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/3737#issuecomment-580824275,https://api.github.com/repos/pydata/xarray/issues/3737,580824275,MDEyOklzc3VlQ29tbWVudDU4MDgyNDI3NQ==,14808389,2020-01-31T17:13:38Z,2020-01-31T17:13:38Z,MEMBER,"looking at the way you pinned the versions (using `>=`) this should be fine as is, no need for switching to pip.

I have also been experimenting with using channels like
```
https://metachannel.conda-forge.org/conda-forge/iris,netcdf4,numpy,numba,pandas,cartopy
```
which in theory should reduce the downloaded `repodata.json` and thus reduce the used memory, but I probably did not use the correct url: the used memory increased instead.

Another way to reduce it could be using `mamba` which reimplements the solver in C++, but it's experimental right now and it's probably also something that has to be installed by RTD, not us (it needs to be installed before creating the environment).","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,558230156
https://github.com/pydata/xarray/pull/3737#issuecomment-580813508,https://api.github.com/repos/pydata/xarray/issues/3737,580813508,MDEyOklzc3VlQ29tbWVudDU4MDgxMzUwOA==,2448579,2020-01-31T16:47:12Z,2020-01-31T16:47:12Z,MEMBER,I'll merge when this build succeeds: https://readthedocs.org/projects/xray/builds/10362524/,"{""total_count"": 1, ""+1"": 1, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,558230156
https://github.com/pydata/xarray/pull/3737#issuecomment-580813313,https://api.github.com/repos/pydata/xarray/issues/3737,580813313,MDEyOklzc3VlQ29tbWVudDU4MDgxMzMxMw==,2448579,2020-01-31T16:46:44Z,2020-01-31T16:46:44Z,MEMBER,"> What do we do about the build for v0.15.0?

Not sure. We could figure out the `pip install` bit and release 0.15.1. In the meantime, once this PR is merged to stable, the docs will still be accurate for 0.15.0. ","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,558230156
https://github.com/pydata/xarray/pull/3737#issuecomment-580806202,https://api.github.com/repos/pydata/xarray/issues/3737,580806202,MDEyOklzc3VlQ29tbWVudDU4MDgwNjIwMg==,14808389,2020-01-31T16:30:13Z,2020-01-31T16:30:39Z,MEMBER,I'd agree with merging. What do we do about the build for `v0.15.0`?,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,558230156
https://github.com/pydata/xarray/pull/3737#issuecomment-580805099,https://api.github.com/repos/pydata/xarray/issues/3737,580805099,MDEyOklzc3VlQ29tbWVudDU4MDgwNTA5OQ==,2448579,2020-01-31T16:28:02Z,2020-01-31T16:28:02Z,MEMBER,The docs build with this branch did work: https://readthedocs.org/projects/xray/builds/10362278/ so shall we merge this to stable and then work to pip install things on master?,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,558230156
https://github.com/pydata/xarray/pull/3737#issuecomment-580804855,https://api.github.com/repos/pydata/xarray/issues/3737,580804855,MDEyOklzc3VlQ29tbWVudDU4MDgwNDg1NQ==,14808389,2020-01-31T16:27:34Z,2020-01-31T16:27:34Z,MEMBER,"also, you can just remove the whole `python` section in `readthedocs.yml`, it is not required and since we use `conda` we use our own python version anyway.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,558230156
https://github.com/pydata/xarray/pull/3737#issuecomment-580803737,https://api.github.com/repos/pydata/xarray/issues/3737,580803737,MDEyOklzc3VlQ29tbWVudDU4MDgwMzczNw==,14808389,2020-01-31T16:25:18Z,2020-01-31T16:25:18Z,MEMBER,"you could also pip install everything except `cartopy` and `cf-units` (which is a dependency of `scitools-iris`, the package name for `iris` on PyPI). The memory usage on my machine is at about 700MB, which should help","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,558230156