[Coin-discuss] inconsistent behavior.

[John J Forrest]

To get an unbounded problem you also need to set lower bounds on rows to
-inf.

Clp Dual (Primal was okay) had a problem if all variables were free -
hopefully fixed.  It also was not checking correctly for dual infeasible
state in dual.  I am fairly sure I have not made things worse and I have
run netlib tests and several nasty problems with free variables so I have
checked in changes to Clp.

How's that for quick reaction to bug reports.  Honestly, I do have other
work to do, but this is more interesting.

John Forrest