Jeff Whitaker

> You may want to consider rebuilding your wheels (for `manylinux` * `x86_64`, anyway). What specifically do you think needs to be done differently when building the manylinux wheels?

curl i s 7.75.0, openssl is 1.0.2u

When you are reading, `nc4_fh["observation"][0][:]` works because `nc4_fh["observation"][0]` returns a numpy array, which is then sliced using `[:]`. Writing is a different story. We can't support all the numpy slicing...

Note that ```python observation[0] = data_0 observation[1] = data_1 ``` does work as expected.

Turns out the the `[][]` slicing syntax doesn't get forwarded to `__setitem__` at all. Not sure why.

OK, here's what's happening with ```observation[0][:] = data_0``` `observation[0]` forwards to `__getitem__` and returns a fully masked array (since no data has yet been written to the variable). data_0 is...

Sorry, but I have no idea. Does it work if you don't import pandas, or if you import netCDF4 last?

Works here too.

Confirmed. The segfault occurs inside the C lib after calling `nc_get_vara` with start,count values of `[11 0 0]` and` [ 5 75 110]`. I will be out of town for...

Scratch that - the segfault is apparently not happening in `nc_get_vara`, it may actually be in the python code. It looks like it may have something to do with the...