At 11:20 30-09-03 -0700, Eich@xxxxxxxxxxx wrote: >Message: 52 >Date: Sat, 27 Sep 2003 13:05:46 +0200 >From: Egbert Eich <eich@xxxxxxxxxxx> >To: xfree86@xxxxxxxxxxx >Subject: Re: [XFree86] Belkin KVM switch and Gnome, console mouse service >Reply-To: xfree86@xxxxxxxxxxx > >Mark Vojkovich writes: > > On Fri, 26 Sep 2003, jmw wrote: > > > > > > Switching VTs and back should help. XFree86's mouse protocol parsing > > just doesn't seem to be smart enough to deal with this sort of thing. > > > >I'm not sure if this is the problem of the mouse protocol parsing. >The mouse needs to send a mouse replug event. There are two >possibilities: > >a: The mouse does not send such an event: in this case the driver has > no chance at all to detect this condition. (I don't know what > happens. I don't own a KVM switch). >b: The mouse does send this event. The sequence sent can also be part > of a normal mouse sequence. Therefore the driver needs to do some > timing to detect this situation. This is not trivial to do from > userland. Currently the code doing this is turned off. > >You could try >Protocol "Auto" >It notices that something the data format changed and will try to >reinitialize the mouse. > >Egbert. thanks for the reply. i don't know, but suspect this MS PS/2 Intellimouse is mostly passive -- no 'replug'. if there were such a command, i expect it would be part of a driver that is in Win 2K (or other Windows O/S). and that could explain why i don't see this problem in a windows environment. if there is similar code in RH9, or XFree86, and it's disabled, it seems my only choice is to restart GPM at each event (what i do now, anyway). i want to try your suggestion, and i poked around in mouse preferences and settings for RH9, but don't find anything like "auto" for protocol choice. could you be more explicit? -- jmw _______________________________________________ XFree86 mailing list XFree86@xxxxxxxxxxx http://XFree86.Org/mailman/listinfo/xfree86