Someone experienced with Fortran and floating point? Need help to fix Gfortran IEEE module

[Sergey Fedorov]

If anyone knows modern Fortran well enough and got a bit of time, I would
greatly appreciate some advice on two issues:

1. Currently gfortran has no support for ieee_arithmetic on PPC. I have
written an implementation, which kinda works – most of tests from Gfortran
suite pass fine. However there are a few unpleasant errors (for example,
related to rounding), which I cannot figure out how to fix – and fixing is
necessary, so that the fix can be merged to GCC upstream (well, I hope Iain
can support this, but I need tests to pass).
I have the code and detailed logs from test runs. It may not be a 5 min
thing to deal with, but it will be a valuable contribution if we can
accomplish this. Many modern Fortran ports rely on ieee_arithmetic (and
yes, those work and pass tests locally for me – so it is meaningful to work
on though completion).

2. Something is broken in HDF5-fortran:
https://github.com/HDFGroup/hdf5/issues/2799
Perhaps should be easy to fix, but not obvious what.

Please ping me here or on GH, if anyone interested to assist with sorting
these out. That would be greatly appreciated.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.macports.org/pipermail/macports-dev/attachments/20230526/75a952ed/attachment.htm>