On 02/05/2018 15:31, Markus Armbruster wrote: >> We do have infrastructure for class properties, we just don't use it much. > ... the sane thing to do would be to limit this command to class > properties, and use class properties for anything that doesn't *have* to > be dynamic. The second half of the sane thing looks like a substantial > chunk of work, though.
It's not that big, but somebody has to focus on it. Paolo