v1.1 [LEGACY] -- Fixed treatment of photo-z and spec-z

This releases fixes a rare bug in the treatment of galaxies which have a photo-z or a spec-z that falls outside of the chosen confidence interval from EAzY. The bug was that such galaxies were never attributed any best fit, because no model was able to satisfy both the confidence interval constraint and the spec-z or photo-z. This is now corrected.

Galaxies with spec-z (or photo-z with FORCE_ZPHOT=1) now have their confidence interval narrowed down to this chosen redshift, so both best-fit parameters and their confidence intervals will correspond to this fixed redshift.

Galaxies with photo-z outside of the confidence interval (with BEST_AT_ZPHOT=1) are now "properly" given a best fit and uncertainties. In this later case, the best-fit solution will be evaluated at the given photo-z, and the confidence intervals for best-fit parameters will instead follow the redshift confidence interval, as for the other galaxies. This will lead to a mismatch between the two, and as such will be difficult to interpret. But this is a problem created by the user in the first place: having a photo-z outside of the confidence interval makes no sense. The only reason FAST++ tolerates this is for compatibility with FAST-IDL, and a warning will be issued so the user can identify such cases and fix them.

NB: This is a "legacy" release. These versions of FAST++ are meant to reproduce the behavior of the old FAST and not introduce any other feature. If you want more fancy options and updated stuff, look for a non-legacy release.