Working for me but has a very noticeable jump

I've been noticing that jump too, but only sometimes (in a Safari window I always get a jump on reload, but on the other window I do not 😰).

This is on 22.6.25

Either way it's unrelated to this specific issue and we should probably open a new issue for that if it persists. The only reason why it could be happening is in globalReady (or due to an unavoidable browser restriction/bug)

is it

ha!

RAAA LOL

 await waitFor(() => document.body); 

That's potentially the issue, but it's not really something new or anything that GitHub or us can affect, that's the weird part.

We can replace it with an observeElement(document.documentElement), assuming that it always exists (can we assume that?)

I forgot already how did we land up with document.body? I would want to check in history why we did not use observe element.

We went though a few options before we came to this

I forgot already how did we land up with document.body?

The content scripts are injected on document_start, which may happen before the document is parsed. This means that the script might find an empty page so we have to wait for it. We use to have requestAnimationFrame to catch the first draw, but that means we can't do anything until the tab becomes visible, even if it loaded a minute ago.

Maybe we can combine the two in a race:

await Promise.race([
	frame(),
	waitFor(() => document.body)
])

If I am not mistaken, requesting the frame broke Firefox. (GitHub froze)

If I am not mistaken, requesting the frame broke Firefox. (GitHub froze)

Almost! It wasn't the frame, but the logic around it:

More pre-bug context/assumptions #4316 (comment)

Full explanation in #4316 (comment)

Basically that's the code I just suggested

Btw there's still some document.body.classList calls in https://github.com/refined-github/refined-github/blob/main/safari/Shared%20(App)/Resources/Script.js, not sure if those should be updated too.

That’s a web view in the native app.