[Bf-committers] Collada -Node- Joints issue?

[Mike S]

Hi all,

Just a quick observation on the current import/export of armatures from 
blender.

At the moment importing a set of NODES of type JOINT treats the last N-J 
as a head of a blender bone rather than the tail of the last. A similar 
issue wioth the export of such files.

If you try the COLLADA test cases for 1.4.1 at the site below 
(skintest_1_ID.dae for example) you will see that blender imports the 
terminating N-J (or N -- should be the same but not nessesarily skinned) 
as the head of a final bone rather than the tail of the last. This 
results in an extra bone compared to the original (see jpg in the folder 
for the test scene in Maya)  This last bone is given the defualt Z-UP 
orientation and a length generated by some means. (default in 2.49 for 
this situation is in the same direction as the last bone -  but still 
generates the extra bone over that desired)

http://collada.org/owl/browse.php?sess=0&parent=41&expand=1&order=name&curview=0&sortname=ASC

It is obvious from the test file and the attached jpg that this is not 
what was intended.


When imported into current 2.56 r34525 this results:
http://www.pasteall.org/pic/8375

when this file is imported into 2.49b this results:
http://www.pasteall.org/pic/8376



Using as DAE generated (2.56) from this blend file:
http://www.pasteall.org/blend/4968

reimported results in this for 2.56 ::
http://www.pasteall.org/pic/8377

and this from 2.49b::
http://www.pasteall.org/pic/8378


Thus I believe the correct interpretation of a set of nodes, 
nodes-joints is that the last (leaf?) is a terminating end of a skeleton 
not the start of another bone.  I may have interpreted this incorrectly 
but to me that seems the intent of the test files and demos on the 
Collada website seem to be.

Perhaps Remi Arnaud may like to comment on the intent of these files?

Cheers

Mike.