Move rectilinear chunk grid from experimental to stable

[maxrjones]

Describe the new feature you'd like

The rectilinear chunks feature was gated behind a configuration flag for the following reasons from #3802:

This gives downstream libraries time to adapt before the API is finalized, and us an opportunity to gracefully finalize the API.

We have not received any complaints about the design in the month since it was released. Some libraries (e.g., xarray) have been slower to adopt the feature than others (e.g., cubed), and I'm curious if removing the experimental gating would help promote further adoption (cc xarray + zarr-python devs @dcherian @jhamman).

Shall we move the feature from experimental to stable by removing the feature gating?

[d-v-b]

works for me!

[TomNicholas]

Can we wait until VirtualiZarr has adopted it? How long has this been out for?

[maxrjones]

Can we wait until VirtualiZarr has adopted it?

Yes, though I'd like to see us moving towards VirtualiZarr adoption if that's going to be a motivating factor here. Would you be able to help with getting pydata/xarray#11279 reviewed and released so that we can address the missing feature in VirtualiZarr? It's been challenging to get into xarray due to unrelated CI failures.

How long has this been out for?

Since 3.2.0 https://github.com/zarr-developers/zarr-python/releases/tag/v3.2.0 (released April 30, 2026).