Release planning: v0.8.0

[jplatte]

There's a milestone here: https://github.com/tokio-rs/axum/milestone/7

I'd like to talk about what makes sense to be in there, and what doesn't, so we can hopefully ship a release soon-ish.

cc @mladedav @yanns @SabrinaJewson

[jplatte]

My personal take on the current open issues in the list:

• IntoResponse implementation for (StatusCode, R) where R: IntoResponse overrides error status codes and might be a footgun #2287 - nice to have, though I don't think I'm gonna pick it up myself. Does somebody else want to?
• Remove or change impl FromRequest* for Option<impl FromRequest*>? #2298 - nice to have, but maybe delay it to ship a bit faster? I worked on it a bit today.
• Consider disallowing root nesting #2651 - nice to have, I just pushed a PR for it that should not take long to review
• Consider changing the behavior of / route when nesting #2659 - nice to have, but also hard so unless somebody has a good idea soon I'd prefer to postpone
• Routing percent-encoded paths #2678 - nice to have, but probably not all that important?
• Make serve generic over the listener and IO types #2941 - would really like to have it, probably the most requested out of the bunch (though still a niche thing); would be nice if folks could chime in re. the latest review
• Spoofable extractors are used with the knowledge of the risks #2998 - 🤷🏼

[jplatte]

No need to apologize, I just wasn't entirely convinced by your arguments, but couldn't write a good counterargument either, and was generally quite busy. Your reviews are much appreciated!

[yanns]

About

* [`IntoResponse` implementation for `(StatusCode, R) where R: IntoResponse` overrides error status codes and might be a footgun #2287](https://github.com/tokio-rs/axum/issues/2287) - nice to have, though I don't think I'm gonna pick it up myself. Does somebody else want to?

I've looked into #2468.
I'm ready to spend more time on it if we first decide that this is the right solution.

[jplatte]

It's maybe a little bit of a "tacked-on" solution, but I still think it would be an improvement over the status quo. So I'm still in favor. Only reason it may be worth delaying is that 0.8.0 is already going to be a rather big release, so maybe it might be better to have subtle changes like that separated out in the upgrading process?

[yanns]

I'm ok for it. Should we remove #2287 from the 0.8 milestone then?

[jplatte]

If you don't care about it much, sure (edit: done).

[yanns]

About #2941, I took a look, but I don't feel qualified enough to assess the comments and the possible alternatives.

[jplatte]

Thanks for looking anyways. On second thought I did like mladedav's suggestions, so have implemented them (w/ slight adjustments).