- From: Bernard Aboba via GitHub <sysbot+gh@w3.org>
- Date: Wed, 31 May 2017 16:32:46 +0000
- To: public-webrtc-logs@w3.org
aboba has just created a new issue for https://github.com/w3c/webrtc-pc: == Section 4.3.1 == >From EKR: https://lists.w3.org/Archives/Public/public-webrtc/2017May/0166.html (operations queue) This algorith seems to imply that operation is run synchronously if the queue depth is one (step 6) whereas if it's not, it's run asynchronously. This definitely isn't invisible in the callback version (reentrancy) and isn't actually invisible in the promises version because these operations have side effects. Is it really the expectation that we should (for instance) start sending media before the promise resolves? That seems odd. (setDescription) The says: To set an RTCSessionDescription description on an RTCPeerConnection object connection, enqueue the following steps: 1. Let p be a new promise. I may be confused, but I don't think that the promise creation and resolution is in fact queued. 2. If elements of the SDP were modified, then reject p with a newly created InvalidModificationError and abort these steps. This only applies to local decriptions. This algorithm is extremely prescriptive about which error checks you need to perform and in which order. This is externally observable because I throw different errors. For instance, why is it the spec's business whether I check for being in the right state before or after I check for the SDP being modified. 5. If the content of description is invalid, then reject p with a newly created InvalidAccessError and abort these steps. What would invalid be here? Anything else? 2. If description is applied successfully, the user agent must queue a task that ru ns the following steps: Didn't we already queue above? Please view or discuss this issue at https://github.com/w3c/webrtc-pc/issues/1277 using your GitHub account
Received on Wednesday, 31 May 2017 16:32:52 UTC