New issue
Advanced search Search tips

Issue 680211 link

Starred by 0 users

Issue metadata

Status: Duplicate
Merged: issue 626748
Owner:
Closed: Apr 2017
Cc:
Components:
EstimatedDays: ----
NextAction: 2017-02-14
OS: ----
Pri: 3
Type: Bug

Blocking:
issue 619103



Sign in to add a comment

Css styles on :hover state of the element do not work as expected

Reported by 039...@gmail.com, Jan 11 2017

Issue description

Chrome Version       :  55.0.2883.87
URL :  https://bugs.webkit.org/show_bug.cgi?id=166618
Behavior in Safari 4.x/5.x: not tested
Behavior in Firefox 3.x/4.x: works as expected

What steps will reproduce the problem?
Please read the https://bugs.webkit.org/show_bug.cgi?id=166618, as I initially filed the issue I encountered there.


Demonstration of the issue : http://jsbin.com/dogoveziji/edit?html,css,output

 

Comment 1 by woxxom@gmail.com, Jan 11 2017

Seems just like the known bug caused by partial raster in Compositor (see  issue 492754 ). The problem is not present if I disable the feature by running Chrome 55.0.2883.87 with "--disable-partial-raster" in command line. No problem in the latest Chrome canary 57 out of the box.
Components: Blink>Compositing
Status: Untriaged (was: Unconfirmed)
NextAction: 2017-02-14
Status: Unconfirmed (was: Untriaged)
Dump from the WebKit bug (note this does not affect Safari).

Chrome and Opera do not reflow as expected on the hover state of an element (Demo: http://jsbin.com/dogoveziji/edit?html,css,output)


The browser where the problem appears 
1) navigator.userAgent 
Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/55.0.2883.87 Safari/537.36

WebKit Version =>537.36


2) Windows and Linux platforms were tested
Comment 1 Alexey Proskuryakov 2017-01-02 18:54:48 PST
I just opened the test page in Chrome 55.0.2883.95, and it looks the same as the attached screenshot. Could you please clarify what the problem is?

Also, please note that Chromium has forked WebKit years ago, so this issue may or may not occur in modern WebKit.
Comment 2 039706 2017-01-10 09:54:31 PST
Created attachment 298475 [details]
The behavior in Chrome
Comment 3 039706 2017-01-10 09:55:10 PST
Created attachment 298476 [details]
Behavior in Firefox
Comment 4 039706 2017-01-10 10:21:08 PST
Above I have attached 2 files to demonstrate the problem I am encountering.

The situation is the following: 
1) instead of using <table> tag to display the tabular data, we are using divs with display properties of "table" and "table-row" , "table-cell" accordingly. (this is done in order to achieve the fixed header effect)
2) it is needed that border-bottom change color when the user hovers over the row (that's where the issue in Chrome appears, as when moving the mouse from top to bottom the border-bottom color does not change, and when going from bottom up - it starts acting weird)
3) In order to make border change color on hover state, I use the following CSS rule (*ISSUE)
el:hover{
     border-bottom-color: blue;
}
4)This is how I currently avoid that problem (*WORKAROUND)
el:hover{
     border: 1px solid blue;
     border-left:none;
     border-right:none;
   }




The problem appears in Chrome/55.0.2883.87. I would be glad to test it in Chrome 55.0.2883.95 (but as far as I could understand that version of Chrome is the current stable for MacOS, and I am using Windows)

Should I rather report the issue to the Chrome browser developers? 
(The reason I reported it here was that the reflow problem occurred not only in Chrome but also in Opera )
expected(firefox).png
8.0 KB View Download
chrome.gif
251 KB View Download
Labels: Needs-Feedback
This does not repro in M56 on Linux or Windows.

Please try installing Chrome Canary (which will not affect your existing Chrome settings) and see if the problem persists.


Owner: wangxianzhu@chromium.org
Status: Fixed (was: Unconfirmed)
Bisected this on linux64 just now since it's fairly easy to reproduce.

---
You are probably looking for a change made after 431741 (known bad), but no later than 431754 (first known good).
CHANGELOG URL:
  https://chromium.googlesource.com/chromium/src/+log/00899e685f753f1e8b0b2452bf374de5f0079feb..13d2c57f10e29de32f916e28ba2eda880f29775c
---


From that list I would suspect that this was fixed by https://chromium.googlesource.com/chromium/src/+/c5cc11e88930c6e3f5e08b330fb14b6b571a0404, but in general schenney@ is correct - this should be fixed in M56 onwards. M56 is currently beta but will be stable in a few weeks.



039706@gmail.com - please reopen if you can reproduce on Chrome M56 or newer.
It should not be https://chromium.googlesource.com/chromium/src/+/c5cc11e88930c6e3f5e08b330fb14b6b571a0404 which fixed the bug because the CL was reverted later. There should be another CL which fixed the bug.

Cc: schenney@chromium.org
Status: Available (was: Fixed)
Good catch. Looks like it isn't fixed - I can reproduce on Linux 56.0.2924.67 beta (64-bit) and Linux 57.0.2984.0 (Official Build) dev (64-bit), so unsure about comment #4.

Reproduction video attached.

Not sure who is best to own this; wangxianzhu@ feel free to remove yourself if you don't think you are that person :).
out-1.ogv
801 KB View Download
I can confirm that it does not reproduce when running Linux 57.0.2984.0 (Official Build) dev (64-bit) with the '--disable-partial-raster' flag.
Blocking: 619103
Components: -Blink>Compositing Blink>Paint>Invalidation
Labels: -Needs-Feedback
Status: Assigned (was: Available)

Comment 10 by 039...@gmail.com, Jan 21 2017

On Version 57.0.2987.4 canary (64-bit) I still have the same issue
Reduced test case: http://output.jsbin.com/muzoqeb


Mergedinto: 626748
Status: Duplicate (was: Assigned)

Sign in to add a comment