Adam Rice

> Can the purpose be fulfilled by something like `pull(c) { c.enqueue(data); return new Promise(r => setTimeout(r, 10)) }`? Another pull won't happen until the promise resolves, right? This doesn't...

This actually protects against a use-after-free attack where the buffer is transferred away in the middle of being decompressed. Calling enqueue can result in executing user JavaScript. Maybe there should...

I have a preference for simple exceptions, but I don't have a good justification for it. I think the compatibility risk for changing this would probably be minimal, but I...

For someone trying to use DecompressionStream for the first time, it might be useful to distinguish between `TypeError` meaning "you gave me something that wasn't a BufferSource" and `DataError` meaning...

We can't change the semantics of `tee()` without breaking existing sites. But as @yonran suggests, we can add an option to base backpressure signalling on the slower branch. While a...

The intent of the test is that it be created in the realm of the constructor. Probably we should add some comments. *Edit:* Incidentally, the point of the test is...

> But [wpt.fyi](https://wpt.fyi/results/streams/transform-streams/invalid-realm.window.html?label=experimental&label=master&aligned) says it passes on Chrome? (Adding a failing test sounds weird, such test should at least be marked as tentative.) The test previously worked in Chrome, but...

My preference is for #1019.

It looks like we have a plan here, and we just need someone to do the work. I don't have time to commit to this in the near future, but...

I like `bytes()` but I thought it might be controversial.