Update CI environment lockfiles

[github-actions]

Lockfiles updated to the latest resolvable environment.

[trexfeathers]

I'm not sure if this explains the current nc-time-axis errors, but I know for certain that the versions of nc-time-axis and cftime proposed here are incompatible. Although not recorded in the nc-time-axis requirements, v1.3 needs a minimum of cftime>=1.5 (currently Iris is pinned cftime<1.3). See SciTools/nc-time-axis#66 (comment).

[bjlittle]

@trexfeathers Cool, I'll issue a patch for nc-time-axis 👍

[trexfeathers]

I can't work out why this is still resolving with nc-time-axis v1.3.1, given that recipe now requires cftime>=1.5. @bjlittle could it still be picking up a previous build, despite conda-forge/nc_time_axis-feedstock#23?

[trexfeathers]

could it still be picking up a previous build, despite conda-forge/nc_time_axis-feedstock#23?

EDIT: resolution definitely is picking up an older build - nc-time-axis-1.3.1-pyhd8ed1ab_0. That probably makes sense, since it's the latest available version on conda-forge that doesn't have the minimum cftime pin. Can we follow the conda-forge broken package steps so that it's not an option for resolution?

[bjlittle]

@trexfeathers I think this is now fixed, see conda-forge/conda-forge-repodata-patches-feedstock#144, which retro-patches the appropriate nc-time-axis recipe metadata for 1.3.0, and 1.3.1 for good measure.

It's pretty cool... I didn't know that was possible, thanks @ocefpaf ❤️

[bjlittle]

So when we unpin cftime in iris, we'll get the correct behaviour as far as nc-time-axis is concerned 👍

[ocefpaf]

It's pretty cool... I didn't know that was possible, thanks @ocefpaf

We were very "afraid" of repo patches due to possible reproducibility issues. In fact I was very against it. However, they are so useful that we got addicted to them and they are here to stay now 😄

[rcomer]

I'm not sure if this explains the current nc-time-axis errors, but I know for certain that the versions of nc-time-axis and cftime proposed here are incompatible.

FWIW, I'm pretty sure it does: SciTools/nc-time-axis#71 (comment)