Adam Rice

Here's a sketch of an approach which reconciles the atomic nature of transfer with the asynchronous nature of streams. I'm going to talk about the WritableStream case because I think...

The standard is (intentionally) ambiguous on how writing is done, so "somehow force the write to happen anyway even though it would normally be disallowed" might be one interpretation.

Yes, you can tick the checkbox for Chrome.

Thanks! I am happy not to have `fill()` since it is not too hard for people to use the more general API.

Yes, `min` is good too. It's a bit more elegant. I don't know whether it's easier to remember or not. `minLength` is another option if we want to be clear...

I prefer the second form, since I don't think it makes sense to have the console locked to a particular source. I'm interested in whether we could give a hint...

Sorry for my slow response. The answer is not straightforward. This does imply that a guide would be useful. I don't know whether it would be better to include it...

Here's what I did for WebSocketStream. I'm not saying this is necessarily correct, in fact I welcome feedback on what the correct behaviour should be. > What signals need to...

/cc @reillyeon who probably thought about these issues in the context of the WebSerial API.

I think we should also resolve the same question for `controller.error()` in this issue, as it is closely related. Currently `controller.error()` does not result in a call to `us.abort()`.