Disable phase 2+3 in fitFluxCal

The current flux calibration task performs the focal plane fitting to correct for fiber offsets with respect to the object centers. A fiber offset introduces a chromatic effect because the seeing is wavelength dependent. The pipeline applies a low-order polynomial as a function of wavelength to correct for the chromatic term. The fit is done against the PS1 photometry, but PS1 extends out to ~950nm and we are extrapolating beyond that. It turns out that the extrapolation is very bad and makes the n-arm spectra useless. I propose to disable phase 2 and 3 in fitFluxCal, which does this flux correction for now. Phase 1 corrects for the normalization and we should keep it.