1326854 - [e10s][APZ] APZ scroll isn't smooth if I move mouse pointer

Status: VERIFIED INVALID

(Core :: Panning and Zooming, defect)

Description

[arni2033]

>>>   My Info:   Win7_64, Nightly 49, 32bit, ID 20160526082509
STR_1:
1. Download and extract attached "testcase 1" in a folder with short full name, then open .htm file
2. Execute in console code [1] (better) or code [2]
3. Keep rotating mouse wheel down and up and moving mouse pointer over the center of the page

AR:  Scrolling is not nearly smooth. Page just teleports up and down by 1 screen
ER:  Scrolling should be smooth, otherwise what's the point of that APZ thing?


STR_2:  (original)
1. Navigate to http://ecat.com/
2. Keep rotating mouse wheel down and up and moving mouse pointer over the center of the page


Notes:
1) Originally encountered this bug on http://ecat.com/  when site was lagging (2016.09.08 20-56-05)
2) I don't like whole APZ thing and don't see the point of creating huge checkerboarding areas

> [1] function F(n){D=Date.now()+700;while(Date.now()<D){}};addEventListener('scroll',F,false);
> [2] function F(n){D=Date.now()+400;while(Date.now()<D){}};addEventListener('scroll',F,false);

Comment 1

[arni2033]

Attachment: testcase 1 - APZ scroll isn't smooth when I move mouse pointer.zip

Comment 2

[Kartikaya Gupta (email:[e-mail])]

I unpacked your attachment to this URL for easier access: https://people-mozilla.org/~kgupta/bug/1326854/

Comment 3

[Kartikaya Gupta (email:[e-mail])]

Both your test case and the ecat.com website have a mousewheel event listener that reimplements scrolling in JS. It's not really using the APZ code. I copied your testcase to this URL:

https://people-mozilla.org/~kgupta/bug/1326854/fixed.html

and made exactly one change, which removes the inclusion of the smoothscroll.js script. On this modified version the bug no longer reproduces, which shows that the cause is the smooth scrolling they reimplemented on the page in JS.

As this is not a browser bug, closing as INVALID.

Comment 4

[arni2033]

I had no idea what is expected, but filed this bug just in case.