Mailing-List: contact dev-help@cocoon.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cocoon.apache.org
Received-SPF: neutral (hermes.apache.org: local policy)
Message-ID: <41B9BA00.1070708@apache.org>
Date: Fri, 10 Dec 2004 16:00:16 +0100
From: Reinhard Poetz <reinhard@apache.org>
User-Agent: Mozilla Thunderbird 0.9 (Windows/20041103)
MIME-Version: 1.0
To: dev@cocoon.apache.org
Subject: Re: Continuation manager modes
References: <20041208114713.31864.qmail@minotaur.apache.org>
 <41B6ED90.3000501@mobilebox.pl> <41B9A2B0.3080403@reverycodes.com>
 <41B9AB33.30109@mobilebox.pl> <41B9AD49.4090802@apache.org>
 <41B9B49A.1050507@mobilebox.pl> <41B9B847.2090106@apache.org>
 <41B9B919.4070004@mobilebox.pl>
In-Reply-To: <41B9B919.4070004@mobilebox.pl>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Leszek Gawron wrote:
> Reinhard Poetz wrote:
> 
>> Leszek Gawron wrote:
>>
>>> Reinhard Poetz wrote:
>>>
>>>> Leszek Gawron wrote:
>>>>
>>>>> Vadim Gritsenko wrote:
>>>>>
>>>>>> Leszek Gawron wrote:
>>>>>>
>>>>>>>
>>>>>>> Previously we have discussed about three continuations manager 
>>>>>>> work modes:
>>>>>>>
>>>>>>> - standard (current functionality)
>>>>>>> - continuations invalidated along with session, still the 
>>>>>>> continuation
>>>>>>> is reachable from other sessions (or no session at all)
>>>>>>> - fully isolated. only the session that created the continuation can
>>>>>>> access it.
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> IIUC before you introduced your changes it was possible to reuse 
>>>> continuations independently from where they have been created. 
>>>> What's the usecase for this so that we still have have to support it?
>>>
>>>
>>>
>>> Hmm after 2nd reading of your post I see I did not understand you.
>>>
>>> There are two orthogonal aspects of continuation visibility:
>>> - interpreter aspect: continuation should always be resumed by the same
>>>   interpreter that created it. If not you could invoke your continuation
>>>   in other sitemap (wrong context, resource not found exceptions,
>>>   security problems).
>>>   This case has been fixed. Still you can enable the old behaviur
>>>   because some users relied on that functionality (although broken).
>>>
>>> - security aspect:
>>>   - OLD MODE: you can make your continuations visible for everyone. One
>>>     user creates a continuation and passes the link to another user. The
>>>     other one invokes it in a browser - it works. This is just as it has
>>>     been from the start.
>>>   - NEW MODE: secure continuations.
>>>     Above behaviour creates following problems for authenticated web
>>>     applications:
>>>     * continuation ids might be stored in browser link history or page
>>>       cache.
>>>     * even though user has logged out and the session has been
>>>       invalidated the continuation might still be valid. As long as
>>>       resuming continuation does not query data from user session it
>>>       will work. This way you can have access to secured part of
>>>       application without even logging in.
>>>     So the following mode has been introduced:
>>>     * continuations are bound to the session.
>>>     * You can lookup the continuation only among the ones you have
>>>       created yourself. This way even though you "steal" a continuation
>>>       id from somewhere it's no use for you.
>>>     * When the session gets invalidated all continuations get
>>>       invalidated too.
>>>
>>> Hope that clears the situation.
>>
>>
>>
>> Thanks for the summary. The only point I still don't understand is: 
>> What's the usecase to resume a continuation in a different sitemap? 
>> What did people try to solve this way? (I'm asking because it sounds 
>> like a bug and not like a feature that we have to maintain.)
>>
> Let's ask the user himself.

Do you remember who is it?
Does she/he monitor cocoon-dev?

-- 
Reinhard