Skip to content

Conversation

@tclune
Copy link
Collaborator

@tclune tclune commented Jan 5, 2026

Types of change(s)

  • Bug fix (non-breaking change which fixes an issue)
  • New feature (non-breaking change which adds functionality)
  • Breaking change (fix or feature that would cause existing functionality to change)
  • Trivial change (affects only documentation or cleanup)
  • Refactor (no functional changes, no api changes)

Checklist

  • Tested this change with a run of GEOSgcm
  • Ran the Unit Tests (make tests)

Description

Long time coming.

Related Issue

tclune added 30 commits January 5, 2026 13:33
New change results in most unused fields being GRIDSET rather than
EMPTY.   It is possible that this is not strictly necessary, but seems
harmless.
- Functions that return allocatable objects must ensure that it is
  always allocated - even when errors occur.

- Improved error message for rank > 4.
These mostly are due to the fact that unused exports now generally
have field status of GRIDSET where before they were EMPTY.   This
is probably fixable if it is important, but a shallow copy of an ESMF
geom seems minor.
Lots of small changes to ensure compatibility with retrieval to/from
state item aspects.
(Functionality is elsewhere now.)
Necessary due to use of VerticalGridManager now.
Might be the fix for #3343.   Empirically seems true.
Was reluctant to delete hard won understanding of useful debugging
info.
Useful bit for debugging. Also should migrate to pflogger.
We need the spec to be properly updated from the payload before
analyzing for extensions.

This line could possibly move elsewhere though.
Still lots of code that should now be deletable.  Next pass.
- updated payload at appropriate point
- new diagnostic procedure (print_spec)
- pass through treatment of deferred aspects (must be in info now)
Some refactoring should reduce redundancy.
- need to use POINTER in some contexts
- more care with detecting mirror case
- Basic use case now functions correctly.
- Reordered tests to isolate the ones that sometimes trigger NAG
  "MAPL_Verify()" bug. (Hopefully now resolved.)
All aspects need superior testing now.
Works differently now with info and grid manager.
New FieldSet and FieldBundleSet replace that functionality.

(Separate PR will delete the legacy layer.)
tclune added 2 commits January 5, 2026 13:33
- eleminated some print statements
- needs more tests
Necessary for use with VerticalGridManager.
@tclune tclune requested a review from a team as a code owner January 5, 2026 18:36
@tclune tclune added 0 Diff The changes in this pull request have verified to be zero-diff with the target branch. 📈 MAPL3 MAPL 3 Related Changelog Skip Skips the Changelog Enforcer labels Jan 5, 2026
atrayano
atrayano previously approved these changes Jan 5, 2026
Library name error was missed due to case-insensive file system in
development environment.
tclune added 8 commits January 6, 2026 09:24
- Not crucial, but found a few irregularities while investigating a
different issue.
Retrieving params in FieldSetGet was causing cyclic dependencies.  So
now we put the burden on the client code to translate to/from an info
object.
Surprisingly only showed up with gfortran + release.
atrayano
atrayano previously approved these changes Jan 8, 2026
@tclune tclune merged commit c6f8f11 into release/MAPL-v3 Jan 8, 2026
33 checks passed
@tclune tclune mentioned this pull request Jan 8, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

0 Diff The changes in this pull request have verified to be zero-diff with the target branch. Changelog Skip Skips the Changelog Enforcer 📈 MAPL3 MAPL 3 Related

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants