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

Issue 612656 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 602959
Owner:
NOT IN USE
Closed: Jun 2016
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux , Windows , Mac
Pri: 3
Type: Bug-Regression



Sign in to add a comment

font-feature-settings & -webkit-scrollbar-thumb:hover/active trigger weird repaint

Reported by adennis...@gmail.com, May 18 2016

Issue description

Chrome Version       : 50.0.2661.102
OS Version: 10.0
URLs (if applicable) :
Code:
http://jsbin.com/hekagugitu/edit?html

Demo (turn on paint flashing):
http://output.jsbin.com/hekagugitu

What steps will reproduce the problem?
1. Use ::-webkit-scrollbar-thumb:hover and font-feature-settings (as seen in code example)
2. mouseenter/mouseleave window with mouse cursor.

1. Use ::-webkit-scrollbar-thumb:active and font-feature-settings (as seen in code example)
2. mousedown/mouseup on window with mouse cursor.

What is the expected result?
There should be no repaint

What happens instead of that?
Repaint

Please provide any additional information below. Attach a screenshot if
possible.

UserAgentString: Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/50.0.2661.102 Safari/537.36



 
Cc: brajkumar@chromium.org
Labels: Needs-Feedback
adennisjin@ - By opening the provided link unable to see any content in the output section of JS Bin. Will you mind in providing the screenshot or screen-cast of the actual and expected behavior of the issue to test it from chrome TE end.

Thanks!
First file is just a html file with the code from jsbin. Open it normally in your web browser.

Second file is a video. In the video, I have dev tools open and have enabled paint flashing. As I move my mouse (and click around), the page seems to repaint.
bug.html
447 bytes View Download
2016-05-18_23-00-24.mp4
143 KB Download
Project Member

Comment 3 by sheriffbot@chromium.org, May 19 2016

Labels: -Needs-Feedback Needs-Review
Owner: brajkumar@chromium.org
Thank you for providing more feedback. Adding requester "brajkumar@chromium.org" for another review and adding "Needs-Review" label for tracking.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
Labels: -Type-Bug -Needs-Review hasbisect OS-Linux OS-Mac Type-Bug-Regression
Owner: ----
Status: Untriaged (was: Unconfirmed)
Able to reproduce on Windows 7, Ubuntu 14.04 and Mac OS 10.11.4 using chrome stable M50-50.0.2661.102.

Bisect Information:
=====================
Good build: 50.0.2628.0  
Bad Build : 50.0.2630.0  

Note: This issue is not seen on latest canary M53-53.0.2746.0

Change Log URL: https://chromium.googlesource.com/chromium/src/+log/a0728f758b8a845c8153b6d54948958c2fdc0d16..7deab7b0ed91ae7e453919ceef37e1b6074b7c1a

v8-autoroll URL:
https://chromium.googlesource.com/v8/v8/+log/3fe9b784..3614a9cf

Unable to find the actual suspect from the above logs, Could anyone please look in to this issue and kindly assign it to the concerned owner for further updates.

Thanks!
Cc: r...@opera.com
Components: Blink>Paint
Possibly rune's change - https://chromium.googlesource.com/chromium/src/+/95a3ba3929113edff393f11d0b7dd830910696e6
Owner: r...@opera.com
Status: Assigned (was: Untriaged)
Sounds like the bug was fixed. Rune could you confirm it was your change, and that
the underlying issue was resolved?

Comment 7 by r...@opera.com, Jun 21 2016

Cc: -r...@opera.com chrishtr@chromium.org
Mergedinto: 602959
Status: Duplicate (was: Assigned)
The underlying issue shows with this stylesheet as well (no scrollbar pseudos):

<style>
    html { font-feature-settings: "kern", "liga" 0, "calt" 0 }
    #nomatch *:active { background: pink }
</style>

Which causes a full document recalc on :active changes. We mark all elements in the document as affected by :active (because we do so during selector matching before we see the #nomatch), but the full recalc is most likely caused by a comparison of font-feature-settings which incorrectly returns false causing an inheritance propagation from the html element down the whole tree.

I think I fixed the underlying issue here: https://codereview.chromium.org/1887613002

Sign in to add a comment