W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 1996

Re: Issues Beyond HTTP/1.1 ???

From: Alex Lopez-Ortiz <alopez-o@barrow.uwaterloo.ca>
Date: Sat, 23 Mar 1996 14:09:04 -0500 (EST)
Message-Id: <199603231909.OAA04614@barrow.uwaterloo.ca>
To: Shel Kaphan <sjk@amazon.com>
Cc: alopez-o@barrow.uwaterloo.ca, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
X-Mailing-List: <http-wg@cuckoo.hpl.hp.com> archive/latest/58
>Alex Lopez-Ortiz writes:
> > 
> > Issue 1:
> > 
> > When a user stops a transmission, the CGI's generated do not get 
> > killed. Am I correct to say that this is so because the HTTP protocol
> > does not contemplate an "transmission interrupted" signal?
> > 
> > Is there any knoen solution known to this problem that does not
> > involve changes to the protocol?
> > 
>Your CGI program ought to get a SIGPIPE signal when the server
>instance goes away, which will happen when the TCP connection goes away.

First off, thanks to all who replied. SIGPIPE works. 

Are there any other situations (with persitent connections and the like) in 
which we might need to explicitly inform the server of the interruption?
Would that be just a waste of bandwith? Comment? Opinions?


Alex Lopez-Ortiz                                       alexlo@opentext.com
Research & Development                               Open Text Corporation
http://daisy.uwaterloo.ca/~alopez-o                     (519)-888-7111x319
Received on Saturday, 23 March 1996 11:14:21 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 2 February 2023 18:42:58 UTC