GeckoView java code wants to know the root content APZC's async scroll offset (and async zoom level). APZ has a mechanism to give GeckoView this information during composites, using [`UiCompositorControllerParent::NotifyUpdateScreenMetrics`](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/gfx/layers/ipc/UiCompositorControllerParent.cpp#214). However, we're a bit confused about what exactly we send it. Here are some of the things we could send it: 1. The latest, hot-of-the-press visual scroll offset which due to the frame delay isn't being composited yet. 2. The visual scroll offset being composited. 3. The main thread's last snapshot of the visual scroll offset. 4. The layout scroll offset. Currently, the WR codepath sends (1) and the Layers codepath sends (4). Bug 1519285 improves things slightly nad has the Layers codepath send (3). Based on discussion [here](https://phabricator.services.mozilla.com/D87163), I believe what we want to be sending it is (2) in both codepaths.
Bug 1659636 Comment 0 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
GeckoView java code wants to know the root content APZC's async scroll offset (and async zoom level). APZ has a mechanism to give GeckoView this information during composites, using [`UiCompositorControllerParent::NotifyUpdateScreenMetrics`](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/gfx/layers/ipc/UiCompositorControllerParent.cpp#214). However, we're a bit confused about what exactly we send it. Here are some of the things we could send it: 1. The latest, hot-of-the-press visual scroll offset which due to the frame delay isn't being composited yet. 2. The visual scroll offset being composited. 3. The main thread's last snapshot of the visual scroll offset. 4. The layout scroll offset. Currently, the WR codepath sends (1) and the Layers codepath sends (4). Bug 1519285 improves things slightly and has the Layers codepath send (3). Based on discussion [here](https://phabricator.services.mozilla.com/D87163), I believe what we want to be sending it is (2) in both codepaths.