1709723 - When starting a scroll in a direction that isn't scrollable along the handoff chain, overscroll the innermost scrollable scroll frame

Status: VERIFIED FIXED

(Core :: Panning and Zooming, defect)

Description

[Markus Stange [:mstange]]

Steps to reproduce:

1. Have a vertically scrollable <div> on a non-scrollable page. For example in the Slack or Matrix channel view.
2. Scroll the div to the bottom, if it's not yet at the bottom.
3. Initiate another scroll towards the bottom, inside the div.

Expected results:
The div should enter overscroll at the bottom.

Actual results:
Nothing happens.

Comment 1

[Botond Ballo [:botond]]

This is a recent regression.

Comment 2

[Botond Ballo [:botond]]

Confirmed to be from bug 1705280. A simple test page is https://theres-waldo.github.io/mozilla-testcases/div.html.

Comment 4

[Hiroyuki Ikezoe (:hiro)]

This will be superseded by bug 1704080. That means that once after we allow overscrolling for non-scrollable containers, in this example case panning on the div will hand off to the root non scrollable container and then the root one starts overscrolling. Since bug 1704080 is marked as a blocker for overscroll-90, I am going to leave this for now.

Comment 5

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1705280

Comment 6

[Botond Ballo [:botond]]

(In reply to Hiroyuki Ikezoe (:hiro) from comment #4)

This will be superseded by bug 1704080. That means that once after we allow overscrolling for non-scrollable containers, in this example case panning on the div will hand off to the root non scrollable container and then the root one starts overscrolling. Since bug 1704080 is marked as a blocker for overscroll-90, I am going to leave this for now.

Is that the desired behaviour, though? That, on a page like this (scrollable subframe + non-scrollable root), if the subframe is scrolled to the top, then scrolling upwards over the subframe overscroll the root and not the subframe?

ni?markus for thoughts on this

Comment 7

[Markus Stange [:mstange]]

I would much prefer the rubber-banding in this case to happen on the nested scroll frame, not on the page. For example, on Matrix, I don't want the sidebar to bounce while the mouse is over the channel contents. Scrolling the innermost scroll frame will also be a more consistent experience if a user scrolls to the bottom with a series of multiple quick scroll gestures: every scroll gesture will scroll/overscroll the same scroll frame, even at the end when the scroll frame has already hit the edge.

Comment 8

[Hiroyuki Ikezoe (:hiro)]

Attachment: Bug 1709723 - Hand off overscrolling to the root content APZC only if the APZC is scrollable in the same direction of the input. r?botond

Comment 9

[Hiroyuki Ikezoe (:hiro)]

I submitted a patch that it's the desired behavior I believe, i.e. overscroll handoff doesn't happen if the root content is not scrollable, it happens if the root content is scrolllable. I will write two gtests for this change, one is the root content is not scrollable either horizontally or vertically, the other is the root is only horizontally scrollable and verticall overscroll handoff doesn't happen.

Comment 11

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/8736e85d1c8f
Hand off overscrolling to the root content APZC only if the APZC is scrollable in the same direction of the input. r=botond

Comment 12

[Iulian Moraru]

https://hg.mozilla.org/mozilla-central/rev/8736e85d1c8f

Comment 13

[Andrei Purice]

Verified fixed on Firefox Nightly 90.0a1 (2021-05-11) (20210511093339) on iMac with OS 10.14.6 and MacBook Pro OS 11.2.3.
Still reproducible on Beta 89.0b10. Waiting for potential Uplift in Beta89 for further verification.

Comment 14

[BugBot [:suhaib / :marco/ :calixte]]

The patch landed in nightly and beta is affected.
:hiro, is this bug important enough to require an uplift?
If not please set status_beta to wontfix.

For more information, please visit auto_nag documentation.

Comment 15

[Hiroyuki Ikezoe (:hiro)]

We've decided to not uplift this.

Comment 16

[Andrei Purice]

Marking this as Verified Fixed based on the last comment from Hiro.