[Cbc] CbcMain1 returns bogus result when it cannot improve on setBestSolution

[Jack Vreeken]

> One problem with indexes is that after pre-processing some variables are 
> removed and the indexes became invalid.

Generating random names, I immediately get "Warning: MIPstart solution is not
valid, column names do not match, ignoring it." Looking at the source, it must
be because I have some variables for which the bounds are equal. If that makes
setMIPStart() no longer a viable option, then I will probably have to resort
to just using setBestSolution (and caching its results myself). I do wonder
why setBestSolution() does not suffer from the same pre-processing issues
that affect setMIPStart you mention.