Gardening log entries from 2017
2017-12-26
Tried to fix build on Debian bot:
-- clopez@igalia.com - Fri Dec 26 2017
|
2017-12-22
Reported:
Landed:
-- mcatanzaro@igalia.com - Fri Dec 22 2017
|
2017-12-20
2017-12-14
2017-12-13
2017-12-04
2017-11-29
- Filed bugs:
- Gardening bug: https://bugs.webkit.org/show_bug.cgi?id=180136
- Stats:
- GTK: 9 failures 4 flakes 8 python tests failed api tests ⇒ 1 failures 2 flakes 8 python tests failed api tests ⇒ Failed 1 failures 2 flakes api tests
- WPE: 5 failures 1 new passes 2 flakes 8 python tests failed ⇒ 8 python tests failed ⇒ Build successful
-- Ms2ger@igalia.com
|
2017-11-27
2017-11-23
2017-11-22
Done:
Found a few flakes:
webkit.org/b/179944 imported/w3c/web-platform-tests/html/browsers/the-window-object/apis-for-creating-and-navigating-browsing-contexts-by-name/open-features-negative-innerwidth-innerheight.html [ Pass Failure ]
webkit.org/b/179947 imported/w3c/web-platform-tests/html/semantics/text-level-semantics/the-bdi-element/bdi-auto-dir-default.html [ Pass Timeout ]
webkit.org/b/179948 [ Release ] fast/hidpi/filters-reference.html [ Pass ImageOnlyFailure ]
Gardening patch: https://bugs.webkit.org/show_bug.cgi?id=179949
-- aboya@igalia.com - Wed Nov 22 2017
|
2017-11-20
2017-11-16
2017-11-15
2017-11-09
2017-11-08
Done:
Reported and updated expectations:
webkit.org/b/127280 media/audio-concurrent-supported.html [ Failure Timeout Pass Crash ]
webkit.org/b/157731 fast/events/touch/touch-inside-iframe-scrolled.html [ Timeout Pass Failure ]
webkit.org/b/179443 imported/w3c/web-platform-tests/html/semantics/text-level-semantics/the-bdo-element/bdo-child.html [ Pass Timeout ]
webkit.org/b/179444 http/tests/security/contentSecurityPolicy/upgrade-insecure-requests/upgrade-insecure-audio-video-in-main-frame.html [ Pass Crash ]
webkit.org/b/162017 http/tests/navigation/ping-attribute/anchor-same-origin.html [ Timeout Crash ]
webkit.org/b/178792 fast/events/touch/touch-inside-iframe.html [ Timeout ]
Reported (deliberately kept expectations unchanged):
https://bugs.webkit.org/show_bug.cgi?id=179430
I also made several improvements to https://github.com/ntrrgc/auto-gardener:
- Regression ranges are now also computed when the first failing revision is the oldest known for that test. Now, instead of only that revision being marked as the origin of the failure, every unbuilt revision until the immediately older built one is accounted too.
- Flakes are considered to have ocurred for a long time if they have existed for 75% of their available history (this was previously hardcoded to 40 entries).
- Ranges are no longer reported for flakiness, as they are not very meaningful there. Instead, the first revision known to have failed is indicated, e.g. "Flaky since at least r223640".
- "Skipped" tests (such as when the bots don't run for whatever reason) are now ignored when searching for the first failing revision in a flaky test.
- Added "r" prefix to revision numbers.
- less -SR is used as pager automatically if stdin and stdout are a tty.
-- aboya@igalia.com - Wed Nov 08 2017
|
2017-11-07
Noticed that building with ENABLE_VIDEO=OFF is broken, so worked on that.
-- aperez@igalia.com
2017-11-06
2017-11-02
2017-11-01
2017-10-30
2017-10-25
2017-10-24
-- aperez@igalia.com
2017-10-23
Done:
— mcatanzaro@igalia.com
|
2017-10-18
Done:
— ms2ger@igalia.com
|
2017-10-03
2017-09-28
Done:
— clopez@igalia.com - Thu Sep 28 2017
|
2017-09-27
2017-09-25
2017-09-22
2017-09-20
Done:
— ms2ger@igalia.com
|
2017-09-14
Done (not exactly on Thu, but since my last gardening duty day):
— clopez@igalia.com - Thu Sep 14 2017
|
2017-09-13
2017-09-12
Done:
— mcatanzaro@igalia.com
|
2017-09-11
Done:
— mcatanzaro@igalia.com
|
2017-09-08
Done:
- Nothing, due to no-op svn.webkit.org servers. Will suppplement over the weekend/on Monday.
— zdobersek@igalia.com
|
2017-09-06
Done:
— ms2ger@igalia.com
|
2017-09-05
I worked on WPE today,
Done:
- fast/canvas/webgl/context-attributes-alpha-depth-stencil-antialias.html
- fast/dom/HTMLOptionElement/option-label-quirksmode2.html
- Passing since history allows
- fast/xmlhttprequest/xmlhttprequest-nonexistent-file.html
- fast/xmlhttprequest/xmlhttprequest-no-file-access.html
- imported/w3c/canvas/2d.line.cap.round.html
- scrollbars/scrollbar-iframe-click-does-not-blur-content.html
- imported/w3c/web-platform-tests/WebCryptoAPI/encrypt_decrypt/rsa.worker.html
- fast/canvas/webgl/drawingbuffer-test.html
- These were marked fail by a gardening commit that missed this one.
- fast/events/input-event-insert-replacement.html
- fast/events/before-input-prevent-insert-replacement.html
- js/dom/global-constructors-attributes.html
Squash: https://bugs.webkit.org/show_bug.cgi?id=176370
— cturner@igalia.com - Tue Sep 05 2017
|
2017-08-31
Done:
— clopez@igalia.com - Thu Aug 31 2017
|
2017-08-30
2017-08-29
Done:
— aperez@igalia.com - Tue Aug 29 2017
|
2017-08-28
Done:
-- mcatanzaro@igalia.com - Mon Aug 28 2017
|
2017-08-24
Done:
-- magomez@igalia.com - Thu Aug 24 2017
|
2017-08-23
Done:
- Filed and marked consistent crashes:
- Filed intermittent crashes:
- Skipped CSS Regions tests that are skipped on other platforms:
- Test expectation updates:
- https://bugs.webkit.org/show_bug.cgi?id=175878
- Marked fast/mediastream/captureStream/canvas2d-heavy-drawing.html as flaky since it started failing as well as passing and crashing.
- Added a port-specific expectation for beacon-async-error-logging.html due to a differing error message.
- Updated global-constructors-attributes.html for GTK and WPE.
- Fixed TestWebExtensions failure:
-- Ms2ger@igalia.com
|
2017-08-22
Done:
-- cturner@igalia.com - Tue Aug 22 2017
|
2017-08-16
Done:
- Reported and fixed regression in 4 related tests:
- http/wpt/beacon/cors/cors-preflight-arraybufferview-failure.html
- http/wpt/beacon/cors/cors-preflight-arraybufferview-success.html
- imported/w3c/web-platform-tests/fetch/api/basic/request-headers.any.html
-
imported/w3c/web-platform-tests/fetch/api/basic/request-headers.any.worker.html
imported/w3c/web-platform-tests/fetch/api/basic/request-headers.any.html.
Probably also aff
https://bugs.webkit.org/show_bug.cgi?id=175635
- Bisected and reported regression in
imported/w3c/web-platform-tests/fetch/api/basic/scheme-about.any.html
Actually that test file has been failing since earlier than reported,
possibly for unrelated reasons, but it was in r220751 when the test
output changed for the last time.
https://bugs.webkit.org/show_bug.cgi?id=175641
-- aboya@igalia.com - Wed Aug 16 2017
|
2017-08-15
Done:
Next gardener:
- Focus on flakes, unexpected passes, and API test failures
-- mcatanzaro@igalia.com - Tue Aug 15 2017
|
2017-08-08
Done:
- Fixed some crashes on debug: https://bugs.webkit.org/show_bug.cgi?id=175323
- That side-tracked me for today's duties.
- r220352 is causing quite a lot of crashes in the garden I guess since Monday. I've let the author know on one of the existing gardening bug reports. Don't have time to mark all the crashes in bugzilla.
- r220352 was fixed by r220416
Next gardener:
- Focus on getting WPE green for now rather than wrestling with GTK+
-- cturner@igalia.com - Tue Aug 08 2017
|
