10 Nov
2021
10 Nov
'21
10:06 p.m.
It is certainly true that NML might be choking on legitimately bogus
monitor data.??? However, one hesitates to observe that signaling
incorrect input by crashing is a misfeature...
The monitor call would be nice to know.? You can trap both UUO's and
JSYi on Tops-20 in a number of ways to see what is actually happening,
whether or not you have the source.? I can't remember whether Tops-10
has a similar feature.? I thought not, although I'm fairly certain it
does have address break.
If the crash dump is Galaxy based, then it will have the AC's preserved
in a block in the crash dump.
------------------------------------------------------------------------
On 11/10/21 1:11 PM, Robert Armstrong wrote:
I don't suppose DEC supplied source for NML?
??Doesn?t look like it ? there?s nothing in [10,7,NML] except for the
NML.EXE.? Oddly there are sources for FAL, NFT and even NCP but not
NML.? I wonder if there?s some ?trick? for building NML.EXE from some
other source?
? In any case, following the rule of ?no matter what bad data you give
in the monitor call, it shouldn?t crash the system? I don?t think NML
is the problem.? It might be interesting to figure out exactly what
call NML is trying to execute, but the bug is in the monitor.
Bob