>At the end of the display:
>Heap Summary:
>
> flags: 8000 (What does this represent?)
Probably the heap Flags.. you can read about these in the Memory Manager Docs.
> size: 1E7EF2 (Total size of the heap?)
Well, it's about 2Meg - 96k, so I'd say yes, and you're on a Palm V. Am I
right? Maybe a 3... I can't remeber if they bumped the dynamic heap to 96k
in the III or in OS 3.1... But I think it was 3.1.
>
> numHandles: #200 (Total number of handles...in use?
>Available?)
Dunno... doesn't matter TOO much for you.
> Free Chunks: #12 (1E58FE bytes)
> (Does this number represent the total amount of free memory
>in the
>heap? If there's a memory leak, is this where I should see memory
>decrease over time?)
This is how much FREE SPACE there is in the heap. I don't know what the #
is exactly.
> Movable Chunks: #51 (0022C4 bytes)
>
> (These chunks are movable. Are they free? Or could they
>be in use,
>but still movable? What would an increasing/decreasing # mean?)
This is how many HANDLES there are allocated. MOVABLE chunk. DmNewRecord,
DmNewHandle.
>
> Non-Movable Chunks: #0 (000000 bytes)
This is how many NON-MOVABLE Chunks (pointer-based chunks, MemPtrNew) are
allocated. Since this is a DATA heap, there are NONE.
>1) Is it safe to run heap commands like hd or ht while my application is
>running? I've found that this can make CodeWarrior unstable...am I doing
>something incorrectly? What about while Gremlins are running? Or should I
>do this before the app is loaded?
I have noticed instability in doing this if my app isn't 'idle'. I always
STOP gremlins temporarily.
>
>2) Using the console, then debugging, then recompiling with changes and
>using the console again can render the whole IDE unstable. Am I missing
>something here? If I close the console before re-compiling, it's usually
>fine.
Dunno.
>3) Do I have to mess with the settings of mdebug before using heap
>commands? Why?
>
>4) What is hchk actually checking? And are hchk's results based on
>mdebug's settings?
I think it checks the flags and makes sure the heap isn't corrupt. It gets
called automatically by MDEbug if heap check is activated. Do a MDebug to
see the current settings.
>5) Occasionally, I'll get a message that the heap is corrupted. For
>example, if there's a FrmHandler in a loop that
>creating/locking/unlocking/freeing handles and I issue "ht 0", I sometimes
>get a "corrupted heap" message:
>
>#ERROR: Invalid Handle field
>#### ERROR: Heap Corrupted!! #####
>
>I assume that this is because the hd is issued at some inappropriate time
>in the loop relative to the state of handles. What is this "corrupted
>heap" message referring to? Under/overlocked handles? Bad pointers? And
>how does it make this determination?
>
>5) mdebug's primary function seems to be to make the IDE unstable. ;-) Is
>this normal? mdebug -min seems like a useful thing to use, given that
>we've been talking about the minimum space on the dynamic heap. But using
>it can cause an "Unhandled Exception" error in the IDE, not unlike the
>infamous error about Palmrez in CWR5 before the patch was issued. I must
>be missing something.
Yes, i"ve noticed this, too. I do my MDEBUG testing on an actual device.
It'd be sweet if it worked better in POSE.
Alan Pinstein
Synergy Solutions, Inc.
http://www.synsolutions.com
1-800-210-5293
