Update post and regional_workflow hashes, QOL improvements#167
Conversation
…o proper hash once PR is merged
jwolff-ncar
left a comment
There was a problem hiding this comment.
Couple of questions on the regional_workflow fork/branch set in this PR.
|
Ran a number of tests on Cheyenne. The following WE2E tests passed:
|
|
Ran another set of tests on Hera, all succeeded that were expected to (there are known failures in the testing system right now due to data being shuffled around). The following WE2E tests passed:
On Jet, I was not able to run any tests, as the old testing data has all been moved. I did build the code successfully. |
|
@mkavulich Can you elaborate on what the BUILD_ALWAYS flag does? What is the undesired behavior that this fixes? Thanks. |
|
@gsketefian The BUILD_ALWAYS flag ensures that when making changes to the source code in src/ and rebuilding, |
* Fix externals reference for GSI in Externals.cfg The ncdiag was removed from GSI and replaced by a submodule. This requires the GSI to be checked out with the ncdiag submodule, which requires addition of a separate Externals_gsi.cfg file to specify which GSI submodules are needed. * Add newline to last line of Externals_gsi.cfg
DESCRIPTION OF CHANGES:
EMC_post builds are currently failing on Hera due to an update to NCEPLIBS. While that is annoying, we also discovered that the App's EMC_post hash is from December 2020...way out of date! Updating to the latest hash fixes the build issue. In addition, regional_workflow must be updated and an additional PR must be applied (https://github.com/NOAA-EMC/regional_workflow/pull/581) in order to accommodate the new executable and filenames for the newer EMC_post hash.
In addition, some long-overdue quality-of-life improvements are added:
TESTS CONDUCTED:
Cheyenne
Hera
DEPENDENCIES: