[rifmj]

New tests in cookies/. I found no existing coverage for these cases:

• samesite/methods-beyond-get-post.html: PUT, DELETE, PATCH and OPTIONS carry a SameSite=Strict cookie on a same-origin request.
• ordering/equal-segment-count.html: two paths with the same number of segments sort by length, not by creation time. ordering.sub.html covers paths with different segment counts only.
• attributes/expired-at-read.html: an expired cookie is gone with no Set-Cookie in between, checked separately for the Cookie header and document.cookie, with a control cookie. Each check waits 4 seconds, so the file has timeout=long.
• attributes/last-attribute.html: Max-Age before Expires in both directions, a last Path without a leading /, and an empty Domain.
• attributes/last-attribute.tentative.html: a last unparsable Expires and Domain=.. Browsers differ here from RFC 6265bis 5.6.1 step 2 and 5.6.3, so these two are tentative.

Results under ./wpt serve with Playwright 1.64.1 builds, three identical runs. I did not run ./wpt run.

Test	Chromium 156	Firefox 157	WebKit 27.2
methods-beyond-get-post	4/4	4/4	4/4
equal-segment-count	1/1	1/1	1/1
expired-at-read	2/2	2/2	2/2
last-attribute	4/4	4/4	3/4
last-attribute.tentative	2/2	2/2	2/2

The WebKit failure is the Path=foo case: the cookie API reports the literal path foo, and the cookie does not come back on the default path. I have not checked Safari.

Related: whatwg/fetch#1965, whatwg/fetch#1826, httpwg/http-extensions#3581, httpwg/http-extensions#3582, httpwg/http-extensions#3583.

Drafted with AI help.