Do we want to return the fluxError vector?

[DougBurke]

In Sherpa we ignore the fluxError component, but we should probably support it for the small (?) number of models that use it.

It would complicate things - e.g. maybe we'd want to return some form of named, or structured, tuple?

[DougBurke]

HEASOFT 6.29 / XSPEC 12.12.0 only has two models with flag values that are not 0 and they are

smaug
pileup

both of whih have flags of [0, 1] - ie they do not use fluxErrors but can not be easily cached

So it's only for usermodels.

% egrep 'pileup|smaug' ~/local/heasoft-6.29/spectral/manager/model.dat
smaug         22   0.0E+00    1.0E+20        c_xsmaug    add 0  1
pileup         7  0.001      1.0e20          C_pileup    acn 0  1