Aurelia Molzer

Generally I'm more comfortable with having a _type_ here instead of needing to tag each usage site with an attribute. Nevertheless, a wrapper with the attribute is probably also possible....

When returning a Rust value whose type has `#[wasm_bindgen]` annotation on it, where is that value stored? As far as I can tell the representation on Js side ends up...

This is in reference to the section here: https://github.com/HeroicKatora/oxide-auth/blob/303e85d46f4c6220731c22733b1406492f469f71/oxide-auth/src/primitives/registrar.rs#L786-L800, correct? The issue is that the argument is never utilized as the return value. This is a bug introduced with the...

Coincidentally the fix to this logic might also unblock the addition of wildcard registrations.

The main effort should be expanding the test suite such that the new match is considered. The fix proposed seems already promising and implementable as is, changing the line (https://github.com/HeroicKatora/oxide-auth/blob/acf956bb98225bbd469f585ab4c7e9f9cb24034d/oxide-auth/src/primitives/registrar.rs#L797)...

> Clippy complaining about Result since it doesn't want a unit type in the position of the error. Doesn't really matter, can be muted. The largest reason this isn't some...

> Missing default implementations for types with an fn new() without parameters: Is this desired? This might be a matter of accepting whether it should be thought of as a...

> Large size difference between variants: This applies to both errors and the internal state machines. This one actually makes sense. We shouldn't really have the state machines and errors...

On a general note: I'd actually rather replace the actix bindings with fully asynchronous ones rather than have two separate packages for it, if that sounds appropriate.

The dispatch need not be byte-compatible with previous version, imo, since in-memory data should be treated as ephemeral. Well, any new design could be validated by trying to swap the...