Description:
After returning to the insert grid after identifying a record insert problem during reverse engineering, Workbench sometimes crashes or displays a dialog indicating it has crashed. However in many cases, the program has not crashed, does not shutdown as indicated and one can continue to work on inserts until it does crash many operations later. 

How to repeat:
I have been working on a l10n module with tables containing foreign key constraints and have been experiencing these crashes hundreds of times per day due to the large number of inserts. The crashes occur when attempting to select the column containing the erroneous data after a failed reverse engineering run. As soon as one clicks the field in the grid, the crash occurs. 

I know the tables themselves as well as the foreign key relationships are correct because the end product does view, create, update and delete correctly.

Suggested fix:
Because of some unrelated problems resulting in odd grid behavior that I will report shortly, I think the grid widget is the source of the problem.

Thank you for the bug report. Your version 6.0.8 is quite older please update to 6.2.5 or 6.3.2 rc. Thanks.