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/4597#issuecomment-733254440,https://api.github.com/repos/pydata/xarray/issues/4597,733254440,MDEyOklzc3VlQ29tbWVudDczMzI1NDQ0MA==,2448579,2020-11-24T21:53:12Z,2020-11-24T21:53:12Z,MEMBER,LGTM too. Thanks @aaronspring ,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,748094631
https://github.com/pydata/xarray/pull/4597#issuecomment-731646114,https://api.github.com/repos/pydata/xarray/issues/4597,731646114,MDEyOklzc3VlQ29tbWVudDczMTY0NjExNA==,6628425,2020-11-21T22:27:34Z,2020-11-21T22:46:08Z,MEMBER,"Thanks @keewis I didn't realize this was also happening in CI.   Yeah, the version is definitely the problem:
```
cftime                    1.0.4.2          py36hc1659b7_0    conda-forge
```
My vote for now would be to skip these tests on versions less than 1.1.0.  Perhaps we could also think about adding a nicer error message if the version is not up-to-date enough.  

~~I guess our test coverage of `infer_freq` wasn't good enough to trip this up already.~~ Oops we already have a precedent of skipping the `infer_freq` tests that require this, e.g.:
https://github.com/pydata/xarray/blob/a2192158e3fbb94b2d972ff3e1693fffa65e50be/xarray/tests/test_cftimeindex.py#L1180-L1192
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,748094631
https://github.com/pydata/xarray/pull/4597#issuecomment-731645464,https://api.github.com/repos/pydata/xarray/issues/4597,731645464,MDEyOklzc3VlQ29tbWVudDczMTY0NTQ2NA==,14808389,2020-11-21T22:22:27Z,2020-11-21T22:24:04Z,MEMBER,"@spencerclark, see the [build logs of the `py36-min-all-deps` CI](https://dev.azure.com/xarray/xarray/_build/results?buildId=4263&view=logs&jobId=ba13898e-1dfb-5ace-9966-8b7af3677790&j=78b48a04-306f-5a15-9ac3-dd2fdb28db5e&t=feaf9696-a084-5b4b-a3e1-cb6a97ea3c30), we can't use version `1.1.0` unless we bump the version of `cftime`","{""total_count"": 1, ""+1"": 1, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,748094631
https://github.com/pydata/xarray/pull/4597#issuecomment-731645114,https://api.github.com/repos/pydata/xarray/issues/4597,731645114,MDEyOklzc3VlQ29tbWVudDczMTY0NTExNA==,6628425,2020-11-21T22:19:12Z,2020-11-21T22:19:12Z,MEMBER,"`freq` would be consistent with pandas:
```
In [1]: import pandas as pd

In [2]: pd.date_range(""2000"", periods=3)
Out[2]: DatetimeIndex(['2000-01-01', '2000-01-02', '2000-01-03'], dtype='datetime64[ns]', freq='D')
```

@aaronspring regarding the local test failure -- which version of `cftime` are you using in your environment?  The `daysinmonth` attribute did not always exist; it was added in version 1.1.0.  ","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,748094631
https://github.com/pydata/xarray/pull/4597#issuecomment-731643958,https://api.github.com/repos/pydata/xarray/issues/4597,731643958,MDEyOklzc3VlQ29tbWVudDczMTY0Mzk1OA==,14808389,2020-11-21T22:09:15Z,2020-11-21T22:09:15Z,MEMBER,"the parameter is named `freq` so I think it should be fine to use that, but I don't have a strong opinion on this","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,748094631
https://github.com/pydata/xarray/pull/4597#issuecomment-731638970,https://api.github.com/repos/pydata/xarray/issues/4597,731638970,MDEyOklzc3VlQ29tbWVudDczMTYzODk3MA==,14808389,2020-11-21T21:25:30Z,2020-11-21T21:28:45Z,MEMBER,"> somehow in some envs `xarray` (but not in others)

Actually, there's no inconsistency here: `py38-flaky` fails, but it is allowed to fail so we still get a green check mark (#4584 slightly changes that). `py36-min-nep18` does not install `cftime` so it doesn't fail, either. `Doctests` fails because it creates and prints `CFTimeIndex` objects with 1 or 2 values (and I think there's also a test that does something similar).

To fix the CI, you might need to fix both the `daysinmonth` and allow `CFTimeIndex` objects with less than 3 values.","{""total_count"": 1, ""+1"": 1, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,748094631