Magnus Ulimoen
Magnus Ulimoen
The usage of `unsafe` in this crate is minimal, which is great for memory safety. We should not have problems with this, as long as all indices are in bounds...
@rwl That seems like it could fit very well with this crate!
Are you looking for a python wrapper around this (#200) or something different? The crate is generic on storage type so you should be able to regenerate the matrix on...
A `TryInto` impl? A direct implementation seems out of scope for this crate, as we have to pull in so many dependencies. This should be as simple as calling [new_unchecked](https://docs.rs/sprs/0.10.0/sprs/struct.CsMatBase.html#method.new_unchecked)...
I think we use a combination of `Default` and `Zero` mixed around the codebase, but they serve two different purposes. `Default` should be exposed when we need to build a...
Either we have to require `Mul` for `MulAcc`, or we can require `Zero`. You can safely add `Zero` as a trait bound for `MulAcc`. I think places we have `Default`,...
In short: `F` must support addition, and the error message is poor. Longer: To support addition for arbitrary types, we must restrict `F` by applying the appropriate trait bounds. Which...
Oh, I see my `cargo add` used `0.9` of `sprs`. That does not look good, I'll check again to see what is going on here
My next attempt involves adding this bound: ```rust for
Paging @vbarrielle, this was introduced in d2f0da76. Any ideas in why type-checking would go haywire here?