- From: Henrik Frystyk Nielsen <frystyk@w3.org>
- Date: Thu, 03 Apr 1997 14:41:51 -0500
- To: Lukas Faulstich <faulstic@inf.fu-berlin.de> (by way of Henrik Frystyk Nielsen <frystyk@w3.org>), www-lib-bugs@w3.org
At 02:32 PM 4/3/97 -0500, Lukas Faulstich wrote: >when trying to load web pages from www.taz.de, w3c get's into a endless >loop which prevents it from retrieving the HTML source. Netscape doesn't >have any problems with this server. An explanation could be the >expiration time of these pages (see listing below). Do you know a >work-around for this problem? Thanks in advance - The problem was that the response was a redirection message with no content-length and hence relied on the connection to be closed in order to terminate the response. If Apache by default closes the connection on a redirect on the same server then this is a mistake which slows the system down by an order of magnitude. You should then notify this to the Apache folks. Libwww saw the closed socket as a sign of a bad response and tries to reopen the connection to repeat the operation. I have a fix ready but need to do some testing before I can release it together with some other fixes. PS: You should update to the latest Apache beta release as this has more HTTP/1.1 performance features than beta 6. Henrik -- Henrik Frystyk Nielsen, <frystyk@w3.org> World Wide Web Consortium, MIT/LCS NE43-346 545 Technology Square, Cambridge MA 02139, USA
Received on Thursday, 3 April 1997 14:41:41 UTC