After belatedly moving to system 5 testing I'm getting a "free Handle" 
error in system code between events (needless to say, the problem 
doesn't occur in system 4 testing).

I'm using CodeWarrior 7 with the Palm OS 5 SDK (68K) R2 on Window 2000 
(5.00.2195)

Although I don't doubt that my code is to blame, several things are 
making this difficult to debug:

The usual behavior is for the simulator to hang.  It may be trying to 
give me a message (when I kill it with Task Manager I'm sometimes told 
it is waiting for a response from me).  It does log a free Handle 
message in the Error Log file.

If I step through the code the hang occurs after a FrmGotoForm call.  
Breakpoints set in PilotMain or anywhere reasonable in my code don't get 
hit.

I did manage to come up with a test case that actually gives an error 
message rather than hanging but if I switch to the debugger, the stack 
traceback doesn't include any of my code.  The PilotMain at the top 
isn't mine.  The error message in the simulator (as in the error log 
file) is

memorymgr.c line 3583 Free Handle

and the traceback is

__Startup__
PilotMain
ProcessNotify
LaunchDBHandleSetInfo
LaunchDBHandleSetInfo
PrvUpdateLaunchDBEntry

What I'm doing just before the FrmGotoForm call is copying a database 
with some changes.  Attempts to pinpoint the cause by leaving out 
various things (and the above trace) suggested the problem had to do 
with the AppInfo block.  Leaving out the update of the AppInfo block did 
eliminate the failure in that test, but it reappeared later (in about 
the same place in my code -- updating a database but without updating 
the appinfo block) in its more usual form (i.e., just a hang).

Further investigation suggests the connection with the appinfo block may 
be spurious, but...

Any suggestions?  I suppose updating my copy of CodeWarrior might be a 
good one.

-- 
For information on using the Palm Developer Forums, or to unsubscribe, please see 
http://www.palmos.com/dev/support/forums/

Reply via email to