Insert artificial delay into IMAP server responses (to workaround OL2002)

[Peter Friend]

On Apr 13, 2004, at 10:17 AM, Doug Koobs wrote:

> I should have said that Outlook is too slow. The problem here is 
> definitely
> not with Cyrus, I was just hoping for a workaround in Cyrus to 
> compensate
> for Outlook's problem.
>
> Here is a thread from a NG between a couple of people suffering from 
> this
> problem, and a tech at MS that is working to resolve the problem:

I wasn't able to view this link. At a protocol level the problem 
doesn't make much sense to me, since Outlook has to read everything 
that Cyrus is sending it, regardless of how fast it is being sent. The 
data would have to actually be ditched by Outlook , or there is some 
nasty thread collision problem that is causing the data to be lost and 
not processed. If this isn't causing any protocol errors it would seem 
that the data loss is specific to certain items. I have a lot of 
Outlook users talking to a Cyrus server (hacked extensively, but not at 
protocol level) so I am interested in what is actually happening here. 
Putting an artificial load on the system should not alter the order 
that Cyrus writes data. Perhaps this load is delaying writes enough to 
cause a larger number of smaller packets to be sent over the wire, 
which really shouldn't matter. I would be doing tcpdump comparisons at 
this point.

Regards,

Peter

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html