Joe Lee

Good catch... Might be worth noting that not-null assertions in kj code are typically done by wrapping expressions in `KJ_REQUIRE_NONNULL(expr, "some message")`, which may explain why the issue hadn't become...

Good to know. Yeah, I was curious if it would work for pointers, so I dug a little into the implementation, and it looks like it should, although I guess...

I'm not sure exactly how the new field is intended to be consumed, but it might be worth noting that `scriptId` is an internal ID, which isn't intended to be...

Not sure of any exact plans, but: - There is a [long-standing TODO](https://github.com/capnproto/capnproto/blob/5303d877/c%2B%2B/src/kj/async.h#L341-L342) to eventually provide something like that. - The code has changed quite a bit since PR #452,...

Looks like Josh's branch is in a personal repo... I couldn't find a way in the github UI to update the PR to use a different branch, so I opened...

Thanks! This looks good to me, but probably worth keeping in mind that the internal `web-socket-proxy` test in edgeworker currently covers the behavior difference between the two autogate settings, so...

(Not a problem, but just FYI... I think if PRs are opened in "draft" mode, it allows iterating on them without generating reviewers notifications? And, I guess, would be another...

> @jclee Can you resolve conflicts? - then we can merge this today I've rebased, resolving the conflict between the workers changelog entry and the changelog update that went in...

> I think my issue is related or at least related to the fix. I see the PR was merged but I am assuming this is not live yet. Correct......

@stayallive Thanks for the details. As far as this bug goes, I would expect that code to work OK, on both the new and old versions of the runtime, so...