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

Issue 766836 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 755871
Owner:
Closed: Sep 2017
Cc:
Components:
EstimatedDays: ----
NextAction: 2017-10-04
OS: Windows
Pri: 1
Type: Bug-Regression



Sign in to add a comment

TD elements with opacity less than 1 render incorrectly

Reported by jscla...@gmail.com, Sep 19 2017

Issue description

UserAgent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/62.0.3202.18 Safari/537.36

Example URL:
https://jsfiddle.net/Waxen/tygkw668/

Steps to reproduce the problem:
1. View the linked jsfiddle
2. Note that in rows 3 & 4, columns 2 & 5 duplicate content from the previous column.

Note that all of the rows in the example have the same content in the TDs, only the styling differs. The exception being the last row, with has one empty TD and one with a literal space rather than nbsp

What is the expected behavior?

What went wrong?
It appears that TDs render with content from other elements when they:
 * contain only an nbsp
 * have an opacity less than 1

Does it occur on multiple sites: Yes

Is it a problem with a plugin? No 

Did this work before? Yes 60

Does this work in other browsers? Yes

Chrome version: 62.0.3202.18  Channel: beta
OS Version: 10.0
Flash Version: 

This was originally noticed with the jQueryUI datepicker.
On the "restricted dates" demo (https://jqueryui.com/datepicker/#min-max), view a month that ends on a day of week other than Saturday.

I had a coworker test this in 60 which worked, then update to 61 and begin seeing the issue.  I can reproduce in 62.0.3202.18
 

Comment 1 by kochi@chromium.org, Sep 20 2017

Labels: Needs-Feedback
NextAction: 2017-10-04
Thanks for the report.

I am having difficulty reproducing the problem as you describe.
I've tested on Windows (60 stable and 63 canary), and on Linux (ToT build
of content_shell), and they look same.
See the attached screenshot for the ToT build of content_shell on Linux.

Could you attach a screenshot when things went wrong?
screenshot.png
85.5 KB View Download

Comment 2 by kochi@chromium.org, Sep 20 2017

Cc: kochi@chromium.org
Cc: jmukthavaram@chromium.org
Labels: -Type-Bug -Pri-2 -Needs-Feedback hasbisect-per-revision M-63 Needs-Triage-M62 Pri-1 Type-Bug-Regression
Owner: enne@chromium.org
Status: Assigned (was: Unconfirmed)
Able to reproduce the issue on Windows 10 using chrome reported version#62.0.3202.18 & Canary#63.0.3219.0 as per below steps:

1. Launch chrome
2. Navigate to below URL
      https://jsfiddle.net/Waxen/tygkw668/
3. Observe 3rd & 4th rows 
4. Values 1 and 4 are populated & rendered in 2nd and 5th columns of 3rd & 4th rows 

Please find the attached screenshot of the exact issue.
    
Manual bisect info:
-----------------
Good-62.0.3176.0 - Revision-491883
Bad-62.0.3177.0 - Revision-492183

Per revision bisect info:
-----------------------
You are probably looking for a change made after 491914 (known good), but no later than 491915 (first known bad).
CHANGELOG URL:
The script might not always return single CL as suspect as some perf builds might get missing due to failure.
https://chromium.googlesource.com/chromium/src/+log/29eff3ab49027f70d179c783247568a3ffaacdac..e3d9eee91f3dfe5503bde5f3f4fc776e1552f857

Possible suspect:
-----------------
https://chromium.googlesource.com/chromium/src/+/e3d9eee91f3dfe5503bde5f3f4fc776e1552f857

enne@,Could you please take a look and reassign to the right owner if it is not related to your change.

Note: No issue observed on Windows 7, mac 10.12.6 & Ubuntu 14.04
Thanks.
766836.PNG
54.4 KB View Download

Comment 4 by kochi@chromium.org, Sep 20 2017

Thanks for the reproduction, jmukthavaram!
It's very appreciated.

Comment 5 by enne@chromium.org, Sep 20 2017

I am still trying to find a machine to reproduce this.

Can you let me know what about://gpu says about the win10 machine that reproduces this? Can you also try --disable-gpu-raster and see if it reproduces.  Similarly, can you try --disable-d3d11?

Comment 6 by jscla...@gmail.com, Sep 20 2017

I've attached my about://gpu output.

I don't seem to be able to run chrome with *any* CLI flags, but changing chrome://flags/#enable-gpu-rasterization from "Default" to "Disabled" does appear to fix the issue.  I'm not seeing a D3D11 flag on that page though, so I did not test that.



gpu.html
209 KB View Download

Comment 7 by jscla...@gmail.com, Sep 20 2017

Not sure what was going wrong before but I was able to test with the CLI flags now.

The issue does not reproduce with either --disable-gpu-raster or --disable-d3d11 set.

Comment 8 by enne@chromium.org, Sep 20 2017

Cc: bsalomon@chromium.org jmad...@chromium.org
Ok, so this sounds like another ganesh only / d3d11 only / intel bug.

Comment 9 by enne@chromium.org, Sep 20 2017

Components: -Blink Internals>GPU>Rasterization Internals>GPU>ANGLE
Labels: GPU-Intel
I can reproduce this on the same laptop as  Issue 755871 .
Mergedinto: 755871
Status: Duplicate (was: Assigned)
This is another manifestation of the driver clear bug in  Issue 755871 .
The NextAction date has arrived: 2017-10-04

Sign in to add a comment