clarify dead code situation in constraints.py

[lukasheinrich]

Question

see TODO in there

[kratsg]

I don't see anything?

[matthewfeickert]

I don't see anything?

It is in PR #519: https://github.com/diana-hep/pyhf/pull/519/files#diff-300adf40cce4eaca8a6a25858fac96d9R155

[kratsg]

Ahh, ... huh. It was introduced in #298 which I didn't catch (sorry!) did catch in this comment. This is also affecting the gaussian combined constraints as well as there's a similar try/except in that block.

I guess it's a question of whether someone can provide a gaussian/poisson constraint without the sigmas/factors assigned... Likely, it doesn't make sense if they don't, so we should just make sure the code crashes if that happens somehow...