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/issues/1935#issuecomment-368687352,https://api.github.com/repos/pydata/xarray/issues/1935,368687352,MDEyOklzc3VlQ29tbWVudDM2ODY4NzM1Mg==,2443309,2018-02-26T23:22:37Z,2018-02-26T23:22:37Z,MEMBER,"I'm going to close now seeing that the upstream/pypy issue has been closed.
@tlechauve - please keep us updated on the pypy integration of xarray and the development of your REST API. ","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,299346082
https://github.com/pydata/xarray/issues/1935#issuecomment-368173087,https://api.github.com/repos/pydata/xarray/issues/1935,368173087,MDEyOklzc3VlQ29tbWVudDM2ODE3MzA4Nw==,67096,2018-02-23T23:54:26Z,2018-02-23T23:54:26Z,NONE,Sounds interesting!,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,299346082
https://github.com/pydata/xarray/issues/1935#issuecomment-368171446,https://api.github.com/repos/pydata/xarray/issues/1935,368171446,MDEyOklzc3VlQ29tbWVudDM2ODE3MTQ0Ng==,221526,2018-02-23T23:43:07Z,2018-02-23T23:43:07Z,CONTRIBUTOR,"@tlechauve One way or another, I'm sure @lesserwhirls would be interested to hear your thoughts on THREDDS.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,299346082
https://github.com/pydata/xarray/issues/1935#issuecomment-368129811,https://api.github.com/repos/pydata/xarray/issues/1935,368129811,MDEyOklzc3VlQ29tbWVudDM2ODEyOTgxMQ==,1102577,2018-02-23T20:29:06Z,2018-02-23T20:29:06Z,NONE,@jhamman I'm working on a light/python version of Thredds. it's not planned for now to be Open-source but i will think about it when a stable version will be available.,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,299346082
https://github.com/pydata/xarray/issues/1935#issuecomment-368082271,https://api.github.com/repos/pydata/xarray/issues/1935,368082271,MDEyOklzc3VlQ29tbWVudDM2ODA4MjI3MQ==,2443309,2018-02-23T17:37:42Z,2018-02-23T17:37:42Z,MEMBER,"@tlechauve - If your REST API is open source, would you mind sharing a link? I'd be interested to see what you're up to. ","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,299346082
https://github.com/pydata/xarray/issues/1935#issuecomment-367946371,https://api.github.com/repos/pydata/xarray/issues/1935,367946371,MDEyOklzc3VlQ29tbWVudDM2Nzk0NjM3MQ==,1102577,2018-02-23T08:45:48Z,2018-02-23T08:45:48Z,NONE,"HI @benbovy,
Actually its not really an issue and its not related to xarray, i'm using [falcon](https://falconframework.org) to make a REST API which provide netCDF data with xarray/dask.
According to falcon [benchmarks](https://falconframework.org/#sectionBenchmarks), pypy has better performance. But maybe in this case i should use CPython.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,299346082
https://github.com/pydata/xarray/issues/1935#issuecomment-367843351,https://api.github.com/repos/pydata/xarray/issues/1935,367843351,MDEyOklzc3VlQ29tbWVudDM2Nzg0MzM1MQ==,4160723,2018-02-22T22:24:34Z,2018-02-22T22:24:34Z,MEMBER,"> With the standard python everything works, but i have performance issues.
Can I ask what kind of performance issues do you have? I don't have much experience with PyPy, I'm just curious on cases where using xarray with PyPy represents a real advantage over using it with CPython.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,299346082
https://github.com/pydata/xarray/issues/1935#issuecomment-367815913,https://api.github.com/repos/pydata/xarray/issues/1935,367815913,MDEyOklzc3VlQ29tbWVudDM2NzgxNTkxMw==,1102577,2018-02-22T20:44:17Z,2018-02-22T20:44:17Z,NONE,Reported https://bitbucket.org/pypy/pypy/issues/2689/_threadlock-is-not-weakrefable,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,299346082
https://github.com/pydata/xarray/issues/1935#issuecomment-367809113,https://api.github.com/repos/pydata/xarray/issues/1935,367809113,MDEyOklzc3VlQ29tbWVudDM2NzgwOTExMw==,1102577,2018-02-22T20:19:44Z,2018-02-22T20:19:44Z,NONE,"@shoyer you are right !
```
Python 2.7.8 (2.4.0+dfsg-3, Dec 20 2014, 13:30:46)
[PyPy 2.4.0 with GCC 4.9.2] on linux2
Type ""help"", ""copyright"", ""credits"" or ""license"" for more information.
>>>> import weakref
>>>> import threading
>>>> lock = threading.Lock()
>>>> locks = weakref.WeakValueDictionary()
>>>> locks['foo'] = lock
Traceback (most recent call last):
File """", line 1, in
File ""/usr/lib/pypy/lib-python/2.7/weakref.py"", line 108, in __setitem__
self.data[key] = KeyedRef(value, self._remove, key)
File ""/usr/lib/pypy/lib-python/2.7/weakref.py"", line 274, in __new__
self = ref.__new__(type, ob, callback)
TypeError: cannot create weak reference to 'thread.lock' object
```
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,299346082
https://github.com/pydata/xarray/issues/1935#issuecomment-367802779,https://api.github.com/repos/pydata/xarray/issues/1935,367802779,MDEyOklzc3VlQ29tbWVudDM2NzgwMjc3OQ==,306380,2018-02-22T19:58:55Z,2018-02-22T19:58:55Z,MEMBER,+1 on reporting upstream if convenient,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,299346082
https://github.com/pydata/xarray/issues/1935#issuecomment-367795719,https://api.github.com/repos/pydata/xarray/issues/1935,367795719,MDEyOklzc3VlQ29tbWVudDM2Nzc5NTcxOQ==,1217238,2018-02-22T19:37:51Z,2018-02-22T19:37:51Z,MEMBER,"It's exciting to see people testing xarray with PyPy! Once we get this working, we should consider adding a continuous integration test.
The specific problem here seems to be `dask.SerializableLock`, which xarray initializes upon import:
https://github.com/dask/dask/blob/de6c2a49d76066abb51085570816322f063fc5c5/dask/utils.py#L737
I think this could be reproduced with something like the following using only the standard library:
```
import weakref
import threading
lock = threading.Lock()
locks = weakref.WeakValueDictionary()
locks['foo'] = lock
```
This works with CPython, but maybe not with pypy.
CC @mrocklin in case he has any ideas. Otherwise if my code to reproduce works, I would suggest reporting this upstream to PyPy.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,299346082