G.A. vd. Hoorn

And because I feel it's important: we're still grateful @Danfoa took the time to figure out what was missing and submit this PR. The code this touches is very old,...

Conceptually, the changes needed would be (similar to) the ones included in #179. I've personally used those as part of the APC setup mentioned in the comments.

@jmarsik: https://github.com/ros-industrial/motoman/compare/kinetic-devel...gavanderhoorn:rebased_two_arms_on_a_rail is I believe the minimum set of changes needed to resolve the `LOST` goal problem for setups like yours. Those changes are from #179, but actually come from...

I believe you mean the `KR 6 R700-2` variant, correct? In any case: we don't have support for that specific variant yet, so you'd have to create the required files....

The meshes look really nice @Briancbn. Could you comment on the file size difference? Good looking meshes are important, but if it increases the size of the files too much...

While this certainly improves the smoothness of the mesh, it increases their size 2x (wrist 3 about 9 times). Personally I feel that's a bit too much, but I'm just...

The other two may be more efficient to reduce. They are the main cause of the size increase I believe.

I've not investigated this further, but I guess there is some unitialised memory being used to set or copy values for the missing elements of the arrays.

@mxgrey: thanks for the extensive descriptions and analyses. As I wasn't involved in any of this I'm going to need some help. First: @hartmanndennis: would you be ok with this...

Additional comment: there appear to be quite a few places where magic nrs are introduced (`5`, `6`). Those should probably be replaced with constants or values should be dynamically determined...