Okay, here's the second version. Like I said there was a slight
misunderstanding in such that #addedOrRemovedSubmorph: is an invalidation
message and not a notification. In addition:
* HandMorph>>dropMorph:event: keeps the owner set itself (which is the same
behavior as before) but a comment has been added to explain why exactly this
happens
* Morph>>slideBackToFormerSituation: has been fixed (referring to the
event's world which is the Right Way to do it)
* Morph>>addAllMorphs: and Morph>>addAllMorphs:after: have been folded into
#privateAddAllMorphs:atIndex: to keep the (needed) complexity in a single
place.

As far as I can tell, this should do it. It certainly solves the reported
problems and still passes the tests ;-)

Cheers,
  - Andreas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: MorphRemoval-ar.5.cs
Type: application/octet-stream
Size: 14434 bytes
Desc: not available
Url : http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20030812/24107aff/MorphRemoval-ar.5.obj