@certik : Please remove the file dmchar.f from the src directory of your Minpack distribution on Github. It was originally used for determining machine characteristics (such as epsilon, huge, tiny, etc.), and DMCHAR is not called from any of the other modified files in the src or examples directories.
Furthermore, if someone becomes tempted to try using DMCHAR, they will find that most of the loops in the subroutine are endless loops when compiled with an optimizing compiler such as Ifort/Ifx.
Good point.
@jacobwilliams let’s join forces and just maintain one version of modernized minpack under fortran-lang as a community. As discussed a few times in the past, as long as we keep the original code reasonably similar (minimal changes, just modernized), then there won’t be any objections and everybody can use it. Why don’t we do a short video meeting to brainstorm this to ensure you are happy with it, and I am happy with it, and then maintain one version under fortran-lang?
@certik We can move the discussion here about moving minpack under fortran-lang. I’m certainly open to it!
Consolidation of the two Minpack Githubs is a timely and interesting topic. May I suggest that we start a new thread for it? The title of this thread (Gfortran bug for ARM) has nothing to do with this new topic!
Sorry, I forgot to post the link in my reply: Move to fortran-lang? · Issue #8 · jacobwilliams/minpack · GitHub We can discuss there?
I split this discussion into its own topic, so we can discuss here.
We have two minpack implementations, so let’s discuss how we can merge them and move it under fortran-lang.