Subject Debugging Memory Growth
Author lester@lsces.globalnet.co.uk
I have debugged my previous problem - the threaded timer module from LMD
tools was creating a new thread - switching back to the basic TTimer
cleared that problem.

Now I have found how to make Memproof do just what I want, I am looking
at a couple of problems which may be due to Memproof it's self, or could
be a real problem.

I have switched to IBO36CeTest ( from 36A ). On W2k if I leave
IBO36Cex.BPL in the monitoring list, then it crashes the program at the
point where IBO is first contacting gds32.dll.

On all three versions give a similar pattern. ( 98 loads dll's in
reverse order )

Exception 80000003 BREAKPOINT at <os specific>
after loading IBO-BPL

Exception 80000003 BREAKPOINT at 00441102
before loading gds32.dll

This is the point W2k collapses at with a
Exception C0000005 ACCESS_VIOLATION at 62050AAD
instead - which is in memproof

I then get an
Exception 0EEDFADE Unknown exception at <os specific>

but that is a different bug.

(Separate Post on that)

Since I am all set up with three similar configurations of machine with
each os running, ( I have my slave station software running on each ),
is there anything I can look at closer to get to the bottom of this?

I am only using 98 and NT in production, so the fixes so far cover them,
when I have packaged that up I will go back to a simple IBO only project
and have a look anyway.

--
Lester Caine
-----------------------------
L.S.Caine Electronic Services