NOAA operations and research computers

[rgrumbine]

Additions to run regression tests on NWS operational system 'phase2' and NOAA R+D system 'theia':

• Developer(s):
  Robert Grumbine

• Are the code changes bit for bit, different at roundoff level, or more substantial?
  Bit for bit

• Does this PR create or have dependencies on Icepack or any other models?
  No

• Is the documentation being updated with this PR? (Y/N)
  If not, does the documentation need to be updated separately at a later time? (Y/N)

No, no

• Other Relevant Details:

Added to .gitignore to ignore the testsuite files and directories.
Though present, the theia_gnu and theia_pgi configurations have not been confirmed to pass the full regression suite.

[apcraig]

The automated testing is failing. It cannot find the icepack code. It looks like the icepack submodule was removed on this branch. Maybe that was done temporarily and for testing, but we'll need to fix that. I think we just need to revert the .gitmodules file. Otherwise, the changes are largely just machine ports. Please revert the .gitmodules file and push and then lets confirm the automated testing is passing.

[rgrumbine]

I restored the .gitmodules and Travis still fails. It looks like it still can't find the icepack. Is .gitmodules case sensitive?

[apcraig]

I had another look and triggered another build just to make sure it wasn't travis. I'm not entirely sure what is happening or how this came about, but the icepack submodule is still not there. You can see at the bottom of the diffs that "icepack" has been removed. So maybe it wasn't enough to revert the .gitmodules file, maybe more needs to be done. Maybe the icepack submodule needs to be formally added some other way again. If you need some help, let me know. One option might be to create a new branch and then move the changes on this branch to the new branch, delete this PR, then create a new PR on the new branch. Just make sure icepack is not removed on the new branch. I'm sure we can also bring icepack back on the current branch somehow.

[eclare108213]

There are some examples for working with Icepack as a submodule here:
https://github.com/CICE-Consortium/About-Us/wiki/Git-Workflow-Guidance#submodules

[apcraig]

It's still not running in travis. The .gitmodules now is different again. And the diff suggests that 46 files were changed in Icepack. I think we need a new PR. I have added a section in https://github.com/CICE-Consortium/About-Us/wiki/Git-Workflow-Guidance#pull-requests about how to refresh the PR by shifting to a new branch. One problem is that the current branch is "rgrumbine:master". We recommend not using master to do development and as a result, you will not be able to follow the process outlined in the document. This is what I suggest. I can create a branch in my fork and pull your "master" mods onto my branch. I can then create the PR. At that point, @rgrumbine can either delete the CICE fork and create a new one clean again or revert the master in the current fork to the version in the CICE-Consortium and then try not to do development on master again. That should both clean up the rgrumbine fork and get the PR onto the trunk. I will setup a branch in my fork now.

[rgrumbine]

Tony, I've separately made a pull request for my version of Icepack. That includes the theia and mac modifications as well, and passes all tests on travis. That reassures me that I haven't done something too extravagantly incorrect in setting up my repositories. But I could easily have established the CICE repository in some odd way that it refuses to recognize the icepack repo. Thanks, Bob

…

On Wed, May 29, 2019 at 1:27 PM Tony Craig ***@***.***> wrote: The automated testing is failing. It cannot find the icepack code. It looks like the icepack submodule was removed on this branch. Maybe that was done temporarily and for testing, but we'll need to fix that. I think we just need to revert the .gitmodules file. Otherwise, the changes are largely just machine ports. Please revert the .gitmodules file and push and then lets confirm the automated testing is passing. — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#317?email_source=notifications&email_token=AFQ65HLUQVSQCSNJEIEPT6LPX24JJA5CNFSM4HQMPDBKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODWQB5QI#issuecomment-497032897>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFQ65HM7I6OHZYPZOLH3NLDPX24JJANCNFSM4HQMPDBA> .

-- Tel: 301-683-3747 Coupling and Dynamics Group

[rgrumbine]

That's it. I had done something bad to the submodule status. Now that this is repaired, travis is trying to run -- but failing at: The command "wget ftp://ftp.cgd.ucar.edu/archive/Model-Data/CICE/CICE_data_forcing_gx3_all.tar.gz && tar xvfz CICE_data_forcing_gx3_all.tar.gz -C ~" failed and exited with 4 during . maybe some system maintenance at ucar? Bob

…

On Thu, May 30, 2019 at 12:47 PM Elizabeth Hunke ***@***.***> wrote: There are some examples for working with Icepack as a submodule here: https://github.com/CICE-Consortium/About-Us/wiki/Git-Workflow-Guidance#submodules — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#317?email_source=notifications&email_token=AFQ65HJ4MFQC6QKURPM5EIDPYAAILA5CNFSM4HQMPDBKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODWS3JGI#issuecomment-497398937>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFQ65HOXQMRW6L765C54CT3PYAAILANCNFSM4HQMPDBA> .

-- Tel: 301-683-3747 Coupling and Dynamics Group