Tom de Geus

Ah right, you had the computation in mind, not the allocation

I would happily review a PR. However, I don't have time to do it myself soon, sorry.

@gouarin I'm not a big expert here so I cannot estimate what it takes to fix this. A question though: Should be aim to include this in the next release.

The error makes sense. You are using the following overload https://github.com/xtensor-stack/xtensor/blob/308458574baa95b0dabb981a4bae97fa29ff673f/include/xtensor/xadapt.hpp#L161 In this case the `size` has to match up `prod(shape) = 4`. The error is telling you exactly that:...

P.S. The fastest way to getting this implemented is to open a PR ;)

Per [NumPy's behaviour](https://numpy.org/doc/stable/reference/generated/numpy.sort.html) `xt::sort` always sorts along an axis and by default along the last axis (see also [docs](https://xtensor.readthedocs.io/en/latest/api/xsort.html)). To sort the full object you indeed should ravel, sort, and...

Do you have a minimal example where we can debug from ?

I agree that it looks that you might be on to possibly two bugs. In order to debug, it would be very helpful to have the full error stream, or...

It sounds great! Indeed I see a mutual benefit: Flashlight can rely on a well-developed CPU backend with convenient NumPy-like syntax, and xtensor benefits from benchmarking which could lead to...

Thanks for the suggestions @NAThompson . I must say that I get a bit confused ;) `pip` belongs to Python. I suppose that it can install `xtensor`'s headers in Python's...