- New interpolators
- New duration
- Removed workspace translation
- Removed overlapping fades -- are now sequential
See video in the bug.
Bug: 227745955
Test: manual
Change-Id: I19fa19a3806e963a04c5e694a52d3c1b921cf51f
- Avoid overriding interpolator in AllAppsTransitionController.setStateWithAnimation as it's no longer needed and it'll wrongly override interpolator for ANIM_ALL_APPS_FADE
- Override ANIM_ALL_APPS_FADE to FINAL_FRAME in QuickstepAtomicAnimationFactory for tap deadzone to dismiss animation, also added EMPHASIZED_ACCELERATE for the dismiss animation
- Tuned dismiss animation across form factors to 300ms
Fix: 220336617
Test: manual
Change-Id: I4b3e827b503dcb1dd39f0bd99d4c1dd5ffdba0f3
Context: there was a bug where you could get stuck in HintState if you
did the following (timing is critical):
1. Short swipe from nav region towards HintState, but not far enough or
fast enough to commit before letting go; this cancels the state
animation, returning towards Normal (but, crucially, StateManager
still has state set as Hint)
2. While previous animation is animating back to Normal, swipe up again,
but this time faster/farther to actually reach Hint; this time, the
animation does go towards Hint, but gets stuck there. The reason it
gets stuck is because StateManager thinks we're already in Hint from
step 1, so doesn't call onStateTransitionEnd(Hint) in step 2. Thus,
we never get QuickstepLauncher#onStateSetEnd(Hint), which is what we
rely on to return to Normal.
Fix is to have StateManager change its internal state to
mCurrentStableState (the state the transition started from) if the
animation is canceled. (Also need to keep that state if restarting the
animation, which AbstractStateChangeTouchController does in onDragEnd,
regardless of whether it ends up going to mFromState or mToState.)
Test: short swipe followed immediately by fast fling from nav region on home successfully goes to HintState and back to Normal
Fixes: 228276181
Change-Id: I2e3aeac06d482b57729416d5de55cc6ffc9df23c
- Make AllApps bottom sheet solid and appears from bottom
- Teleport AllApps bottom sheet as user drag to reduce drag range
- Consider teleport interpolation for state transition sdetection
- Tuned workspace motions for AllApps bottom sheet (no translate, shrink)
- Add portrait vertical translate for tablet portrait including taskbar AllApps
- Updated bottom sheet handle and created common variables for other bottom sheets
Bug: 208599118
Test: manual on tablet AllApps, taskbar Allapps and handheld AllApps
Change-Id: I69dba5f155914cd012cc8ef3be1ef71fb2be5a40
* Copied logic mostly from
Base/RecentsViewStateController
* There's one KI crash specific to 3P
(b/209694087)
* Swipe to home animation is TODO, but
then again it also seems incorrect for
full-screen app. 3P has many cosmetic issues
throughout.
Bug: 195607777
Test: Tested w/ Nova launcher with grid
and non-grid, able to invoke split screen with
same animations as 1P launcher
Change-Id: I35a1bc92a51caccac4214289e614000cd47bb503
- SysUI removes SYSUI_STATE_IME_SHOWING when starting a gesture from an app, but because unstashing has implications on the gesture transition (e.g. clips the bottom of the app), we defer handling the ime hiding until the gesture settles. Repurposed the flow that swaps the taskbar background during the gesture to support this case as well.
- Delay the unstash when IME is closing, to align with the end of the IME exit transition
- Remove TaskbarViewController.ALPHA_INDEX_IME now that we stash when IME is opening, since stashing already hides the taskbar icons
- Also support passing a starting progress to the stashed handle reveal animation, to allow it to be reversed when cancelled. For example, when returning to an app that has IME showing, we first start unstashing because we're in an app, but then we get the signal that IME is attached so we stash again almost immediately (within a frame or two).
Test: In both 3 button and fully gestural, open a keyboard in an app, ensure taskbar gets out of the way and then reappears at the end when the keyboard is dismissed
Bug: 202511986
Change-Id: I93c298a98ba369ea6310466ff3f802231c582687
- Uses overscroll damping logic to reduce the velocity
- The start to target rect interpolation can be from the
start, center, or bottom of the rect depending on where the item
is on the workspace. This reduces the amount of distance needed to
travel between, which helps further reduce gesture feedback.
Bug: 173107751
Test: test closing app that is on:
- top row of home screen
- middle of home screen
- in hotseat
Change-Id: I055dd61ca3491807109ff2f6c501bf710c8d340f
Bugs tracked in b/181704764
Bug: 181704764
Test: Tested on phone and large screen in multiple orientations
Change-Id: I07509006ae3d1f4425dc5119d0c8ed52b41a3bc2
- Update ic_corp_off icon to outline
- Remove shadow from work toggle
- Translate work fab when keyboard is shown
Screenshot: https://screenshot.googleplex.com/593tEg7bE4kSS4y
Bug: 191251404
Bug: 191250785
Test: local
Change-Id: Ie7dddfd17eb90575a1e1f67e281070dd8d268f8d
Previously we were only updating visibility in onAnimationSuccess(), which doesn't handle the canceled case.
Test: swipe up from nav bar twice in all apps, ensure all apps is invisible/non-interactable
Bug: 190315266
Change-Id: Ic59e6d4b404de4ee2d17c45d6fa59e2aee70c65f
Override onAnimationEnd(Animator), rather than (Animator, boolean)
which may not be called.
Test: Swipe right to quick switch from home, then swipe back left
(in the same gesture), ensure next quick switch isn't broken.
Bug: 185554764
Change-Id: I53e7526ed5b55fbae9aa654ace70a6ef7561e1d8
Bug: 175329686
Test: long press on apps where container opens above & below
and where container is aligned left & right
Change-Id: I09bbf57b6e19cc6f022b9a96fb48d8c4ae080813
Make it possible to set the color of workspace scrim to a different color
per state. Motivated by making Overview Scrim and All Apps scrims different
colors.
Bug: 186253733
Test: Local build and flash
Change-Id: Id7c38ce3c9173308eedfcb7592ececa7bd6bf220
- We remove the fling completely
- We set damping to 1 for x/y springs so there is no
bounciness, and updated the stiffness and bounds to tighten
up the path to the final location.
- During the animation, we translate all launcher content
down
- When the x/y springs, and rect animator all finish,
we use a spring to bounce the launcher content back up
- Added AppCloseConfig so that the entire animation can
be defined in one location, with getter methods so that
all the involved parties of the animation can access the
current value.
- The animations are all defined linearly, and then
interpolated over using a 3 point curve.
Building behind feature flag as we tune the values.
Bug: 173107751
Test: manual, visual
Change-Id: I83ad0fa2c4234cf30004240d43e191354595adc8
- Play the nav bar fade-out animation at the same time when the app
launching animation starts and make the fade-in animation ends at the
same time when app launching animation ends.
- To make the nav bar fade-in animation looks like it's attached to the
app, apply crop rect and translation that the app targets apply to the
nav target.
Bug: 181638132
Test: manual: click app icon on launcher to launch an activity and
observe the navigation bar animation.
Change-Id: If7e610eca5fccbb747a76c87335a600b018195a6
- Grid scaling has problem that scales task icon and menu, which makes it hard to control icon size and text size in the manula
- Replaced the whole concept with dedicated Task size calculation in grid layout
- Support different icon size in TaskView in grid and removed task_thumbnail_top_margin
- Removed grid progress in TaskViewSimulator as well
- Refactored how ClearAllButton scroll and translations are calcualted to align clear all properly in grid
- Make page center calculation aware of PagedView pivot and scaling
Bug: 174464863
Test: Manual on two screens
Change-Id: I47b13ef6e55c6e16c52ea04225f5bde02ed82fc2
This would simplify reusing the same code for 3-button mode
Test: Verified on device
Bug: 175137718
Change-Id: Ia757eec069efc251475baa6af38c690431325f66
- Map to progress 1 instead of dividing by 0 if expectedDuration == 0
- Map to starting progress one frame ahead based on velocity
- Change direction of velocity so it's in the same direction as the
task progress (0 is down and 1 is up).
Change-Id: I5ac3f8a0d6c616bd303ac1a902242964bb33d3c7
- Rename "pullback" to "resistance" to reduce confusion.
- Remove mDragLengthFactorStartPullback & mDragLengthFactorMaxPullback
- Add AnimatorControllerWithResistance, which has 2 controllers, one
for the normal shift to overview, then one to apply the resistance
when swiping beyond that.
- Don't hack animator interpolators/progress; insteaad, allow progress
to go > 1 (which will run the separate resistance animator).
- Don't start launcher controller separately from window controller;
instead, both are controlled by mCurrentShift in updateFinalShift().
- The resistance animation logic is shared by both the active window
and launcher (RecentsView).
Bug: 149934536
Change-Id: Ib0f9da18e10cc9ddf1a2f82ed767f237c89d3a41
This reverts commit a8c08584a7.
Reason for revert: "caused a regression with quick switch from home: if you start the gesture then swipe back to the left, it ends up launching the task anyway"
Change-Id: I8e12e2de46b6fc6a3faeb0336762da08080c61d6
With the second swipe, we never complete the swipe to Overview
NoButtonNavbarToOverviewTouchController#maybeSwipeInteractionToOverviewComplete
- mReachedOverview = true
- mDetector.isSettlingState = false
And then the second swipe starts the state transition to Hint but then
it never gets completed because:
1. The animation starts
2. Gets cancelled
3. Starts again
4. Finishes, but is not marked as success since the cancel in #2 was never
set back to false
Bug: 160759508
Change-Id: I8c3972e6209c3d5a4a0bdd9f9b7683de18105d57
Sunny Goyal
Luca Zuccarini
TreeHugger Robot
Alex Chau
Tony Wickham
Josh Tsuji
Vinit Nayak
Jonathan Miranda
y
Pat Manning
Winson Chung
Zak Cohen
Andy Wickham
shawnlin
Hyunyoung Song
Raman Tenneti