cursors are still not behaving properly on the fit wrapper

[srkline]

The logic needs to be defined a bit better as to what the "proper" behavior should be... and then what procedures should be responsible. The checkbox can't do it all, if the data set popup is changed, or the graph(s) are closed. The moment of dispatch can be a last point of defense, but is not optimal.

[srkline]

I think I've got some of this worked out, at least the last defense check just before the fit. But it'll take more use before it's deemed safe.

[srkline]

Added a bit more so that the action for the cursor check box does the work, and is called as needed (r394). Folks will have to get used to the behavior of the cursors disappearing if they don't belong, or moving to the newly selected (or Loaded) data set if possible.