Re: Why Thin handler uses "Chunked" Trasfer-Encoding?

["Iñaki Baz Castillo" <ibc@aliax.net>]

El Domingo, 6 de Diciembre de 2009, Ryan Tomayko escribió:
> On Sat, Dec 5, 2009 at 4:40 PM, Iñaki Baz Castillo <ibc@aliax.net> wrote:
> > Hi, Thin adapter looks like:
> >
> > -------------------------------------
> > require "thin"
> > require "rack/content_length"
> > require "rack/chunked"
> >
> > module Rack
> >  module Handler
> >    class Thin
> >      def self.run(app, options={})
> >        app = Rack::Chunked.new(Rack::ContentLength.new(app))
> >        server = ::Thin::Server.new(options[:Host] || '0.0.0.0',
> >                                    options[:Port] || 8080,
> >                                    app)
> >        yield server if block_given?
> >        server.start
> >      end
> >    end
> >  end
> > end
> > -----------------------------
> >
> > Why to force "Chunked"?
> 
> That only applies chunked encoding if no Content-Length response
> header exists at that point in the middleware chain. I believe all
> Rack adapters that don't support chunked encoding internally follow
> the same pattern.

Thanks for the explanation.

-- 
Iñaki Baz Castillo <ibc@aliax.net>