De-template COM declarations#4555
Conversation
|
I'm trying to build an experiment with this branch now and setup_workflow.py is breaking. Looks like there are some templated variables in dev/workflow/rocoto/gfs_tasks.py that need to be de-templified as well. |
|
@TravisElless-NOAA There is an error analysis report posted above that might help with some of those templated variable issues. |
dev/jobs/JGLOBAL_FORECAST
Outdated
| if [[ 10#${ENSMEM:--1} -ge 0 ]]; then | ||
| # Turn ensmem into 3-digit format with leading zeros | ||
| memdir=$(printf "mem%03i" "${ENSMEM}") |
There was a problem hiding this comment.
This is failing for ensemble members 8&9 with the following error message.
line 45: printf: 008: invalid octal number
There was a problem hiding this comment.
You might need that 10# in front of ${ENSMEM}
There was a problem hiding this comment.
@TravisElless-NOAA @CoryMartin-NOAA I think this is fixed now. I successfully reran an enkfgdas_fcst job for member 1.
Travis, could you try rerunning your test case now?
There was a problem hiding this comment.
It's working correctly for mem 9 in the forecast job now. Thanks @DavidHuber-NOAA!
|
Restarting tests on WCOSS2. |
|
All tests passed on WCOSS2 except the |
|
Gaea's gitlab runner seems to be down and Ursa's runner scripts seemed to crash. I will carry on the Ursa tests manually. |
|
All tests passed on WCOSS2: |
TravisElless-NOAA
left a comment
There was a problem hiding this comment.
I haven't tripped over these yet in my test, but probably need to update the $ENSMEM call in these places too?
Co-authored-by: Travis Elless <[email protected]>
Description
This removes all calls to
declare_from_tmpland makes allCOMdeclarations explicit. It also adds a utility that can be used to de-template COM declarations.Resolves #4522
Type of change
Change characteristics
How has this been tested?
All cases passed on C6. Further testing on other platforms will commence shortly.
Checklist