Billy Tierney wrote in a message to TOM TORFS:
 TT> I'm having a problem here with a program that works fine on 286 and
 TT> above, but  never fails to crash on an XT with a 8088 CPU (it seems to
 TT> get in some sort of  infinite loop, and if I try to press ctrl-alt-del
 TT> it hangs completely).
 BT> Some versions of the 8088 chip have a bug in the way the stack is
 BT> managed... If you modify the stack registers (SS or SP), first
 BT> clear the interrupt flag...as sometimes a hardware interrupt will
 BT> get generated. 
I did that already.
 BT> Avoid use of the string instructions
 BT> (err...STOSB/STOSW/CMPSB/CMPSW/LODSB/LODSW/etc), as in many cases
 BT> they weren't accepted on some systems (My old TEC (Toshiba) had
 BT> that problem... they would work, but the next instruction would be
 BT> corrupted). 
Well, I use string instructions all over the place and I'm not going to omit 
them just for that. They are valid instructions therefore I am allowed to use 
them :-) They weren't the cause either...
 BT> Are you playing with the interrupt vector table, out of curiosity? 
Yes, and with many other things also :-)
The cause was in my so-called "running-line" system. On the 8088, the int 1 
trapping seems to be activated just one instruction later and it seemed to 
have some trouble stepping over things like FNINIT if no coprocessor is 
available, int 3 calls, etc. I eliminated all these and now it works fine.
Thanks for the suggestion though.
greetz,
Tom
tomtorfs@mail.dma.be
--- timEd/2 1.10+
---------------