Fix Fermi catalog flux point upper limits #858
@adonath - Thank you!
Travis-ci passed, I'm merging this now.
In the test removed here you had
In any case, the int -> diff flux conversion should be pointed out to users, as it's something to be aware of when plotting and analysing flux points.
The docs example with an upper limit is nice.
Can we change the source at to one that has upper limits:
@adonath - Thanks!
I've been thinking about whether
I think I'd be +1 to "yes" on both questions, but I don't see strong pros and cons yet. So do as you think best, just wanted to mention it, or get an opinion from @jjlk or @joleroi on what we should do, before you spend the time to implement it and modify docs / tests.