Tom Nicholas

> it can get tricky to essentially reimplement many parts of xarray and then also maintain them with upstream xarray changes. I completely agree, which is why I've been looking...

Note that the way we have been doing this so far is good in that we haven't committed any large files to git, so we don't have to do any...

> On this and https://github.com/zarr-developers/VirtualiZarr/pull/235, should we consider moving the tests outside the source code directory or explicitly excluding the data files from the release manifest? I don't think there...

Thanks for raising. I think the problem is that the scheme `'abfs://'` is not in [`VALID_URI_PREFIXES`](https://github.com/zarr-developers/VirtualiZarr/blob/a8e0ac7a74704c4d84dc204a55474683f4a96c89/virtualizarr/manifests/manifest.py#L12), so `open_virtual_dataset` is assuming that it's a local path. This is obviously bad behaviour,...

I think you're right @rsignell . In fact I'm not sure we should have `VALID_URL_PREFIXES` at all.

Isn't the solution to this error to pass `authorize_virtual_chunk_access={prefix: credentials}` when opening the icechunk repo?

Yes good plan. Our parser design means there is nothing to enforce compliance across 3rd-party parsers, but we should at least try to be compliant for the parsers we ship...

>One practical workaround for a lot of cases would be to load the encoded variables into memory (i.e. "inline" them) See also https://github.com/TomNicholas/VirtualiZarr/pull/73#issuecomment-2042915164, which implies that kerchunk / fsspec doesn't...

> Concating data with different encoding becomes identical to concating data with different codecs. Is this supported by VirtualiZarr or not? It requires a different abstraction than the one currently...

> Thinking forward to https://github.com/TomNicholas/VirtualiZarr/issues/5 should we consider including encoding information at the `ChunkEntry` level rather than the array level? If we put encoding in the `ChunkEntry`s then serialization of...