Hi,
My travel to XDC 2008 is finalized now, I am collecting the topics I should bring up there. I already have these:
*) Mouse input handling. I've CC'ed Vitaliy. All I know so far is that Vitaliy said opening the core pointer as XInput device would be good. Vitaliy, do you have any proof of concept code / test apps? Do you have time to discuss the details? I think this is the most important topic, and I don't know anything about it, so I should be well-prepared.
*) Tablet support. Dan mentioned something about it, but I have no details.
*) Graphics support: I'll talk to the X11 devs about including the wine d3d tests in their driver QA. Lobby for widespread support of some rare extensions(GL_NV_half_float) and help on adding some others, like for the flat shading issue. Others?
*) Discuss the opportunities we have with DRI2
Does anyone have other items?
Hi,
My travel to XDC 2008 is finalized now, I am collecting the topics I should bring up there. I already have these:
*) Mouse input handling. I've CC'ed Vitaliy. All I know so far is that Vitaliy said opening the core pointer as XInput device would be good. Vitaliy, do you have any proof of concept code / test apps? Do you have time to discuss the details? I think this is the most important topic, and I don't know anything about it, so I should be well-prepared.
*) Tablet support. Dan mentioned something about it, but I have no details.
*) Graphics support: I'll talk to the X11 devs about including the wine d3d tests in their driver QA. Lobby for widespread support of some rare extensions(GL_NV_half_float) and help on adding some others, like for the flat shading issue. Others?
*) Discuss the opportunities we have with DRI2
Does anyone have other items?
Most xorg devs aren't a big fan of pbuffers but they are part of the GLX 1.3 spec which drivers want to support these days. Some drivers are considering to just fail on pbuffer creation. Encourage them to offer pbuffer support on some visuals/fbconfigs. A lot of windows apps rely on their presence including Wow in opengl32 mode and lots of other (older) programs.
Roderick
Roderick Colenbrander wrote:
Hi,
My travel to XDC 2008 is finalized now, I am collecting the topics I should bring up there. I already have these:
*) Mouse input handling. I've CC'ed Vitaliy. All I know so far is that Vitaliy said opening the core pointer as XInput device would be good. Vitaliy, do you have any proof of concept code / test apps? Do you have time to discuss the details? I think this is the most important topic, and I don't know anything about it, so I should be well-prepared.
I don't have any code for this unfortunately.
What Wine needs is the physical mouse movement events, that work even when pointer itself "stuck" on the window/screen border. This should also work for the entire desktop not just one window. Absolute/relative events - doesn't matter, it's easy to convert from one to another.
In short - in the current xorg code there are checks to explicitly prohibit opening of the primary pointer and keyboard device as XInput device.
There is some capturing extension don't remember the name of it. It requires two server connections and works for the specified window only. This won't work for us.
Vitaliy.
Am Freitag, 4. April 2008 18:55:03 schrieb Roderick Colenbrander:
Most xorg devs aren't a big fan of pbuffers but they are part of the GLX 1.3 spec which drivers want to support these days. Some drivers are considering to just fail on pbuffer creation. Encourage them to offer pbuffer support on some visuals/fbconfigs. A lot of windows apps rely on their presence including Wow in opengl32 mode and lots of other (older) programs.
Can we emulate pbuffers via FBOs, or are there some problems? Pbuffer support is troublesome on MacOS and the ATI binary driver as well, so even if the xorg devs implement it we'll be stuck with a partial fix only
Most xorg devs aren't a big fan of pbuffers but they are part of the GLX 1.3 spec which drivers want to support these days. Some drivers are considering to just fail on pbuffer creation. Encourage them to offer pbuffer support on some visuals/fbconfigs. A lot of windows apps rely on their presence including Wow in opengl32 mode and lots of other (older) programs.
Can we emulate pbuffers via FBOs, or are there some problems? Pbuffer support is troublesome on MacOS and the ATI binary driver as well, so even if the xorg devs implement it we'll be stuck with a partial fix only
In theory you might be able to offer this emulation but I think it will be very tricky and it might not work well at all. Remember pbuffers are WGL things while FBOs require a GLX/WGL context. You would have to override a bunch of GL calls, states .. I don't think this works well. Second WGL allows you to query the capabilities of a pbuffer, we would have to fake most settings which might not result in optimal performance (e.g. color bits, stencil ..).
Mainly overriding stuff and the requirement of a GLX/WGL context is problematic I think. I fear a half baked implementation which works for some of the easier apps. The drivers could also emulate pbuffers using FBOs and they might have less issues.
Roderick
-
Roderick Colenbrander -
Stefan Dösinger -
Vitaliy Margolen