Pass pullAlgorithm while initializing [[readable]]?

[saschanaz]

https://w3c.github.io/webrtc-encoded-transform/#stream-creation

Set up this.[[readable]]. this.[[readable]] is provided frames using the readEncodedData algorithm given this as parameter.

But it defines no pullAlgorithm, and nothing explicitly calls readEncodedData outside the stream, which means nothing will enqueue any data at all. Should it pass a pullAlgorithm that calls readEncodedData?

[youennf]

Right, there is a mismatch here.

The intent was to use the start algorithm that would basically wait for a frame to be available and enqueue it as soon as available, independently on the actual queue size or whether pull was called (no backpressure like on writer side).

We should revise this section to align and could use readEncodedData as part of pulling data.

Maybe we could revise the back pressure mechanism here, depending on what UAs actually do (Safari has no back pressure here).

[saschanaz]

The intent was to use the start algorithm that would basically wait for a frame to be available and enqueue it as soon as available, independently on the actual queue size or whether pull was called (no backpressure like on writer side).

Sounds also like it could be helped by setting highWaterMark to infinity, as doing so will start pulling immediately and indefinitely.