Skip to content

Skip cosmo5.01_fresh and CLM3.5 during recursive submodule clone - #136

Merged
s-poll merged 5 commits into
masterfrom
dev_gitsub-shallow
Jul 29, 2026
Merged

Skip cosmo5.01_fresh and CLM3.5 during recursive submodule clone#136
s-poll merged 5 commits into
masterfrom
dev_gitsub-shallow

Conversation

@s-poll

@s-poll s-poll commented Jul 28, 2026

Copy link
Copy Markdown
Member

models/cosmo5.01_fresh points to a repository that require authentication. Because of that, git clone --recurse-submodules and git submodule update --init --recursive fail (or hang waiting for password) for anyone who doesn't have credentials for this mirror, even if they don't need this component.

I would suggest that in this case the TSMP1 components (COSMO and CLM3.5) are not cloned to avoid confusion.

If one specifically need cosmo5.01_fresh or CLM3.5, one can still clone these repositories:

git submodule update --init models/cosmo5.01_fresh
or
git submodule update --init models/CLM3.5

@s-poll
s-poll requested review from jjokella and kvrigor July 28, 2026 16:41
@kvrigor

kvrigor commented Jul 29, 2026

Copy link
Copy Markdown
Member

CLM3.5-ParFlow-PDAF build is failing; perhaps we should remove it? Is CLM3.5-PDAF still being used @jjokella ?

@jjokella

Copy link
Copy Markdown
Contributor

Actually I am still running old testcases for backward compatibility with

  • CLM3.5-PDAF
  • and more importantly CLM3.5-ParFlow-PDAF

However, as I understand, this will still be possible with the explicit commands and that is completely fine with me to restrict the general command to TSMP2 component models.

@jjokella

Copy link
Copy Markdown
Contributor

However, I think this problem with the CLM3.5-ParFlow-PDAF build should be resolved before merging: https://github.com/HPSCTerrSys/TSMP2/actions/runs/30379409691/job/90343260280

seems the update: none somehow breaks also the update in the CI, I could not immediately find the reason, though.

@s-poll

s-poll commented Jul 29, 2026

Copy link
Copy Markdown
Member Author

I see that the CI should be fine before merging. I will take a look.

@jjokella

Copy link
Copy Markdown
Contributor

I just investigated a little more into it and I think that

git submodule update --init models/CLM3.5

is not enough to overried the update = none, seems there is --checkout for this reason:

git submodule update --init --checkout models/CLM3.5

From the manpage:

   --checkout
      This option is only valid for the update command. Checkout the commit recorded in the superproject on a detached
      HEAD in the submodule. This is the default behavior, the main use of this option is to override
      submodule.$name.update when set to a value other than checkout. If the key submodule.$name.update is either not
      explicitly set or set to checkout, this option is implicit.

@s-poll
s-poll merged commit e7c3cee into master Jul 29, 2026
8 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants