For fission, we might have an unscrollable OOP iframe inside an inactive scrollframe in the parent document. The hit-test would target the OOP iframe's root scrollframe, and APZ would be responsible for handing off any scroll attempts on that to the enclosing scrollframe. If the enclosing scrollframe is inactive, APZ might not even really know about it, and hand off to the wrong thing.

So when building the display list we should make sure to add metrics for inactive scrollframes that contain OOP iframes, so that APZ knows the right handoff target. That scrollframe doesn't need to be activated though; we can activate it lazily from the repaint request after APZ tries to scroll it. This scenario presumably won't get hit often enough to warrant the memory hit from activating the scrollframe.

Tentatively moving all bugs whose summaries mention "Fission" (or other Fission-related keywords) but are not assigned to a Fission Milestone to the "?" triage milestone.

This will generate a lot of bugmail, so you can filter your bugmail for the following UUID and delete them en masse:

0ee3c76a-bc79-4eb2-8d12-05dc0b68e732

Botond, is this still broken?

Yes, this is still an issue. I plan to look at this as part of the M7 APZ work after work on M6c issues wraps up.

I'm going to keep this as an M7 because it only affects scroll handoff, i.e. it doesn't actually prevent you from scrolling or interacting with anything (except perhaps in a pathological case of a scrollable subframe whose viewport is taken up entirely by an OOP iframe).

I am going to use this bug to add a (at least one) test for this issue, this issue itself has been fixed by bug 1675547 I assume.

Depends on D105698

https://hg.mozilla.org/mozilla-central/rev/73824cc3c66a
https://hg.mozilla.org/mozilla-central/rev/81bec1bb5b4e