Brian Ward

> Would it make sense to return a tuple of arrays rather than an array of tuples I think the issue is that both of these things (tuples of arrays...

Structured dtypes don't really help this as far as I can tell: ```python >>> import numpy as np >>> x = np.array([(1.2, 3.4), (4.5, 6.5)], dtype='f,f') >>> x[:,1] Traceback (most...

Ah, there an option using structured dtypes: ```python >>> import numpy as np >>> x = np.array([(1.2, 3.4), (4.5, 6.5)], dtype='f,f') >>> x.dtype dtype([('f0', '

This was resolved by https://github.com/stan-dev/stan/pull/3179

A combination of https://github.com/stan-dev/cmdstan/pull/1106 and https://github.com/stan-dev/stan/pull/3179 have allowed this

https://github.com/stan-dev/cmdstan/pull/1107

I intend to take a pass at this via annotations sometime soon

These tests were re-written in https://github.com/stan-dev/cmdstan/pull/1017 and this issue was fixed as part of that

I don’t believe we have made any intentional change, but it’s possible OS updates have improved the M1’s parity with x86 in ways that helped this

We have been using these newer RTools for a bit now, so I think we can close this issue