Re: Rack::Request#params EOFError

[Eric Wong <e@80x24.org>]

> John Smart <smartj@gmail.com> wrote:
> > When multipart > 112KB, I noticed that Unicorn tees the input stream
> > to a temporary file.  I was wondering: might Unicorn::TeeInput raise
> > an EOFError as part of normal operation when reaching the end of the
> > input stream?  If so, this would break the Rack spec.  I only tested
> > this on Darwin, still working on a self-contained repro.

Ping on this.

Eric Wong <e@80x24.org> wrote:
> No, it should not raise EOFError unless a client sent less than
> the Content-Length it declared in the header, or if it sent a
> short chunk with "Transfer-Encoding: chunked".
> 
> EOFError should be raised to break out of the application
> processing entirely if and only if the client decides to disconnect
> prematurely.  This is needed to allow unicorn to move onto other
> clients.
> 
> What unicorn could (and maybe should) do is raise a different
> error which is not a subclass of EOFError; to prevent the error
> from being caught by Rack (or any other middlewares).

I'm still considering this, but I'm also wondering if it'll
break any existing code which relies on Unicorn::ClientShutdown
being a subclass of EOFError

> What client are you using?
> 
> Is it sending "Transfer-Encoding: chunked" or a Content-Length?
> 
> Is nginx in front of unicorn?  If not, does it happen when nginx
> is in front of unicorn?