@lgrahl note that with the same mechanism (reducing the `a_rwnd` when the data is sent to the js app and not when passed from the sctp stack to the dc stack) would work when using your WATHWG api proposal and you would not need any extra per-datachannel buffer size back pressure mechanism. You will decrease the `a_rwnd` when the `read` operation is done, so the only way of one datachannel draining the queue and stalling it is that the js app does not read from the datachannel (on purpose or due to a bug) . If any case, stalling is the correct behavior there IMO. -- GitHub Notification of comment by murillo128 Please view or discuss this issue at https://github.com/w3c/webrtc-pc/issues/2086#issuecomment-458147084 using your GitHub accountReceived on Monday, 28 January 2019 14:14:45 UTC
This archive was generated by hypermail 2.4.0 : Saturday, 6 May 2023 21:19:46 UTC