I have a multiview documented based application and (not surprisingly) I get
a set of KERN-EXEC 3 panics before BaseConstructL of AppUI class. After
reading most of the earlier posts, I figured that this sort of system panic
is occuring to most of the applications I followed the FAQ-1254
http://www3.symbian.com/faq.nsf/0/E7FB1E0EDCEA88E380256FD600216593?OpenDocument
and my application seems to be behaving as described there.. So, I hope
that if I send the proper description with the debug logs, this can be
waived off.
Besides this, I'm trying to ensure that the application itself has no
errors, so I continued running Lowmem, after increasing the "Min Heap Size"
enough to get rid of the initial KERN-EXEC,3 panics. After few seconds it
presents the [D2] Continue dialog, if I select "No" then lowmem closes
properly, but if I select "Yes" (which I think I should, becuase the
applcation screen is not yet started) then lowmem seems to hang. . . .
with the progress bar "Testing app xxxxxb ....." If I switch to the
application using the task list, I see the OOM dialog, which seems to pop up
irrespective of the number of times I dismiss it. and lowmem stays at the
same amount of Free Heap size as displayed in the last progress bar. . The
only way to exit out from lowmem is to press "cancel".
Is this is a problem becuase of application faulty code or anything to do
with lowmem ? Please share if anyone has encountered this before.
Btw, I tried uninstalling and reinstalling both lowmem and the application
and have also started the application once before testing in lowmem.
thanks,
Kishore.
Hi Kishore,
You�ll see from my earlier posts that i too was having problems with
KERN-EXEC 3s from lowmem.
One thing that i was told to try by a symbian signed expert was to check
that the destructor of the AppUi class wasn�t referencing things that might
not be allocated before the KErrNoMemory leave occurs. This solved a lot of
my problems. the main thing is that now i can tell the difference between
when the code is leaving and when it is panicking, this means that i have
been able to fix issues that occur in my code.
Sadly on certain devices i still get a KERN-EXEC 3 on the lowest memory
condition which is coming from the BaseConstructL so i still need a waiver
but at least i know that there are no gremlins in the startup code that i
have written.
I don�t have any comment on the end of test issue that you raise for lowmem
but maybe you can eliminate KERN-EXECs as a potential candidate for the
problem now.
Best regards
PL
"Kishore Raja" <[email protected]> escribi� en el mensaje
news:tk20ipgUGHA.1188@extapps30...
>I have a multiview documented based application and (not surprisingly) I
>get a set of KERN-EXEC 3 panics before BaseConstructL of AppUI class.
>After reading most of the earlier posts, I figured that this sort of system
>panic is occuring to most of the applications I followed the FAQ-1254
>http://www3.symbian.com/faq.nsf/0/E7FB1E0EDCEA88E380256FD600216593?OpenDocument
>and my application seems to be behaving as described there.. So, I hope
>that if I send the proper description with the debug logs, this can be
>waived off.
>
> Besides this, I'm trying to ensure that the application itself has no
> errors, so I continued running Lowmem, after increasing the "Min Heap
> Size" enough to get rid of the initial KERN-EXEC,3 panics. After few
> seconds it presents the [D2] Continue dialog, if I select "No" then lowmem
> closes properly, but if I select "Yes" (which I think I should, becuase
> the applcation screen is not yet started) then lowmem seems to hang. . . .
> with the progress bar "Testing app xxxxxb ....." If I switch to the
> application using the task list, I see the OOM dialog, which seems to pop
> up irrespective of the number of times I dismiss it. and lowmem stays at
> the same amount of Free Heap size as displayed in the last progress bar. .
> The only way to exit out from lowmem is to press "cancel".
>
> Is this is a problem becuase of application faulty code or anything to do
> with lowmem ? Please share if anyone has encountered this before.
> Btw, I tried uninstalling and reinstalling both lowmem and the application
> and have also started the application once before testing in lowmem.
>
> thanks,
> Kishore.
>
>
>
hey PL,
Thanks for your comments. Fortunately I have no problems in my app and it
works fine. The issue was with the specific phone device I have here. When
I tested my app (with lowmem) on a different phone it runs properly and
closes the app after it has opened. I spent quite few days breaking my head
with lowmem and it gets difficult when there is no emulator version. .
The initial panics definitely seem to be coming from system code. You are
right that, any referece made by a non-allocated and non-initialized pointer
in the AppUi destructor will cause a panic. The same rule applies to
destructors of all classes (not just AppUi). After adding proper sanity
checks in all destructors, I'm no more getting any panic.
Btw, after pressing "No" on the [D2] Continue test dialog, lowmem pops up
[D4] dialog as it failed to close the app and requests an explicit "kill".
Is this anything to do with the way app is behaving? My application does
have exit implementation on EEikCmdExit in both AppUi and view.
Kishore.
"petxolobo" <[email protected]> wrote in message
news:%23sLMYo9UGHA.2208@extapps30...
> Hi Kishore,
>
> You�ll see from my earlier posts that i too was having problems with
> KERN-EXEC 3s from lowmem.
>
> One thing that i was told to try by a symbian signed expert was to check
> that the destructor of the AppUi class wasn�t referencing things that
> might not be allocated before the KErrNoMemory leave occurs. This solved
> a lot of my problems. the main thing is that now i can tell the
> difference between when the code is leaving and when it is panicking, this
> means that i have been able to fix issues that occur in my code.
>
> Sadly on certain devices i still get a KERN-EXEC 3 on the lowest memory
> condition which is coming from the BaseConstructL so i still need a waiver
> but at least i know that there are no gremlins in the startup code that i
> have written.
>
> I don�t have any comment on the end of test issue that you raise for
> lowmem but maybe you can eliminate KERN-EXECs as a potential candidate for
> the problem now.
>
> Best regards
>
> PL
>
>
> "Kishore Raja" <[email protected]> escribi� en el mensaje
> news:tk20ipgUGHA.1188@extapps30...[color=green]
>>I have a multiview documented based application and (not surprisingly) I
>>get a set of KERN-EXEC 3 panics before BaseConstructL of AppUI class.
>>After reading most of the earlier posts, I figured that this sort of
>>system panic is occuring to most of the applications I followed the
>>FAQ-1254
>>http://www3.symbian.com/faq.nsf/0/E7FB1E0EDCEA88E380256FD600216593?OpenDocument
>>and my application seems to be behaving as described there.. So, I hope
>>that if I send the proper description with the debug logs, this can be
>>waived off.
>>
>> Besides this, I'm trying to ensure that the application itself has no
>> errors, so I continued running Lowmem, after increasing the "Min Heap
>> Size" enough to get rid of the initial KERN-EXEC,3 panics. After few
>> seconds it presents the [D2] Continue dialog, if I select "No" then
>> lowmem closes properly, but if I select "Yes" (which I think I should,
>> becuase the applcation screen is not yet started) then lowmem seems to
>> hang. . . . with the progress bar "Testing app xxxxxb ....." If I
>> switch to the application using the task list, I see the OOM dialog,
>> which seems to pop up irrespective of the number of times I dismiss it.
>> and lowmem stays at the same amount of Free Heap size as displayed in the
>> last progress bar. . The only way to exit out from lowmem is to press
>> "cancel".
>>
>> Is this is a problem becuase of application faulty code or anything to do
>> with lowmem ? Please share if anyone has encountered this before.
>> Btw, I tried uninstalling and reinstalling both lowmem and the
>> application and have also started the application once before testing in
>> lowmem.
>>
>> thanks,
>> Kishore.
>>
>>
>>
>
>
>[/color]