srrrse

Hey, not to be a bother but I just wanted to mention that now that the auto linking to docs.rs has existed for a while, it has become very common...

`NewService` is a viable approach, but I'm not clear if its intended to be the approach. My impression was that `NewService` was a potentially expensive operation which would be called...

This definition of `Service` would behave how I would expect it to & compiles today, but it forces a dynamic dispatch at every middleware level (in practice, I fear that...

> To clarify some things, it is expected that the response future should not be bound to the service value itself. Some reasons for this would be to allow the...

I also made the connection to scoped threads. :-) > However, when we worked through all this in the past, the basic feeling was that the complexity wasn't worth it;...

I would also be surprised if its a huge performance hit. I think my feelings are largely puritanical rather than practical, and (until this becomes easier to implement) y'all have...

#13 changed `Service::call` to take `&mut self`. That has significant implications for this issue. There are multiple ways to combine two services, but the specific combination at issue is sequentially...

I'm still confused (even concerned) about this issue. Is it correct now to wrap services in a `Mutex` when you need to chain their responses? Is the potential "scoped threads...

@cramertj I think it may be more complicated than that; IIRC many of the executor APIs are not designed to run non-static futures, so this wouldn't necessarily be useful for...

I'm about to go on vacation for the rest of the week and then the next week is a Rust all hands, so I'm not sure when I'll be able...