Make TestSDKDistribution crash if p values are close to 1?

[alexandrebouchard]

E.g. of concerning case:

Running ExactInvarianceTest on model MarkovChain [IntSliceSampler] -- currentRandStatus(1316499326)
Corrected pValue = 0.005
All tests passed:
MarkovChain	IntSliceSampler	ListHash	0.9999999818884424	6.02010000000001(0.01891045635451333)	6.0242999999999896(0.018759161538143488)

[alexandrebouchard]

Could be simply caused by the strategy used in ListHash, etc. Revisit those choices.