Ben Plommer
Ben Plommer
Good question! I remember needing #5229 to get type inference working satisfactorily, but I don't remember whether covariant messages without covariant encoders actually made things worse or just didn't gain...
Yikes, this calls for IntelliJ's magic merge tool 🪄
This triggered a compiler bug with scala 3.0.2 but seems to be ok with 3.1.0 😅
Is the principle of this controversial or is it just the implementation that needs scrutiny? I can break it into smaller PRs if that helps.
Sorry for taking so long to get to this! It's purely down to my own laziness - I'll definitely give this and your other PR a proper review this side...
> The case-sensitivity of the parameters is the toughest case, if some, but not all, parameters are case sensitive. That's a significant modeling challenge. If case-insensitive is getting a breaking...
I don't like the idea of automatically encoding input strings. There's a *lot* of places where the spec mandates a `token` but http4s uses plain `String`. Would it be too...
Looking at the specs again, I *think* this should actually be `Header.Single` rather than `Header.Recurring`. [rfc7230 s.3.2.2](https://datatracker.ietf.org/doc/html/rfc7230#section-3.2.2) says > A sender MUST NOT generate multiple header fields with the same...
Now if only I'd thought to compile the tests
Apparently I messed something up when running mima locally, because this is super binary incompatible. (It could still go in 1.0 though, if bincompat can't be salvaged.)