changelog

[martindurant]

No description provided.

[d-v-b]

can you explain why the fsspec dep is fsspec==2025.5.0, rather than fsspec>=2025.5.0?

I think the == requirement makes the upstream fsspec incompatible with the upstream s3fs according to pip's solver. see https://github.com/zarr-developers/zarr-python/actions/runs/15218451870/job/42826570416?pr=3087.

[martindurant]

@d-v-b , see #966 ; only zarr cares about this?

For reference, distributed depends on an exact pin of dask ( https://github.com/dask/distributed/blob/main/pyproject.toml#L31 ).

[d-v-b]

@d-v-b , see #966 ; only zarr cares about this?

For reference, distributed depends on an exact pin of dask ( https://github.com/dask/distributed/blob/main/pyproject.toml#L31 ).

Anyone who attempts to install

    'fsspec @ git+https://github.com/fsspec/filesystem_spec',
    's3fs @ git+https://github.com/fsspec/s3fs',

in the same environment with a solver that cares about a consistent set of dependencies will care about this.