On Wednesday 29 September 2004 19.23, John D wrote:[...]

Thanks for your efforts John.

I believe what you found is really a work-around rather than the problem. We 
are trying to track it down - it seems to be a subtle multi-threading bug, 
probably introduced by the changes we made to eliminate the delays in HTTPS 
responses. Please wait a bit longer - hopefully Monday - and we'll announce 
it officially.[...]

Hi Robert,

I am glad to hear you are tracking the real culprit! Since I hadn't heard of
any progress being made on these issues, I was offering my 2 cents.

This may not be important, but your message talks about HTTPS... We are (were)
seeing these issues in both HTTP and HTTPS. The problem you are referring to
may affect both, but I just wanted to make sure.

John D

********** Original Email *********
** To:   pound(at)apsis.ch
** From: Robert Segall <roseg(at)apsis.ch>
** Date: Fri, 1 Oct 2004 15:33:54 +0200
**********

On Wednesday 29 September 2004 19.23, John D wrote:[...]

Thanks for your efforts John.

I believe what you found is really a work-around rather than the problem. We 
are trying to track it down - it seems to be a subtle multi-threading bug, 
probably introduced by the changes we made to eliminate the delays in HTTPS 
responses. Please wait a bit longer - hopefully Monday - and we'll announce 
it officially.[...]

On Friday 01 October 2004 16.46, John D wrote:[...]

We are certainly grateful for the effort.
[...]

I may have expressed myself in a confusing manner. What I meant is that the 
problem was introduced by the fix for the HTTPS problem (as is so often the 
case: you fix one bug and introduce another). It clearly affects both HTTP 
and HTTPS.

Once we release the fix we'll ask you to test it with and without SSL, as well 
as making sure that the HTTPS delays are also fixed.[...]

> > Hi all,[...]
work [...]
have [...]
[...]
Off).[...]

Just for the record, I've tried this too, but they do not go away, they
just appear less often. Both e500's, blank pages, and garbled content
(possibly from another session).
[...]
let [...]

I don't believe that to be the case, as I experience this behavior with
releases as early as 1.4

Kind Regards,

John Hansen