Update externals to what is expected to be in cesm2_3_beta17 fix some conda issues... #2385
Conversation
0f884bfec Merge pull request ESCOMP#205 from jedwards4b/sunset_svn_git_access 82a5edf79 merge in billsacks:svn_testing_no_github 17532c160 Use a local svn repo for testing 9c904341a different method to determine if in tests 539952ebd remove debug print statement cc5434fa7 fix submodule testing 1d7f28840 remove broken tests 04e94a519 provide a meaningful error message 38bcc0a8c Merge pull request ESCOMP#201 from jedwards4b/partial_match b4466a5aa remove debug print statement c3cf3ec35 fix issue with partial branch match git-subtree-dir: manage_externals git-subtree-split: 0f884bfec8e43d0c02261de858d6ec3f6d855e51
…o needed to update numpy version for it to work
… are main and develop, with the new externals the testlist wasn't working, it failed trying to setup cases
ekluzek
added
the
enhancement
|
@samsrabin I added you as reviewer for this. The main thing to look at is py_env_create. |
|
The change in answers is just because the USUMB test failed previously , and the FatesMimics test is expected to fail. So it's just the ERI tests that I need to figure out. |
|
I now have a cime tag that fixes the ERI tests, so they are fine. So aux_clm should be good, but I need to rerun as I just updated. The issue now is that the python system tests are failing for the lilac tests. With a message like this for example... |
There was a problem hiding this comment.
I can confirm that py_env_create works on both Derecho and Izumi, thanks! I saw the same "local variable 'machine' referenced before assignment" error when I tried the Python system tests on Derecho, but interestingly I didn't see that on Izumi. Maybe a clue?
…D gnu, use one or the other, roughly alturnate between them
|
Testing is fine. The test SMS_D_Lm1_Mmpi-serial.CLM_USRDAT.I1PtClm50SpRs.derecho_intel.clm-USUMB_nuopc is labeled as different answers, but that's just because the baselines don't have all the files. So I consider this fine. |
|
I thought I had fixed the lilac issue, but in testing for #2416 I see that it's still an issue. So I need to get make stest working again here. |
|
OK, from talking with @billsacks it looks like the problem is this one filed in cime. And @billsacks has a simple fix for that -- that will likely fix the problem coupled with some other simple changes that we talked about. |
|
This small change to cime solves the issue! |
|
This also brings up something that we will need to do sometime, but can get away with for now: #2419 |
Fix undefined machine on cime6.0.217_httpsbranch branch Initialize machine to None to prevent not initializing it to anything. This is cherry-picked from #4596 Testing: Run CTSM python testing (which was failing before this change) ./scripts_regression_tests.py (currently running) Test status: bit-for-bit Fix for #4588 on the httpsbranch branch. This fixes issues in CTSM discussed here: ESCOMP/CTSM#2385 (comment) User interface changes?: None Update gh-pages html (Y/N)?: N
Description of changes
Update externals to what's currently planned for cesm2_3_beta17.
Also update manage_externals. This is only important if you add CAM to your externals.
This was important on the Dust PR, so this will allow that PR to show less differences
to master when this comes in.
Specific notes
Contributors other than yourself, if any:
CTSM Issues Fixed (include github issue #):
Fixes #2351
Fixes #2380
Fixes #2331
Fixes #1910
Fixes #1658
Are answers expected to change (and if so in what way)? No
Any User Interface Changes (namelist or namelist defaults changes)? No
Testing performed, if any: standard
ERI tests currently fail
This one might have trouble: ERP_P256x2_Ld30.f45_f45_mg37.I2000Clm51FatesRs.derecho_intel.clm-mimicsFatesCold
This one shows different answers so I'll fix that: SMS_D_Lm1_Mmpi-serial.CLM_USRDAT.I1PtClm50SpRs.derecho_intel.clm-USUMB_nuop