Added a warning to the discretemeasure function

[davibarreira]

I noted that there is an "inconsistency" with our discretemeasure function. If I pass musupp = [[0],[1],[2]], the function is dispatched to the multivariate case, instead of passing to the DiscreteNonParametric. I can see why that is, since the type required would be a vector, and not a vector of vectors.

I've been trying to fix this issue, but all solutions I came up with require that we check the length of the vector inside the vectors of vectors. And I'm guessing this would cause instabilities to the compiler (this is a supposition based on past code I wrote that @devmotion required changes).

So instead, I decided to add a warning inside the discretemeasure function, so that the user is informed that he should not pass a vector of 1D vectors, and should instead flatten into a single vector in case his support is 1D.

[coveralls]

Pull Request Test Coverage Report for Build 987000086

• 1 of 2 (50.0%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-0.2%) to 96.092%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/utils.jl	1	2	50.0%
<!--	Total:	1	2	50.0%	-->

Totals
Change from base Build 976321222:	-0.2%
Covered Lines:	541
Relevant Lines:	563

---

💛 - Coveralls

[davibarreira]

This PR is very simple, I guess it can be merged before #92 .