html_url,issue_url,id,node_id,user,created_at,updated_at,author_association,body,reactions,performed_via_github_app,issue
https://github.com/pydata/xarray/pull/5622#issuecomment-886869099,https://api.github.com/repos/pydata/xarray/issues/5622,886869099,IC_kwDOAMm_X8403Ixr,35968931,2021-07-26T16:58:23Z,2021-07-26T16:58:23Z,MEMBER,"My 2 cents:

> How important is the argument order for the plotting functions?

Obviously it would be nice to be able to get the arguments in the same order across functions, but I think we probably care more about not suddenly breaking backwards compatibility - any change to the order should technically require a deprecation cycle... That said standardizing something more consistent would be good.

> How important are figure legends in facetgrid? Some tests breaks on this now but I'm not sure it's a good idea to change to figlegends instead of legend per plot.

Not really sure what the best thing to do is.

> We seem to be lacking tests regarding how the plot should look like.

My understanding is that testing the displayed output of plotting functions is [notoriously tricky and unreliable](https://stackoverflow.com/questions/27948126/how-can-i-write-unit-tests-against-code-that-uses-matplotlib), hence when we currently test we interrogate properties of matplotlib objects. There are libraries that check images are correct, and `@pytest.mark.flaky` helps, but I'm not sure if it's worth it. Also we are merely wrapping matplotlib here - if we make efforts to check that the objects' properties are as expected then at some level we obviously have to trust that that corresponds to the same image.

> For example an acccidentally inverted plot didn't crash the tests.

Is there no obvious object property test that would have caught this?","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,948049609