David Ramos

I'm leaning toward accepting this change at the same time as #237 since they both require a major version bump. After looking closely at `http::Request` and `http::Response`, I agree that...

One other breaking change I think we should make in the same release is to define `AsyncHttpClient` and `SyncHttpClient` traits instead of [using `FnOnce`](https://github.com/ramosbugs/oauth2-rs/blob/ed6312689ce533c48a9ae58a9fd5f9e271fe5d37/src/lib.rs#L1309), which someone previously found confusing while...

I've started merging breaking changes into `main` for the future 5.0 major release, so now would be a good time to start making these changes

This is now released in [5.0.0-alpha.1](https://crates.io/crates/oauth2/5.0.0-alpha.1). If you can, please try it out and let me know if you find any issues that should be fixed before stabilizing the 5.0...

> I _assume_ that any stringified `url::Url` should be parsed successfully as a `http::Uri`, and we can just return `oauth2::RequestTokenError::Other` in the unlikely event that `http::Uri` parsing fails. Turns out...

JWTs add significant complexity that I would like to avoid in this crate. See earlier [comment](https://github.com/ramosbugs/oauth2-rs/pull/217#issuecomment-1555397088). Just as I implemented OpenID Connect in a [separate crate](https://github.com/ramosbugs/openidconnect-rs) on top of this...

This crate's MSRV goes back even further than `openidconnect` to Rust 1.45, which I guess is incompatible with this version of `base64`'s use of the 2021 edition. Whenever version 5.0...

> Hey, just checking in the progress around this crate, would the [db0ea44](https://github.com/ramosbugs/oauth2-rs/commit/db0ea44657bd6c1130b83ce135ac2691ba091fad) commit replace this PR as it's already upgraded in main now? Yes, both `base64` 0.21 and 0.22...

Very reasonable question! Most of the methods in this crate that accept `String` need to store the string (in `self`) beyond the lifetime of the method call. That requires storing...

Looks like the current API is in line with the official [Rust API Guidelines](https://rust-lang.github.io/api-guidelines/flexibility.html#caller-decides-where-to-copy-and-place-data-c-caller-control): > If a function requires ownership of an argument, it should take ownership of the argument...