> On Wednesday 13 November 2002 1:18 am, [EMAIL PROTECTED] wrote: > > > On Tuesday 12 November 2002 12:56 pm, Dr Andrew C Aitchison wrote: > > > > On Tue, 12 Nov 2002, John Tapsell wrote: > > > > > On Tuesday 12 November 2002 10:35 am, Dr Andrew C Aitchison wrote: > > > > > > I think it would need either two event queues, or a new parameter > > > > > > in the event queue to say which input group created the event. > > > > > > I don't know that area, but I'm not sure that such a change could > > > > > > be done without changing the protocol. > > > > > > > > > > Hmm, can this be worked around? Apps don't need to know which input > > > > > group provided the input, and to them they could see just one input > > > > > group. (If two input groups have focus on the same window, perhaps > > > > > merge the two event queues.) > > > > > > > > The WM is another application, and it needs to know which input > > > > group triggered the event if it is to maintain two foci. > > > > > > Sure, but it is just about feasible to require that the WM has to be > > > changed to use some new/existing extension, as opposed to changing ever > > > app. > > > > > > Can anyone tell me what extension gimp uses for graphic tablets etc? > > > Would that be applicable here? > > > > I haven't looked at the GIMP code, but almost certainly it is using the > > XInputExtension. That's the point of the XINPUT extension, to add support > > for graphics tablets and other additional input devices. > > Can WM's be hacked to use the XInputExtension, have multiple focused windows, > and do everything we have mentioned?
Maybe. I'm not sure. Having the WM be aware of the XInputExtension would certainly help, but I don't know if you could do it all without making any changes to the client apps themselves. Perhaps, by using the XTEST extension too, you could do it. That is, you have the WM get the events via the XInputExtension and then turn around and feed appropriately modified synthetic events into the queue via the XTEST extension (so they don't appear to be synthetic). > > > > > > Besides, if a game or a CAD or similar program uses two > > > > mice/trackballs/joysticks I'd expect it to want to know which one > > > > sent the event, but maybe we could do that by using different events in > > > > the same input group. > > > > That does make it difficult to know how much to merge two inputs in one > > > > group though. > > > > > > _______________________________________________ > > > Xpert mailing list > > > [EMAIL PROTECTED] > > > http://XFree86.Org/mailman/listinfo/xpert > > > > _______________________________________________ > > Xpert mailing list > > [EMAIL PROTECTED] > > http://XFree86.Org/mailman/listinfo/xpert > > _______________________________________________ > Xpert mailing list > [EMAIL PROTECTED] > http://XFree86.Org/mailman/listinfo/xpert > _______________________________________________ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
