Tom Nicholas

Q: Should this utility expose `max_chunk_lengths` or be changed to expose `max_num_chunks` instead?

For some reason the new strategy is not showing up on the [API docs page](https://dask--9374.org.readthedocs.build/en/9374/array-api.html#hypothesis-testing-strategies) 

`mypy` is failing with ``` dask/array/strategies.py:11: error: Untyped decorator makes function "block_lengths" untyped [misc] dask/array/strategies.py:45: error: Untyped decorator makes function "chunks" untyped [misc] Found 2 errors in 1 file (checked...

Thanks for your comments @jsignell ! > The mypy failure definitely feels like a hypothesis issue. I noticed there are recent changes to the code for that decorator in hypothesis...

Is it going to affect us in a way that is different to how it would affect a naked `dask.array` user? We don't have many custom HLGs in xarray, so...

> This isn't about HLGs. My thought was that a user might have a custom HLG whose performance could be affected by suddenly having different chunking structures than the previous...

I was thinking I want it to work periodically even in integer index space too - surely I need isel for that? I might be misunderstanding something fundamental here though.

Thank you for the clear explanation @benbovy! @keewis you were right, as ever, but now I understand why!) > There's a big difference now between isel and sel: So `Dataset.sel`...

That's very helpful, and should definitely go in the docs somewhere! On Wed, 14 Sep 2022, 11:31 Benoit Bovy, ***@***.***> wrote: > My understanding from reading the docs was that...

I had another go and now I have this (the `.sel` method is just copied from `PandasIndex.sel` with minor changes): ```python from xarray.core.indexes import ( PandasIndex, is_scalar, as_scalar, get_indexer_nd, IndexSelResult,...