"ASCII serialisation of an origin" and "domain to ASCII"

[SimonSapin]

https://html.spec.whatwg.org/multipage/#ascii-serialisation-of-an-origin includes:

Apply the domain to ASCII algorithm to each component of the host part of the origin tuple, and append the results — each component, in the same order, separated by U+002E FULL STOP characters (.) — to result. [URL]

… but domain to ASCII in https://url.spec.whatwg.org/#concept-domain-to-ascii takes and returns a full domain, not a domain component. It already splits on U+002E FULL STOP and joins back internally.

I think the quoted text should at least be replaced with something like:

Apply the domain to ASCII algorithm to the host part of the origin tuple, and append the result to result. [URL]

Additionally, assuming the origin comes from https://url.spec.whatwg.org/#concept-url-origin, its host part is the host of an URL record that is the result of host parsing https://url.spec.whatwg.org/#concept-host-parser, and therefor has already gone through domain to ASCII.

Is it useful to call domain to ASCII again? Is it not idempotent?

CC @nikkisquared

[annevk]

It's not useful to call it again and I don't think a specification should call that algorithm directly anyway. This should be easier to fix once the terms introduced by #678 land.