New issue
Advanced search Search tips

Issue 755817 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 798535
Owner: ----
Closed: Aug 2017
Cc:
Components:
EstimatedDays: ----
NextAction: 2017-08-31
OS: Windows
Pri: 2
Type: Bug



Sign in to add a comment

mouseleave fired in error during rapid click events of child elements

Reported by johnsch...@gmail.com, Aug 16 2017

Issue description

UserAgent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/60.0.3112.90 Safari/537.36

Steps to reproduce the problem:
1. see prepared jsfiddle https://jsfiddle.net/janga_jack/9oqvcqd1/17/
2. furiously click on elements within the blue area.
3. watch all the click events register in the log. Eventually, if you follow a similar pattern of furious clicking on the li, label, input, and hr elements similar to the attached gif an errant mouseleave event will fire. Legitimate mouseleave events also fire and are picked up.

** Not found in Firefox-stable, or Edge-stable **

What is the expected behavior?
all mouse events triggered within the blue area are still child events of the div.wrapper and should bubble without triggering mouseleave, as I understand the internet.

What went wrong?
mouse events and actions within the div.wrapper element are causing a mouseleave event on that element. I don't believe this should happen, as all events are being triggered on child elements of the wrapper.

Did this work before? N/A 

Does this work in other browsers? Yes

Chrome version: 60.0.3112.90  Channel: stable
OS Version: 10.0
Flash Version:
 
ezgif.com-video-to-gif.gif
529 KB View Download
updated jsfiddle to https://jsfiddle.net/9oqvcqd1/20/ to support Edge, and any other browsers that do not support Element.prepend().

Comment 2 by hayato@chromium.org, Aug 16 2017

Components: -Blink>DOM Blink>Input
Blink>Input might be better for mouseleave.
Labels: Needs-Feedback
You are correct in what you expect here however I'm not able to reproduce it.

Are you able to provide us with a Input Latency trace from: chrome://tracing 

Have you tried this on different hardware? One explanation of such behavior is sometimes some hardware will generate phantom touch moves and if the touch position differs from the mouse position you'd see something like this.

NextAction: 2017-08-31
Sure. Heads up, this is my first filed chrome bug, and my diagnostic knowledge of the sub-system is garbage.

I tested a single system with 2 input devices.
- Logitech 810-002182
- Microsoft X08-71118

They are cheap mouses, but they have a wide distribution.

I included 2 traces per mouse,
- input-only: was recorded under the requested 'input latency' profile
- full: was recorded with extra features that may be helpful such as, blink, event, ui, v8, etc.

I made sure that each trace includes several invalid mouseleave events, but they will also have 1 legitimate mouseleave where the mouse exited the demo target to terminate the recording.

If these traces are not helpful I can test another system.
trace_mouseleave-Logitech-input-only.json.gz
1.1 MB Download
trace_mouseleave-MS-input-only.json.gz
1.4 MB Download
trace_mouseleave-Logitech-full.json.gz
1.6 MB Download
trace_mouseleave-MS-full.json.gz
1.3 MB Download
Project Member

Comment 6 by sheriffbot@chromium.org, Aug 17 2017

Cc: dtapu...@chromium.org
Labels: -Needs-Feedback
Thank you for providing more feedback. Adding requester "dtapuska@chromium.org" to the cc list and removing "Needs-Feedback" label.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
2 more important points:

1. I was able to recreate this bug in Edge today.
2. The speed of these mouse clicks is frantic, however, this report started with a list similar to my demo that was hiding itself on mouseleave. Very normal paced clicking in the list was causing spurious mouseleave events and causing the list to hide. So under different circumstances, the pace of mouse events is not an issue.
Labels: Needs-Feedback
Sounds to me this is a hardware issue. You might examine the native OS windowing events using something like Spy++. Since it is reproducible in Edge as well there isn't much we can do.
The trace clearly indicates that a mouse leave is being generated but isn't from a touch event. 

I wonder if you have any software that is generating these phantom leaves. Have you tried removing any non-standard software (like logitech drivers/plugins).
The NextAction date has arrived: 2017-08-31
Status: WontFix (was: Unconfirmed)
I also tested this and couldn't reproduce the issue on the latest stable. Closing this for now. Feel free to file a new bug or comment on this one if you found a more reliable way of producing this.
Mergedinto: 798535
Status: Duplicate (was: WontFix)

Sign in to add a comment