Remove (or standardize) the callback-based RTCPeerConnection#getStats() |
||
Issue descriptionThis is not in the spec, it has only the promise-based getStats(): https://w3c.github.io/webrtc-pc/#dom-rtcpeerconnection-getstats() Everyone would like for the callback-based variant to go away, and this is a tracking bug for that. In addition to getStats() itself, related bits of IDL are the callback interface RTCStatsCallback and the interfaces RTCStatsResponse and RTCLegacyStatsReport. The callback-based variant is not terribly interoperable, but if removal should be impossible due to compat constraints, then standardization of some sort should still be on the table. But that's plan B.
,
Apr 7 2017
This should not be standardized since it returns a big blob of non-standard stats. The standardization effort should go into migrating useful stats from the callback-based to the promise-based. The callback-based should be removed, but it will take time for products to migrate to the new stats. This bug will be stale for some time.
,
Apr 7 2017
Thanks! Can you link to bugs which must be resolved before this becomes actionable?
,
Apr 7 2017
When usage is low enough. The better the new getStats gets the more likely people will switch, root bug for new getStats: https://bugs.chromium.org/p/webrtc/issues/detail?id=7060
,
Mar 3 2018
,
Mar 5 2018
In order to get anywhere on this we need insensitive to switch. We should ban adding new stats to the old signature and formalize the stats process such that it is a viable option. This is on my TODO list but there seem to always be higher priority items in the way.
,
Mar 5 2018
A viable option to experiment with the new API and to go through the standards process without getting blocked.* |
||
►
Sign in to add a comment |
||
Comment 1 by guidou@chromium.org
, Apr 6 2017Status: Assigned (was: Untriaged)