Re: [fetch] Introduce Response.body. (#118) (#128)

> +<code title=concept-ReadableStream>ReadableStream</code> with given a byte stream
> +<var>bstream</var>, run these steps:
> +
> +<ol>
> + <li><p>Let <var>strategy</var> be an object. The user agent may choose any object.
> +
> + <li>
> +  <p>Let <var>pull</var> be an action that runs these substeps:
> +
> +  <ol>
> +    <li><p>Let <var>bytes</var> be a byte sequence which is 
> +    <span data-anolis-spec=encoding title=concept-stream-read>read</span> from <var>bstream</var>.
> +
> +    <li><p>If <span data-anolis-spec=encoding title=concept-stream-read>read</span> returns
> +    <span data-anolis-spec=encoding>end-of-stream</span>,
> +    <span title=concept-close-ReadableStream>close</span> <var>stream</var> defined below.

I will use ReadableStream instead of byte stream in request and response.

For extracting Strings, ArrayBuffers and FormData objects, they can be converted to byte sequences and thus easily enqueued to a ReadableStream object.

For extracting Blobs, it would be possible to define an operation returning a ReadableStream object consisting of the blob contents (At least, as [blob] (https://w3c.github.io/FileAPI/#blob) refers to a byte sequence, enqueuing it to the ReadableStream object is possible. Allowing the UA to split the sequence to chunks would be beneficial, though).


---
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/fetch/pull/128/files#r41135406