From investigations so far, the main thing that throws out nested groups is the way group anchor points need to be explicitly defined, rather than being defined by the average of the children’s coordinates. It used to be you would add object1, 2 and 3 into Group1 then define where the anchor point (reference point) was out of TopLeft, TopCenter etc… But you didn’t need to define where the group was in x and y - it knew where it was as defined by the contents placement.
Now in G2, in order to move (to a definite point rather than self.x=self.x + z), scale, rotate or add transitions, you have to state that anchorChildren = true. As soon as you do, the contents of a group are moved so that their anchor point go to the groups default position 0,0. Now it becomes necessary to define the actual placement of the group which is the average placement of its contents, something which G1.0 already did.
Another useful thing with reference points is you could change them on the fly without upsetting the position of the group’s contents. Now with G2.0 if you need to do that, you’d also have to formulate a way to move the contents back to where they were doing something like group.x = group.x + (newAnchor - oldAnchor) * group.width or something like that…
So with G2.0 there has been introduced many more lines of code to achieve the same thing as G1.0. How is that an improvement? It is one thing to have new api design introduce migration headaches, it is another to create more work when nested groups are required.
As an animator, nested groups has been one of the great productivity time savers and has enabled all kinds of uses. Now I’ve read here that Tom from Corona Staff doesn’t even think anchor points should be used with groups, especially not nested groups. Well it’s worked absolutely perfectly in G1.0, why kill something so useful?
http://forums.coronalabs.com/topic/42781-questions-about-anchorchildren/?hl=groups
If G.0 suddenly can’t handle nested groups, what is the point? Or more usefully, what is the alternative?