Content-Range header parsing

[annevk]

I looked at https://wicg.github.io/background-fetch/#extract-content-range-values and one problem I found is that this will lead to different results between

Content-Range: x, y

and

Content-Range: x
Content-Range: y

which is something we should avoid (with the notable exception of Set-Cookie). Please use https://fetch.spec.whatwg.org/#concept-header-list-get to get the value instead.

The implications for this algorithm are that either you need to handle values containing , in a special way or you let them be treated as errors. Not sure what matches implementations better.

The outcome here might also impact https://github.com/annevk/orb/pull/16 as that has rather lax header checking.

(I suspect we might also want to move this into Fetch if this is to be shared between this, media elements, and ORB.)

[jakearchibald]

I should just be able to switch to https://fetch.spec.whatwg.org/#concept-header-list-get, and step 3 will still do the right validation.

[annevk]

Yeah, though the second example above might pass today (for a correct x) and would no longer after that change. Worth testing as implementations are buggy around stuff like this.