I finally found the problem. I believe the Oracle HTTP Server/OC4J combination sends a bad HTTP Content-Location Header:
marco@linux:~> wget -S http://sles.sonic.net:7777/oratest --22:41:57-- http://sles.sonic.net:7777/oratest => `oratest' Resolving sles.sonic.net... 192.168.1.107 Connecting to sles.sonic.net|192.168.1.107|:7777... connected. HTTP request sent, awaiting response... HTTP/1.1 301 Moved Permanently Date: Sat, 28 Jan 2006 06:42:03 GMT Server: Oracle-Application-Server-10g/10.1.2.0.2 Oracle-HTTP-Server Cache-Control: private Location: http://sles.sonic.net:7777/oratest/ Connection: close Content-Type: application/octet-stream Location: http://sles.sonic.net:7777/oratest/ [following] --22:41:57-- http://sles.sonic.net:7777/oratest/ => `index.html' Connecting to sles.sonic.net|192.168.1.107|:7777... connected. HTTP request sent, awaiting response... HTTP/1.1 200 OK Date: Sat, 28 Jan 2006 06:42:03 GMT Server: Oracle-Application-Server-10g/10.1.2.0.2 Oracle-HTTP-Server Content-Location: http://sles.sonic.net:12501/oratest/Page1.jsp Set-Cookie: JSESSIONID=7f00000230d5675863965c6b4979a0ebc9ce3af82d9c.e34Ma3uPbhqMe34RbxaPbNaRc310n6jAmljGr5XDqQLvpAe; path=/oratest Cache-Control: private Content-Language: en-US Connection: close Content-Type: text/html; charset=ISO-8859-1 Length: unspecified [text/html] [ ] 2,016 --.--K/s 22:41:57 (73.95 MB/s) - `index.html' saved [2016]
See the last Content-Location above, that’s the communication port between the HTTP server and OC4J:-(
And KDE‘s Konqueror believed that Content-Location when it tried to request more pieces and finally for the POST on the button press:-(
Mozilla/FireFox/IE seem not to care too much about that error. They work.
That solves the last mystery with the OC4J;-)
Have fun,
— Marco
Update Sat Jan 28 12:36:29 PST 2006:
See KDE bugs 51185 and 82747. Apparently Konqueror did the right thing but too many servers don’t send the right information in that HTTP Header element.
So now I only have to wait for a released version of KDE 3.5.
Thanks,
— Marco