Issue 826438  has been merged into this issue.

Same issue on Windows/Linux.
The issue did not occur in Firefox/Safari, no infinite loop.

Able to reproduce this issue on reported version 65.0.3282.140 and on latest canary 67.0.3381.0 using Mac 10.13.3.

This issue is seen in 64.0.3282.186 and from M60 onwards.  Attaching screencast for reference.

@Reporter: Please check the video and let us know if it is the same behavior you are observing . 

Deleted: 826437.mp4 5.8 MB

	826437.mp4 5.8 MB View Download

Correct, the behaviour is same and the key thing here that this is not only in Mac OS this bug also on Windows, Linux.

Please take a look at this issue on other OS.

Thank you for providing more feedback. Adding the requester to the cc list.

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

Able to reproduce this issue on Linux and Windows. As this issue is seen from M-60 considering this issue as Non-Regression, removing Needs-Bisect and marking as Untriaged.

Thanks! 

chaopeng@ please investigate.

The Infinite scroll is from auto scroll animation:

#2 0x7f8e51df868f blink::EventTarget::DispatchEvent()
#3 0x7f8e51da041b blink::ScriptedAnimationController::DispatchEvents()
#4 0x7f8e51da08d7 blink::ScriptedAnimationController::ServiceScriptedAnimations()
#5 0x7f8e51cc3356 blink::Document::ServiceScriptedAnimations()
#6 0x7f8e529b2eb8 blink::PageAnimator::ServiceScriptedAnimations()
#7 0x7f8e529b8a3d blink::PageWidgetDelegate::Animate()

I think we should not dispatch this scroll to js.
https://cs.chromium.org/chromium/src/third_party/WebKit/Source/core/dom/ScriptedAnimationController.cpp?rcl=2ed30ffa70000045d888892851a3e7d69b181a7b&l=175

But why does it DispatchEvents? Presumably ServiceScriptedAnimations should not have ScheduledItems if we're  not currently doing an animated scroll. When I repro this, I'm using trackpad scrolling and I see the scroll event dispatched when I stop scrolling.

Because the JS change the item height then layout and scroll position updated by PageAnimator.

Rhino, Nitro and Chakra engines don't have this issue, but I haven't compared them with V8.

Ah, changing the height alone shouldn't change the scroll offset but we do try to anchor the scroller which means we might change the scroll offset due to layout. Is this where the scroll change is coming from?

I thought Scroll Anchoring tries to avoid this kind of infinite loop. skobes@?

+skobes@ ^^^

Scroll anchoring avoids some infinite loops, but not all.  If a scroll occurs while the first <li> is fully offscreen, we will perform an anchoring scroll in response to the height change.

But this behavior isn't new, so if it worked in M64 we might need a bisect.

It also not works in M64, Tested in 64.0.3240.0.