Add test case for verifying correct absolute scaling with both sasview and internal/sim models.

[toqduj]

Absolute scaling seems to work fine now for both, but to prevent the previous issues from recurring, with the weird scaling factor necessary for the sasmodels, tests would be good.
Quicktestdemo can be used for this, spheres, SLD of 33.4 (e-6 1/A^2) should give around the following volume fractions for the populations:
total range: 0.9
2.5-25 nm radius: 0.114
25-75 nm radius: 0.382
75-250 nm radius: 0.404