Mathieu Amiot

Seems it works now 🤔 I'll merge it for v0.2.0

Would be super cool indeed! We are using requests timeouts in prod, but since nitox is wrapped by `actix-nats` and we use the timeouts provided by `actix`, the leaf future...

Maybe we could simply enable lock fairness as described here: https://docs.rs/lock_api/0.1.5/lock_api/struct.RwLockWriteGuard.html#method.unlock_fair I think it's worth trying when the lockguard is held for too long.

Also, I want to apologize for the relative inactivity, we are currently super swamped on R&D here and can't afford extra time to maintain code that actually already works super...

I've looked a bit into it and: - The usages of `RwLock` in `connection.rs` are not a problem since the non-blocking `try_{read/write}` versions are used - The other usages of...

@thedodd Hey! I'll get back to it soon, but feel free to write tests, I'd get them merged against this branch. If I can make a recommendation on how to...

No problem! Yeah I know I skipped on this feature, I went down the opinionated route to make stuff easier to implement but seems your use case was not covered....

Reporting @thedodd's comment on the todo here for information; > * resolves the current (as of 2019.01.16) merge conflict in `feature/nats-server`. > * fixes a bug where streaming client IDs...

I'd *really* like to have tests ready before merging it. Even if the feature is opt-in, it would be super cool to be confident that it works 100%.

I just added the support for the verbose mode! It works by holding off future resolution until a `+OK` message comes down the line. Next step will be writing an...