Justus Magin

I now think that while raising would be the cleanest way, it would be inconsistent with the other `numpy` functions dealing with truthiness (for `np.all` and `np.any` see #965). To...

the `DimensionalityError` raised for `(-270 * ureg.m) % 360` is a bit surprising to me: dividing a quantity by a dimensionless quantity works, so modulo (`%`) and floor division (`//`)...

You're right, `dir(obj)` returns a sorted list because that makes look-ups faster. Thus, the only way I can see to actually do this is by removing the units from `__dir__`,...

> `TypeError: object of type 'int' has no len()` Note that you can work around this by setting `ureg.force_ndarray_like = True` or `ureg.force_ndarray = True`

I'd also recommend using `autosummary` to document the API: it creates overview tables that are much easier to read than the output of `autodoc` directives

ping, @jules-ch. Is there anything I can do to help finishing this? Otherwise, should we release now and include the new documentation in the next release? There's a few bug...

@jules-ch, any updates on this? I think this was close enough, and the only thing left to do is to fix the linter job.

> I'll try to look into it soon. sure, no pressure. FYI the merge conflicts seem to be pretty small (I'd think the only one caused directly by the facets...

From a quick glance, this looks good to me. I'd probably fix the the footer and merge afterwards: this change is big enough already, so I think anything further would...

not really, besides pushing pydata/duck-array-discussion#3 forward. For now, the code for `fully_qualified_name` could be something like: ```python def fully_qualified_name(obj): t = type(obj) module = t.__module__ name = t.__qualname__ if module...