Open NeoLegends opened 8 years ago
This is definitely under review. I can't promise that tokio-rs will be supported, but it is definitely a goal.
I would also like to register my interest here...
I actually went yesterday through all the code to understand it(and try to port it to futures
) but it was to much for a single evening to understand how it works.
Next weekend I'll get back to it...
Thank you! I'm interested in your interest.
Futures (as in performing the handshake and stuff as part of obtaining the Sender
) would also be nice. :)
@lilianmoraru: any progress?
@lilianmoraru: any progress (it's ok if you didn't have time)?
@brunoqc I started working on a new interesting project at work, so I completely left all of this and started doing Qt-related stuff in my free time: https://www.youtube.com/watch?v=WIQy0oI6Ijc&t=16s. I don't like that this issue turned into a chat, so I'll remove the previous comments...
TL;DR: For tokio support in the near-term, check out tungstenite.
With the advent of tungstenite, I believe that people looking for WebSockets with tokio now have a good option. My plan for WS-RS is to focus first on better integration with raw Mio, so I would recommend tungstenite for people who need or would prefer tokio support today. I did some basic testing of tungstenite, and I am very pleased with it. I ran the little WS-RS bench tool on it and it's currently slightly faster than WS-RS! (test for yourself if you care about the details). Anyway, I believe that that means that tungstenite may be even faster than what the bench tool shows because the bench tool can never really run faster than WS-RS client. It's exciting.
@housleyjk they don't, tungstenite still does not support permessage-deflate
As part of supporting mio 0.6.0, how about integrating
tokio-rs
? The ability to provide an external event loop to the library would make composing and using it together with other asynchronous code much easier.