Updates:
Status: Fixed
Comment #14 on issue 842 by alexandr...(a)gmail.com: ROTranslatingShape mouse
hotspot mis-alignment
http://code.google.com/p/moose-technology/issues/detail?id=842
Hi!
Thanks for your post. It was inspiring. I spent a lot of time on think
about the implication about having a translating shape and an offset.
Neither approach has convinced me so far.
The reason is the following: shapes are closely dependent from each other,
meaning that the element is the only entity subject to layout, looking up
subnodes, and responsible for its bounds. This model is quite simple, easy
to understand, easy to debug. Adding a concept of offset breaks this nice
small model.
As soon as you want shapes to be next to each other, you need a kind of
layout, at least to say that shape1 is on the left-hand side of shape2.
However, layouts work only on elements.
I came to the conclusion to not have translation or offset for shape. The
source code has been reverted to the equivalent of what it was a couple of
weeks ago.
This is not I am against this idea, however I could not find a clean and
appealing way to implement it without redesigning the shapes hierarchy.
Subclassing ROElement should easily solve your problem although.
I consider this issue as fixed so far.