I still don't see a need for this property. All we do is something gtk can easily do itself. We are working around a gtk bug in all window managers. I think that's wrong.

Furthermore we decided to be feature Frozen on X11.

This is a really good idea. Nothing to lose and tons to gain, especially considering we get to use our existing code
+1

+1, sometimes just a little bit of flexibility has a big effect.

From my side a -1 as we are feature frozen. Gtk can easily check whether the window manager supports this property and if not do the same as this patch does. This is clearly a gtk bug and if gtk fixed it, it works for everyone and there is no need to incorrectly claim support for the property. Has this issue been reported to Gtk devs?

https://gitlab.gnome.org/GNOME/gtk/issues/499

Gtk can easily check whether the window manager supports this property and if not do the same as this patch does.

Whilst I would agree GTK should have kept their fallback, re-adding it couldn't be the same (at least, not by following focus)
The flag itself does solve a real problem that we would hit in Qt if we tried to have the same visual effects.

If I had a different window palette for unfocused windows as soon as I open a combo box (and grab focus) the window would appear to be inactive.
Breeze style doesn't have different palettes for unfocused windows, but that's a chicken and egg problem.

We also shouldn't just think GTK, it's exposed in at least SDL as well.

We quite literally have nothing to lose by landing this.

(And whether or not GTK developers want to or need to tackle this issue has very little to do with the price of tea in China, especially considering that this would help us too in case we decide to implement these effects and wouldn't hinder in the slightest)

bump, this is a change I do want to push for.

I am happy to have a technical discussion as long as it addresses the comments that I have already mentioned.

I did look at patching GTK before I went for kwin. It is far from trivial to do correctly in the client because of all the edge cases. Kwin's mantra has always been "window manager knows best" when clients try and inferring things about state, this is another classic occasion.
It doesn't make sense to make a bigger patch that won't work as well as the simple one.

+1, big improvement for almost no code change. I know we're technically feature frozen but I think this is a case where a little flexibility can go a long way.