Should distributions have == defined?

[nickrobinson251]

This took me by surprise, so i wondered if this was intentioned behaviour or if it was a missing feature?

julia> mv = MvNormal(rand(3), let x=rand(3,3); x*x'+I; end);

julia> mv2 = deepcopy(mv);

julia> mv == mv2
false

[johnczito]

It's intended in the sense that it's a general phenomenon not specific to distributions: [1], [2], [3].

[nickrobinson251]

right, makes sense, so i think we file it under "missing feature"?

I presume that we want mv == mv2 in the example above, and it's just that this is a case where Julia doesn't give us it for free.

[nickrobinson251]

xref #696