1562793 - docs.oracle.com - Scrolling is janky

Status: RESOLVED WORKSFORME

(Core :: Layout: Scrolling and Overflow, defect, P3)

Description

[Karl Dubost💡 :karlcow]

(this might not be the right component)

This is a strange issue initially reported on https://webcompat.com/issues/27844

1. With Firefox Android
   Or with Firefox Desktop + RDM, make the window width (411 x 731)
2. Go to https://docs.oracle.com/cd/B28359_01/server.111/b28310/archredo002.htm
3. Scroll slowly to the bottom

Actual:
It jumps at a point to the bottom
When trying to scroll up after this, the site is becoming janky, and doesn't want to let you scroll up again, except if you do it very fast.

Expected:
It scrolls smoothly

It took me quite a lot of time to find the culprit.
The page is loading this script, which has a couple of functions.
https://docs.oracle.com/cd/B28359_01/dcommon/js/jqfns.js

After bisecting the script, I came up with this line creating the issue.

/*
Copyright 2014, 2016, Oracle and/or its affiliates. All rights reserved.
Version: 2016.6.4
*/

$(document).ready(function() {
  $(window).scroll(function() {
    var width = $("#NAV").width() + "px";
  });
});

This is really a reduced version of the script. It's possible to test it by using Arai-san Addon DITM which can help you replace a script by another.

Not sure this is a performance issue into computing the width, but this is happening.
Chrome doesn't have issue with this page.

Comment 1

[Emilio Cobos Álvarez (:emilio)]

Smells a lot like a scroll anchoring issue. Does the issue repro with layout.css.scroll-anchoring.enabled=false?

Comment 2

[Daniel Holbert [:dholbert]]

I can reproduce, and it does seem to me that it goes away if I turn off layout.css.scroll-anchoring.enabled. karl, can you confirm that this pref-flip makes the issue go away?

Comment 3

[Karl Dubost💡 :karlcow]

Daniel,
bingo! Yes it stops it.

Comment 4

[Emilio Cobos Álvarez (:emilio)]

Doesn't seem to be the case anymore, probably fixed by bug 1561450 or related. Mind confirming?

Comment 5

[Karl Dubost💡 :karlcow]

This is solved indeed.
Thanks emilio.