The following layout test is flaky on MacOS and iOS Release http/tests/resourceLoadStatistics/cap-cache-max-age-for-prevalent-resource.html Probable cause: Test became flaky very recently around r240360. It suddenly failed across several platforms. Attempted repro by just iterating the test 5000 times but this caused no failures. Flakiness Dashboard: https://webkit-test-results.webkit.org/dashboards/flakiness_dashboard.html#showAllRuns=true&tests=http%2Ftests%2FresourceLoadStatistics%2Fcap-cache-max-age-for-prevalent-resource.html Diff: --- /Volumes/Data/slave/highsierra-release-tests-wk2/build/layout-test-results/http/tests/resourceLoadStatistics/cap-cache-max-age-for-prevalent-resource-expected.txt +++ /Volumes/Data/slave/highsierra-release-tests-wk2/build/layout-test-results/http/tests/resourceLoadStatistics/cap-cache-max-age-for-prevalent-resource-actual.txt @@ -4,11 +4,12 @@ PASS loadedUrl is "http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=1234" -PASS loadedUrl is "http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=4321" -PASS loadedUrl is "http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=" -PASS loadedUrl is "http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=" -PASS loadedUrl is "http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=" +FAIL loadedUrl should be http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=4321. Was http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=1234. +FAIL loadedUrl should be http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=. Was http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=1234. +FAIL loadedUrl should be http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=. Was http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=1234. +FAIL loadedUrl should be http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=. Was http://localhost:8000/resourceLoadStatistics/resources/echo-query.php?value=1234. PASS successfullyParsed is true +Some tests failed. TEST COMPLETE
Created attachment 359981 [details] test list this test list run with one child process was able to reproduce the failure local
used command: run-webkit-tests --root testbuild-240366 --test-list test1 --child-processes 1 This reproduced the failure on the first attempt. I have not reproduced repeatedly yet. This indicates that this failure may have been caused by a test earlier in the run.
test is flakey when reproducing as well. attempting to bisect the failure
This tests may also be related, began failing around the same time: http/tests/resourceLoadStatistics/capped-lifetime-for-cookie-set-in-js.html history: https://webkit-test-results.webkit.org/dashboards/flakiness_dashboard.html#showAllRuns=true&tests=http%2Ftests%2FresourceLoadStatistics%2Fcapped-lifetime-for-cookie-set-in-js.html
<rdar://problem/47553061>
I believe this test http/tests/resourceLoadStatistics/cap-cache-max-age-for-prevalent-resource.html became flakey with r240360. with my repro steps above I bisected down to this revision. I am able to get a failure on r240360 but have yet to get a failure on 240359 after about 20 attempts locally.
(In reply to Truitt Savell from comment #6) > I believe this test > http/tests/resourceLoadStatistics/cap-cache-max-age-for-prevalent-resource. > html became flakey with r240360. with my repro steps above I bisected down > to this revision. I am able to get a failure on r240360 but have yet to get > a failure on 240359 after about 20 attempts locally. You can’t skip the test (referencing this bug) if you want. I have a few more steps in the refactoring aimed at resolving these issues.
(In reply to Brent Fulgham from comment #7) > (In reply to Truitt Savell from comment #6) > > I believe this test > > http/tests/resourceLoadStatistics/cap-cache-max-age-for-prevalent-resource. > > html became flakey with r240360. with my repro steps above I bisected down > > to this revision. I am able to get a failure on r240360 but have yet to get > > a failure on 240359 after about 20 attempts locally. > > You can’t skip the test (referencing this bug) if you want. I have a few > more steps in the refactoring aimed at resolving these issues. By can't I meant CAN. Also: One of the causes of flaky timeouts should be resolved in r240509.