Bug 106454
Description
Pedro Bezunartea
2003-10-07 10:05:39 UTC
That in itself is not a bug - it may be that the applet is buggy. Can you capture a complete network trace between the applet and the server (and applet and proxy) with "tcpdump -w" and attach that? It may be you need a BrowserMatch to force an HTTP/1.0 response to the applet, try something like the existing: BrowserMatch "Java/1\.0" force-response-1.0 BrowserMatch "JDK/1\.0" force-response-1.0 Created attachment 94976 [details]
This is a tcpdump from the client with the browser
Created attachment 94977 [details]
tcpdump from the server to the camera
Created attachment 94978 [details]
tcpdump from the server to the camera
Both dumps were made with apache version 2.0.40-21.5.
Sorry, that needs to be "tcpdump -s0 -w" otherwise tcpdump only captures a few bytes of packet data. Can you redo? Created attachment 94987 [details]
Tcpdump of comm. from the client to the server
I added the line 'BrowserMatch "Java/1" force-response-1.0' first, and
'BrowserMatch "Java/1.4.2_01" force-response-1.0' later, but neither made a
difference.
Created attachment 94988 [details] Tcpdump of comm. from apache server to the camera. This file is related to the previous attachment #94987 [details]. Try "downgrade-1.0" in the BrowserMatch line as well. The traces seem to show the server acting correctly, but the Java client is closing the connection as soon as it receives the chunked response. Added the following line to httpd.conf: BrowserMatch "Java/1" force-response-1.0 downgrade-1.0 But the applet still doesn't work. Is there a way to make the ProxyPass work like the previous version? Can you attach the tcpdump between client and server in that configuration? Created attachment 95067 [details]
Tcpdump client-server with downgrade-1.0 directive
Created attachment 95068 [details]
Tcpdump server-camera with downgrade-1.0 directive
The bottom line is that the proxy is compliant with the HTTP specification, but the Java applet is not at all compliant. We'll look into adding a workaround for this to preserve the Content-Length header if possible; apparently the "Windows Update" client has the same, severe, problem. There are packages here which might fix this problem: http://people.redhat.com/jorton/9-httpd/ can you try these out? Tryed and worked just fine :) Thanks, Pedro. An errata has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2003-320.html |