Maybe it is helpful for everyone understanding to provide verbose info by
"-v" or trace log by "--trace" to show the all protocol handshaking
procedures?

Song Ma

2007/5/7, Dan Fandrich <dan_at_coneharvesters.com>:
>
> On Mon, May 07, 2007 at 09:11:14AM -0600, Doug McNutt wrote:
> > Assuming TCP-IP and not UDP does that convict the server of committing
> the error? Is there an observable difference between. "I answered you but
> the mail didn't get delivered" and "I'm sorry, I just never answered before
> I closed the connection"?
>
> A FIN from the server to close the connection won't get delivered before
> any intervening data. It's possible that the data from the server didn't
> get delivered (due to a network crashing somewhere), but if the connection
> then closes it would be due to that condition persisting.
>
> > Would some obscure TCP logs have something to look at? A retry count or
> something like it?
> >
> > Might network address translation have anything to do with it? A packet
> getting lost or mangled on the return path through a local NAT router?
>
> One lost/mangled packet shouldn't be enough to close the connection
> because
> TCP will retry lost packets.
>
> > Does an empty packet satisfy the HTTP requirement to respond?
>
> The server needs to respond with an error code.
>
> >>> Dan
> --
> http://www.MoveAnnouncer.com The web change of address
> service
> Let webmasters know that your web site has moved
>
Received on 2007-05-08