Layout Test virtual/scalefactor150/fast/hidpi/static/*-picker-appearance.html is failing on Linux, Win debug |
|||||||||||||
Issue descriptionI'm not clear on how long this has been failing, or what the cause is ... https://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=virtual%2Fscalefactor150%2Ffast%2Fhidpi%2Fstatic%2Fcalendar-picker-appearance.html%2Cvirtual%2Fscalefactor150%2Ffast%2Fhidpi%2Fstatic%2Fdata-suggestion-picker-appearance.html%2Cvirtual%2Fscalefactor150%2Ffast%2Fhidpi%2Fstatic%2Fpopup-menu-appearance.html oshima@, can you take a look? This seems like it's related to your work in bug 567837 , maybe?
,
Mar 10 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/b25c5403ae4917db672cc95b131905e0bd2b8e5e commit b25c5403ae4917db672cc95b131905e0bd2b8e5e Author: dpranke <dpranke@chromium.org> Date: Thu Mar 10 06:15:12 2016 Suppress some virtual/scalefactor150 layout test failures. TBR=oshima@chromium.org, tsergeant@chromium.org BUG= 593567 Review URL: https://codereview.chromium.org/1782783002 Cr-Commit-Position: refs/heads/master@{#380354} [modify] https://crrev.com/b25c5403ae4917db672cc95b131905e0bd2b8e5e/third_party/WebKit/LayoutTests/TestExpectations
,
Mar 15 2016
,
Mar 15 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/debc2d3cdb14b9d7a356218127e0ac26a273b9ea commit debc2d3cdb14b9d7a356218127e0ac26a273b9ea Author: oshima <oshima@chromium.org> Date: Tue Mar 15 20:24:45 2016 Rebaseline expectatiosn for virtual/scalefactor150 BUG= 593567 R=wkorman@chromium.org Review URL: https://codereview.chromium.org/1801273002 Cr-Commit-Position: refs/heads/master@{#381299} [modify] https://crrev.com/debc2d3cdb14b9d7a356218127e0ac26a273b9ea/third_party/WebKit/LayoutTests/TestExpectations
,
Mar 15 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/63a1926a7956f48d94fb7f5550d522a9c5aecf03 commit 63a1926a7956f48d94fb7f5550d522a9c5aecf03 Author: Rebaseline Bot <blink-rebaseline-bot@chromium.org> Date: Tue Mar 15 21:48:38 2016 Auto-rebaseline for r381299 https://chromium.googlesource.com/chromium/src/+/debc2d3cd BUG= 593567 TBR=oshima@chromium.org Review URL: https://codereview.chromium.org/1801363002 . Cr-Commit-Position: refs/heads/master@{#381328} [modify] https://crrev.com/63a1926a7956f48d94fb7f5550d522a9c5aecf03/third_party/WebKit/LayoutTests/TestExpectations
,
Mar 15 2016
It seems the rebaseline didn't think the results on the build bots were different from what's already in the tree: https://codereview.chromium.org/1801363002 Does running these tests locally on Linux/Win fail?
,
Mar 16 2016
Hmm, I wonder why tests were failing then. Running locally did create slightly different images (I didn't go and check into details. I'll run it again shortly), although I've been told that the bot may generate slightly different baseline due to difference in environment.
,
Mar 16 2016
So this is failing locally. The major difference seems to be coming from font rendering, although there seems to be other difference such as arrow. I also looked at the latest result, and it looks to me that only debug builds are failing. wkorman@ am I reading it right? If so, I'll skip the test on debug (at least for now)
,
Mar 16 2016
Running locally with Debug: virtual/scalefactor150/fast/hidpi/static/calendar-picker-appearance.html virtual/scalefactor150/fast/hidpi/static/data-suggestion-picker-appearance.html virtual/scalefactor150/fast/hidpi/static/popup-menu-appearance.html I get attached difference, which shows two failing: virtual/scalefactor150/fast/hidpi/static/calendar-picker-appearance.html - here, the gray shadow at bottom of calendar picker box looks darker in the 'actual' virtual/scalefactor150/fast/hidpi/static/data-suggestion-picker-appearance.html - here, the gray shadow at right of picker looks darker in 'actual' but yes it's my understanding that Linux workstations have varying configs, fonts, etc. and can lead to some set of test differences. These tests all pass when run with a Release build. I think in general we don't want to have baselines that differ from debug to non-debug, and if we do have them, we want it to be a temporary state. +ojan for any further commentary.
,
Mar 16 2016
Generally speaking we should try to match the same fonts locally and on the bots. It is possible that people have different fonts installed, but I haven't really many issues where that happens. It is also possible that tests will produce slightly different results in debug and release, in which case the correct thing to do is to just leave the suppression in. However, it would be good to try and figure out exactly which case is which and, if it is some sort of environmental difference, find out what the difference is.
,
Mar 17 2016
So I run it Goobuntu and the difference looks same as what whatwkorman@ had. (I run my test in #8 at home, and my PC probably had different font settings than corp machines) I'll Skip the debug while I investigate why.
,
Mar 17 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/75bacd2faadde60c2dbe83e02c6507a6a6439695 commit 75bacd2faadde60c2dbe83e02c6507a6a6439695 Author: oshima <oshima@chromium.org> Date: Thu Mar 17 02:08:07 2016 Skip scalefactor150 tests that are failing on debug build. BUG= 593567 R=wkorman@chromium.org Review URL: https://codereview.chromium.org/1813603002 Cr-Commit-Position: refs/heads/master@{#381641} [modify] https://crrev.com/75bacd2faadde60c2dbe83e02c6507a6a6439695/third_party/WebKit/LayoutTests/TestExpectations
,
Mar 17 2016
The difference comes from the factor that compositor's default background is blue on debug build (see LRenderer::ClearFrameBuffer in cc/output/gl_renderer.cc for the reason), and at fractional scale factor there is a half pixel region that weren't drawn. I got the same result if I change not to clear the framebuffer with blue. Since this is one of things that will be addressed in crbug.com/485650, I'll close this. I confirmed that this does not happen in new use-zoom-for-dsf mode (I haven't added scalefactor150withzoom yet, but will do soon).
,
Apr 12 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/36d028eda81c6a8e36c99b7cc7c5aa713b6f0be9 commit 36d028eda81c6a8e36c99b7cc7c5aa713b6f0be9 Author: kjellander <kjellander@chromium.org> Date: Tue Apr 12 13:22:48 2016 Restore skipping of virtual/scalefactor150/fast/hidpi/static/ webkit_tests Three tests were no longer skipped as part of https://codereview.chromium.org/1857083002 which caused them to start failing on Linux and Windows debug builds again. I'm moving these back to be skipped, awaiting further investigation. BUG= 593567 TBR=eae@chromium.org, oshima@chromium.org Review URL: https://codereview.chromium.org/1880943003 Cr-Commit-Position: refs/heads/master@{#386664} [modify] https://crrev.com/36d028eda81c6a8e36c99b7cc7c5aa713b6f0be9/third_party/WebKit/LayoutTests/TestExpectations
,
May 18 2016
,
May 20 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/aaf3b898c1665889c3aed2d0e439f6ec74c1bdc5 commit aaf3b898c1665889c3aed2d0e439f6ec74c1bdc5 Author: oshima <oshima@chromium.org> Date: Fri May 20 20:08:38 2016 Enable these tests only on debug build. They're expected to fail on debug build. BUG= 593567 Review-Url: https://codereview.chromium.org/2003643002 Cr-Commit-Position: refs/heads/master@{#395147} [modify] https://crrev.com/aaf3b898c1665889c3aed2d0e439f6ec74c1bdc5/third_party/WebKit/LayoutTests/TestExpectations
,
May 20 2016
Just noticed typo in CL description, my bad. It should read: "Disable these tests only on debug build. ...."
,
Jul 13 2016
,
Jul 13 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/68a356c37d43acc2e21a43bb500211bef227cfe2 commit 68a356c37d43acc2e21a43bb500211bef227cfe2 Author: engedy <engedy@chromium.org> Date: Wed Jul 13 09:44:08 2016 Mark calendar/data picker tests @150% failing on Linux Dbg. virtual/scalefactor150/fast/hidpi/static/calendar-picker-appearance.html virtual/scalefactor150/fast/hidpi/static/data-suggestion-picker-appearance.html BUG= 627754 , 593567 TBR=bsep@chromium.org NOTRY=true Review-Url: https://codereview.chromium.org/2139293003 Cr-Commit-Position: refs/heads/master@{#405109} [modify] https://crrev.com/68a356c37d43acc2e21a43bb500211bef227cfe2/third_party/WebKit/LayoutTests/TestExpectations
,
Jul 13 2016
Reopening, as there are some image inconsistencies between Release and Debug configurations on Linux. Please see Issue 627754 for details.
,
Jul 13 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/68a356c37d43acc2e21a43bb500211bef227cfe2 commit 68a356c37d43acc2e21a43bb500211bef227cfe2 Author: engedy <engedy@chromium.org> Date: Wed Jul 13 09:44:08 2016 Mark calendar/data picker tests @150% failing on Linux Dbg. virtual/scalefactor150/fast/hidpi/static/calendar-picker-appearance.html virtual/scalefactor150/fast/hidpi/static/data-suggestion-picker-appearance.html BUG= 627754 , 593567 TBR=bsep@chromium.org NOTRY=true Review-Url: https://codereview.chromium.org/2139293003 Cr-Commit-Position: refs/heads/master@{#405109} [modify] https://crrev.com/68a356c37d43acc2e21a43bb500211bef227cfe2/third_party/WebKit/LayoutTests/TestExpectations
,
Jul 13 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/652f1abd082dde8d8d775a3f72f80358e14fce04 commit 652f1abd082dde8d8d775a3f72f80358e14fce04 Author: bsep <bsep@chromium.org> Date: Wed Jul 13 20:19:47 2016 Mark editing/spelling/inline-spelling-markers-hidpi-composited.html as flaky on Windows 7 and fix TestExpectations for picker appearance tests. BUG= 627782 , 593567 TBR=engedy Review-Url: https://codereview.chromium.org/2150523003 Cr-Commit-Position: refs/heads/master@{#405268} [modify] https://crrev.com/652f1abd082dde8d8d775a3f72f80358e14fce04/third_party/WebKit/LayoutTests/TestExpectations
,
Jul 13 2016
Like it says in #13 this will be fixed by launching use-zoom-for-dsf on Linux, so there's no reason to keep this bug open.
,
Jul 22 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/ddb31050950f1ccf79b03269dea44926c5e02661 commit ddb31050950f1ccf79b03269dea44926c5e02661 Author: Bret Sepulveda <bsep@chromium.org> Date: Fri Jul 22 22:44:13 2016 Mark calendar/data picker tests @150% failing on Linux Dbg. virtual/scalefactor150/fast/hidpi/static/calendar-picker-appearance.html virtual/scalefactor150/fast/hidpi/static/data-suggestion-picker-appearance.html BUG= 627754 , 593567 TBR=bsep@chromium.org NOTRY=true Review-Url: https://codereview.chromium.org/2139293003 Cr-Commit-Position: refs/heads/master@{#405109} (cherry picked from commit 68a356c37d43acc2e21a43bb500211bef227cfe2) Review URL: https://codereview.chromium.org/2178673003 . Cr-Commit-Position: refs/branch-heads/2785@{#316} Cr-Branched-From: 68623971be0cfc492a2cb0427d7f478e7b214c24-refs/heads/master@{#403382} [modify] https://crrev.com/ddb31050950f1ccf79b03269dea44926c5e02661/third_party/WebKit/LayoutTests/TestExpectations |
|||||||||||||
►
Sign in to add a comment |
|||||||||||||
Comment 1 by osh...@chromium.org
, Mar 10 2016Status: Assigned (was: Untriaged)