You are viewing a plain text version of this content. The canonical link for it is here.
Posted to modperl@perl.apache.org by Eric Lenio <mo...@lenio.net> on 2006/05/18 18:15:45 UTC
Re: POST form hanging at new CGI
Enno, you might want to try applying this patch:
http://issues.apache.org/bugzilla/show_bug.cgi?id=37145
I had a problem similar to yours and this fixed it.
Eric.
On Thu, May 18, 2006 at 06:02:22PM +0200, Enno wrote:
>
>
> On Thu, 18 May 2006, Thomas Hilbig wrote:
>
> >
> >
> > --- Enno <bu...@xs4all.nl> wrote:
> >
> > > Let me guess, they are using apache 2.0.55 as
> > > proxy...
> > > mod_proxy in 2.0.55 has problems with properly
> > > delivering chunked
> > > encoding.
> > > The CGI.pm error sounds very familiar to the one I
> > > had with apache 2.0.55
> > > proxy.
> > >
> > > 2.0.54 and 2.2.2 dont have the chunked encoding
> > > prob.
> > >
> > > Enno
> > >
> >
> > Just so I'm clear on this, you think the customers
> > that are having the problem are using a 2.0.55-based
> > proxy and my end is simply waiting for the rest of the
> > data? I am assuming the one client is using a
> > Symantec proxy because of the HTTP_VIA env variable --
> > but I don't know if that product is based on apache or
> > not.
> >
> > HTTP_VIA="1.0 WEBSENSE01SA, 1.0 Symantec_Web_Security
> > (3.0.0.52)"
>
> hmm, I wouldnt go so far to say symantec is selling apache
> as a proxy... but it is possible that the symantec proxy has
> the same bug...
>
> also, please note that I located that bug in systems with reverse
> 2.0.55 proxies, using ssl on both ends.
>
> > Is there a third-party URL that I can ask my (big
> > company) users to test against to prove to them it is
> > not at my (small company) end?
>
> any form where you can post a large amount of kbs (eg 30kb)
> of data will do just fine. (or even upload a 100kb file using a form)
>
> > Would I see the problem at my end if I used something
> > like tcpdump to capture the packets? What am I
> > looking for?
> >
> > Tom
>
> I'd guess so, I never had to go through the troubles as I
> was able to locate the problem quite quickly without it.
>
> Enno
>