SS> Since I use my OS/2 machine for programming DOS (and 
 SS> hopefully soon, OS/2)
 SS> apps, my system is a little more vulnerable to crashes due to code in
 SS> progress.  For example, I've had some bad code (e.g. stray pointers, bad
 SS> addresses, etc.) that locked OS/2 (2.11) tight as a drum.  Once I crashed
 SS> OS/2 so bad that my boot drive was fouled where OS/2 said a kernel file
 SS> was missing upon reboot.  I had to reformat and reinstall after that one.

I find this distressing;  if you can reproduce it make sure IBM finds out about it.

I don't use many DOS programs,  but I've had several that will lock
real-dos up tight (where OS/2 would kill the vdm).

Are you sure it is not a hardware problem?  I have a dos program that will
crash os/2,  but only because it tells a piece of hardware to do nasty
things on my bus.

Are you running a DOS debugger when these things happen?   A dos debugger
will really stretch OS/2's crash protection to the limit.  You might
consider Watcom,  which allows you to use the OS/2 hosted debugger to debug
dos apps.


--- Maximus/2 2.01