Benjamin Saunders
Benjamin Saunders
Presently, all quinn futures are `Unpin`, allowing them to be moved around in memory even after they've been polled. This gives users maximum flexibility, but is rarely needed, though it...
[Required Operations on Connections](https://quicwg.org/base-drafts/draft-ietf-quic-transport.html#required-operations-on-connections): > When implementing the server role, applications need to be able to: > - if Early Data is supported, embed application-controlled data in the TLS resumption...
See [the draft](https://quicwg.org/base-drafts/draft-ietf-quic-transport.html#preferred-address) for details.
Reduces confusion over the correct way to reference a type, particularly due to documentation and error messages. A downside to this is that all the typenums show up on the...
Currently, `cast` is resolved via `Unit`'s `Deref` impl, which leads to it confusingly stripping the `Unit` from `UnitVector`s. Add an inherent impl which takes precedence, similar to the existing specialization...
32-bit seeds aren't great, and pervasive use of mostly-0 seeds may be causing odd results already. I propose that `impl Seedable for T` be replaced with `impl Distribution for Standard`...
Web stuff bloats the dependency graph for downstream crates who don't need it.
Processing bodies generically (for example, in a force generator) may requires access to body part specific information, such as center of mass. Presently there doesn't appear to be any way...
It would be useful to support bodies that have restitution and friction per-shape for bodies with compound shapes. ncollide's `TriMesh` already stores uv coordinates, so you could conceivably even have...