Re: nvidia and the wayland display server
sorry for warming up a relatively old thread but the subject is not closed yet.
Assuming that Wayland gets an abstraction layer like DDX I think it should be easy
for NVidia to provide a driver that hands out a buffer to the client. But how does a
client draw into that? The open source applications will use some non-NVidia library.
So the only way to have clients independent of the vendor of the graphics hardware
is to use OpenGL for rendering. The latter beeing available from NVidia in a version
compatible with their Wayland driver buffers.
Is OpenGL generally suitable for the 2D-part of traditional GUIs? Or are other libraries
foreseen for this task in the Wayland architecture? A related issue is that toolkits like
Qt need to have two underlying backends in the same library if the applications shall
be startable on X11 and on Wayland. Otherwise the choice between X11 and Wayland
is spilled into all the applications which need to be provided for both cases.
So my current understanding is that the difficulty for NVidia is the render library choosen
on application level. Their current aim is to provide accelerated OpenGL. Or am I misreading