> I just tested the new behavior: enter \frac{}{b}, backspace from > numerator, \frac{}{b} is gone all together!Sure, it replaces the whole inset with the contents of the cell. If the cell is empty, the inset will vanish entirely. It is exactly what I want. As it has been like that for ages I'd expect it is what quite a few people expect.
I do *not* like this behavior because it deletes something unexpectedly. Bug 686 calls to remove *empty* cell with backspace or del, but in this case, it removes the whole structure because one of the cells is empty. Why should operations in the superscript of x^a_b removes x and its subscript? Following this logic, you should remove the whole matrix if you do backspace in a matrix cell. This does not make sense to me at all. And there is this 'del' problem. Should it behave the same as 'bs' for empty cell? Bo
