[ODE] Negative penetration depths

[Russ Smith]

> Would it be possible to have a new kind of constraint that replaces
> "positive separation velocity" with "separation velocity greater than X,"
> and allow X to be negative?  In other words, is this a limitation of ODE's
> contacts, or a limitation of ODE constraints?  Also, would this do what
> Tom wants?  I could also use such a thing, although for different reasons.

yes, this is quite easy, it is just a change in the 'lo' (and/or 'hi')
vectors filled in by the contact constraint. (although i believe the
current LCP solver might limit you to lo<=0 and hi>=0). in other words,
the above is an artificial limitation of the contact joint code. this
flexibility (in lo and hi) is available to all constraints, but of course
each constraint just exposes a small amount of flexibility through its
API. errmmmm ... not sure if this is tom's solution or not.

russ.

--
Russ Smith
http://www.q12.org