I've mentioned in another thread how the lpt driver bogs my system down.
In particular, X becomes very jumpy.  This brought to my attention another
thing that I thought that I should ask if others have seen:

The X mouse-pointer _graphic_ seems to be about 1 update behind the X
internal mouse-pointer _state_.  (This is made noticable, as I imply, by
heavily burdening the system with /dev/lpt interrupts.)

I first noticed this when I found that if I moved the mouse quickly, then
stopped, the mouse would stop (surprise, surprise), but when I made a
minor adjustment to the mouse, the first screen-update of the position
would be a largish jump in the direction that I had previously been
(quickly) moving the mouse.

Further, when running Navigator-4.51, I noticed the following: I moved the
mouse to where it should have been over one of the toolbar buttons, and
stopped.  I noticed that, because of fortuitous timing, the mouse stopped
well OFF of the button.  HOWEVER, Navigator received from XFree86 the
information that the mouse was now over the button, so it highlighted the
button.  The picture looked something like this:


         ************************
         ** Highlighted Button **
         ************************




     X  <=- The mouse pointer was about here.


(The above roughly conveys the situation, though the X graphics looked
nicer, and it should be scaled down by a factor of 2.  *grin*)

I waited a moment to see if the mouse pointer would jump to the button, as
I recall.  It didn't.  But when I moved the mouse a little, the onscreen
pointer jumped over the button.


Can anyone else reproduce at least the ``big jump'' effect, if not clearly
show that X applications have different (and more updated) notion of the
mouse than does the X graphic display?


  "I probably don't know what I'm talking about."  --rkr@rkr.kcnet.com