MNT: rerender

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-webservice.

I've rerendered the recipe as instructed in #24.

Here's a checklist to do before merging.

• Bump the build number if needed.

Fixes #24

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[beckermr]

@conda-forge-admin rerender

[beckermr]

@conda-forge/proplot can you look at this?

xref: proplot-dev/proplot#309
xref: #23

[lukelbd]

Thanks for the help!

Does this update the conda recipe for all proplot versions or just 0.9.5? While we're at it we might as well update the recipe for all proplot versions (...assuming that's trivial to do?).

[lukelbd]

Or perhaps that's non-trivial... and what I should be doing is pinning every proplot version to the latest matplotlib version at the time of release, rather than using a naked matplotlib dependency?

[beckermr]

We can push a fix for all versions to our repodata patches. However, if proplot is using non-declared apis within matplotlib, a pin to a specific version (maybe major.minor) is not a horrible idea.

[lukelbd]

Cool -- then I assume that global fix would be outside of this PR? So, the best course would be to close this PR, edit conda's repodata patches (...don't really know what that means by the way 🤷‍♂️), then in future proplot releases I'll restrict the matplotlib version in setup.cfg rather than meta.yml (always testing against the latest version)?

The general rule proplot follows is "use matplotlib's public API whenever possible"... but occasionally it's not possible. Now, if I am using private API, I try to also do quiet/gentle failures with warnings in case it the API changes, but previously I didn't always do that consistently.

Matplotlib also does not follow semantic versioning and occaionally deprecates the public API even in minor releases, and they can be fairly quick to remove deprecations, so restricting the mpl version is probably a good practice going forward.

[beckermr]

You have to do it in the meta.yaml too. Conda does not know about the setup.cfg. Plus there is the setuptools issue I identified. This PR needs to be merged.

[beckermr]

Thank you!

I'll make the patch to the repodata and tag you on the PR so you can see it.

[beckermr]

see here: conda-forge/conda-forge-repodata-patches-feedstock#201

PR to be made