[ODE] Constraint Force Mixing

[Kenny Erleben]

I have tried to work through the formulas in section 3.8.1 of the user
guide, and I can not get the same result? Could somebody please point out
to me if I doing something wrong?

Using a forward euler step on the newton euler equation I get

  M (v(t+h) - v(t)) / h  = J^T lambda

from which I isolate v(t+h)

  v(t+h) = v(t) + h M^-1 J^T lambda

substitution into the constraint equation (which I assume to hold at time
t+h), I get:

  J v(t+h) = c + CFM lambda

which yields

  J v(t) + h J M^-1 J^T lambda = c + CFM lambda

Solving for lambda I get:

  (J M^-1 J^T - CFM/h) lambda = c/h - J v(t)/h

Ehh, this is not what it says in the documentation, here the result is

  (J M^-1 J^T + CFM/h) lambda = c/h

What I have looks similar, but I am particular concerned about the minus
sign on the CFM term, is there not a typo in the constraint equation in
the documentation?

regards

Kenny