RE: Potentially serious error in unix se
dkearns{TCNET/HR/dkearns}@klaven.tci.com
Thu, 17 Nov 94 08:49:00 -0600
>From: RDENNY @ SMTP (Bob Denny) {[email protected]}
>Date: Wednesday, November 16, 1994 6:31PM
>
>>[...]
>>The problem comes in handling cases where a file request is aborted
>>by the client by the client dropping its end of the connection.
>>This case is sometimes erroneously logged as being successful even
>>when the disconnect is detected; however, for a large class of cases
>>(including he one I was actually interested in, demmit), the client
>>disconnection is never discovered by the server.
>
>Neither of these happen with the Windows Web server. It immediately
>detects a closure by the client. No entry is made in the access log
>unless the transaction completes, including successfully sending the
>server's response (whatever that is) back to the client.
>
I'm not sure that's a plus. If a NetScape user stops a transmission simply
to abort graphics load, but stll has received my text, I'd consider that
a 'completed' transaction for access log purposes....
-dave