Wasn't handling split-screen properly. Now it checks if the
home/recents is the top of its "visibility" stack vs top
of root-task. This means it must be the top sibling in either
the displayarea or the root-split. This check is necessary
for dealing with visible windows that are behind a translucent
task.
Bug: 159138687
Test: Enter split with 2 apps, close the top app, return to home
and verify that it receives touches in top half
Change-Id: Ie6fed2dd945952b83116721a98fc6d7c8396729f
If a carrier-associated app is added to the system image after the
device is initially launched (e.g. Q -> R OTA), the logic to disable
apps without a corresponding SIM will not run, leaving the app in an
enabled state in more cases than there should be.
To safely account for this, we only reevaluate on SDK change for apps
whose addedInSdk value falls within the range since the last time we
evaluated carrier apps' statuses. Apps that set lower (or higher)
SDK versions are ignored. We choose *not* to use minimum or target SDK
version here because it's more likely to change with each OTA /
prebuilt, which could cause unintended disable operations (or lack
thereof).
Bug: 154872019
Test: manual, QA, atest FrameworksTelephonyTests:CarrierAppUtilsTest
Change-Id: I97ddfa03f34e7d00e66ce7cd508d6d2a5e83fbc2
Previously, the sysconfig wasn't capable of understanding
carrier-associated apps that were added after a device's initial launch
(i.e. via OTA) because the logic in CarrierAppUtils explicitly avoids
disabling such apps a second time.
Most of this change is just plumbing everything through. For now, it's
all @hide due to R API deadlines. It will be made public in S.
Bug: 154872019
Test: manual, QA, atest FrameworksTelephonyTests:CarrierAppUtilsTest
Change-Id: I530a4f73146b09879547ca2e0c26428957fef37a
When resumption is disabled, we should only remove players when
timedOut is true.
Fixes: 158721393
Test: manual
Change-Id: I18b36c95ac72d5728ed203ac2b5d344b58676dc4
1. decouple Bubble from NotificationEntry
2. save title from Bubble into BubbleEntity
3. copied boolean values from NotificationEntry into Bubble for UI
variants (e.g. isVisuallyInterruptive, isClearable,
shouldSuppressNotificationDot... e.t.c)
Bug: 151474524, 159061670
Change-Id: I606c6ff93b3dc3867b4d0a6129d7117d9999c170
Test: atest android.app.cts.NotificationManagerTest
This reverts commit 2e73117c0e.
Reason for revert: issue fixed, rolling forward again
Change-Id: I1180954a76bee9782df46de0f202c614f05d560f
This change adds a null check on PackageSetting.pkg before passing it on
to the method that checks for the QUERY_ALL_PACKAGES permission on the
package.
Test: atest AppEnumerationTests AppsFilterTest
Bug: 150405193
Change-Id: Ic92e29c685ff215f1c54dbb4aa8a0c6377a30fa7
Setting an authentication IntentSender on an Augmented Autofill
suggestion was being ignored. With this change, it will trigger
the auth flow as documented.
Fix: 157863999
Test: Manual
Test: atest android.autofillservice.cts.inline
Change-Id: Id21c8f074bd0f49992e01445d50b1503af4720b6
+ Correctly synchronize a couple of methods
(used for debugging) which were missing the locking.
+ Log entries deleted during idle maintenance window
which happens at most once every 24 hrs.
Fixes: 158351456
Test: atest --test-mapping apex/blobstore
Change-Id: I5d1a1ffc28b009274d981d1ccdac30092d4c3732
* For inline autofill, we always need to get an IME callback before
calling the augmented autofill service. This is because although
autofill keeps a list of undestroyed sessions, the IME side
always discards the old callback for autofill when a new one
is created. So in case of switching back to an existing autofill
session, we need to make sure we ask the IME to return a new
callback.
* we current have the check which missed a case where we are
re-entering an existing autofill session, but the standard
autofill was returning null the previous time,
so it goes to trigger augmented autofill directly. In this case,
we need to call IME onCreateInlineSuggestionsRequest to get a
callback before calling the augmented autofill. We didn't, and
this patch fixes it.
Test: atest android.autofillservice.cts.inline
Test: manual
Bug: 158877106
Change-Id: Ie15cf0763ae49a204ad09c2eaac798890388622e
Enable permissions needed to write the trace file in the device.
Bug: b/155805394
Test: make MultiUserPerfTests
Change-Id: I650307133c5325aa92aa6e12f2a77566f314f174
This ensures that we only animate the stack temporarily invisible if we're not dragging something or a flyout is still showing.
This also adds a 2s delay before hiding the stack if it was visible due to drag/flyout, since a) it looks better b) that way you can open the stack back up if you didn't mean to let go.
Fixes: 159064066
Test: open fullscreen youtube, show status bar, drag around until status bar goes away, then fling bubbles somewhere and observe
Test: get a flyout while in fullscreen, touch the flyout until the status bar goes away, then collapse flyout
Change-Id: I06bc148161a1d51ee047b3861c1d01e7685c2943
Fixes the NullPointerException. Unknown permissions was skipped in
restorePermissionState() when handling changing regarding platform
permissions, so it's safe to skip it when generating missing
(platform) permission state as well.
Test: Flash a recent build without this fix
Test: Install an app which uses an unknown permission
Test: Manually remove app entry in runtime-permissions.xml with root
Test: Reboot device and confirm system server crash and stacktrace
Test: Flash a build with this fix, without clearing data
Test: Confirm system boots, log shows missing permission state was
Test: generated for the app and unknown permission was skipped
Bug: 158860316
Change-Id: Idfc9b5bf6c69772d41ccf781b84a4d101ed1277d
