[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kfm-devel
Subject:    Re: [PATCH] disable keep alive connections when using SSL
From:       "Dawit A." <adawit () kde ! org>
Date:       2002-03-31 18:43:09
[Download RAW message or body]

On Sunday 31 March 2002 03:23, Matthias Welwarsky wrote:
> Waldo Bastian wrote:
> > I think the better solution would be to make sure that non-idempotent
> > methods (basically everything but GET) should start a new connection.
> > (Patch attached)
>
> I've attached a modified patch that also closes the connection before a GET
> if the request carries a query.

No, this will immediately make the idea of persistent connections almost 
useless in my opnion.  How many websites today are dynamically generated ?  
Many.  Read RFC 2616 section 9.1.1 paragraph 3.  The user cannot be 
responsible for actions (s)he did not deliberatly and knowingly take.  I mean 
if you click on a link that says go to order page and you end up with six 
packages of something without inputting any data, you cannot be reponsible 
for it.

BTW, only POST is non-idempotent.  The meaning of idempotent is given in 
section 9.1.2:

"A sequence is idempotent if a single execution of the entire sequence always 
yields a result that is not changed by a reexecution of all, or part, of that 
sequence."

So repeating a DELETE or a PUT or any of the other method does not result in a 
different behavior if that request is executed once or X number of times.  
The contradiction in that section where it states:

"However, it is possible that a sequence of several requests is non-
idempotent, even if all of the methods executed in that sequence are
idempotent"

are only applicable to pipelining over persistent connections IMHO since it 
assume HTTP/1.1 protocol.  Otherwise, it does not make sense how using the 
same persistent link for multiple GET actions could result in a 
non-idempotent condition...

Regards,
Dawit A.

[prev in list] [next in list] [prev in thread] [next in thread]