- If any call to finish the recents animation comes in before
onRecentsAnimationStarted, we'll clear the callbacks but the listener
registered in startRecentsAnimation will persist. Since the call
to actually finish the animation with the system is posted on the
main thread, if the animation started callback returns in that time,
the controller will be set (via the listener) but the callbacks will
be unset. Instead, we can clear all the listeners from the previous
callbacks when cleaning up the animation.
Bug: 145758818
Change-Id: I29dab94f5fb25f762a823d59ef318a361282b595
* Adds additional fields to launcher_log.proto to capture smart folder related information.
* Uses ProtoLite to generate log object using builder pattern and converts to nano version before writing to clearcut. Hence not making drastic change to existing logging pattern.
Change-Id: I89b10da8d4e35e3abc7ddb553046946f91b43445
- Dump OverviewComponentObserver (to check if we're using 3P launcher)
- Dump current enabled and disabled plugins
- Dump all debug FeatureFlags instead of a few handpicked ones
Bug: 146174627
Change-Id: Ic8df899433b0e42d0e05eb1c80c87f62d3b936b6
Removed the code to handle table copying across dbs. We don't really need this since per ag/10298497, we can just copy db files directly.
Bug: 149236106
Test: N/A
Change-Id: I90bd7b6779fcaa841c71f5d0f27ec91907f926fc
When computing applying the rotation matrix to the TaskView thumnail,
use the rotation of the window configuration instead of the displays.
The forced rotation (ag/10176951) does not yet provide overrided values
for the display orientation when the requesting app is being forced
rotated, which is the case of the launcher. So
getDisplay().getRotation() would return the true rotation of the display
and not the rotation of the activity.
Test: with "adb shell settings put global forced_rotation 1"
and ag/10176951 and parents applied
1. Open landscape application
2. Display recents view
3. Check that the thumnail is in the same rotation as the opened
application
+ tests from I060a5dfcbcbb2ac21624d2ac778f76562bd60f40
Bug: 143053092
Bug: 143892437
Change-Id: Iee250459050b3c75dc3b79deed0682f809c7168b
It tests UiDevice.pressRecentApps(), which we don't use, so we don't
need to test it.
Bug: 131115533
Change-Id: I3a1b31f3b8368e3c848c888a5a27577faa940832
* Introduce FolderNameInfo class for passing down the folder name
suggestions from FolderNameProvider.
* Use FolderNameInfo for storing the serialized suggested names for
Folders. It is parsed and used in FolderEdit.
Bug: 148417030
Bug: 148916551
Bug: 148432151
Change-Id: Idaa81e203cc42889be15d0845230b4508521041c
The flake had disappeared, perhaps because of this logging, or,
hopefully, for some other reason.
Bug: 148313079
Change-Id: I636783d5fc71474dd443c143289c3ff74651835e
- Notifies prediction client when a folder is converted into workspace
item as a result of losing all items but one.
- Notifies predction client when a folder is created from an item in the
workspace with no duplicates.
Bug: 148749638
Test: Manual
Change-Id: Ifd51a6fe2a40c8baf4d88e2dba1e5bdc925e1608
I saw flakes when logcat didn't return records that are 300ms after the
specified time. I hope that moving the start time 1 sec back will
workaround this.
Change-Id: I6a4b66094d38f555d10284f19a71152a8be47b2e