It would be great to have various test cases for zarr v3 datasets. Those can be used for testing by different implementations, and also to illustrate the spec more concretely. Both positive (valid, according to the spec) and negative (invalid) examples might be helpful IMO. Probably it's useful to wait for a more finalized version of the spec, which is expected at the end of this month (Feb '23).
It would be great to have various test cases for zarr v3 datasets. Those can be used for testing by different implementations, and also to illustrate the spec more concretely. Both positive (valid, according to the spec) and negative (invalid) examples might be helpful IMO. Probably it's useful to wait for a more finalized version of the spec, which is expected at the end of this month (Feb '23).