llyyr
> if this is grouped together with BIT_BORDER, does that work well enough for mpv? It being part of BIT_BORDER or its own flag, either way works for mpv. >...
> squeezing blur into an existing flag might work out ok I don't agree with this interpretation. I believe my PR is more akin to a bug fix than introducing...
> If we introduce a new bit, it will require a newer libass _and_ a newer mpv. This is fine. I'm not opposed to any of the ideas above, as...
> actually mark it as deprecated and officially discourage any further use This would need to introduce a new API to pass like `ass_set_style_overrides`, but it would take in ASS_Style...
> macOS does internal dithering so the old behavior was correct. There's no real harm to dithering down to 8 bpc in mpv itself if the content isn't wide-gamut. And...
Can't reproduce, when I exit mpv or toggle fullscreen in step 6, my screen 2 stays disabled as it should and my screen 1 remains enabled as it should.
> it's slow Why does this need to be faster than a Lua script? > not builtin mpv. Why does such a non-standard use case need to be built in?...
I still don't understand why this is so performance critical that `fuzzydir.lua` isn't fast enough.
> I don't see any point in merging that to be honest, there weren't any important fixes on libplacebo side recently. I half agree, there weren't any important API changes...
Fixes #15120