inline post restart issue with gnu compiler

[junwang-noaa]

Description

This PR is to fix the issue with inline post in restart runs with gnu compiler.
Is a change of answers expected from this PR? The fix does not changes results.

Issue(s) addressed

Link the issues to be closed with this PR, whether in this repository, or in another repository.
(Remember, issues should always be created before starting work on a PR branch!)

• fixes noaa-emc/fv3atm/issues/<320>

Testing

How were these changes tested?
Run the control restart test with gnu compiler

What compilers / HPCs was it tested with? GNU
Are the changes covered by regression tests?
NO, the current control restart does not have inline post. The global tests update (ufs-weather-model PR#561) has inline post turned on in the control and restart.
Have the ufs-weather-model regression test been run? On what platform?
Tested on hera. Will test on other tier-1 platforms

• Will the code updates change regression test baseline?
  This change does not change results, but the ufs-weather-model PR#561 does require new baseline.

• Please commit the regression test log files in your ufs-weather-model branch

Dependencies

If testing this branch requires non-default branches in other repositories, list them.
Those branches should have matching names (ideally)

Do PRs in upstream repositories need to be merged first?
If so add the "waiting for other repos" label and list the upstream PRs

• ufs-weather-model/pull/<561>
• noaa-emc/fv3atm/pull/<321>

[climbfuji]

The Fortran compiler specifications do not state that the tests in an if clause will be executed ini the order written. You need to do the following to be fully compliant:

if (associated(paramset)) then
  if (size(paramset)>0) then

[junwang-noaa]

Actually the size(paramset) is 1 but associated(paramset) is false with gnu compiler.

[climbfuji]

Same here

[climbfuji]

How? If associated(paramset) (means it's a pointer) is false, how can the size be 1?

[climbfuji]

That is what I got when I tried with gnu compiler. The issue happened when it tried to get the size of paramset(1)%param.
…

I think it would be best to make the same changes as in lines 161-162, 165-166.

[junwang-noaa]

That is what I got when I tried with gnu compiler. The issue happened when it tried to get the size of paramset(1)%param.

…

On Tue, Jun 1, 2021 at 1:12 PM Dom Heinzeller ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In io/post_regional.F90 <#321 (comment)>: > @@ -160,11 +160,13 @@ subroutine post_run_regional(wrt_int_state,mypei,mpicomp,lead_write, & if(mype==0) print *,'af read_xml at fh00,name=',trim(filenameflat) else if(ifhr > 0) then filenameflat = 'postxconfig-NT.txt' - if(size(paramset)>0) then + if(associated(paramset) .and. size(paramset)>0) then How? If associated(paramset) (means it's a pointer) is false, how can the size be 1? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#321 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AI7D6TPGKGXOVDWYMES7AETTQUIHJANCNFSM45XB5XNQ> .

[climbfuji]

Thanks!