Answer by acr for Integrals of compactly supported functions of positive type

I asked a (near-) duplicate of this question on mathoverflow, where it was answered; indeed, it turns out that the triangle is optimal. The proof of this is a simple and very pretty application of the Poisson summation formula.