Define the processing model for dns-prefetch

[noamr]

Uses the exported "resolve an origin" from Fetch.

Note that this still doesn't include the CSP bits, which are also not implemented anywhere for this & preconnect.

Closes #9015

• At least two implementers are interested (and none opposed):

  • Already supported by all engines

• Tests are written and can be reviewed and commented upon at:

  • Currently untestable with WPT infrastructure, need to think about this further in the future.

• Implementation bugs are filed:

  • N/A

• MDN issue is filed: …
  https://developer.mozilla.org/en-US/docs/Web/Performance/dns-prefetch

(See WHATWG Working Mode: Changes for more details.)

---

/infrastructure.html ( diff )
/links.html ( diff )

[domenic]

LGTM after one fix.

I assume this is pretty much untestable, at least at this point. (Maybe if WPT had its own DNS server??)

Can you fill in the implementation status?

[domenic]

LGTM. Will give a day or so for anyone else to comment, in particular @annevk might be interested because of the Fetch integration.

I'm overall really happy with how quickly you bang out pieces of specification that have been missing for many years :)

[annevk]

This seems slightly incorrect. I would suggest something like:

As the results of this algorithm can be cached, future fetches could be faster.

[annevk]

Looks great modulo nit. I was thinking we might need obtaining the partition key to move to Fetch, but it doesn't seem like a big deal. We can do that if we ever need a bigger refactoring.

[noamr]

Looks great modulo nit. I was thinking we might need obtaining the partition key to move to Fetch, but it doesn't seem like a big deal. We can do that if we ever need a bigger refactoring.

Fixed. We can think of refactoring it into fetch in the next patch (using prefetch-diractive CSP for dns-prefetch and preconnect)