Conversation
See below for the new language, to clarify that documentation is only necessary for "user visible changes." I added "including notable bug fixes" to indicate that minor bug fixes may not be worth noting (I was thinking of test-suite only fixes in this category) but perhaps that is too confusing.
|
I think that's reasonable I was leaning to "everyone add their name", as it's a way of giving contributors credit, and some visible ownership — even for small changes which might be someone's first PR. But I agree that's not always consistent with making the whatsnew better for readers. |
I agree, this is a nice idea. My goals were to make things simpler for new contributors and also to make whats-new a little bit for readers. Two ways we might still achieve this:
|
Now I think of it, we do this already: So maybe we paste that into the whatsnew too, and we're covered? |
Seems like this was the consensus at this meeting. So let's add another step to |
|
OK, revised. |
…o-combine * 'master' of github.com:pydata/xarray: (81 commits) use builtin python types instead of the numpy alias (pydata#4170) Revise pull request template (pydata#4039) pint support for Dataset (pydata#3975) drop eccodes in docs (pydata#4162) Update issue templates inspired/based on dask (pydata#4154) Fix failing upstream-dev build & remove docs build (pydata#4160) Improve typehints of xr.Dataset.__getitem__ (pydata#4144) provide a error summary for assert_allclose (pydata#3847) built-in accessor documentation (pydata#3988) Recommend installing cftime when time decoding fails. (pydata#4134) parameter documentation for DataArray.sel (pydata#4150) speed up map_blocks (pydata#4149) Remove outdated note from datetime accessor docstring (pydata#4148) Fix the upstream-dev pandas build failure (pydata#4138) map_blocks: Allow passing dask-backed objects in args (pydata#3818) keep attrs in reset_index (pydata#4103) Fix open_rasterio() for WarpedVRT with specified src_crs (pydata#4104) Allow non-unique and non-monotonic coordinates in get_clean_interp_index and polyfit (pydata#4099) update numpy's intersphinx url (pydata#4117) xr.infer_freq (pydata#4033) ...
* upstream/master: (21 commits) fix typo in error message in plot.py (pydata#4188) Support multiple dimensions in DataArray.argmin() and DataArray.argmax() methods (pydata#3936) Show data by default in HTML repr for DataArray (pydata#4182) Blackdoc (pydata#4177) Add CONTRIBUTING.md for the benefit of GitHub Correct dask handling for 1D idxmax/min on ND data (pydata#4135) use assert_allclose in the aggregation-with-units tests (pydata#4174) Remove old auto combine (pydata#3926) Fix 4009 (pydata#4173) Limit length of dataarray reprs (pydata#3905) Remove <pre> from nested HTML repr (pydata#4171) Proposal for better error message about in-place operation (pydata#3976) use builtin python types instead of the numpy alias (pydata#4170) Revise pull request template (pydata#4039) pint support for Dataset (pydata#3975) drop eccodes in docs (pydata#4162) Update issue templates inspired/based on dask (pydata#4154) Fix failing upstream-dev build & remove docs build (pydata#4160) Improve typehints of xr.Dataset.__getitem__ (pydata#4144) provide a error summary for assert_allclose (pydata#3847) ...
See below for the new language, to clarify that documentation is only necessary
for "user visible changes."
I added "including notable bug fixes" to indicate that minor bug fixes may not
be worth noting (I was thinking of test-suite only fixes in this category) but
perhaps that is too confusing.
cc @pydata/xarray for opinions!
isort -rc . && black . && mypy . && flake8whats-new.rstfor user visible changes(including notable bug fixes) and
api.rstfor new API