Antoine Pitrou

I've posted a [comment](https://github.com/OpenMathLib/OpenBLAS/issues/4361#issuecomment-2983471801) there but I'm not sure we should count on the issue being actually fixed in OpenBLAS. I'm not sure how deeply OpenBLAS is wired into NumPy,...

> I don't think this is a regression in Numpy 2.0 - we're seeing the same symptoms in 1.26.4: Hmm, perhaps OpenBLAS wasn't packaged by default before in manylinux wheels?

> I should have mentioned that we use the builds from conda-forge rather than the wheels. But I think in the example you showed, it also looks like it's happening...

Hi, can we perhaps find another term than "extension" for this feature? The reason is that we may also want to add support for extension types to Parquet (where "extension...

> For the extensions in the discussion, I do not have a lot of context: what's the advantage of such extensions vs adding the logical types to the spec? The...

> It would be beneficial to avoid confusion. I can't think of a great name for it though. If we name them "experiments" it will make them look risky and...

The reason to prefer "extension types" over "third-party types" or "external types" is that at some point some of them might get standardized inside Parquet, like [Arrow does](https://arrow.apache.org/docs/format/CanonicalExtensions.html). Though we...

> @pitrou wdyt about calling the extensions in this PR "Metadata Extensions" vs the other ones "Type Extensions"? Would that clear it? Perhaps "Binary protocol extensions" or "Unparsed protocol extensions"?...

Doesn't this require a vote before merging?

@github-actions crossbow submit -g cpp