Georg Baum wrote: > > Please test. > > It fixes 2139 for me, and I could not find problems (I did not test this > one).
Thanks. The fix is obvious. The cursor tries (in both bug 2139 and Helge's case) to enter a cell of the scriptinset which does not yet (or no longer) exist, and fails. MathScriptInset::ensure checks if the cell exists, and creates it, if not. Can I apply it? Jürgen
