On 02/07/2015 13:42, Emil Velikov wrote:
On 27 June 2015 at 12:21, Jon TURNEY wrote:
This smells to me like there's something not quite right about the
partitioning of stuff between libmesa and libglapi, but I don't have any
insight into what exactly the problem is.
Quite likely. Although it would be better to resolve it, as (if my
understanding is correct) libGL should not know anything about mesa's
internals. It ought to interact with DRI and glapi alone.
Mildly related:
Jon what happened to the applegl-vtable branches ? I really liked the
direction that they were going.
I'd quite forgotten about those.
Currently, there is a build-time choice between using apple-dri, or
indirect and swrast. This is not ideal (I believe it makes remote
clients under OSX useless), so they should all be runtime alternatives.
Those branches are really just experiments of mine to try to understand
what that might look like before I did a similar implementation for Windows.
But I think I came to the conclusion that it was less work to use the
existing __GLXDRIdisplay implementations as a model, rather than convert
appple-dri to that interface and then use it as a model for my
windows-dri implementation (see [1])
[1]
http://cgit.freedesktop.org/~jturney/mesa/commit/?h=windows-dri&id=84b173c63727099ec74edc74a15c6ea953312360
I think it's perfectly possibly to put apple-dri into a similar form,
it's just someone has to do it :)
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
