Adam Rice

We need interest from one other browser besides Chrome. I should also review the tests before landing, but that's not strictly a blocker.

> I updated tests in [web-platform-tests/wpt#39520](https://github.com/web-platform-tests/wpt/pull/39520). Should we merge these WPT tests first so that I update this PR to use the WPT commit? Yes, please.

My interpretation is that in the second case, all the data has been read, so the underlying source doesn't need to know about the cancellation. But in the first case...

I don't think it's the promises that are causing you to run out of memory, but the large number of pending writes holding onto buffers. So awaiting the results of...

I was hoping that we could just require the underlying source to do cleanup in this case. But the underlying source doesn't know it is being used in async iteration....

Yes. See also https://github.com/WICG/compression/issues/22.

Sorry, it is implementation-dependant at the moment. I don't think there's a good workaround. We should make it a priority to fix #22 once the migration from WICG to WHATWG...

It is very likely that HTTP/2.0 connections will be reverse-proxied on to legacy servers. For example, https://example.com/shiny-service/ws and https://example.com/legacy-service/ might be served from different backend servers. It is the responsibility...

I wasn't there when this was decided, but I think the intent is that the server can treat these status codes as "trusted" when the client is a browser. How...

1. I'm not sure I fully understand this proposal. How would "custom-type" be set in a WebSocket frame? Changes to the protocol would require an update to the RFC, which...