We should switch from testRunner.installStatisticsDidModifyDataRecordsCallback() to testRunner.installStatisticsDidScanDataRecordsCallback() in http/tests/resourceLoadStatistics/partitioned-and-unpartitioned-cookie-deletion.html to make it not be a flaky timeout.
<rdar://problem/37364749>
Committed r228294: <https://trac.webkit.org/changeset/228294>
Good news: the test isn't a flaky timeout anymore. Bad news: it is a flaky failure. --- /Volumes/Data/slave/highsierra-debug-tests-wk2/build/layout-test-results/http/tests/resourceLoadStatistics/partitioned-and-unpartitioned-cookie-deletion-expected.txt +++ /Volumes/Data/slave/highsierra-debug-tests-wk2/build/layout-test-results/http/tests/resourceLoadStatistics/partitioned-and-unpartitioned-cookie-deletion-actual.txt @@ -27,13 +27,13 @@ -------- After removal, should receive no cookies. Did not receive cookie named 'firstPartyCookie'. -Did not receive cookie named 'thirdPartyCookie'. -Client-side document.cookie: +Received cookie named 'thirdPartyCookie'. +Client-side document.cookie: thirdPartyCookie=value -------- Frame: '<!--framePath //<!--frame4-->-->' -------- After user interaction, should receive no cookies. -Did not receive cookie named 'firstPartyCookie'. +Received cookie named 'firstPartyCookie'. Did not receive cookie named 'thirdPartyCookie'. -Client-side document.cookie: +Client-side document.cookie: firstPartyCookie=value https://build.webkit.org/results/Apple%20High%20Sierra%20Debug%20WK2%20(Tests)/r228318%20(2007)/results.html
I think I understand why this happens. This is down to flakiness of cookies. Earlier, we sometimes didn't get the notification of data records removal so we timed out. Now we instead wait for a notification of data records scanning which sometimes fires before removal has finished underneath us. So cookies are still there. :(
Marked as flaky in https://trac.webkit.org/changeset/228403/webkit