Justus Magin

another option would be to manually raise [here](https://github.com/HypothesisWorks/hypothesis/blob/8005910d4160778c08abb974f41b52b00e19324b/hypothesis-python/src/hypothesis/extra/ghostwriter.py#L458) if the signature detected by `inspect.signature` is `*args, **kwargs` and the tricks below are likely to work.

I ran into this, as well. It appears this is only triggered by certain repositories: in my case I need to fetch at least 17 pages from [xarray](https://github.com/pydata/xarray) – and...

well, me neither, which is why I wasn't able to investigate any further than that.

should this have been closed by #740? I'm not seeing this issue anymore.

the difference here is that `npcompat.isdtype` translates the string to a `numpy.dtype` superclass, then uses `isinstance` to perform the check, while `np.isdtype` explicitly raises if it receives anything other than...

for reference, the reason this error is raised is because `numpy._core._type_aliases.allTypes` contains a explicit list of allowed dtypes, so any new dtypes that are not in that list – like...

> Can you open a NumPy issue about it? See numpy/numpy#27545 > In the meantime, would it make sense to simply continue falling back into the xarray implementation `npcompat.isdtype` As...

this actually looks like a bug in `virtualizarr`. The problematic part is this: https://github.com/zarr-developers/VirtualiZarr/blob/43209b32b8d84f7f57f91942c5b8d3b6cf2db0f4/virtualizarr/manifests/array.py#L241-L247 where the error message says that `None` would be fine as a `tuple` element but the...

I believe this may be intentional (I may be wrong, though): it is often not so useful to reduce the coordinates with the same operation as the data, and so...

just as a note, I think this should be done using a combination of `set_coords` / `reset_coords` and the `agg` proposed in #9479 (if we can figure out a well...