Re: PROPOSAL: i99 Pipelining Problems

Agreed; listing the bugs present in implementations isn't useful. What  
may be useful is noting that there do seem to be a high proportion of  
bugs in deployed implementations, and so clients using pipelining need  
to exercise a fair amount of caution.

So, here's a straw-man to add to p1 7.1.2.2:

"Some early implementations of HTTP/1.1 servers and proxies have been  
noted to implement pipelining incorrectly, and some commonly-deployed  
(if not spec-compliant) devices may interfere with its correct  
operation. Clients choosing to send pipelined requests on the open  
Internet should, as a result, do so cautiously."



On 05/04/2008, at 5:17 AM, Roy T. Fielding wrote:

>
> On Apr 4, 2008, at 6:00 AM, Julian Reschke wrote:
>>
>> Hmm.
>>
>> Any volunteers for providing spec-ready text?
>
> "Shit happens.  Deal with it."
>
> Seriously, there is no reason to specify all the possible ways
> in which messages might get lost on an unreliable connection.
> The only thing I would change is to resurrect my original design
> for the Keep-Alive header, which indicates how many more requests
> are allowed on a given persistent connection.  The others are just
> single point, non-reproducible bugs.
>
> ....Roy
>
>


--
Mark Nottingham     http://www.mnot.net/

Received on Monday, 7 April 2008 02:21:00 UTC