Jan Wassenberg

Nice, thanks for making the updates! I like your NDArray idea. CI has noticed an issue, looks like #include is required in aligned_allocator. > I could do it but I...

The tests have uncovered some issues: > AlignedVector result(n); Only [u]intXX_t are supported types for lanes. Technically size_t could be a different type, and certainly differs between platforms and is...

> I also tried to integrate random in bazel, from command line crypto test does not build huh, the Bazel change looked like it should work, though we are overwriting...

Nice, u64 change looks good. No worries :) Don't worry about the Bazel issue with :unroller, this is unrelated and we'll take care of it. The remaining issue is that...

The 'unroller' issue/comment above is still causing internal CI to fail :)

Thanks for updating :) Yikes, a long series of comments because we are seeing issues one by one as the CI raises them. `#include ` must actually be `#include "hwy/*"`,...

Oh, now it looks like libstdc++ is shooting us down :) ``` uniform_real_distribution.h:102:17: error: static assertion failed due to requirement '__libcpp_random_is_valid_urng::value': 102 | static_assert(__libcpp_random_is_valid_urng::value, ""); ``` Probably we're missing one...

Interesting, I did not know it is no longer required. Today's harvest: 😿 1) wasm is throwing an exception, perhaps in an infinite loop hence stack overrun: ``` [ RUN...

Nice, thanks for updating :) The current issue we are seeing is the extraneous ; after {} on line 136, which actually fails the build due to our -Werror.

Hi @oshaboy , it's an interesting topic. I personally do not know Rust, Go and Ada very well. What seems very feasible is to have C bindings for larger groups...