iraf.net

There is a bug in dispcor, which appears if using more than one reference spectrum, thus having more than one dispersion function. There is no space between the last coefficient of the first function and the weight value for the next function. This is no problem if the coeff looks like 0.001 and the weight is 0.1, then they are interpreted as 0.0010 an .1
But if the last coeff is 1.0E-4, and weight is 0.1, then we end up having coeff 1.0E-40 and weight .1, which is not good.example excerpt from the header:
spec1 = "1 65 2 8701.4507623941 0.054886902605075 2043 0. 504.27 509.64 0.49098182 0. 1 5 1. 2043. 8758.27795653822 56.0486188941174 -0.789337307722521 -0.00979961852427856 8.79201599084727E-40.50901818 0. 1 5 1. 2043. 8758.27996142626 56.0507627044079 -0.788054200855183 -0.0105519575156028 -5.74790837023030E-6"I have reported this before, and it was reported being solved in a beta version. The above example comes from 2.14.1, Linux version.

Hi Michi !
I have reported in March 2008 this problem still not being solved on 2.14 (with detailed analysis of its consequences on precise RV work and citing your original bug report),
https://iraf.net/phpBB2/viewtopic.php?mode=reply&t=86994
but it was not reproduced by Frank Valdes and I wanted to prepare the more detailed report but then I have probably forgotten about it due to some more urgent problems - sorry for that.Anyway the bug is tricky and it seems to influence the solution only if the NON-LINEAR dispersion is used. So I would be glad if this is solved in 2.15 as it may be the source of tiny, however nasty errors in precise RV measurementsPetr Skoda