Rob Ede

I think it's important to note that this is a degenerate case, and as such may actually break some expectations of other consumers. When we say "an empty gzip", we...

My personal view is that it's a bug in Deno that lead to this request, it shouldn't be trying to decompress something it knows is empty.

I was kinda waiting for someone on the Deno team to comment on that issue. It's nicer if this is a conversation rather than just saying "not our problem" immediately....

drafting til we hear from Deno team on that issue, i'm subscribed to it so will see updates when they happen

Initial investigation reveals that reliance on `mime` crate is causing this. [Reduced playground](https://play.rust-lang.org/?version=stable&mode=debug&edition=2021&gist=69126039df984a90c193a7a5afc885c4)

FWIW the most recent docs point this out: https://docs.rs/actix-web/4/actix_web/dev/struct.ConnectionInfo.html#method.realip_remote_addr

Worth mentioning we do have a relevant test: https://github.com/actix/actix-web/blob/f4ee39cdd707ab4973ec4fe8650e6c78d09f6008/actix-web/src/route.rs#L359-L395 The only difference to OP's code sample is the ordering of `.wrap` and `.to`.

The warning is more about specifically the auto reload use case where I found signal handling to be more difficult to integrate correctly with the systemfd tool than desirable for...

`.next()` is from the `StreamExt` trait, but you're right that the import of it isn't shown on the README

hey @Andrew15-5, sorry 😬 i made a mess trying to omit the readme changes from your PR, carrying on and collating the work from various contributions here: #69 i cherry...