Jean-Marc Lasgouttes wrote: > A single undo().recordUndo(it) at the right place in renameBranches > should suffice.
OK, thanks. > A more annoying problem is that your existing code only iterates over > top-level insets. And this is at the price of code that is more > complicated than a forwardInset-based one. > > Have a look at mutateNotes in InsetNotes.cpp. I see. I have to think a bit. > I should revive my inset-foreach patch. The question I am not sure about > is whether it should iterate into child documents. You chose to do it > here, but I can see arguments for or against such behaviour (having the > change done only in the local buffer is the least surprise, like > search/replace). It is related to our other master/child discussions. > And if one runs the function from a child, should it update the master > and all the siblings? I think it should iterate into childs from within master, but not vice versa. > Would it be enough to have a 'treat compound document as a single one' > checkbox and try to honor it in all our code? As already mentioned, I think we should have some "standalone" or "child mode" checkbox within childs. Jürgen
