This new error is probably b/c you're reading off the end of a chunk.
That's why it's rare; it only happens if you read too far. Often happens
with arrays and reading from either off-by-one array index access or often
if you use flags such as -1/0xFFFF to mean noSelection then sometimes you
will read from array[0xffff] which is obvisouly not going to be in your
allocation block.
Alan Pinstein
Synergy Solutions, Inc.
http://www.synsolutions.com
1-800-210-5293
>On Thu, 3 Jun 1999, Bill Kirby wrote:
>
>>
>> >ExpandFieldHeight <- here comes the problem every time
>> >
>> >When I unchecked 'Single Line' and compiled again, warning has dissapeared
>> >but the new type of warning started to show up
>> >"MyApp read directly from memory manager data structures."
>> >This message appears maybe 3 times in 100000 events generated by gremlin
>> >0.
>>
>> It sounds like you have the 'dynamic-size' attribute set for your field.
>> This should not be set for single-line fields. I ran into the same problem,
>> and after turning off the 'dynamic-size' attribute, the memory warnings
>> went away.
>> --
>> Bill Kirby
>
>
>Yes, I had 'dynamic size' attr. set for that field (it was written down in
>my first mail). It's interesting that Constructor (CW R4) actually lets me
>have both set.
>
>I have turned off 'single-line' and the message didn't show up again. The
>question is why the other message "MyApp read directly from MemMgr data
>structures" is still here (but very rare - 3 times within 100000
>gremlin steps).
>
>
>I wonder why such warnings show up so irregulary. Why not *every* time or
>not at all?
