Alex,
> > What exactly do you mean by updating connections? When does that happen?
> >
> > Why would data_update_extents() and layer_update_extents() have to scan
> > all the other layers when the objects in the other layers can't be
> > moved or selected (only connected to or disconnected from)? I can see
> > how it would become very complicated if objects in different layers
> > could be copied & pasted, grouped & ungrouped, or edited together
> > in some other ways. That's why I'm only suggesting to allow connections
> > be made or disconnected between layers. Once that's down, then maybe
> > we could look at how complicated & slow it would make things to try to
> > manipulate objects in different layers together in other ways, but I
> > really don't want to get into that right now.
>
> If objects in the active layer are connected to objects in another layer
> (which in turn may be connected to objects in yet other layers), and the
> object is moved, resized or otherwise changed this might trigger a change
> of the position, size and orientation of the connected objects (and
> therefore *any* object in *any* layer).
You're right that moving or resizing an object in one layer can cause an
object its connected to move or change in the other layer (I didn't realize
this before). But that's not going to make updating connections slower,
because it would be no different than if they were in the same layer:
(in your message on Mon, 14 Jun 1999 20:40:06 +0200 (CEST)):
> ... It breaks the separation between
> layers, which makes updating connections slower...
Sorry about being a pest, but I still don't understand how updating
connections would be made slower if inter-layer connections could be
made.
> And then when you need to recalculate the extents after having moved one
> object you can't do as usual, recalculate the extents for the active layer
> by looping over all objects in it and doing rectangle_union and then merge
> it with the extents of the other layers. Because the extents of the other
> layers may have changed too.
Now that I understand that objects in other layers can get moved, I see
that recalculating the extents is necessary.
> Yes, wait till it's done. Now it's full of bugs and core-dumps.
About how long do you think this will take before you want dia alpha or
beta tested for the new undo functionality?
-Ben
--
Ben Hochstedler http://www.mei.com GE Marquette Medical Systems
[EMAIL PROTECTED] Phone: 414-362-3317 Fax: 414-362-3389