[fa.info-vax] DBG lossage **EXTRODINAIRE**

info-vax@ucbvax.ARPA (07/31/85)

From: *Hobbit* <AWalker@RUTGERS.ARPA>

Under 4.x, DBG has changed considerably for the worse in its handling of
registers.  Instead of referring to then as Rn, they use this asinine %Rn
construct, which is a DBG-permanent symbol that refers to someplace in 
memory where DBG presumably stores your user registers.  You can deposit
and examine them in the usual fashion,  **UNTIL** ....

The little routine you were just executing blams into an error and returns
with some non-normal status.  Then, the useless trash will *not* let you
modify registers!! You can deposit away to your heart's content, and DBG
just returns like it  did it, but you'll find that your registers contain
something dumbass like 80010FFE and there it stays.

What I'm doing here is RUN/DEB on some image or other, with the idea of looking
for a few longwords here and there.  So I type in this little 4-instruction
thing to compare the contents of r7 with some address pointed to by r8, and
return [to DBG] on a match.  DBG does eventually return, with some message
that VMS can't associate a string with, and then I can't do squat with my
registers after that.

How do I convince DBG that I still want to do things, and it should behave
itself so I *don't* wind up throwing my manuals across the room??

_H*
-------