I’m porting the library to Fortran but I’m not the one who ran those tests - I clicked on the link and it goes to this post in the forum, apparently the “error” is comparing Bessels.jl against another Julia library.
IMHO this type of tests should be run against an analytical solution, but of course we’re all lazy and end up comparing against just another library. In the post it looks like it’s a quad-precision library, so probably not bad?
You’re fine, so far the Fortran part is just a port, i.e., no attempts at changing the accuracy of the original library. If you could contribute your test case either here or on the github page, it would be a nice addition.
Right, the original Julia library has insanely long lines and some parts haven’t been fully refactored yet - as you see on the page you need to run with -ffree-line-length-none with gfortran, or equivalent commands from other compilers