Use URL serializer

[npm1]

Fixes #62 and #63
@annevk

---

Preview | Diff

[annevk]

It seems resolved URL is the wrong concept here. If this holds an image request, you want https://html.spec.whatwg.org/multipage/images.html#img-req-url, right?

As that's already a string, you'll have to parse it to determine the scheme. (Although also, this seems like a flaw there, it should have been a URL.)

Do we need to strip things like username/password/fragment? It probably doesn't matter here, or are these also reported outside of this specific object in some way?

A URL's scheme doesn't include the colon. (Existing error.)

[npm1]

It seems resolved URL is the wrong concept here. If this holds an image request, you want https://html.spec.whatwg.org/multipage/images.html#img-req-url, right?

Yea, that sounds good.

As that's already a string, you'll have to parse it to determine the scheme. (Although also, this seems like a flaw there, it should have been a URL.)

Done

Do we need to strip things like username/password/fragment? It probably doesn't matter here, or are these also reported outside of this specific object in some way?

Not reported outside of this object.

A URL's scheme doesn't include the colon. (Existing error.)

Done

[annevk]

Looks good, except for a minor linking issue.

[npm1]

I guess I need to wait a day or so for bikeshed to pick up the new exports? I tried bikeshed update just now and it still does not pick up the new exports added on HTML. I can try again on Monday.

[annevk]

Yeah, generally it takes about a day. If it continues failing I recommend reaching out to Tab Atkins.

[npm1]

Does this look good @annevk?

[annevk]

Yeah thanks.

It's worth filing a follow-up bug for the other references to HTML. They ought to either be exported or some refactoring should be done, depending.

[npm1]

Sounds good, filed #67