Mailing-List: contact dev-help@cocoon.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cocoon.apache.org
Received-SPF: neutral (athena.apache.org: local policy)
Message-ID: <47A462F9.10503@apache.org>
Date: Sat, 02 Feb 2008 13:32:57 +0100
From: Grzegorz Kossakowski <gkossakowski@apache.org>
User-Agent: Thunderbird 2.0.0.9 (X11/20070801)
MIME-Version: 1.0
To: dev@cocoon.apache.org
Subject: Re: Run latest from trunk
References: <47A2C4B8.2070404@otego.com> <47A2D93D.7090204@apache.org>
 <47A366DC.9030205@apache.org> <47A385A4.4000703@apache.org>
In-Reply-To: <47A385A4.4000703@apache.org>
Content-Type: text/plain; charset=ISO-8859-15
Content-Transfer-Encoding: 7bit

Grzegorz Kossakowski pisze:
> Grzegorz Kossakowski pisze:
>> Felix Knecht pisze:
>> I can confirm now that I see this problem locally. By mistake, I was testing a copy without this bug
>> fix before committing so I didn't spot it.
>>
>> I'm just having a look at this.
>>
>> Sorry for breaking your work at the end of the week! :-(
> 
> I have looked at this and found the cause for this bug. Class
> StatusRetrievableBufferedWrappedResponse passes Content-Length setting to the wrapped response
> _before_ final settlement of this value which leads to situation where Content-Length is set to 0.
> 
> I tried to fix this but whole issue looks to be more troublesome that I previously thought and I
> need to rethink whole approach. Basically, I opt to writing our own, full-blown implementation of
> HttpServletResponse that will buffer *everything* until we can be 100% sure that everything is
> established in final shape.
> 
> This looks like quite a lot of work to implement meaningful implementation of HttpServletResponse
> that can copy to another one and will certainly increase complexity of SSF.
> 
> If you have any thoughts on this I'm willing to hear.

I regret to say that I probably won't be able to fix this until final release of SSF.

I reverted commit blowing up SSF but this does not solves the whole issue, of course. The original
bug is still there and it will bite people even in more obscure way because its real consequences
can be seen in complicated setups with servlet inheritance only. Nevertheless, it's still better
than completely broken SSF.

-- 
Grzegorz Kossakowski
Committer and PMC Member of Apache Cocoon
http://reflectingonthevicissitudes.wordpress.com/