Rework of Coliny's option processing. I'm not sure how it worked
before, but the old testing scheme probably didn't rigorously evaluate
the option processing itself.
Misc fix in the Sinusoidal1 test function.
Updates of Coliny test scripts. The use of the 'python coliny_test'
syntax is the biggest change in most of these.
Rework of coliny_test script to support independent parameterization of
hybrid solvers. These parameters are inserted into an AMPL *.mod file
now, which avoids the separate parameter file used in 'runexp'.
( Links to Acro-related mail archives at http://software.sandia.gov/Acro/ )