Bug 300050 – REGRESSION(300617@main): [macOS wk1] imported/w3c/web-platform-tests/infrastructure/server/wpt-server-http.sub.html is failing

• Status: RESOLVED
• Resolution: WONTFIX
• Component: Tools / Tests

[Marta Darbinyan]

The following test is constant failure on Sequoia and flaky failure on Sonoma on wk1 after 300617@main

imported/w3c/web-platform-tests/infrastructure/server/wpt-server-http.sub.html

Diff:
 PASS HTTP protocol, non-existent domain, port #1
 PASS HTTP protocol, non-existent domain, port #2
 PASS HTTPS protocol, no subdomain
-PASS HTTPS protocol, www subdomain #1
-PASS HTTPS protocol, www subdomain #2
-PASS HTTPS protocol, www subdomain #3
-PASS HTTPS protocol, punycode subdomain #1
-PASS HTTPS protocol, punycode subdomain #2
+FAIL HTTPS protocol, www subdomain #1 assert_true: expected true got false
+FAIL HTTPS protocol, www subdomain #2 assert_true: expected true got false
+FAIL HTTPS protocol, www subdomain #3 assert_true: expected true got false
+FAIL HTTPS protocol, punycode subdomain #1 assert_true: expected true got false
+FAIL HTTPS protocol, punycode subdomain #2 assert_true: expected true got false
 PASS HTTPS protocol, non-existent domain, port #1
 PASS HTTPS protocol, non-existent domain, port #2

Test results:
https://build.webkit.org/#/builders/1233/builds/5548

Test history:
https://results.webkit.org/?suite=layout-tests&test=imported%2Fw3c%2Fweb-platform-tests%2Finfrastructure%2Fserver%2Fwpt-server-http.sub.html&flavor=wk1

[Radar WebKit Bug Importer]

<rdar://problem/161851589>

[EWS]

Test gardening commit 300906@main (2a530014269e): <https://commits.webkit.org/300906@main>

Reviewed commits have been landed. Closing PR #51709 and removing active labels.

[Sam Sneddon [:gsnedders]]

This makes absolutely no sense for it to be flaky; it's very worrying if we're flaking this because it implies we're not always obeying the DNS server, and that would imply flakiness throughout all of WPT — not just on this specific test.

We have an expectation for LayoutTests/platform/mac-sonoma-wk1/imported/w3c/web-platform-tests/infrastructure/server/wpt-server-http.sub-expected.txt where we seemingly expect this not to work on Sonoma WK1 (similarly worrying).

If we can't get the functionality working on Sonoma on WK1, maybe we should simply not enable the DNS server there, because otherwise we're gonna be gardening Sonoma WK1 specifically, and we don't really have great EWS coverage for that.

[Jonathan Bedard]

Closing bug to match associated radar