776310 - Tune flinging friction and logic in AsyncPanZoomController

Status: RESOLVED FIXED

(Core :: Graphics: Layers, defect)

Description

[Doug Sherk (:drs) (inactive)]

Flings run very slow in no-opt/debug builds and sometimes have logical errors like applying too much resistance during very large time steps that the fling actually goes backwards.

In addition, the frictions values are very far off from being optimal. Now that gralloc has landed on B2G, I can properly tune it. This will probably need additional work in the future.

Comment 1

[Doug Sherk (:drs) (inactive)]

Attachment: Proposed patch

Comment 2

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Comment on attachment 644714 [details] [diff] [review]
Proposed patch

>diff --git a/gfx/layers/ipc/AsyncPanZoomController.cpp b/gfx/layers/ipc/AsyncPanZoomController.cpp

>-  if (!mX.FlingApplyFrictionOrCancel(aDelta) && !mY.FlingApplyFrictionOrCancel(aDelta)) {
>+  bool shouldContinueFlingX = mX.FlingApplyFrictionOrCancel(aDelta),
>+       shouldContinueFlingY = mY.FlingApplyFrictionOrCancel(aDelta);
>+  // If we shouldn't continue the fling, let's just stop and repaint.
>+  if (!shouldContinueFlingX && !shouldContinueFlingY) {

Let's do

  if (mX.Blah() || mY.Blah()) {

Comment 3

[Doug Sherk (:drs) (inactive)]

I believe it may have been getting short-circuited.

Comment 4

[Doug Sherk (:drs) (inactive)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/8aa24637d3fa

Comment 5

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/8aa24637d3fa