Notes.txt

Notes:

When compiling Fortran codes to a Python library, it is highly recommended to
avoid using Unicode or any other non-ASCII characters, except in rare cases
where it is absolutely necessary.

The presence of non-ASCII characters in Fortran files can lead to
compatibility issues, especially when using numpy-f2py and compiling with
Conda.

You can run in your files the following command:

grep --color='auto' -P "[^\x00-\x7F]" <filename>

One common error message that may occur is:

Exception: readfortrancode: reading src/fortran/EvalFilters.f90#2705 failed with
  'ascii' codec can't decode byte 0xce in position 1739: ordinal not in range(128).

To ensure a smooth compilation process and avoid such errors, it is best to
restrict the usage of non-ASCII characters in Fortran codes. Stick to ASCII
characters whenever possible.

If, for some specific reasons, the use of non-ASCII characters is unavoidable,
make sure to thoroughly test and validate the Fortran code before
compiling. Additionally, installing the charset_normalizer package might help
f2py determine the input file encoding correctly, but it is not a guaranteed
solution.

During the compilation process, if you encounter the following error:
error: subprocess-exited-with-error
× python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> See above for output.

Please double-check the Fortran files for any non-ASCII characters and
consider removing or replacing them with ASCII equivalents whenever possible.

This cautionary approach will help prevent potential compilation errors and
ensure a smoother compilation process for Fortran codes to Python libraries.