#MungWall bug??
9 messages in this thread
Does anyone have any scenario under which MungWall might decide to not
free some memory and NOT output anything? I have a case (repeatable) in
which I free the memory but it doesn't get freed, and I get no information
from MungWall. This one really has me baffled guys. HELLLLLLLLPPPPP!!
What memory free routine are you calling? It's not a pooled memory
scheme, is it?
–Doug
Doug, nope, it's FreeMem(). It's not on all the calls, only some, I… can't
find out what's different (at the AllocMem()/FreeMem() level). … I watch what
happens to the blocks being 'freed' and most of them get… munged just fine
(and disappear from the munglist list)./exit
Are you calling FreeMem from a Disable or a Forbid ? That might have an
effect.
andy
Andy,
Talked with Carolyn Wed morning, got the problem resolved in the
afternoon, to make a long story short, the last time we installed all the
system sw I accidently installed the serial version of mungwall instead of the
parallel <sigh>. I've found, tho, a definite bug in MungWall (which is now
fixed) and memory leak in MetaScope. Boy, one can sure get sidetracked when
one starts playing with these debug tools. I _DID_ at least find the problem
that triggered all this.
So why did you install the serial version when you wanted the parallel port
version?
:-0 (running away)
John,
You can run, but you cannot hide.
When I was 'fixing' a couple minor bugs in mungwall, I managed to do
the same darn thing. Took me 20 min to discover why my new fix didn't
work <sigh>. Must be too little sleep. BTW, I sent my fixes to Carolyn
and hopefully she'll release new versions shortly (I'm particularly
pleased with my enhancement to munglist).
Sounds good, then.
andy
Andy,
Yup, all solved, now I only have 1023 bugs left to find <g>