Py3

[doutriaux1]

All done. Py3 fails because it runs with esmf 7.1 which is different form 7.0 (py2 fails both from master or this with esmf 7.1)

[durack1]

@doutriaux1 time for a new baseline with ESMF7.1 (if we get comparable results using regrid2 for the atmos grids)

@gleckler1

[doutriaux1]

@gleckler1 thiking of switching the tests to use linear at least until CDAT/cdms#231 is fixed. What do you think?

[durack1]

@doutriaux1 I think we really need to work through this one with @dnadeau4 and make sure that the defaults are giving the results that a user expects - it's a powerful regridder, but default outputs are VERY important

[doutriaux1]

@durack1 agreed, that's why we could also disable conservative completely at the moment and raise an error, until this bug is addressed. Thoughts?

[durack1]

@doutriaux1 this is not released, right? A warning would be a good idea until the behaviour is acceptable

[gleckler1]

@doutriaux1 @dnadeau4 @durack1 Seems like a good thing to do if not too onerous. I have told the E3SM people using the PMP to not use ESMF conservative until we resolve this, but maybe better is to disable for now.

[doutriaux1]

ok will disable on Monday and merge this in master.