Regression : "CLOSE (X)" icon disappears after moving mouse over it, on awwwards.com.
Reported by
avsha...@etouch.net,
Mar 22 2017
|
||||||||
Issue descriptionChrome Version : 59.0.3048.0 (Official Build) 1af111915f36842841682091b8c6882cae410831-refs/heads/master@{#458590} 32/64 bit OS : Windows (7,8,10), Mac (10.11.6, 10.12.1, 10.12), Linux (14.04 LTS) Test URL : https://www.awwwards.com/22-experimental-webgl-demo-examples.html What steps will reproduce the problem? 1. Launch chrome and navigate to above URL. 2. Click on 'MENU' icon seen at the top left corner of the page and hover mouse on "CLOSE (X)" icon 2-3 times. 3. Observe the (X) icon. Actual : "CLOSE (X)" icon disappears after moving mouse over it. Expected : "CLOSE (X)" icon should not disappear after hovering mouse over it (i.e "X" icon should stay visible). This is a regression issue broken in ‘M-59’, below is the Manual Regression range and will soon update other info. Good build : 59.0.3037.0 Bad build : 59.0.3038.0 Note : Above issue is not seen in other browser like Safari and Firefox.
,
Mar 22 2017
,
Apr 4 2017
,
Apr 4 2017
I can reproduce on mac with --disable-prefer-compositing-to-lcd-text. I cannot yet reproduce on Linux. I think the difference has to do with the size of the raster tiles.
,
Apr 4 2017
I can repro on Mac but not on Linux via Chromoting. I suspect it will repro in a regular Linux box when accessing localy, with --force-gpu-rasterization. Tien-Ren could you try that for me?
,
Apr 5 2017
ping on this
,
Apr 6 2017
I have redone the bisect and it pointed to #460521 (Re-enabling SPInvalidation). As it is just turning on the flag, it gives little information how the bug was introduced. Some important timeline: 445778: SPInvalidation first enabled 445955: 59.0.3037.0 456244: 59.0.3038.0 456857: SPInvalidation disabled 458451: SPInvalidation enabled 459201: SPInvalidation disabled 460521: SPInvalidation enabled I'll take a look at chrishtr's CL (which is supposed to introduce no behavior change) to see if there is a typo. At the same time redo the bisect between 445778 and 456856.
,
Apr 6 2017
,
Apr 6 2017
My latest bisect matches previous result. (https://codereview.chromium.org/2745563004) I reviewed the CL, although I feel the use of m_tempRect looked dangerous, I didn't find anything wrong. I'm guessing it could be related to the handling of infinite rects? It looks like the X button disappears only during the opacity transition. During the transition the CLOSE (X) button get promoted to a compositing layer. It feels like the culling logic somehow omitted painting the X. Assigning back to chrishtr for further investigation.
,
Apr 10 2017
Can repro on my Linux box. Saved the page locally, am now reducing it.
,
Apr 11 2017
As expectd, it does not reproduce without SPInvalidation.
,
Apr 12 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/64eb4f544d1b3b778f5b915de00e1e6cdced0aa3 commit 64eb4f544d1b3b778f5b915de00e1e6cdced0aa3 Author: chrishtr <chrishtr@chromium.org> Date: Wed Apr 12 03:03:39 2017 Fix return value for "fallback" case of inverted clips. Also changed LocalToAncestorVisualRect to return a const reference, which was an oversight in previous CLs. BUG= 704077 CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2 Review-Url: https://codereview.chromium.org/2816533002 Cr-Commit-Position: refs/heads/master@{#463910} [modify] https://crrev.com/64eb4f544d1b3b778f5b915de00e1e6cdced0aa3/third_party/WebKit/LayoutTests/TestExpectations [modify] https://crrev.com/64eb4f544d1b3b778f5b915de00e1e6cdced0aa3/third_party/WebKit/LayoutTests/paint/invalidation/svg/feImage-target-attribute-change-with-use-indirection-2-expected.txt [modify] https://crrev.com/64eb4f544d1b3b778f5b915de00e1e6cdced0aa3/third_party/WebKit/LayoutTests/paint/invalidation/svg/marker-viewBox-changes-expected.txt [modify] https://crrev.com/64eb4f544d1b3b778f5b915de00e1e6cdced0aa3/third_party/WebKit/Source/platform/graphics/paint/GeometryMapper.cpp [modify] https://crrev.com/64eb4f544d1b3b778f5b915de00e1e6cdced0aa3/third_party/WebKit/Source/platform/graphics/paint/GeometryMapper.h [modify] https://crrev.com/64eb4f544d1b3b778f5b915de00e1e6cdced0aa3/third_party/WebKit/Source/platform/graphics/paint/GeometryMapperTest.cpp
,
Apr 12 2017
,
Apr 12 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/e83f086321c0dab4f9562a8d0140ea3b878d535f commit e83f086321c0dab4f9562a8d0140ea3b878d535f Author: Rebaseline Bot <blink-rebaseline-bot@chromium.org> Date: Wed Apr 12 04:38:16 2017 Auto-rebaseline for r463910 Build: https://build.chromium.org/p/chromium.infra.cron/builders/rebaseline-o-matic/builds/660252 https://chromium.googlesource.com/chromium/src/+/64eb4f544d1b3 BUG= 704077 TBR=chrishtr@chromium.org Review-Url: https://codereview.chromium.org/2814723003 . Cr-Commit-Position: refs/heads/master@{#463930} [modify] https://crrev.com/e83f086321c0dab4f9562a8d0140ea3b878d535f/third_party/WebKit/LayoutTests/TestExpectations
,
Apr 18 2017
Tested the issue on Mac 10.12.4,Windows 7 & Ubuntu 14.04 using chrome version# 59.0.3071.9. Observed that the "CLOSE (X)" icon is not disappeared after hovering mouse over it . Hence adding TE-Verified labels. Please find the attached screen cast for the same. Thanks!! |
||||||||
►
Sign in to add a comment |
||||||||
Comment 1 by rbasuvula@chromium.org
, Mar 22 2017Labels: hasbisect-per-revision ReleaseBlock-Stable
Owner: chrishtr@chromium.org
Status: Assigned (was: Unconfirmed)