Michael Waskom

I'm confused about how what you're describing there is not represented in the notebook you shared? You seem to be plotting `heatmap` just fine there?

Oh that's interesting! I'd say this sounds like an issue on the ipympl side of things. I wish I could offer a hypothesis based on the fact that it seems...

This would be a very welcome addition for downstream libraries. For documentation of specific parameters, the admonition produced by versionadded directive takes up a lot of vertical space, so I...

Ah thanks missed that original discussion. It seems like @jklymak made the opposite argument from point 3 here, that logging is preferable because it's easier for knowledgable users to disable...

Well the warnings module has a context manager for temporarily changing warning filters. I’m not aware of anything similar for logging?

> For the RGB message, that seems something a user is more likely to get wrong than a library, in which case the log message seems the best place to...

Another case where it is useful to turn warnings into errors is in test suites.

I think you can use warnings and satisfy your desire for a `plt.helpful` mode by defining a custom warning class (`MatplotlibSharpEdgeWarning` 😉 ) and setting filters on it. I realized...

> The point was if someone does something and it behaves unexpectedly we can ask them to turn logging on. I still think logging is the wrong mechanism here, but...

More context in which the problem of "auto" producing catastrophically small bin widths is discussed: - https://github.com/matplotlib/matplotlib/issues/16403 - https://github.com/numpy/numpy/issues/15569 Numpy seemed open to adding a `bins="auto-capped"` (or whatever) option, but...