that makes sense to me. whereas the root zone thing - you’d…

They should, but whatever Netscape intern was in charge of implementing the cookie jar in 1997 didn't realize host-matching should have been more complicated than a strcmp() against the struct field until it was far too late

1 4

thought agent jabsco 🏁15 Dec 2023

like, without looking it up, I know there's gotta be at least one issue about this on crbug.com resolved WONTFIX because it would break some SAP on-prem server 2003 that exploits it as a key feature of the implementation and now Hixie's made it part of the HTML Living Standard

1 2

thought agent jabsco 🏁15 Dec 2023

while we're on the subject, you can do the same thing with sites that don't consistently redirect to/drop the "www." prefix because, unlike the explicit-root-zone, they're not the same domain

1 3

Samuel

@samuel.fm

that makes sense to me. whereas the root zone thing - you’d expect it to be implicit/explicit versions of the same thing

1 1

thought agent jabsco 🏁15 Dec 2023

yeah, like... a cursory search on Monorail and bugzil.la didn't turn up any issue like the one I described, so I'm starting to think this is worth instrumenting, just so it doesn't turn into another edge case that gets encoded because "nobody wants to break legacy compatibility with IE"

1 1

thought agent jabsco 🏁15 Dec 2023

(ie. instrumenting so the case can be made in six months that it can be safely consolidated without breaking something because nothing in the wild uses it, before some Clever Jack goes and upstreams a patch to jQuery to add "free multiplexing")