Add fortran subprogram gblevents.f from w3emc to UFS_UTILS to solve circular dependency problem in NCEPLIBS?

[edwardhartnett]

@GeorgeGayno-NOAA we have a fortran subprogram in NCEPLIBS-w3emc which should probably be in UFS_UTILS.

According to @kgerheiser:

It's that gblevents.f program which uses nemsio and sigio.

https://github.com/NOAA-EMC/NCEPLIBS-w3emc/blob/develop/src/gblevents.f

It seems to me that it should be moved out of w3emc. Into it's own repo? UFS_UTILS? I think it was put there just because that repo is the dumping ground for miscellaneous things. Then, w3emc would have no dependencies.

[kgerheiser]

It's not a program though. Everything else here is an executable.

[edwardhartnett]

I think the fact that it's a library is not a problem.

But perhaps it should be in NCEPLIBS-bufr? @jbathegit and @jack-woollen do you guys know about this program?

[kgerheiser]

I say we track this in the w3emc repo

NOAA-EMC/NCEPLIBS-w3emc#71

[edwardhartnett]

If added to UFS_UTILS it will introduce a dependency on NCEPLIBS-bufr, which is new. All other dependencies are already in use by other parts of UFS_UTILS. So that does look like a good fit...

[edwardhartnett]

What I wonder is - who calls this function?

[kgerheiser]

That, I don't know. I know @BoiVuong-NOAA worked on it.

[kgerheiser]

I also don't think it would be appropriate here because we don't install UFS_UTILS. Depending on what applications use gblevents.f it wouldn't be available pre-installed.

[edwardhartnett]

Well I really don't want to create a new repo for one subprogram!

[jbathegit]

I think the fact that it's a library is not a problem.

But perhaps it should be in NCEPLIBS-bufr? @jbathegit and @jack-woollen do you guys know about this program?

For my part, I don't have any knowledge of or familiarity with this program. But if the main docblock is correct, it looks like it also has dependencies on NCEPLIBS-sigio, NCEPLIBS-sp, and NCEPLIBS-nemsio.

[edwardhartnett]

Although there are certainly some valid objections to putting glbevents into UFS_UTILS (including the fact that UFS_UTILS is not built by hpc-stack), it seems to be the best available option.

@Hang-Lei-NOAA will do this in coordination with @GeorgeGayno-NOAA .