Hi, I've been using this as a Tampermonkey script for a while now, but an issue is that it won't trigger unless the YT page is reloaded. I've tried changing some of the Settings parameters like "Run at" but it doesn't help.

If I watch the Tampermonkey icon I can see the overlay "1" only appears after a page load.

Other than that it seem to work really well but it would be nice if this could be figured out.

This is using Vivaldi 3.8.

Include #4 (comment), then tag, release, build, and publish to the Chrome Web Store.

Damn, I should have tested this more thoroughly before submitting it to the Chrome Web Store.

I didn't realise there'd be a difference between running it through Tampermonkey vs as an extension, but here we are! It works fine in Tampermonkey, but when run as an extension (using the CRX file), it mistakenly detects the leaving fullscreen transition as the entering fullscreen one, thus when leaving fullscreen, it logs:

[YT-FFT] Entering fullscreen; fluff has been hidden
[YT-FFT] Finished entering fullscreen transition; showing fluff again

So it immediately shows the fluff again, nullifying the performance improvement of having hidden it.

Perhaps the addEventListeners run at a different time, and so we end up with a different order of event handlers firing (mine and YouTube's)...

I'm detecting whether we're fullscreen (and so what the direction of the transition is) by the presence of the fullscreen attribute on ytd-watch-flexy. I guess by the time the extension checks that, it's too late, as it's already been removed?

Similar to #3, I realised that pressing Escape to leave fullscreen is still slow. I'd forgotten to handle that.

I had a bit of a go, and it seems like neither detecting Escape keydown nor the fullscreenchange event happen soon enough (or at all?). If this is even possible to fix, it's going to be difficult. I'd rather not dive deep into how YouTube's fullscreen toggling works to delay the actual transition.

Handling F11 is related. I haven't tried to do that yet, but I suspect it'll require the same mechanism.

My YouTube transition lag situation seems to have improved. I've just updated to Chrome 109.0.5414.74, so maybe it's that. Or maybe from recently installing hardware acceleration packages on Arch Linux that I didn't realise were necessary: libva-utils, mesa-vdpau, and libva-mesa-driver... Anyway, I can now barely notice a difference with and without this extension when viewing a single video; and the delay for a video playing from a massive playlist is much reduced (~1 second on my ~5000 video Watch Later list). However, the F key is no longer triggering this extension's fast transition, unfortunately. Clicking the fullscreen button and double-clicking the video still work though.

I think I need to change it so the key's listened to when only the video or any of its controls have focus - maybe this can be done by instead adding the event listener to a parent element of the video? I want to make sure it's on anything that the F key normally works with.

It would appear that Vivaldi needs a longer delay before revealing the fluff again when entering fullscreen. It looks like it might be doing an extra repaint, with the toolbars introducing another step in the transition. I'll have to make the code smarter to somehow poll until it's finished transitioning. Users, in the meantime, you could use the script with Tampermonkey and try adjusting the delays in the code yourself - on my new Linux laptop, it looks like the setTimeout(showFluff, DELAY) needs to have the delay be at least 150 (ms). Too high a delay though, and leaving fullscreen will take longer unnecessarily. Hence I'll need to again separate the codepaths for entering and exiting fullscreen to use different delays.

I just discovered tonight that I'd forgotten about double-click xP I tend not to do that since it takes slightly longer to register.

I'm not sure yet how to reliably hook into this; investigation required.

Upon entering fullscreen, there is now about 20 pixels of space you can scroll down to below the video. Due to this, there is a scrollbar overlaid. This extra space is causing the extension to be unable to detect that that video size has become fullscreen; so the detection is timing out and so unnecessarily waiting a few seconds before revealing the fluff again. At this point, the scrollbar goes away.

YouTube must have changed something to cause this - looks like I need to keep up with their changes.

There are a lot of them