On 15/06/12 12:48, jerro wrote:
Joseph's pull request already contains a fix for this bug but I'm guessing it won't be merged until other issues with randomSample are resolved.
Assuming that everyone agrees that a random sample should always lazily evaluate to the same values, a different fix will be needed; probably that if no RNG is specified, RandomSample should contain a generator set to Random(unpredictableSeed).
On the other hand, if RNGs are to be passed to randomSample by reference, it might make more sense to always have the sample evaluate differently.
