OpenMDAO 3.43 Support

[dzalkind]

Purpose

I ran into an issue in WEIS using the latest version of openmdao (3.43): https://github.com/dzalkind/WEIS/actions/runs/23260484956/job/67627345448#step:9:233

It seems like in v3.43, openmdao is checking connections and making graphs: https://github.com/OpenMDAO/OpenMDAO/releases. I traced the error to this new feature, which seems to run regardless of whether you use the new command, and it's getting stuck around the ca_usr_grid and other member load inputs/outputs.

These inputs seem to be connected/promoted more than once, and this change seemed to make my example on the WEIS-side work.

Type of change

• Bugfix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (non-backwards-compatible fix or feature)
• Code style update (formatting, renaming)
• Refactoring (no functional changes, no API changes)
• Documentation update
• Maintenance update
• Other (please describe)

Testing

Checklist

• I have run existing tests which pass locally with my changes
• I have added new tests or examples that prove my fix is effective or that my feature works
• I have added necessary documentation

[gbarter]

Seems like the issue was that the old code had "cd_usr" and "ca_usr" etc that were added to the promote list each time in the k-for-loop. Not sure if deleting all together is the right solution. . .

[dzalkind]

I think it's promoting for each member, and the tuple is (old_name, new_name)

I decided to remove this because I became convinced that the issue was that these outputs were being promoted both here and here.

[gbarter]

I see, okay. Thanks for explaining

[dzalkind]

I should have noted in the PR that this is a starting point, and I'm not sure this is the correct way to fix this. It's one way to bypass the error I was getting.

[gbarter]

I've finally been able to devote time to this, recreate the bug, and convince myself that the fix here is correct. I'll look into why the tests are failing now.

[dzalkind]

Great...thanks!

I'm guessing it's a combination of

1. my heavy-handed "fix," (the water density issue) and
2. openmdao checking things in a bit more detail now, that may have previously been swept under the rug (possible example).

I can keep an eye on your new branch and have some time to tinker early this week, too.

[gbarter]

I've got it covered- no worries. Should be able to merge the fix into develop by the evening.

Great job with the debugging and detective work. OpenMDAO may be doing more rigorous testing, but it's error messages provide no help what-so-ever.