env["rack.io"] + WebSockets notes

env["rack.io"] + WebSockets notes

[Eric Wong]

Eric Wong <normalperson@yhbt.net> wrote:
(in <20091216221427.GA21033@dcvr.yhbt.net>)
> With a env["rack.io"] that exposes an IO-ish object, then nothing
> prevents developers from using EM.attach or Rev::IO.new on it, either.
> If people like that, I'd be happy to include it EM and Rev-based
> concurrency models for Rainbows! and add support for it in Sunshowers.

Of course the IO object was already attached to the event loop by the
server processing the request.  I suppose we could just stick the
EM::Connection or Rev::IO-based object in env["rack.io"] and then
have the Rack application do:

  env["rack.io"].extend(MyWebSocketApp)

-- 
Eric Wong

Re: env["rack.io"] + WebSockets notes

[James Tucker]

On 16 Dec 2009, at 22:58, Eric Wong wrote:

> Eric Wong <normalperson@yhbt.net> wrote:
> (in <20091216221427.GA21033@dcvr.yhbt.net>)
>> With a env["rack.io"] that exposes an IO-ish object, then nothing
>> prevents developers from using EM.attach or Rev::IO.new on it, either.
>> If people like that, I'd be happy to include it EM and Rev-based
>> concurrency models for Rainbows! and add support for it in Sunshowers.
> 
> Of course the IO object was already attached to the event loop by the
> server processing the request.  I suppose we could just stick the
> EM::Connection or Rev::IO-based object in env["rack.io"] and then
> have the Rack application do:
> 
>  env["rack.io"].extend(MyWebSocketApp)

Bye bye method cache.

> 
> -- 
> Eric Wong