Re: New NTLM code - First feedback

From: Serassio Guido <guido.serassio@dont-contact.us>
Date: Tue, 10 Feb 2004 09:38:28 +0100

Hi Henrik,

At 01.46 10/02/2004, Henrik Nordstrom wrote:

>On Mon, 9 Feb 2004, Serassio Guido wrote:
>
> > - no more authentication pop-ups at all
>
>Excellent.
>
> > - the authentication process seems to be more fast
> > - the helper's usage seems lower
>
>Do you have challenge reuses enabled? Well, I suppose not as you said you
>had been running the patch for 3 days. Challenge reuses was broken until
>today...
>
>So you are comparing with challenge reuses disabled. Save from the bugs in
>the old code there is no speed improvements in Squid here. Only slowdowns
>if challenge reuses are enabled.

Native Windows helper doesn't support challenge reuse at all, so I can't
run with them enabled.

>I suspect you are using a broken MSIE version which caused the old code to
>feel much slower than needed.
>
>MSIE6 SP1 + latest security hotfix is broken or at least very confused
>about what it is doing with it's own connections, and they still do not
>get the connection management state correct. This version of MSIE quite
>often sends yet another NEGOTIATE packet after successful authentication
>which before the patch triggered a new NTLM handshake. The new code
>silently ignores these as the connection is already authenticated.

Correct, I forgot to say that my IE6 SP1 has the latest patches applied.
The resulting user perception is a more responsive Internet Explorer usage.

> > Henrik: a very beautiful work !!!, the best compliments :-)
>
>Thanks ;-)
>
>Also have patches fixing always/never_direct, cache_peer_access etc in the
>queue thanks to a user who contracted us for having these fixed, and while
>fixing these I stumpled over the problems fixed by the patch you have been
>testing.. Waiting for confirmation from the customer before commit.

Whit this other patches, I think that the Windows integration level of
Squid will be comparable to Microsoft ISA Server, but with more flexibility.

>There is still a few memory leaks to hunt down, and assertion failures on
>shutdown. But all of these did exists before the patch as well so I don't
>care that much about these yet.

Yes, I remember this old asserts at Squid shutdown.
It would be very nice to fix they before STABLE 5 release.

Regards

Guido

-
========================================================
Guido Serassio
Acme Consulting S.r.l.
Via Gorizia, 69 10136 - Torino - ITALY
Tel. : +39.011.3249426 Fax. : +39.011.3293665
Email: guido.serassio@acmeconsulting.it
WWW: http://www.acmeconsulting.it/
Received on Tue Feb 10 2004 - 01:38:29 MST

This archive was generated by hypermail pre-2.1.9 : Mon Mar 01 2004 - 12:00:04 MST