venugopalpaikr wrote:
Hi,
I am modifying Live555 to respond to Windows Media Player. While trying
to establish the session WMP sends SET_PARAMETER. The actual response to
SET_PARAMETER should be
RTSP/1.0 200 OK
Content-Type: application/x-rtsp-udp-packetpair;charset=UTF-8
Content-Length: 43
Date: Mon, 01 Dec 2008 05:28:27 GMT
CSeq: 3
Session: 13114845032829158031;timeout=60
Server: WMServer/9.5.6001.18000
type: high-entropy-packetpair variable-size*
The end of message should be at d asterisk sign as shown above.
Do you mean it shouldn't have a CRLF on the end? That seems mighty
fussy of the client!
I have written the response for SET_PARAMETER as shown :
void RTSPServer::RTSPClientSession::Call_SET_PARAMETER(char const* cseq) {
unsigned char ResponseBuffer[10000];
snprintf((char*)ResponseBuffer,sizeof ResponseBuffer,"RTSP/1.0 200
OK\r\nContent-Type:
application/x-rtsp-udp-packetpair;charset=UTF-8\r\n%sCSeq: %s\r\nSession:
%d;timeout=60\r\nServer: WMServer/9.5.6001.18000\r\n\r\n",dateHeader(),cseq,
fOurSessionId);
This "\r\n\r\n" is correct - the blank line splits the header section
from the response body.
memset(fResponseBuffer,0,RTSP_BUFFER_SIZE);
send(fClientSocket,(char const*)ResponseBuffer,
strlen((char*)ResponseBuffer), 0);
snprintf((char*)ResponseBuffer,sizeof ResponseBuffer,"type:
high-entropy-packetpair variable-size\r\n");
(*) You're adding a CRLF here.
send(fClientSocket, (char const*)ResponseBuffer,
strlen((char*)ResponseBuffer), 0);
memset(fResponseBuffer,0,RTSP_BUFFER_SIZE);
Not sure why you are resetting this buffer twice (or at all, actually!).
}
the response is as shown below
RTSP/1.0 200 OK
Content-Type: application/x-rtsp-udp-packetpair;charset=UTF-8
Date: Fri, Jan 02 1970 22:53:22 GMT
CSeq: 3
Session: 1;timeout=60
Server: WMServer/9.5.6001.18000
type: high-entropy-packetpair variable-size
*
the message ends at the asterisk sign which i don't want. Am not able to
replicate the response properly. How should i modify the code? if i don't
use "\r\n\r\n" the response gets appended to the Teardown response and is
sent along with it when a Teardown request is received.
You're explicitly sending the final CRLF at (*) above; do you think
you're getting another one as well? I can't see how not sending the
double CRLF earlier could affect what the server sends back, but I guess
WMP's response parser might get confused if it doesn't see the end of
headers - where are you tracing this output?
Regards
Paul
--
Paul Clark
Packet Ship Technologies Limited
http://www.packetship.com
_______________________________________________
live-devel mailing list
live-devel@lists.live555.com
http://lists.live555.com/mailman/listinfo/live-devel
