> From owner-wrec@cs.utk.edu Mon Aug 2 15:44:16 1999
> Delivery-Date: Mon, 02 Aug 1999 15:44:16 -0700
> Return-Path: owner-wrec@cs.utk.edu
> Date: Mon, 2 Aug 1999 16:43:04 -0600
> From: Duane Wessels <wessels@ircache.net>
> To: Joe Touch <touch@ISI.EDU>
> cc: wrec@cs.utk.edu
> Subject: Re: oh the irony
> List-Unsubscribe: <mailto:wrec-request@cs.utk.edu?Subject=unsubscribe>
>
>
>
> On Mon, 2 Aug 1999, Joe Touch wrote:
>
> > Wouldn't that more correctly show up as an error on the socket,
> > rather than an EOF? I'm bothered by the notion that you have
> > to check length, if you correctly check for an EOF.
>
> Perhaps, yes. But socket errors are not normally propogated
> along by proxies.
>
> A proxy is reading bytes from an origin server. In the middle of the
> transfer it gets a read error on that socket. What should the proxy do
> to its connection with the client to "pass on" this error?
You mean it's not passed on? That'd be a proxy (or protocol) bug, IMO.
Joe
This archive was generated by hypermail 2b29 : Thu Nov 18 2004 - 11:21:26 MST