@isuruf just call the stacktrace function, just like you do for abort.
and for exception it has to be called at the point of the exception I think, which is tricky. But I thought it was working before using catch. Was it because if exception gets uncaught, the stacktrace is set for where the exception was raised?
So perhaps we can just remove the custom abort/exception handler in catch?
But we want Catch to do the accounting which tests fail/succeed.