Pradeep Garigipati
Pradeep Garigipati
> I think the clippy errors are inherited from the current HEAD. I took care of those, thanks for letting me know.
> @9prady9 I think this is good to go. I made additional changes for f16 for other functions but I was having a problem with the negate function. I think...
> Are the traits just implementation detail ? I don't think they are just implementation detail. Generic Array trait bounds help with validating the input type statically at compile time....
I was about to add a modified example of the code stub from your original description to show why I think such usage of traits: `Convertable` and `ImplicitPromote` for `Array`...
https://github.com/arrayfire/arrayfire-rust/pull/265#issuecomment-739799995
~I think the recent change of `Convertible` has an unexpected issue with different kind of function. May be it should be reverted but I am not sure yet. Will post...
Leaving this as a place holder issue to improve generic programming using rust wrapper. Any future user, kindly showcase your use case if facing issues. We will try to improve...
@BA8F0D39 I think I didn't refactor the scalar argument(when Array was refactored to be generic) to avoid an up-cast at library level from `T` to `f64`. Given that ArrayFire C-API...
@BA8F0D39 I was able to reproduce the issue. Shall update here once I have more info on the root cause.
Found the problem, the `u64` happens to be quite large that it isn't accurately represented using `double`/`f64`. If I reduce the constant by a thousand or so, then the value...