Janus

@TomAugspurger: I have a potential fix for part of the `groupby` issues, see c882db554888b141fa95d1891f9cb84ab06b95f4. The code was originally included in PR #35688 but was taken out to allow a separate...

> > I have a potential fix for part of the groupby issues, see [c882db5](https://github.com/pandas-dev/pandas/commit/c882db554888b141fa95d1891f9cb84ab06b95f4). The code was originally included in PR #35688 but was taken out to allow a...

Seeing a similar effect here, and it seems to be a regression: My existing lockfiles were generated with `conda-lock` a while back and agree with `conda create`, but when running...

Reduced this to the smallest reproducible form I could find: The following `environment.yml` is not processed correctly by `conda-lock` v 0.13.2: ``` name: example-project channels: - defaults dependencies: - python>=3.8,

> The mamba solver tends to behave a bit better in almost all cases. Would recommend looking there first > The solver behavior for conda-forge vs defaults may also differ...

I have previously contributed fixes to Hugo, but I am no golang expert and do not know enough about reflection/introspection in golang to offer any help on this. As a...