[martinstark]

Tested and working.

[joeyparrish]

I don't understand this PR or the associated issue, and it seems really odd to me. Let's discuss on the issue, but for now, please do not merge the PR. Sorry for the inconvenience.

[martinstark]

Thanks @joeyparrish, I'll look at adjusting according to your feedback and add a test as well as docs

[martinstark]

@avelad @joeyparrish I've addressed the feedback, introduced unit tests and added a short blurb in the documentation.

I opted for not restricting the unit test to only run on Safari. The functionality that's being tested is just overriding and restoring browser globals.

I don't know of a way of checking strict equality on the mediaKeys getter. The test does a best effort check to see if it looks like the value is no longer overridden, but it doesn't guarantee that it was restored.

I have not yet run the latest PR updates on a Mac, I'll update the PR once I have.

Edit,
tested and working

[avelad]

@martinstark It seems that the tests have failed, can you check it?

[martinstark]

@avelad managed to finally get through linting 😅

 1) cues ending exactly now
     TextDisplayer layout using browser-native rendering
     native-end-time-edge-case: Expected 0.9133498736570255 not to be less than 0.95.

I don't think this is related to my changes, what do you think?

[avelad]

I'll re-run the failed jobs

[martinstark]

@avelad the safari test still fails, I've rebased on the lastest main to see if that makes a difference

[martinstark]

1) cues ending exactly now
     TextDisplayer layout using browser-native rendering
     native-end-time-edge-case: Expected 0.9133498736570255 not to be less than 0.95.

Keeps failing with exactly the same number. I've had a brief look at the test in question, as well as the text displayer, but no ideas spring to mind.

[avelad]

I'll ignore the fail test to approve the PR. Please send the final code please (no comments in the tests)!

[martinstark]

@avelad I tried to remove the unit tests to see if that's what was causing the Safari test suite failure. I've reverted this and squashed everything.

Thanks for your time!

[avelad]

A minor nit, otherwise LGTM

[martinstark]

@avelad not sure what your process is, I just squashed and force-pushed after you approval, but that dismisses your review. Any input?

[avelad]

The squash it's done by github, so really it's not a problem :) I'll approve again

[avelad]

@joeyparrish can you review it? Thanks!

Changes addressed, reviewed by Álvaro