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/6728#issuecomment-1167548711,https://api.github.com/repos/pydata/xarray/issues/6728,1167548711,IC_kwDOAMm_X85Fl2En,2448579,2022-06-27T16:10:20Z,2022-06-27T16:10:20Z,MEMBER,Thanks @kmuehlbauer ,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1285001323
https://github.com/pydata/xarray/issues/6728#issuecomment-1167405546,https://api.github.com/repos/pydata/xarray/issues/6728,1167405546,IC_kwDOAMm_X85FlTHq,352673,2022-06-27T14:11:46Z,2022-06-27T14:11:46Z,NONE,"> Possibly related issue: #5706

Oops, looks like my issue is a dupe. Sorry! I looked but I somehow missed that one.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1285001323
https://github.com/pydata/xarray/issues/6728#issuecomment-1167398217,https://api.github.com/repos/pydata/xarray/issues/6728,1167398217,IC_kwDOAMm_X85FlRVJ,5821660,2022-06-27T14:06:27Z,2022-06-27T14:06:27Z,MEMBER,Possibly related issue: https://github.com/pydata/xarray/issues/5706,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1285001323
https://github.com/pydata/xarray/issues/6728#issuecomment-1167344851,https://api.github.com/repos/pydata/xarray/issues/6728,1167344851,IC_kwDOAMm_X85FlETT,352673,2022-06-27T13:20:14Z,2022-06-27T13:20:14Z,NONE,"> Can you reproduce this without xarray? It should probably be reported upstream? (In the netCDF4 repo?)

Yeah, it can be triggered with:

```python

from netCDF4 import Dataset
import h5py
import numpy as np
import operator

filename = ""foobar.h5""
with h5py.File(filename, ""w"") as fp:
    fp[""test""] = np.array([""foo"", ""bar"", ""baz""], dtype=""|S3"")

rootgrp = Dataset(""foobar.h5"", ""r"", format=""NETCDF4"")
# This works
print(rootgrp[""/test""][slice(None)])

# This segfaults
operator.getitem(rootgrp[""/test""], slice(None))
```","{""total_count"": 1, ""+1"": 1, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1285001323
https://github.com/pydata/xarray/issues/6728#issuecomment-1167339640,https://api.github.com/repos/pydata/xarray/issues/6728,1167339640,IC_kwDOAMm_X85FlDB4,14808389,2022-06-27T13:15:25Z,2022-06-27T13:15:25Z,MEMBER,"the segfault should definitely be reported upstream ([Unidata/netcdf4-python](https://github.com/Unidata/netcdf4-python), I think), but we can also try to avoid hitting this or the `UnicodeDecodeError`: I can successfully read the dataset with `xarray=2022.03.0` and `netcdf4=1.5.8`.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1285001323
https://github.com/pydata/xarray/issues/6728#issuecomment-1167320520,https://api.github.com/repos/pydata/xarray/issues/6728,1167320520,IC_kwDOAMm_X85Fk-XI,10194086,2022-06-27T12:57:52Z,2022-06-27T12:57:52Z,MEMBER,Can you reproduce this without xarray? It should probably be reported upstream? (In the netCDF4 repo?),"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1285001323
https://github.com/pydata/xarray/issues/6728#issuecomment-1166664112,https://api.github.com/repos/pydata/xarray/issues/6728,1166664112,IC_kwDOAMm_X85FieGw,352673,2022-06-26T22:29:11Z,2022-06-26T23:33:03Z,NONE,"I can confirm that if I downgrade netCDF4 from 1.6.0 to 1.5.8 I get `UnicodeDecodeError`

Edit1: I checked that if I install netCDF4-1.6.0 from git, it works (i.e. I get the `UnicodeDecodeError`). Only with `pip install netCDF4==1.6.0` I'm able to trigger the segmentation fault.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1285001323
https://github.com/pydata/xarray/issues/6728#issuecomment-1166635630,https://api.github.com/repos/pydata/xarray/issues/6728,1166635630,IC_kwDOAMm_X85FiXJu,14808389,2022-06-26T20:22:01Z,2022-06-26T20:49:46Z,MEMBER,"Can you check if that's due to the new version of `netcdf4`? I can't seem to install that from `conda-forge`, and if I execute your MCVE using `engine=""netcdf4""` and `netCDF4==1.5.8` I get a `UnicodeDecodeError` (which is not perfect but definitely better than a segfault).

Edit: I can successfully reproduce if I use `netcdf4` and `h5py` from PyPI, which I guess means that the combination of the new versions and the change you mentioned triggers the segfault
Edit2: this is curious, all my new environments get the segfault but my old development environment raises a `UnicodeDecodeError`","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1285001323