Quote Originally Posted by RussianNeuroMancer View Post
Soo... maybe there is Gala or/and Cinnamon issue? Maybe you have to contact with Gala and Cinnamon developers and ask them to look into this issue? If there is really driver issue, they can say what exactly fail, contact with Catalyst Team and ask for fix - that will be easier and faster solution for everybody, isn't it?
Gnome Shell developers already fix that but Catalyst Team probably need to get doesn't crash previous builds of Gnome Shell (that doesn't include fix). Also Catalyst developers is writing patches for KWin (Jammy Zhou) - that what called "proper support of catalyst" on Phoronix.
No, it's not a Gala/Cinnamon issue, when I have it on other window managers too and only under Catalyst driver. And as mentioned doesn't happen on other drivers. Or does mesa/gallium also have wrong GLSL and whatever else specs.
I've asked to at least include Gnome Shells fixes on Gala and Cinnamon as well by default(note, I don't use Cinnamon, only test now and then). If that gets some response, I'll go on with more things to report, but I don't see much that traction in non game bugs.
I mentioned KWin to show you one(of many) example where developers have to treat Catalyst as a special case. Compiz also does that a LOT. Don't you think something is wrong with that?
In "no hacks mode" I can't even take proper screenshots(they might show the background, they might show an application that was previously open, weird stuff). Is that a window manager problem too? Again Catalyst specific.
Certainly everyone is broken and Catalyst follows everything to the letter, but most applications are apparently also developed in a broken way. Do you think that's how it is?

I am not an opensource driver fanatic, but at least the open drivers seem to (mostly) work without bugs when a spec is implemented. At least they are implemented properly I mean and no application specific hacks are used. Certainly I can't say quite the same for nvidia. But can you claim that such a different behavior just by renaming an executable file is not rational from a properly working driver. Past experiments have shown certain OpenGL extensions to be available or not depending on the name of application(let's say compiz). I haven't tested much on that now(not enough knowledge on my part), but I hope it's not as bad as it used to. Because having GLX_EXT_texture_from_pixmap or whatever(I don't completely recall the name of the extension) only when you're named compiz(probably a few other names as well) is pretty wrong from my point of view at least. (And racist, lol)

Btw, that's a bugfix update for Shell/libmutter, still it should be backported, but who am I who know whatever genius versioning system they'd use.

May I suggest an experiment. Copy "gnome-shell" executable to a file with another name. Run "/path/to/new-name --replace". See how well it runs on Catalyst now that application specific bug was fixed. You can try it with other wm's too. Maybe it runs fine, maybe not, who knows.