> This is hard to autotest since what was happening is that all the delegates where being created and then immediately > destroyed, so at the end the number of delegates we'd have in a tryCompare would be the same.
Should be quite easy to detect I think...
Get the pointer to the first item that would be destroyed without this patch, do the overshoot and verify that the pointer didn't change. In case its destroyed and recreated, you'd have a different pointer, no?
> This is hard to autotest since what was happening is that all the delegates where being created and then immediately > destroyed, so at the end the number of delegates we'd have in a tryCompare would be the same.
Should be quite easy to detect I think...
Get the pointer to the first item that would be destroyed without this patch, do the overshoot and verify that the pointer didn't change. In case its destroyed and recreated, you'd have a different pointer, no?