Thanks again Chris. I'll keep exploring with the n_cst_error and the like. From what I can see so far, no code changes have been made to these events. The pfc_system_error, being ancient code, varies only slightly in the error message configuration to more recent pfc code.

Did you update the PFC layer version with the one that matches your PB version or is this an old PFC version that has just migrated forward again and again?

Hi Olan,



This is contrived code to force an array bound being exceeded - same freezing problem with or without the TRY / CATCH.



The freezing also occurs with an Set or Get Item for a datawindow/store row if the row or column-name is in error.



The code under is the pfc_SystemError even in the application manager ....





string ls_a[]



TRY

If ls_a[10] = 'test' Then

li_rc = 1

End If



CATCH (Exception exception)

RETURN -1

END TRY





//////////////////////////////////////////////////////////////////////////////

//

// Copyright © 1996-1997 Sybase, Inc. and its subsidiaries. All rights reserved.

// Any distribution of the PowerBuilder Foundation Classes (PFC)

// source code by other than Sybase, Inc. and its subsidiaries is prohibited.

//

//////////////////////////////////////////////////////////////////////////////

string ls_message

string ls_msgparm[1]



ls_message = 'Error Number '+string(error.number) &

+'.~r~nError text = '+Error.text &

+'.~r~nWindow/Menu/Object = '+error.windowmenu &

+'.~r~nError Object/Control = '+Error.object &

+'.~r~nScript = '+Error.objectevent &

+'.~r~nLine in Script = '+string(Error.line) + '.'



if isvalid(inv_error) then

ls_msgparm[1] = ls_message

inv_error.of_message('pfc_systemerror', ls_msgparm)

else

messagebox('System Error',ls_message, StopSign!, Ok!)

end if



this.event pfc_exit()

Thanks again for your interest, Chris.

I put a TRY CATCH around a contrived array bound error, and a breakpoint in the (unedited) pfc_SystemError event of the app manager.

The step-through still froze within the TRY/CATCH on the array bound error. The pfc_systemerror event did not fire.

My only other observation was that the local variable tab pane under the code went blank when the error row 'freeze' occurred. The same occurs on my test machine.

I should also add the pfc_library remains the same back to our original use of PB - copyright date 1996-1997 in the pfc_systemerror event.

Mulling this over, should there be a 'system error' even - an original event ex-pfc_system_error? The only system error event I can find in my configuration is the pfc_system_error.