I know this is a broad and somewhat open-ended question, but this is being very difficult to figure out. When using RTMFP to stream video to VideoDisplays, what could cause the videos to occasionally not show anything when pretty close to everything seems normal with the objects and connections? The code involved is not really situated so that I can copy and paste it very well for an example.
Essentially though my VideoDisplays are showing video about 80% of the time that they successfully connect peer-to-peer through RTMFP, they're not showing video the other 20%. I've compared NetStream.info.toString()'s, checked on the number of children in the VideoDisplays, checked on the NearIDs / FarIDs, etc., and everything checks out.
One piece of solid, easy-to-describe information though: After hours of experimenting, about the only thing that started to come up was that when videos would fail (and it's always one side at a time by the way), the receiving NetStream object associated with the failing video would have a value of 0 in its info.playbackBytesPerSecond field. So I tried using a timed event to keep testing that and keep trying to recreate the NetStream and the video whenever it was set to 0. That worked for a short time, but then it proved to not be reliable. The last test I ran with that attempted hackjob-fix, when one of the sides failed to show video, the playbackBytesPerSecond was equal to 0 initially. But after that video and NetStream got restarted, not only was the video continuing to fail (which I know would have had a chance of happening), but despite its failure, the playbackBytesPerSecond suddenly jumped up to a normal value.
This project was previously configured to use RTMP, and reverting back to RTMP, this problem doesn't come up.
I know this may not be a pristine description of the problem. It's just that it's probably some sort of learning curb thing, and I'm having a very difficult time figuring it out myself. Again this is peer-to-peer RTMFP communication. Thanks!