New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 678516 link

Starred by 3 users

Issue metadata

Status: Fixed
Owner:
Last visit > 30 days ago
Closed: Mar 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux , Windows
Pri: 1
Type: Bug-Regression



Sign in to add a comment

Regression:Unwanted white patches are observed after exiting the full screen in "files.unity3d.com"

Reported by adha...@etouch.net, Jan 5 2017

Issue description

57.0.2972.0 (Official Build) b1f498915edb2115c8f993e60e19728afda80acb-refs/heads/master@{#441559}(32/64-bit)
OS: Windows(7,8,8.1,10),Linux (14.04 LTS)

TEST URL:http://files.unity3d.com/jonas/webglManualTests/

What steps will reproduce the problem?
(1)Launch chrome and navigate to the above URL.
(2)Click on 'OK' button of the dialog box that appears.
(3)Now click on full screen icon and then press Esc.
(4)Repeat step(3) 2-3 times and observe.

Actual:Unwanted white patches are observed after exiting the full screen.

Expected:No such Unwanted white patches should be seen after exiting the full screen.

This is Regression issue broken in M-56,will soon update other info:

Good build:56.0.2918.0
Bad build:56.0.2919.0
 
Actual result.mp4
560 KB View Download
Expected result.mp4
233 KB View Download

Comment 1 by adha...@etouch.net, Jan 5 2017

Labels: hasbisect
Owner: khushals...@chromium.org
Status: Assigned (was: Unconfirmed)
Narrow Bisect info:
https://chromium.googlesource.com/chromium/src/+log/db94ad755cdc8dd3e4ee8cf4c7e7d43f52117ab7..93d3a4a026d6ad24981f394883b9947edda80bdf?pretty=fuller&n=100

Suspecting:r431772?

Kindly help to re-assign if your change is not the cause for this issue.
Cc: khushals...@chromium.org
Owner: pdr@chromium.org
That patch only removed code that was never used in Chromium anyway.

+pdr, is there a blink expert who could take a look at this? Nothing in that regression range looks related to this bug.

Comment 3 by pdr@chromium.org, Jan 9 2017

Cc: pdr@chromium.org
Labels: -hasbisect Needs-Bisect
Owner: ----
Status: Untriaged (was: Assigned)
I suspect this bisect range is wrong. Can you please re-bisect?

Comment 4 by adha...@etouch.net, Jan 10 2017

Labels: -Needs-Bisect hasbisect
Owner: mcasas@chromium.org
Status: Assigned (was: Untriaged)
Above issue is reproducible on latest chrome version i.e 57.0.2977.0

Re-Bisected the issue,below is the narrow bisect range:
https://chromium.googlesource.com/chromium/src/+log/83034cc5fb53f6d66f73f3a8afbfdd0d14b42943..12596183761900660c2f5a8208a861e0ba19e7e6?pretty=fuller&n=100

Suspecting:r423859?

Kindly help to re-assign if your change is not the cause for this issue.

Comment 5 by mcasas@chromium.org, Jan 10 2017

Owner: dalecur...@chromium.org
Status: Untriaged (was: Assigned)
My patch in the range only affected real time capture (i.e.
WebCam).  dalecurtis@ for further triaging.
Issue 679673 has been merged into this issue.
Cc: mlamouri@chromium.org
Owner: foolip@chromium.org
Status: Assigned (was: Untriaged)
=>foolip,mlamouri who have been working in this area lately. I don't think the bisect is correct.
foolip@, Gentle Ping!

Comment 9 by foolip@chromium.org, Jan 18 2017

I haven't had time to look into this I'm afraid. I will prioritize it after  issue 677827 .
Labels: Needs-Reduction
I will not have time to look into this before BlinkOn. But if by any chance it was caused by https://codereview.chromium.org/2573773002 then I am planning to revert that due a non-trivial regression that will need some redesign.

#4, are the repro steps for this 100% reliable? If they are not, then the identified range could be bogus. Adding Needs-Reduction to check if the repro is reliable, and if not reduce until it is and re-bisect.
Owner: wkorman@chromium.org
Re-bisected the issue,below is the Narrow bisect info:
https://chromium.googlesource.com/chromium/src/+log/8e0b2a7306378c41550a589dfbb8cefa0ebabca0..3c74f1cb2742e9b050fbc830f1ac0e06163c2370?pretty=fuller&n=100

Suspecting:r416191?

Kindly help to re-assign if your change is not the cause for this issue. 
Cc: junov@chromium.org
Components: Blink>Canvas
Status: Fixed (was: Assigned)
Apologies for the delay in looking into this. This looks fixed at ToT. Bisecting to find the fixing change produced:

https://chromium.googlesource.com/chromium/src/+log/da07d58a5d041395d3d6317ab3aef31c0087285e..9205e65c7dbfa96352e305ed7210946fc9c9e93c

From there the two most likely culprits are:

https://chromium.googlesource.com/chromium/src/+/7cfe269af863d9726b70c316ffbcc1f9f41dfe69
Fix box shadow rendering on opaque WebGL canvases

https://chromium.googlesource.com/chromium/src/+/5f32dc55b47bf2ca7eee0857ecd9eacdd9171116
Roll src/third_party/skia/ 7adde145d..905a94ff3 (5 commits).

It looks like 57.0.2981.0 is the first Chrome release that incorporates the above changes. Current stable Linux Chrome is 57.0.2987.110.

Given the fix is in stable, I don't believe there is any further work required to merge. Please let me know if otherwise or any questions.

Sign in to add a comment