Perhaps we should fix the bug that requires synphot templates to require 4 or more wavelength points instead of these kludgey workarounds.
The tests I added here do not give any flux when only 2 or 3 points are used. I think it might be a problem with determining overlaps. Not sure.
Originally posted by @hugobuddel in #74 (comment)