r271703 from bug 220566 seems to have regressed a few tests in WPT: https://wpt.fyi/results/payment-request?q=seq%28status%3Apass%7Cstatus%3Aok%20status%3Apass%7Cstatus%3Aok%20status%3Apass%7Cstatus%3Aok%20status%3Apass%7Cstatus%3Aok%20status%3Apass%7Cstatus%3Aok%20status%3A%21pass%26status%3A%21ok%20status%3A%21pass%26status%3A%21ok%20status%3A%21pass%26status%3A%21ok%20status%3A%21pass%26status%3A%21ok%20status%3A%21pass%26status%3A%21ok%29&run_id=5703710288117760&run_id=5730791449427968&run_id=5176095785615360&run_id=5761234664161280&run_id=6289110253699072&run_id=6295692878282752&run_id=5717790667309056&run_id=6252961795670016&run_id=5112363554439168&run_id=5651017565732864
Are you sure this was not caused by r271735? If it is r271735, then this is expected (see the comment in the ChangeLog of r271735 as to why).
rdar://problem/73464404 Okay, so the basic answer is that we just outright ignore PaymentShippingOption.selected now, and always use the first item, given the current spec cannot be implemented on macOS/iOS with PassKit? And that the tests previously passed because we had disagreement between the JS API and the UI presented by PassKit? If the spec is unimplementable with a major OS payment API, that sounds like feedback we should send to the WG. Leaving this issue open to: 1. Send feedback (assigned to Andy to do so, per Devin) 2. Update the spec and tests, thereby resolving the issue by making us compliant with the spec; or fix our implementation v. the current spec if PassKit adds a suitable API, and to document our current violation of the spec
Created attachment 443595 [details] Patch
Comment on attachment 443595 [details] Patch If there is a regression, it should be fixed separately from new feature work.
Created attachment 443637 [details] Patch
Committed r285521 (244038@main): <https://commits.webkit.org/244038@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 443637 [details].