Mailing-List: contact tomcat-dev-help@jakarta.apache.org; run by ezmlm
Precedence: bulk
Reply-To: tomcat-dev@jakarta.apache.org
Sender: bojan@binarix.com
Message-ID: <3B80E328.E767DC1B@binarix.com>
Date: Mon, 20 Aug 2001 20:15:04 +1000
From: Bojan Smojver <bojan@binarix.com>
Organization: Binarix Corporation Pty Ltd
MIME-Version: 1.0
To: tomcat-dev@jakarta.apache.org
Subject: Re: mod_jk 1.2.0 (CVS) + TC 3.3 B1
References: <D84B43C5F94DD511877400609457497C01C58A@css4.cs>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

GOMEZ Henri wrote:
> 
> Hi to all,
> 
> I'm back from hollidays :(

Welcome back! Hope you had a good one...

> >After finally getting over the 'graceful restart' issues...
> 
> Could you developp it more please ?

This is related to the previous thread "Problem with mod_jk 1.2.0
(latest CVS snapshot)" where mod_jk would pass NULL or incorrect
pointers to some functions after graceful restart of Apache. You'll find
a few 'gdb where' dumps in the thread. I'm not certain is this problem
has actually been fixed or did it just 'accidentally' become something
else...

> >Is this for some reason a 'forbidden' combination? I've observed some
> >really strange stuff going on with this combo - parameters don't get
> >passed correctly, even session stuff tends to be screwed (ie. same
> >session data appears in two totally different browsers).
> >
> >Is this something that's known not to work or did I just screw things
> >up? Things work just fine when mod_jk shipped with TC 3.3 B1
> >is used. Is
> >mod_jk 1.2.0 only to be used with the latest TC 3.3 CVS snapshot?
> 
> mod_jk in JTC is an evolution of mod_jk in CVS. Some refactoring
> was done to share code between ajp13 and ajp14.
>
> >If this is not a known thing, I can send some more data...
> 
> Yes, please, strace could help in that case.

OK. I'll run up a few examples and send you the strace.

Bojan