VLC 2.0.5 RTSP problem

[digitalzy]

Hi all,

I am making a Java-based RTSP proxy program, which supports RTP-over-UDP forwarding. The original streaming server is live555. My proxy program works well with VLC 2.0.2, but doesn't work with VLC newer than 2.0.2. From my debug information I can tell that it seems that the newer versions of VLC doesn't recognize the SETUP response message from my proxy program. I'm attaching the debug information as follows:

Code: Select all

RTSP/1.0 200 OK CSeq: 4 Date: Tue, Jan 22 2013 23:53:17 GMT Transport: RTP/AVP;unicast;destination=127.0.0.1;source=127.0.0.1;client_port=56470-56471;server_port=7970-7971;ssrc=1A23C5E4 Session: DE9EDB74

Both my proxy program and VLC get stuck at this step.

The weird thing is that my program works properly with VLC 2.0.2. So, there must be some changes on the support of RTSP, but I cannot figure out what the changes are and what goes wrong.

Anyone has ideas to fix it?

Thanks so much and apologize for my English.

[Rémi Denis-Courmont]

Recompile your VLC version with the same live555 library version as the one that works...

[digitalzy]

Recompile your VLC version with the same live555 library version as the one that works...

Thanks for your response!

Actually I can make it work, but I'm wondering why my program doesn't work with new VLC releases. It's good to know that my program is tested under VLC 2.0.2, so I think the RTSP messages from my program are legal and understandable for old releases. After changing to VLC newer than 2.0.2, although all RTSP messages remain the same, it doesn't work any more.

I checked the VLC debug message, but I couldn't figure out what goes wrong inside, because the debug message is not detailed enough and I cannot see the interaction between the VLC and the streaming server.

[Rémi Denis-Courmont]

It's probably a bug in the embedded live555 version.

[digitalzy]

Problem solved!

Finally I tried openRTSP, which is a tool in the source code of live555. I found it's due to a redundant ending CRLF in the DESCRIBE response. It's quite interesting because this redundant ending CRLF in the previous DESCRIBE response affects the SETUP response.

Anyway

Thanks!