Re: [whatwg/streams] Specify ReadableStream.[[Transfer]] (#623)

I think that might not be so bad... For example, consider a readable stream with HWM = 5 in a worker that gets transferred to the main thread, where it ends up with HWM = 1. The HWM isn't observable from the main thread, since all you have is a reader. Meanwhile, the stream creator over in the worker thread is still getting appropriate backpressure signals, because they still see HWM = 5.

A writable stream is a bit more troubling. Assuming the same setup, the producer in the main thread will always see desiredSize = 1. I wonder if we should consider some kind of asynchronous proxying of desiredSize over to the main thread... that makes things two-directional, which seems weird. This reminds me of when we were discussing whether the caller of getWriter() should determine the HWM, instead of the writable stream creator.

The fact that readables and writables are so different here implies to me I might be missing something. Thoughts welcome.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/streams/pull/623#issuecomment-280795756