Ian Rose

It looks to me like you have covered most of the things I can think of @jacobtomlinson. @ggosiang can you make sure that the `jupyter_server_proxy` serverextension is installed an enabled?...

Looks like you may need to change `jupyterlab-nvdashboard` to `jupyterlab_nvdashboard`. That is to say, the hyphen does not make a valid python import, it needs to be an underscore. You...

Hmm, can you verify that it is, in fact, importable?

No, I'm going to wait for the dust to settle a bit with 0.35 and then do it.

> There have been MANY issues with server implementation, since the "standards" are unclear on every point. Yep, I have also run into issues around gzipping with specifically Esri/ArcGIS servers...

I should also note that when there is a failure to download a block for the cache, it can leave that cache in a broken state, with partially downloaded content...

Hmm, I still see the same issue on master (disabling blocking or setting the headers as above are still functional workarounds though).

> Maybe this only works with the multi-context: > > ``` > with fsspec.open_files(path, "rb", **storage_options) as files: > data = files[0].read() > ``` Yep, I can confirm that the...

No updates from my end, but have you tried the above workarounds for `block_size: 0` or the http headers @albertvillanova ?

The specific problem we were seeing is with the `associated_trips` field. Since it is `null` much of the time, Pandas seems to be defaulting to nullable-`double`, which is not correct,...