Timothy Hodson

It's a bit hacky but have a look at my workaround: https://github.com/hytest-feedstocks/ssebop-feedstock Note the extra zip sugar in the url pattern.

I apologize, my organization is swapping SSL certs and that seems to be preventing me from running the unit test, but this PR is straightforward.

Changelog updated. And I meant that the SSL was preventing me from running tests locally. I don't fully understand why, though I could see this going awry if Julia, pip,...

Well, this is awkward. `xb.get_bitinformation` calls itself such the warning is thrown at least twice (+1 per additional dim). Need to ponder the simplest way to avoid that.

Eh, I don't particularly like my solution, which just loosens the unit test to permit the multiple warnings. The alternative is to import some other module that let's us inspect...

Ah, maybe aspects of this are resolved by #234

Those are good insights. I'll look for lower fruit. Regarding my initial test. I altered the threshold to 1.1 based on performance with my dataset. Later I realized that the...

Exactly right, of course. I'll proceed with `custom_function` for now and see where we get. Along that tack, a possible outcome might be to implement such a function in `xbitinfo`,...

https://github.com/zarr-developers/numcodecs/pull/503#issuecomment-1919916377 Unfortunately, I'm not sure this will work. I'm by no means the expert, but as I look through the respective code bases, a reimplementation might be the simplest option,...

...true, this might reduce direct traffic to `xbitinfo`, but I also think it's a valuable step into mainstream adoption. ...or `numcodecs` might accept `xbitinfo` as an optional dependency, allowing us...