Ryan May
I think this is referring to catalog-level metadata. Currently, all catalog metadata are added as properties as they are found during parsing. Instead: 1. Don't pollute the Dataset object with...
Well, the outcome of [our discussion on MetPy](https://github.com/Unidata/MetPy/discussions/1678) was to move Siphon to focus exclusively on THREDDS, and MetPy should gain remote functionality. So I think the answer to that...
I knew it existed, so I kept searching more when issues didn't give me what I wanted. 😆
@WardF I think @durack1 is referring to a GitHub "project management" project that lists all the relevant issues and shows their status.
Is there actually a problem with defining CC=g++ ? What about manually linking in libstdc++?
Linking to the workshop in general is a good idea. What would make the most sense though is to steal the ideas from the workshop notebook and incorporate them directly...
Another thought: While all the datasets have names, I have yet to see us use the name of a dataset to pull something out of that `datasets` dictionary. That tells...
Search might also be helpful beyond regex with something like: https://github.com/jamesturk/jellyfish
We could consider tying it a `validate` kwarg-only argument to `get_query()`, though right now I think I'd want it to default to True. There's not a lot of point in...
For binary compatibility, all packages in conda-forge are pinned to an exact version of libnetcdf, and they migrate in lock-step. Might be able to request a bump here: https://github.com/conda-forge/conda-forge-pinning-feedstock/issues