Re: [webrtc-stats] Adding minor clarification regarding stats object lifetime.

I worry this comes at the cost of getting an accurate snapshot of the 
current state. We're mixing models IMHO. A "snapshot" to me means just
 that, a snapshot of the *current state*. If every object that ever 
lived hangs around, then that's not a snapshot.

This pushes the problem of reading current state correctly onto the 
caller: Are people supposed to check timestamps on every object to 
deduce which objects are real and which are ghosts? Live rtcp 
timestamps may beat only every ~5 seconds, so getting an accurate 
snapshot can get tricky.

Do tracks, transports ever go away except in response to JS? If not, 
then it seems simple enough for JS to call `getStats` before 
`removeTrack` if it wants to collect everything that happened.

GitHub Notification of comment by jan-ivar
Please view or discuss this issue at 
using your GitHub account

Received on Wednesday, 11 January 2017 16:10:08 UTC