On Sat, 22 Nov 2008 16:57:17 +1000 Peter Hutterer <[EMAIL PROTECTED]> wrote:
> On Sat, Nov 22, 2008 at 01:52:48AM +0100, Christian Beier wrote: > > (1) Master Devices created on the fly: > > XTestFakeDeviceButtonEvent, etc. It injects the event server-sided as if the > event was created by the input driver (although there isn't one for MDs), so > it'll generate the appropriate core/device events. I only tested that to some > extent, so please file a bug if it doesn't quite work right. > Having said that, injecting events into a MD directly is a bit troublesome due > to the design of the MD/SD device hierarchy. This is one more thing that needs > fixing before XI2 actually comes out. Sounds in a way more feasible to me than hacking around with uinput. And x11vnc is doing it via XTEST as well. Okay, i can only submit key/button presses and motion events, but I can't get any more over the wire with RFB anyway. Or what's the catch with using XTEST? > > > (2) Fake Slave Devices > IMHO long-term uinput gives you much more flexibility than doing anything with > xtest. I'd definitely go for that approach. Tried this few minutes ago and it was surprisingly easy. So i have to to take back my remarks on XTEST in a way. The only thing I don't like is that my app then needs root privs... Cheers, Christian -- what is, is; what is not is possible.
pgpzSnBXviifY.pgp
Description: PGP signature
_______________________________________________ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg