Bug, FMD not doing multihypo correctly

[dehann]

FactorMetadata not following multihypo particular to each solve in the same way that the factor evaluation does. Factor evaluation is working correctly, this is part of a more advanced feature for special sampler and fmd in factor arguments. This will likely become a part of revamping FMD which is due to happen anyway.

Also see:

• Discussion: Pass variables to factor evaluation so that we can use variable data? #784

• Towards immutable FactorMetadata(...) #825

• standardize special sampler using new FMd / CalcResidual #927

• more testing of specialSampler #692

• approxConvBinary does not yet support specialSampler  #640

• Manifest VariableTypes (Singleton model) and new FMD #783 (comment)

• Remove Union types from FactorMetadata #383

• Initial test already exists by must be updated as this gets fixed, see new test for fmd bug on multihypo #1024

• Also see Deprecate AbstractRelativeRoots shuffleX, preserve partial #1072