[ODE] dGeomSetOffsetPosition

[David Walters]

> I would suggest that the dBodySetMass structure should warn (or even
> assert) when passed a mass with a non-zero 'c' field, since that value
> is ignored by ODE.

I agree that an assert would be a very valuable addition - rather than
having a system where it builds and runs but behaves unexpectedly (in
varying degrees depending on the configuration of masses and geoms)  -
you would get a simple "STOP! What you are doing is not right"
announcement.

As for the 'dBodySetOffsetMass' suggestion, I think that it would be
useful, but also potentially dangerous, as I'm not sure I like the
idea of it moving the geoms around for me. without some stronger
indication in the function name or possibly just the comments.

I know it's splitting hairs, and I don't want to seem finnicky, but if
it was called "dBodySetMassCentreGeoms" I think this would be clearer
and be a great addition.

Both I and my colleague have common code for creating composite geoms
which is just like this pseudocode you've written, I'm sure something
to
simplify the process would be the right thing to add.

If no-one objects I will put in the assert and new helper function.

Regards,
Dave