Bug in spectralModel / calcSpectraErr

[Michael Keith]

There’s something wrong in the last fit in spectralModel. Please see attached .par and .tim file.

If you run:

tempo2 -gr spectralModel -f J0742-2822.par J0742-2822.tim

and then: 0 1 0.1 6 5 0 1

You’ll see that the green and yellow spectra are completely different. If I do this with an earlier version of tempo2 then they basically align with each other and so there’s a mistake somewhere in the most recent version.

[Michael Keith]

There was a one-line error in TKspectrum that was fitting for a random third parameter and therefore blowing out then fitting. Fixed as of 1fc1a26