httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roy T. Fielding" <field...@liege.ICS.UCI.EDU>
Subject Re: Netscape and Keepalives
Date Thu, 15 Aug 1996 19:41:35 GMT
>> Is this one symptom of the keepalive problem?
> This is the symptom of a different keepalive problem with Netscape. Are
> these pages very small (under 1024 bytes) and are you on a fast local
> network? If so, you may be running into the problem where if Navigator
> reads the entire body of the entity in the same read as the headers, it
> won't recognize keepalive status. We added the fflush() to stop this from
> happening, but on a very fast network, it still can sometimes.

Yep, I get it all the time now with my test server -- it affects the
last in-lined map on any page (in my case, our powered-by gif on the
initial server root page).  I also observed it on that IBM olympics
site that was running 1.0.5, which seems weird.  This is with
Netscape 2.02 on X/Solaris 2.5.

I think we should disable keep-alive for any Mozilla/2 -- with separate
patches for 1.1.1 and 1.2-dev (the latter using browsermatch).


View raw message