Magnus Ulimoen
Magnus Ulimoen
As a temporary measure we could backport the `ndarray` bump to `0.9`.
@vbarrielle I've added a new branch `0.9.4` branching off the `0.9.3` tag. This bumps `ndarray` for `sprs` and `sprs-ldl`.
Thanks @vbarrielle, `0.9.4` should be uploaded now.
I think the original compilation error is gone with the following extra bounds: ```rust fn test(val: F) -> CsMat where F: Float, F: Default + Send + Sync + num_traits::MulAdd,...
@bytesnake This does not seem related to the original issue. I've opened #292 to address this issue edit: This is indeed the same problem
I also see that my ["fix"](https://github.com/vbarrielle/sprs/issues/278#issuecomment-821520964) did not actually work as I replaced addition by multiplication
@StefanUlbrich How should this fix be integrated into `sprs`?
The `hadamard` product is rather simple, but requires left and rigth to have the same sparsity structure. In which case you can iterate over `vec.data()` of lhs and rhs. For...
The `MulAcc` trait was added to avoid a `Copy`, and not for accumulating into another type, but this would be a very interesting extension. As far as I can tell,...
> `T: MulAdd` This trait bound looks right, although care must be taken so it does not involve a `Copy`. At the current state, it should be possible to use...