Developers

[jedashford]

We've released our app with the new scene graph Video node instead of the roVideoScreen. We use HLS. With the roVideoScreen I was able to get the status code and url of every segment as it was downloading. It doesn't appear this is available on the new Video node. I can only get the current segment being played on the screen, but none of the networking details about segments being downloaded and buffered. 

The problem arises as we're getting customer complaints that the player is crashing. I've added external logging to catch the player error, but is devoid of any useful info:

Network error. This could be caused by any of the following problems: (1) The server is down or unresponsive. (2) The server is unreachable. (3) There is a network setup issue on the client.

If I could get more info on the network error mentioned above (url, http status code, etc), we can start to track down the problem. 

Anyone have any ideas?

[jedashford]

bump. Looking for info on how to view incoming (downloading) HLS segments.

[monkeysource]

Hi,

Did you discover anything further for your issue? I'm seeing the exact same error message and finding it hard to debug. 

From my research:

 - it only happens on older devices (Roku 1 for me)
 - On older devices, it happens every time if you set videoNode.control = "stop" before it was playing anything. I had this line in for a safety call to make sure the previous stream wasn't playing. Taking it out virtually resolves the issue.
 - i know only see it after playing, stopping about 10 streams in a row. 

As far as I can see it's a bug in the video node on older hardware, but Im still looking into a resolution for all devices. 

Thanks,

Ryan

[Techie]

Anyone was able to resolve this issue?

[monkeysource]

I was able to resolve our issues by creating a new Video node for each stream I played. 

So instead of adding the Video node to XML I create and add dynamically in code. Whenever the video's stream is changed, I remove the old Video node and create a new one. For now, this has fixed our issues with no noticeable performance hit.

Thanks,

Ryan