"call --videoinput video.rtpdump" won't work when receiver has a camera

[GoogleCodeExporter]

What steps will reproduce the problem?
1. build libjingle and call example using VS 2010
2. run "call.exe --voideinput voice.rtpdump --videoinput video.rtpdump" 
3. vcall xxx@gmail.com

What is the expected output? What do you see instead?
When xxx@gmail.com has a camera connected on his computer, the call cannot be 
established. I get following messages:
   calling...
   call destroyed
i.e. the session is closed immediately and xxx@gmail.com don't even see a video 
call invitation.
However if xxx@gamil.com disconnect his camera from the computer, and call can 
be successfully established and he can see the video and hear the man's voice 
"1,2,3,4,5".

What version of the product are you using? On what operating system?
libjingle-0.6.6
VS 2010
Win 7

Please provide any additional information below.
The VS 2010 solution can be downloaded from Issue 242. Some source files need 
to be updated to version 0.6.6.

Original issue reported on code.google.com by tangketa...@gmail.com on 24 Dec 2011 at 2:18

• Merged into: #242

[GoogleCodeExporter]

BTW, another problem is, "call.exe --videoinput video.rtpdump" won't work. 
--videoinput and --voiceinput have to be used together, and then use command 
"vcall". Only in this way the receiver can both hear the voice and see the 
video, if his computer don't have a camera.
Is this a bug or a special design?

Original comment by tangketa...@gmail.com on 24 Dec 2011 at 2:34

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

I'm not able to repro this issue when building the project with Scons env. This 
must be something to do with the VS specific building env.
We'll look into this when started working on VS specific issues. Centralize 
this issue to #242

Original comment by jun...@google.com on 4 Jan 2012 at 5:18

• Changed state: Duplicate
• Added labels: ****
• Removed labels: ****