Peter Brown schrieb:
Hallo, never mind, it had nothing (or very little) to do with the mouse
driver. The IWF (Integrated Workframe) that comes with VAC3 was so
screwed up that it completely hung my system (obviously even with
disabled interrupts, my system was completely dead).
After reinstalling I had no more severe hangs (except for a single mouse
driver trap, but I know the failing location and cause).
Now to your problem:
it most likely is not a mouse driver problem. The reason for this is
that the graphic mouse cursor and the switch from one OS/2
windowed/PM/DOS windowed to another is entirely managed by PM because
all these apps run in a single session.
The mouse driver however only manages the mouse switch between sessions
(for example, it keeps track of the cursor position for each of these
individual sessions) and this only applies to OS/2 fullscreen/DOS
Therefore the behaviour you describe (switching from PM to a detached
session or a DOS fullscreen/Windows fullscreen gives you an "alive and
kicking" mouse cursor) is a good indication that the mouse driver works.
Have you XWorkplace running ? Just for a test I would deinstall and see
if the problem goes away (XWorkplace is quickly reinstalled, I have done
it a couple of times).
Post by Peter Brown
I noticed something "funny" about mouse drivers a couple of years ago.
I usually use MultiDesk/2 to enable all my systems as multi-user, each
with their own Desktop.
Around about the time of eCS1.1 (MCP2) it became impossible to create
New User Desktops that included a mouse. All user Desktops boot from the
same configuration file as the original Desktop that does have a mouse.
All mouse related INI file entries seem to be present when compared to
the original Desktop INI files.
It seemed to me to be something that had changed in the mouse related
files and possibly also somewhere within SNAP.
Using the amouse driver when I create New User Desktops sometimes they
have a mouse, sometimes they don't and those that did have a mouse on
creation may not have the next time that Desktop is booted...
I have even tried creating those new User Desktops using the VGA driver
instead of SNAP but still had the same problem. Guess that takes the
SNAP code out of the equation but not any of the GRADD updates that SNAP
Interestingly if I Ctr-Alt-W to bring up Watchcat on a "mouseless
Desktop" the cursor can be controlled by the mouse...
Come to think of it I have also had a mouse in a winos2 session when the
eCS Desktop in use did not have a mouse...
Basically, something mouse related seems broken in comparison to Warp4
and eCS1 (MCP1). Yes, the trackball in use has not been changed at all
during that time so I know it is not a "hardware issue" - when I
switched from ps/2 to usb connection (I was using an adapter) shortly
after installing eCS1 there was no problem.
Hopefully you have found and fixed some part of the problem. Maybe my
vague description of mouse problems when trying to create New User
Desktops will help track down any other cause(s).