Adam Rice

My current plan is to do the work in two stages. The first stage will only clone chunks. This will be inefficient for ArrayBuffers and transferring stream-of-streams will still not...

> Hi @ricea - What is the timeline for implementing the plan described in the comment above? I have a PR in progress for stage 1 https://github.com/whatwg/streams/pull/1053 which I hope...

@domenic @tyoshino Suggestions, comments, radical changes all welcome. I uploaded it as a whatwg/streams branch to make it easy to look at the preview, which should show up at https://streams.spec.whatwg.org/branch-snapshots/writable-streams-design-philosophy/...

I added another item, "Queued calls to writer methods such as write() are not cancelled when writer.releaseLock() is called. This makes them easy to use in a "fire and forget"...

> is this so that you can detect if eg deflate, gzip, deflate-raw, brotli is supported and if not you can add (polyfill) them yourself? Yes. > ...if so maybe...

I am expecting to add an options bundle as the second argument to the constructor. So for example we'd have something like ```javascript new CompressionStream('deflate', { level: 0.1, flush: 'always'...

> I imagine that will also include the use of a custom dictionary? Yes, that's on the roadmap, although I don't think I've specifically mentioned it here. I filed issue...

@noell I didn't know about the `quality` argument to `toDataURL` and `toBlob`. That's a good precedent. I feel there should be some kind of restriction on implementations. For example, `level:...

Correct. This is essentially blocked behind https://github.com/whatwg/streams/issues/616. Which itself is blocked on any browser implementing byte streams, as well as the backlog of higher-priority work.

This makes "deflate-raw" awkward. It's something we definitely want to support in future, but it's not used by HTTP, and almost certainly never will be.