+Chunk specifications are judged only by the chunk normalizer: the separate duck-typed check for rectilinear input is gone, so the v2 and sharding restrictions apply to what the normalized grid says was declared. numpy values are accepted everywhere a chunk or shard shape is: 0-d arrays unwrap to their scalar, and the legacy `zarr.create(chunks=np.array(...), zarr_format=2)` no longer fails on the truth value of an array. A non-integer scalar chunk specification raises a `TypeError` naming the problem.
0 commit comments