CCPP Framework Meeting Minutes 2025 02 05

Agenda

• GitHub issues/PRs
• New items
• Updates
• Other business
• notes

---

Attendees: Michael Kavulich, Dom Heinzeller, Jordan Powers, Cheryl Craig, RhaeSung Kim, Dustin Swales, Grant Firl, Soren Rasmussen, Courtney Peverley, Jesse Nusbaumer

GitHub issues/PRs

CCPP Framework (issues, PRs, discussions)

• Add support and testing for equivalent units #571

• Enable multiple schemes to use the same local name for different variables #621

  • Any updates?

• Constituent index lookup #622

  • Approved, but looks like Steve was still going to make some changes?

• unit testing? Discussion#623

• MPI interface for CCPP physics schemes Issue#624

• Bugfix for unpromoted ddt-type scalar variables #626

• Capgen in SCM: Bug fix for scheme ordering in non-run phases #629

  • Merged, surprised this has never caused trouble before!

• Allow for multiple instances of local_name in Cap. Issue#630

  • Capgen in SCM: Multiple instances of local_name in Group Cap #631
    • Just one suggestion and one question, otherwise looks ready to go

• Sub-cycling not working properly Issue#632, Suite-part list is wrong Issue#634

  • Capgen in SCM: Fix to allow for scheme subcycling. #633
    • For merge after #631

• Add diagnostic_name metadata field Issue#635

• Allow for differing module/file names Issue#636

  • Capgen in SCM: Differing source and module name #637

• Update main with develop branch #638

  • Merged, is testing status okay?

• Pass DDT subfields into Group Caps. Issue#639

  • Capgen in SCM: DDT subfields into Group Caps #640

• Code Valor: high-impact violation in ccpp_prebuild code generator Issue#641

  • ccpp_prebuild.py: remove execute function from common.py, clean files the Python way #642

Drafts

• Testing refactor #612
  • Continued discussion

Standard names (issues, PRs, discussions)

• add emissions names #86

  • Just need one final update from Jerome and then ready to merge.

• Second rules update in v1 branch, update names, update write_standard_name_table.py to allow for subsections #87

  • Comments addressed, waiting for feedback

New items for discussion

Updates from last time

Previous notes

Other business

Offer suggestions/comments on development workflow rules:

• https://docs.google.com/document/d/1iMh_iKsUrB_mOMF6lGkFggG85JV62NFkTjKJboqWvV8/edit?tab=t.0

Mike: I have disabled merge commits since it is too easy to accidentally commit them that way. There have been two changes to develop that were not squashed in the past two months. It's easy enough to re-enable them temporarily when merging in main.

Meeting notes

CCPP Framework

• Ed Hartnett unit testing
  • Dom pushed back; we need to wait for Michael W’s test refactor to be complete anyway
  • Michael to create an issue from the discussion
• MPI interface
  • Used by JEDI
  • Will keep to just wrapping broadcast calls
  • Will be optional, but we’ll see how many developers use it
• Multiple local names in group cap
  • Cheryl & Courtney to take a look
• Differing source and module name
  • Design questions
    • Possible future work of new class (beyond scheme, host) that’s parsed first
    • Dustin to open an issue
• Update main with develop branch
  • Merged, but UFS hasn’t tested yet
  • Dom suggests updating template w/ check boxes for each organization
    • Michael to look into this
• DDT subfields in group caps
  • Dustin to move to draft until PR it’s based on is merged
• Remove execute function from common.py (prebuild)
  • Dustin to look at, Michael K to merge

Standard Names

• Almost ready to merge Jerome’s emissions names (#86)
  • XML and metadata still out of sync
• v1 rules update
  • Michael K addressed Dom & Steve’s comments
  • Needs more eyes; Jesse to take a look next week
• Michael will change name to “ESMStandardNames” on 2/13

Discussion

• Squash vs merge discussion
  • Michael K to look at new github branch rules
  • General agreement with Michael K’s decision to disable merge commits and then enable them to merge to main
• Dustin - in group caps, argument lists are really long
  • In debug mode/intel, intel complained about too many arguments (255)
  • Dom - if some schemes want full ddt, and some schemes want just some fields, are we passing both?
    • We should ONLY pass the DDTs
  • Hard limit for arguments - might work to override
    • Somehow works in non-debug mode, though
    • Soren testing flags, Dustin to look into repercussions of adding the flag in the UFS
• Michael - people have been asking about standard names governance
  • Who’s going to be a member?
    • Someone from ESMF team should be involved - Ann Tsay is the ESMF lead