Re: [fetch] Replace byte stream with ReadableStream. (#200)

> @@ -2760,9 +2805,15 @@ <h4 id="should-response-to-request-be-blocked-due-to-nosniff?"><dfn title="shoul
>       <span title="handle content codings">handling content codings</span> given
>       <var>codings</var> and <var>bytes</var>.
>  
> -     <li><p>Push <var>bytes</var> to <var>response</var>'s
> -     <span title=concept-response-body>body</span>.
> -     <!-- XXX xref payload body / streams -->
> +     <li>
> +      <p><span title=concept-enqueue-ReadableStream>Enqueue</span> a <code>Uint8Array</code>
> +      object wrapping an <code>ArrayBuffer</code> containing <var>bytes</var> to <var>stream</var>.
> +
> +      <p class="note no-backref">This operation will not throw an exception.
> +
> +     <li><p>If <var>stream</var> doesn't <span title=concept-ReadableStream-need-more-data>need more
> +     data</span> and <var>request</var>'s <span>synchronous flag</span> is unset, ask the user agent
> +     to <span title=concept-fetch-suspend>suspend</span> the ongoing fetch.
>      </ol>
>  
>      <p class="note no-backref">This means that the `<code title>Content-Length</code>`

I think we should move this note up directly under the "handling content codings" step. That will make it a little clearer. And maybe reword the note to say "This makes the Content-Length header unreliable to the extent that it was reliable to begin with." to address the concern from @domenic.

---
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/fetch/pull/200/files#r51995048