Ondřej Čertík

See a recent discussion about this feature: https://fortran-lang.discourse.group/t/optional-dependencies-with-fpm/8538.

Good catch and analysis. Yes, we need to handle `( i, sigt(j,i), i = 1, ng )` correctly. I think it creates alternating `i, sigt(j,i)` as `i` increases. I am...

If the types are the same, such as in the following example, then it seems to be working: ```fortran MODULE setup_module IMPLICIT NONE INTEGER :: nmat=2 INTEGER :: ng=2 integer,...

To fix this, we could possibly even use a tuple to store the `i, sigt(j,i), sigt2(j,i)` part of the implied do loop `( i, sigt(j,i), sigt2(j,i), i = 1, ng...

Yes, we don't handle `iomsg` yet.

Yes, we need an ASR parser: https://github.com/lfortran/lfortran/issues/2357. It's not difficult, but it's also not completely a small task. The other issue is that we should change the syntax to be...

Thanks! I need to test this one by hand to ensure everything still works. It's really easy to break some things in cmake changes that are not revealed by the...

@serge-sans-paille thank you and my apologies for not finding time to thoroughly test it out yet. It's been on my mind almost every day though. I'll try to find time...

@serge-sans-paille I finally got to test this. There are several features that I extracted into separate PRs, so we can discuss those features there: * https://github.com/lfortran/lfortran/pull/5747 * https://github.com/lfortran/lfortran/pull/5748 The last...

Closing in favor of https://github.com/lfortran/lfortran/pull/5747 and https://github.com/lfortran/lfortran/pull/5748.