inspector/debugger/break-on-* tests are flaky These tests are flaky and seem to fail together on mac-wk1 with a similar diff inspector/debugger/break-on-exception-finally.html inspector/debugger/break-on-exception-native.html inspector/debugger/break-on-exception.html inspector/debugger/break-on-uncaught-exception-finally.html inspector/debugger/break-on-uncaught-exception-native.html inspector/debugger/break-on-uncaught-exception.html https://build.webkit.org/builders/Apple%20El%20Capitan%20Debug%20WK1%20(Tests)/builds/6287 --- /Volumes/Data/slave/elcapitan-debug-tests-wk1/build/layout-test-results/inspector/debugger/break-on-exception-finally-expected.txt +++ /Volumes/Data/slave/elcapitan-debug-tests-wk1/build/layout-test-results/inspector/debugger/break-on-exception-finally-actual.txt @@ -4,6 +4,5 @@ CONSOLE MESSAGE: line 8: TestError Checking pause locations when pausing on all exceptions. -PAUSE #1 AT: doThrow:8:22 Uncaught exception in test page: TestError [break-on-exception-tests.js:8]
<rdar://problem/27087938>
Seems to have started ~r202537, but I do not see an obviously related change.
The first failures that I see are in r202531, in a test run that covered revisions r202531-r202529. While <http://trac.webkit.org/projects/webkit/changeset/202529> does not look like an obvious culprit, it has to be it with fairly high probability. This happens in WK2 too.
One possibility is merely adding a new inspector test causing other to break. Also, these tests all have InspectorTest.reloadPage, which I suspect has always been the source of some instability given how inspector tests work. I'd be fine with having these tests just be skipped while I investigate. Just running the inspector/debugger directory I encounter some sporadic issues that will require more involved investigation.
Skipped tests to get bots back to green while Joe investigates http://trac.webkit.org/projects/webkit/changeset/202720
I just completely rewrote all the break-on-* tests for a number of reasons: - easier to read and understand using new libraries - avoid reload, should make them individually faster - merge tests, instead of 1-file-per-test - eliminates a ton of dead and stale code in these tests That said, I think an issue is that the bots might be running them slowly, so I'm going to mark them as [ Slow ] when first enabling them.
Created attachment 282613 [details] [PATCH] Proposed Fix
Created attachment 282614 [details] [PATCH] Proposed Fix Rebaselined.
Created attachment 282616 [details] [PATCH] Proposed Fix
Comment on attachment 282616 [details] [PATCH] Proposed Fix Rejecting attachment 282616 [details] from commit-queue. Failed to run "['/Volumes/Data/EWS/WebKit/Tools/Scripts/webkit-patch', '--status-host=webkit-queues.webkit.org', '--bot-id=webkit-cq-01', 'apply-attachment', '--no-update', '--non-interactive', 282616, '--port=mac']" exit_code: 2 cwd: /Volumes/Data/EWS/WebKit Last 500 characters of output: t-exception.html patching file LayoutTests/platform/mac-wk2/TestExpectations Hunk #1 succeeded at 294 (offset -2 lines). patching file LayoutTests/platform/mac/TestExpectations Hunk #1 FAILED at 737. Hunk #2 succeeded at 1234 (offset -1 lines). 1 out of 2 hunks FAILED -- saving rejects to file LayoutTests/platform/mac/TestExpectations.rej Failed to run "[u'/Volumes/Data/EWS/WebKit/Tools/Scripts/svn-apply', '--force', '--reviewer', u'Timothy Hatcher']" exit_code: 1 cwd: /Volumes/Data/EWS/WebKit Full output: http://webkit-queues.webkit.org/results/1616548
<https://trac.webkit.org/changeset/202825>