This may be a long shot but the circumstances sound familiar. I had a problem where an HTTP connection was simply not being properly closed and Netscape4 was just hanging around waiting for the response to finish whereas other browsers/platforms had no problem, ie. the other browsers were auto-cancelling the delayed response. Connecting to port 80 over telnet also clearly showed the connection was simply not closing and I modified the java proxy server source in an attempt to hammer blow the connection. Turns out that the JRE needed upgrading on the proxy and the I/O worked properly. My point is it may well be nothing at all to do with Apache/mod_perl. I had been trying to debug the PHP engine on the server but telnetting from the proxy to the server showed the connection was closing correctly so it had to be the proxy's fault.
as I said, this may be a long shot. chris On Thursday 18 Jul 2002 3:21 pm, you wrote: > On 7/10/02 10:37 AM, "David Wheeler" <[EMAIL PROTECTED]> claimed: > > I have a couple of strange bugs showing up in Bricolage that have > > something to do with Apache->print timing out (with Win 95 and Mac OS 9 > > -- yes, these folks need to join the new millennium), and have no idea > > how or under what circumstances print times out. > > Okay, I've got the output of some tcpdump runs in Bricolage Bugzilla. > > > http://bricolage-bugzilla.thepirtgroup.com/showattachment.cgi?attach_id=35 > > All they demonstrate, however, is that Apache/mod_perl never finishes > sending the page to the browser. One page is 32120 bytes, and the other is > 36500 bytes. I really have no clue what could be causing this. Why should > Apache->print ever time out? I didn't even know that timing out was > something that a print method could do. And why does it only happen for > some browser/platform combinations? > > Anyone who has some ideas, or who can suggest other approaches to > diagnosing the problem, will get a great deal of gratitude from me and many > Bricolage users. > > Thanks, > > David -------------------------------------------------------
