Jean-Marc Lasgouttes wrote: > OK, a new version (first the reversion, and then the actual patch). I do > not see why we could not always use recordUndoInset, except that it may > use more memory.
Excellent, this works fine! Thank you very much for the help. Georg
