- From: docfaraday via GitHub <noreply@w3.org>
- Date: Thu, 26 Feb 2026 16:46:10 +0000
- To: public-webrtc-logs@w3.org
> https://websockets.spec.whatwg.org//#garbage-collection -- the websocket server can definetly observe the garbage collection. The difference here is that WebRTC is running the server in JS. Just adding on to this. As a thought experiment, consider a middlebox implmented in some non-browser JS environment, using RTCPeerConnection. How would an implementer of such a thing react if they were told "Oh, the JS engine doesn't actually free up these network resources on GC, because otherwise a browser might notice you GC'ed"? Look, I get it. Webrtc is peer-to-peer, and that means that the browser is doing server stuff, which is *extremely unusual* and flies in the face of many long-held principles in the web model. The fact is that GC is going to be observable for someone determined enough to observe it. The entire point of GC is to free up resources when they are no longer needed, and the freeing of resources is always going to be observable to some extent. The question is how far we are willing to bend over backward to make that difficult. I think I've made it clear where I stand on that particular question, and I think this position will be very common among transport protocol people. -- GitHub Notification of comment by docfaraday Please view or discuss this issue at https://github.com/w3c/webrtc-pc/issues/3090#issuecomment-3967823251 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Thursday, 26 February 2026 16:46:11 UTC