Add consistency check: there should not be two variables with same value reference but different types about fmi-cross-check

[ghorwin]

A side note on the FMI standard: IMHO, it would have been easier to enforce unique value references than to introduce the concept of a "variable index" later in order to uniquely identify variables in the dependency section...

And, if you think of the variable index as something like a hash value to the unique variable name (just to make coding easier and avoid string handling for the c-guys among the FMU coders) a unique value reference would have done the job as well. Oh well... :-)

[t-sommer]

FMI 2.0 does not require value references to be unique across types (or even for the same type).