Nvidia evaluating adding Wayland support, and that was already a year ago (at that time, NVidia was obviously still not evaluating adding Mir support ).
But it's not a surprise, it has always been the case, and it's just what you'd expect from them: they are looking a what it will cost them to support something, and what they will gain, and when. Then they make decisions based on that. But until there is actual support, they'll never confirm anything. If you ask them about mir support, you'll get the same answer.
BTW, is there any write-up which deals with the complexity of an EGL driver for Wayland/Mir compared to an X driver?
AFAIK, some stuff is delegated to the driver, like multi-monitor handling.
OTHO, it sounds like a driver for Wayland/Mir is far easier to support for AMD/NVidia as you have not to deal
with all the 2d-cruft present/required for X (which causes many problems and burns lots of man hours, I guess).
Is this correct?
Last edited by entropy; 06-17-2013 at 05:56 AM.
But, as I see it, what you said is correct by the wrong causes. It doesn't need to deal with X specifics, which are an extra to the support they need to add just for running on Windows (which is their main target in general because of market share), but supporting 2D has nothing to do with the X specifics. I don't know if OpenGL and EGL deals with it, but some cards (I'm not sure if modern cards still ship it) have a 2D engine and it could be managed by a driver even with no X. The problem would be, in the case OpenGL and its derivatives doesn't deal with it, you need another generic interface.
About the complexity, I guess it's the same as an OpenGL driver, since AFAIK is a specification derived from it.
Furthermore, in relation to the KWin team, as I said before they have the right to rip out any code that introduces bugs and vulnerabilities as seen when they dumped their tiling window function much to the dismay of the community.
That said, it would be up to the Canonical team to supply those patches in the first place, but there should be no reason for the KWin team to reject those contributions outright simply because they're distro specific.
And an apology is an apology. Some people can be stubborn and not accept it, but that doesn't negate the fact that it exists.
If you retract one time, you can retract another time and so on. And in case of Ubuntu, they seem able to retract an impressive frequency.
Ubuntu: "We will utilize wayland"
Ubuntu: "We need to retract: we will utilize Mir because wayland repeats X's mistake"
Ubuntu: "We need to retract again: we will utilize Mir, but wayland does not suffers the same X problems"
Ubuntu: "Mir is protocol agnostic (?) and tailored around unity, so API can be changed every times it needs."
Ubuntu: "We need to retract: API no more begin unstable... bla bla bla"
Who can be so stupid to trust in ubuntu's declaration one more time?
I somewhat agree, but:That said, it would be up to the Canonical team to supply those patches in the first place, but there should be no reason for the KWin team to reject those contributions outright simply because they're distro specific.
If Canonical is supplying the patches, and Canonical is the only ones using the patches, what does it even need to go upstream for? The only real reason would be to get upstream to maintain it over time, so Canonical doesn't have to do so anymore. That is asking them to take on more work just to help out Canonical, and it doesn't seem like they are in the mood to do so right now.
Edit: sorry, i guess Rahul already made the same point above.
Last edited by smitty3268; 06-17-2013 at 07:07 PM.