New issue
Advanced search Search tips

Issue 890638 link

Starred by 4 users

Issue metadata

Status: Fixed
Owner:
Closed: Oct 12
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux , Windows , Mac
Pri: 1
Type: Bug-Regression



Sign in to add a comment

Menu stripped off in espn.com

Reported by josh.lin...@gmail.com, Sep 30

Issue description

UserAgent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/71.0.3566.0 Safari/537.36

Example URL:
ESPN.com

Steps to reproduce the problem:
1. Go to espn.com
2. Click on any sport along the top line (tabs/list/menu whatever it is)
3. Drop-down doesn't work

What is the expected behavior?
It SHOULD populate across the screen - so you don't have to click to go to a sport THEN access standings, scores, etc.

What went wrong?
It is CUTTING OFF the menu

Does it occur on multiple sites: Yes

Is it a problem with a plugin? No 

Did this work before? Yes 68

Does this work in other browsers? Yes

Chrome version: 71.0.3566.0  Channel: canary
OS Version: OS X 10.14.0
Flash Version: 

I tried to open the report but someone in INDIA claimed it was working in INDIA.
Well here's a clue - I'm not in INDIA. I don't have cricket and ----- tasting - this is the USA site (default) and IT DOES NOT WORK
 
espn-dropdown-CHROME.png
1.0 MB View Download
Labels: Needs-Bisect Needs-Triage-M71
Labels: TE-NeedsTriageFromMTV Triaged-ET
Unable to reproduce the issue on mac 10.13.3 using chrome reported version #71.0.3566.0.
Attached a screen cast for reference.

Following are the steps followed to reproduce the issue.
------------
1. Navigated to espn.com which got redirected to espn.in
2. Clicked on any sport along the top line (tabs/list/menu whatever it is)
3. Observed that Drop-down worked without any issues.

As per comment #0, it seems that there is no issue with drop down in espn.in(India region)and issue exist in USA site. Hence, forwarding the issue to MTV team for further triaging.

Thanks...!!
890638.mp4
2.0 MB View Download
Here we go again. I've attached Mac-INCOGNITO (Chromium 71.0.3566.0) VERSUS Mac-INCOGNITO (Chrome 69 release 69.0.3497.100)

Note it's the same exact issue on Windows (7 and 10). And same exact thing on other computers (Macs) - I switch between two Macs based on daily needs.
2018 10 01 ESPN Incoginito.png
104 KB View Download
2018 10 01 Chrome Incognito Okay.png
295 KB View Download
Components: -Blink Blink>Animation Blink>Compositing
Can reproduce fairly consistently on Windows 10. The popup menu container is given the correct size yet the layer is cut-off. It uses an animation (transition) to expand and it appears to stop after a frame or two in some cases. Removing the transition avoids the problem.

Labels: -Type-Bug -Pri-2 ReleaseBlock-Stable RegressedIn-70 FoundIn-71 FoundIn-70 OS-Linux OS-Windows Pri-1 Type-Bug-Regression
Owner: chrishtr@chromium.org
Status: Assigned (was: Unconfirmed)
Repros on Linux too with Beta channel (M-70).

Bisects on Linux to this (bisected twice to verify): https://chromium.googlesource.com/chromium/src/+log/1ad5c66907292930c378b882c1634fad9b2b29eb..f5e8bf6941f0fb31103ad926e86729d45b276e01

Of which the most likely candidate is:
 [PE] Stop dirtying compositing inputs on any layout for the root PaintLayer
https://chromium.googlesource.com/chromium/src/+/46520ae13478333fe882e79041b667e5a3112796

Sometimes it works before the page has fully loaded, or maybe before some onload handler has run. Given the bisect, maybe that's because something else is dirtying things at that point in the page load cycle.

Project Member

Comment 6 by sheriffbot@chromium.org, Oct 4

Cc: schenney@chromium.org
This issue is marked as a release blocker with no milestone associated. Please add an appropriate milestone.

All release blocking issues should have milestones associated to it, so that the issue can tracked and the fixes can be pushed promptly.

Thanks for your time! To disable nags, add the Disable-Nags label.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
Labels: Target-70 M-70
Labels: Hotlist-Polish
Cc: ligim...@chromium.org
Labels: -TE-NeedsTriageFromMTV -Needs-Bisect
Summary: Menu stripped off in espn.com (was: Re-reporting bug since was closed without action)
Updating summary and labels.
Just as a reminder, we are less than 1 week away from M70. Any updates yet on this bug?
Update #1: verified that reverting 

https://chromium.googlesource.com/chromium/src/+/46520ae13478333fe882e79041b667e5a3112796

makes the bug go away. Debugging now to find out where we are under-invalidating.
Cc: chrishtr@chromium.org
 Issue 893262  has been merged into this issue.
I found the root cause. It's that visual overflow is changing, and we are correctly
detecting that, but the PaintLayer whose visual overflow is updating is squashed
into another one that is not its descendant, and so the call to SetNeedsGraphicsLayerUpdate
in CompositingInputsUpdater does not result in a rebuild of the squashing layer.

CL in progress: https://chromium-review.googlesource.com/c/chromium/src/+/1274895
Let's hope this fixes a couple of other issues that I suspect are related, particularly http://crbug.com/893439
Fix is done in and in CQ.
Labels: Target-71
Project Member

Comment 17 by bugdroid1@chromium.org, Oct 11

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/2bd342fdf320e2e38bc48ef3a71355c8815b2b21

commit 2bd342fdf320e2e38bc48ef3a71355c8815b2b21
Author: Chris Harrelson <chrishtr@chromium.org>
Date: Thu Oct 11 19:44:37 2018

Set dirty bits for GraphicsLayerUpdater correctly for PaintLayers that are squashed.

Previously, a PaintLayer whose compositing inputs changed didn't propagate those
changes to the squashing layer, if the PaintLayer was squashed.

Bug:  890638 

Cq-Include-Trybots: luci.chromium.try:linux_layout_tests_slimming_paint_v2;master.tryserver.blink:linux_trusty_blink_rel
Change-Id: I0ad3bb79aa15e59eca9efd443f40cd5b66dcd583
Reviewed-on: https://chromium-review.googlesource.com/c/1274895
Commit-Queue: Chris Harrelson <chrishtr@chromium.org>
Reviewed-by: vmpstr <vmpstr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#598894}
[add] https://crrev.com/2bd342fdf320e2e38bc48ef3a71355c8815b2b21/third_party/WebKit/LayoutTests/compositing/squashing/overflow-change-squash-non-ancestor-expected.html
[add] https://crrev.com/2bd342fdf320e2e38bc48ef3a71355c8815b2b21/third_party/WebKit/LayoutTests/compositing/squashing/overflow-change-squash-non-ancestor.html
[modify] https://crrev.com/2bd342fdf320e2e38bc48ef3a71355c8815b2b21/third_party/blink/renderer/core/paint/compositing/compositing_inputs_updater.cc
[modify] https://crrev.com/2bd342fdf320e2e38bc48ef3a71355c8815b2b21/third_party/blink/renderer/core/paint/compositing/compositing_inputs_updater.h

Labels: Merge-Request-70 Merge-Request-71
Requesting merges. Will test on Canary tomorrow and also update.
Project Member

Comment 19 by sheriffbot@chromium.org, Oct 11

Labels: -Merge-Request-71 Hotlist-Merge-Review Merge-Review-71
This bug requires manual review: We don't branch M71 until 2018-10-11.
Please contact the milestone owner if you have questions.
Owners: benmason@(Android), kariahda@(iOS), kbleicher@(ChromeOS), govind@(Desktop)

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
Project Member

Comment 20 by sheriffbot@chromium.org, Oct 11

Labels: -Merge-Request-70 Merge-Review-70
This bug requires manual review: We are only 4 days from stable.
Please contact the milestone owner if you have questions.
Owners: benmason@(Android), kariahda@(iOS), geohsu@(ChromeOS), abdulsyed@(Desktop)

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
Let's verify this in canary tonight. 
Verified working fine on Canary Mac 71.0.3578.0 / Retina.
Also, I think this merge is very safe. It only adds more cases of setting dirty bits to the system, which can only result in additional recalculations and no crashes.
Labels: -Needs-Triage-M71 -Merge-Review-71
Removing merge request for M71. The fix made it in time for the branchpoint.
Labels: -Merge-Review-70 Merge-Approved-70
branch:3538
Project Member

Comment 26 by bugdroid1@chromium.org, Oct 12

Labels: -merge-approved-70 merge-merged-3538
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/4c77f77c82ed718b2e9163a4a7c99c0554e2b466

commit 4c77f77c82ed718b2e9163a4a7c99c0554e2b466
Author: Chris Harrelson <chrishtr@chromium.org>
Date: Fri Oct 12 23:01:03 2018

Set dirty bits for GraphicsLayerUpdater correctly for PaintLayers that are squashed.

Previously, a PaintLayer whose compositing inputs changed didn't propagate those
changes to the squashing layer, if the PaintLayer was squashed.

Bug:  890638 

TBR=chrishtr@chromium.org

(cherry picked from commit 2bd342fdf320e2e38bc48ef3a71355c8815b2b21)

Cq-Include-Trybots: luci.chromium.try:linux_layout_tests_slimming_paint_v2;master.tryserver.blink:linux_trusty_blink_rel
Change-Id: I0ad3bb79aa15e59eca9efd443f40cd5b66dcd583
Reviewed-on: https://chromium-review.googlesource.com/c/1274895
Commit-Queue: Chris Harrelson <chrishtr@chromium.org>
Reviewed-by: vmpstr <vmpstr@chromium.org>
Cr-Original-Commit-Position: refs/heads/master@{#598894}
Reviewed-on: https://chromium-review.googlesource.com/c/1279346
Reviewed-by: Chris Harrelson <chrishtr@chromium.org>
Cr-Commit-Position: refs/branch-heads/3538@{#990}
Cr-Branched-From: 79f7c91a2b2a2932cd447fa6f865cb6662fa8fa6-refs/heads/master@{#587811}
[add] https://crrev.com/4c77f77c82ed718b2e9163a4a7c99c0554e2b466/third_party/WebKit/LayoutTests/compositing/squashing/overflow-change-squash-non-ancestor-expected.html
[add] https://crrev.com/4c77f77c82ed718b2e9163a4a7c99c0554e2b466/third_party/WebKit/LayoutTests/compositing/squashing/overflow-change-squash-non-ancestor.html
[modify] https://crrev.com/4c77f77c82ed718b2e9163a4a7c99c0554e2b466/third_party/blink/renderer/core/paint/compositing/compositing_inputs_updater.cc
[modify] https://crrev.com/4c77f77c82ed718b2e9163a4a7c99c0554e2b466/third_party/blink/renderer/core/paint/compositing/compositing_inputs_updater.h

Labels: Merge-Merged-70-3538
The following revision refers to this bug: 
https://chromium.googlesource.com/chromium/src.git/+/4c77f77c82ed718b2e9163a4a7c99c0554e2b466

Commit: 4c77f77c82ed718b2e9163a4a7c99c0554e2b466
Author: chrishtr@chromium.org
Commiter: chrishtr@chromium.org
Date: 2018-10-12 23:01:03 +0000 UTC

Set dirty bits for GraphicsLayerUpdater correctly for PaintLayers that are squashed.

Previously, a PaintLayer whose compositing inputs changed didn't propagate those
changes to the squashing layer, if the PaintLayer was squashed.

Bug:  890638 

TBR=chrishtr@chromium.org

(cherry picked from commit 2bd342fdf320e2e38bc48ef3a71355c8815b2b21)

Cq-Include-Trybots: luci.chromium.try:linux_layout_tests_slimming_paint_v2;master.tryserver.blink:linux_trusty_blink_rel
Change-Id: I0ad3bb79aa15e59eca9efd443f40cd5b66dcd583
Reviewed-on: https://chromium-review.googlesource.com/c/1274895
Commit-Queue: Chris Harrelson <chrishtr@chromium.org>
Reviewed-by: vmpstr <vmpstr@chromium.org>
Cr-Original-Commit-Position: refs/heads/master@{#598894}
Reviewed-on: https://chromium-review.googlesource.com/c/1279346
Reviewed-by: Chris Harrelson <chrishtr@chromium.org>
Cr-Commit-Position: refs/branch-heads/3538@{#990}
Cr-Branched-From: 79f7c91a2b2a2932cd447fa6f865cb6662fa8fa6-refs/heads/master@{#587811}
Status: Fixed (was: Assigned)

Sign in to add a comment