Jean-Marc Lasgouttes wrote: > I think we should get rid of the test > > if (prevcur.pos() != 0 > && prevpar.isSeparator(prevcur.pos() - 1)) > --prevcur.pos(); > > Does anyone know why it could be useful?
Looks like it tries to prevent that the cursor ends up in a position behind a linebreak inset. However, this case should never happen with the boundary approach, should it? So it can probably be removed. Jürgen
