Fix env_mach_specific validation #1509
Conversation
There was a problem hiding this comment.
The change to env_mach_specific.xsd fixes #1505 - thank you!
I'm not sure I see the need for the other changes here. Are there legitimate times when we expect a user to need to bypass schema validation? Or is this just a mechanism to give users a workaround in case there are not-yet-discovered problems in the xsd files? I'm wondering if it's worth cluttering the user interface with this....
I also don't love the slippery slope of introducing this new global variable, and I also notice that it makes setup_standard_logging_options a misnomer now.
All that said, I'm okay with all of these changes if you feel they're warranted.
|
@jedwards4b I agree with @billsacks . |
|
I agree that the name setup_standard_logging_options is incorrect, but my intent was always that it be setup_standard_options. This option was added based on my perception of @billsacks level of distress when an xml file didn't validate, but there may be other good reasons for it. Do you have any suggestions other than the global for how to implement? |
|
The only other way would be a very-annoying pass-down of a parameter all the way down to GenericXML. |
|
okay - only the schema is fixed |
|
Update the description to match the title. |
Fix the schema of env_mach_specific to allow environment_variables, mpirun and entry to not be specified.
Test suite: scripts_regression_tests.py
Test baseline:
Test namelist changes:
Test status: bit for bit
Addresses #1505
Code review: