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

Issue 787286 link

Starred by 1 user

Issue metadata

Status: WontFix
Owner:
Closed: Nov 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux , Android , Windows , Mac
Pri: 1
Type: Bug-Regression



Sign in to add a comment

The web page uses javascript scrolling by default, which tends to get choppy or Jump

Reported by laiyong...@gmail.com, Nov 21 2017

Issue description

Steps to reproduce the problem:
1. goto https://uland.taobao.com/new_cat.htm?spm=a311n.9159044%2Fc.10803055.1&pid=mm_26632386_9834707_32742010
2. scroll in the page
3. 

What is the expected behavior?
Normal scroll

What went wrong?
The web page uses javascript scrolling by default, which tends to get choppy or Jump

Did this work before? Yes 38.0.2125.102

Chrome version: 63.0.3230.0  Channel: n/a
OS Version: 7.0
Flash Version: 

This page use point event, but the old version of the support is not so good.

When touch move, the chromium produce kTouchScrollStarted event,then the blink will calling PointerEventManager::BlockTouchPointers function and in_canceled_state_for_pointer_type_touch_ set to true,so it can not produce point event to js.

I have a solution, when the pointdown/pointmove event  script handler calling preventDefault,  do not calling PointerEventManager::BlockTouchPointers function.
 
Screenrecording_20171121_104318.mp4
6.4 MB View Download
Screenrecording_20171121_111257.mp4
2.0 MB View Download
Cc: msrchandra@chromium.org nyerramilli@chromium.org ligim...@chromium.org sandeepkumars@chromium.org
Components: -Platform>DevTools Blink>Scroll
Labels: -Pri-2 hasbisect-per-revision M-64 Needs-triage-Mobile Triaged-Mobile OS-Linux OS-Mac OS-Windows Pri-1
Owner: mustaq@chromium.org
Status: Assigned (was: Unconfirmed)
Tested the issue in Android and could reproduce the issue.

Steps Followed:
1. Launched Chrome Browser.
2. Navigated to  https://uland.taobao.com/new_cat.htm?spm=a311n.9159044%2Fc.10803055.1&pid=mm_26632386_9834707_32742010 
3. Observed the scrolling is a bit choppy

Chrome versions tested:
62.0.3202.94, 64.0.3273.0

OS
Android 7.0.0

Android Devices
7.0.0; SM-J710F Build/NRD90M

Below is the bisect info
=======================
Chrome Good Build -- 55.0.2875.0
Chrome Bad Build  -- 55.0.2877.0

Manual CL:
https://chromium.googlesource.com/chromium/src/+log/55.0.2875.0..55.0.2877.0?pretty=fuller&n=10000

You are looking for a change made after 421810(GOOD), but before 421811(BAD).

From the above revision range suspecting the following change have fixed the issue and assigning the issue to the owner concerned--
https://chromium.googlesource.com/chromium/src/+/902a3d6303dfbfd49a24173777afcb4b70234a17

@mustaq -- Could you please look into the issue.

Please navigate to below link for log's and video--
go/chrome-androidlogs/787286

Issue is seen in Desktop as well using #64.0.3273.0.

Thanks!

Comment 3 Deleted

hi,@mustaq.

I have a solution, when the pointdown/pointmove event script handler calling preventDefault, do not calling PointerEventManager::BlockTouchPointers function. Can I submit a patch to solve this problem?

Thanks!

Comment 5 by mustaq@chromium.org, Nov 21 2017

Components: Blink>Input
Labels: -Needs-triage-Mobile Hotlist-Input-Dev
Status: WontFix (was: Assigned)
Hi laiyonggao@: this is not a Chrome bug.  An ongoing PointerEvent stream must be blocked (after firing a 'pointercancel')) when browser action such as scrolling starts.

Looks like the site is calling preventDefault() on a 'pointerdown' to stop browser scrolling.  This is not the correct approach for PointerEvents.  To prevent any browser action (scrolling/zooming), add the style "touch-action: none" to the JS-scrolled element.

Here is the spec if you are interested:
https://w3c.github.io/pointerevents/#declaring-candidate-regions-for-default-touch-behaviors

---

I have verified that adding "touch-action: none" to
<div class="catGroupItem" data-spm="01110" data-spm-max-idx="9">
fixes the site's scrolling problem.


Hi,mustaq.

Thank you for your reply and suggestion.

Sign in to add a comment