This makes the System size more consistent because cache is not
attributed to the system (when the cache is under quota).
Change-Id: I680e70daf5e98b9a205023a218dfd1dcc8ee8334
Fixes: 62623731
Test: Settings Unit Test
Due to issues w.r.t. attribution across multiple users, we originally
duplicated the previous implementation which zeroed out the sizes. This,
however, caused system bloat due to the change in how we calculate the
system size.
By attributing apps which do not exist in the primary profile to the
first user that shows up with it installed, we can avoid accidentally
attributing it to the system size.
Bug: 62623731
Test: Settings unittest & Settings robotest
Change-Id: I9514c9ecef62ea6270723f62e6bf27c69b75f180
If we do not have any recent scan results, show the previous APs from
when the app was last paused, by not triggering a force update.
Accompanying changes are also made in WifiTracker to prevent
onAccessPointsChanged callbacks from being invoked until the tracker
receives a new 'SCAN_RESULTS_AVAILABLE' broadcast (ag/2409026).
Bug: b/38212080
Test: runtest --path
packages/apps/Settings/tests/unit/src/com/android/settings/wifi/WifiSettingsUiTest.java
one minute later.
Change-Id: I4f9b2ec855e057e28235b0253ab42c6b4521bebc
Merged-In: I4f9b2ec855e057e28235b0253ab42c6b4521bebc
Fix several bugs related to storage accounting. Since getDataBytes()
already includes cached data, we need to subtract it to avoid blaming
apps for it.
We also need to blame app code on someone, so we blame it on the
current user. StorageStatsManager was fixed awhile back to only
return the app code size on the requested storage volume, so we can
remove the system app checks.
Subtract "appBytes" from external storage accounting, since it's
already been blamed elsewhere against specific apps.
Pass along storage results from all users on the device, and subtract
them all when estimating size of "system" data. To avoid embarrassing
estimation bugs, make sure that "system" data is at least 1GB.
Bug: 38008706
Test: cts-tradefed run commandAndExit cts-dev -m CtsJobSchedulerTestCases -t android.jobscheduler.cts.StorageConstraintTest
Test: cts-tradefed run commandAndExit cts-dev -m CtsAppSecurityHostTestCases -t android.appsecurity.cts.StorageHostTest
Change-Id: Ide1e6d0690e5ad4e751c87891f63ba1036434619
The override filter for storage to ensure that we filter by UUID and
user id was designed as an override filter. This worked as long as there
never was any other filters which could be used out of Storage Settings.
It turns out the instant apps filter applies here.
As a result, I've moved these filters into being "composite filters"
which are composed with the primary filter instead of overriding the
primary filter.
Fixes: 35245772
Test: Settings Robotests... but as for the composing part it is
b/34768986
Change-Id: I6f5acb909568ab02e81187c09fdda9533d635781
The other storage view used a compound filter with a UUID filter and the
normal filter. The category filters were omitting the compound UUID
filter. This adds it.
Test: Settings unit test
Change-Id: Ic0118abf58dabbf430a81e1fbcefb7acfdef472c
Fixes: 37780836
Static overlays aren't changeable by a user. So we shouldn't list them
as a theme.
This CL fixed b/36812704 as well to run unit tests.
Test: building succeeded and tested on sailfish.
make RunSettingsRoboTests
adb shell am instrument -w -e class \
com.android.settings.display.ThemePreferenceControllerTest \
com.android.settings.tests.unit/android.support.test.runner.AndroidJUnitRunner
Bug: 37480890
Bug: 36812704
Change-Id: I8c13b6956083095dceaab3da6642bef1647d6a30
This reverts commit b2bdc8f17b
and fixes a broken Espresso test case.
Reason for revert: Rolling forward after fix in ag/2133493, to fix b/37324736
Bug: b/37324736
Test: runtest --path
packages/apps/Settings/tests/unit/src/com/android/settings/wifi/WifiSettingsUiTest.java
Change-Id: I9cdf8cb1043af5d527eccbd9cc56d3b02419d7ac
If a user tells a device to forget a storage volume and then
goes back to the PrivateVolumeForget activity from their Recents, it
attempts to re-initialize itself with a volume that it forgot.
And this makes Settings crash.
By gracefully killing the activity when we try to open
PrivateVolumeForget for a forgotten volume, we avoid this crash.
Change-Id: Ib4e881c10f0c872ce6b268b16a573960230ef99b
Fixes: 34856304
Test: Settings unit test
This crash occurs because, even if you finish an Activity during
onCreate, the other lifecycle methods occur. In this case,
onActivityCreated assumes that onCreate ran properly.
By exiting early from onActivityCreated, we can ensure the activity
is finished properly.
Change-Id: Ia354341bf82d295c7dd042668fb11d588ddfebc0
Fixes: 37334861
Test: Instrumentation test
Bug: 35356645
Test: Verify for both config_nightDisplayAvailable true and false,
adb shell am instrument -e class
com.android.settings.display.NightDisplayTest -w
com.android.settings.tests.unit/android.support.test.runner.AndroidJUnitRunner
Change-Id: I7de3eddf62428d75bef621d3387fa33a76205fe7
It may crash if opened during the uninstall of an app. By catching
the exception which may occur, we can just skip the uninstalled app
and avoid crashing.
Change-Id: I1b96b0697f4041be356260d6c675593affc1cb69
Merged-In: If556db7b5a299ba53a29baefbbe9709ba6d12190
Fixes: 36793223
Fixes: 36793372
Test: Settings unit test
The recent upgrade to Mockito 2.7.13 has caused some problems
for grok (and new javac toolchain). The changed files used
wildcard imports to import the same methods from both Matchers
and Mockito.
This was not a problem previously because in the previous
version of Mockito the Mockito class extended the Matchers class
and so even though two methods with the same name were imported
they were not treated as a conflict because they were treated as
being the same method. In 2.7.13 the Mockito class no longer
extends the Matchers class instead they both extend the
ArgumentMatchers class and so for some reason javac no longer
treats them as the same method.
This was not detected by presubmit builds because jack is
still treating the methods as being the same.
Bug: 32912773
Test: m -j32 ANDROID_FORCE_JACK_ENABLED=disabled FrameworksWifiTests
Change-Id: I8c2b1dd39d2d653fe0fba19ed6421b7de25bb82f
This CL adds the APCT tests within this project to
a similar suite as CTS known as device-tests.
The current method of running APCT tests in the infrastructure
is unaffected.
Bug: 35882476
Test: `make dist device-tests -j` and local builds of
continuous_instrumentation_tests & continuous_native_tests
Change-Id: Ie1b2c6aefbe1497a2e90077a3ed8ab75721dc87b
The old Games flag was being ignored for categorization.
This caused the very few games which used the flag to be
miscategorized as "Other apps."
Change-Id: Ib8650d1c43579bc1698576c09c884e5a7bc4626d
Fixes: 36109122
Test: Settings robotest
The system apps' base size was not being properly attributed
to the system. This was causing the storage to go totally
unattributed (and causing roughly a 1GB gap in sizes.) On my
test device, the sizes now sum properly.
Change-Id: Ibc13015e40f854090a7a3ec09eef4c5c52918e0f
Fixes: 35144717
Test: Settings robo
Add calls into the OverlayManagerService to update the overlays
directly and remove reboot dialog.
Test: make RunSettingsRobotTests
Change-Id: I92841e643eab70b4b194f60cce95ce3cedb75972
Fixes: 34701451
BackupSettingsActivity should be the only activity listenting to the
BACKUP_AND_RESET_SETTINGS intent. It then decides which acitivity to
launch.
Bug: 35237721
Test: make SettingsUnitTests
Change-Id: I2021ec57e08c305771dd040c52725065597fd042
This replaces the calculation of the apps category
and add in a calculation for the games category using
the faster GID calculation. This should result in app
sizes loading roughly 10 times after.
Bug: 34204877
Test: Settings unit & robo tests
Change-Id: I78044a8d50f695f8c0a7e04183030232a9719260
In preparation for removing junit classes from the Android API
the legacy-test target will be removed from the
TARGET_DEFAULT_JAVA_LIBRARIES. This change adds explicit
dependencies on junit and/or legacy-android-test to ensure that
modules will compile properly once it is removed.
Bug: 30188076
Test: make checkbuild
Change-Id: I5d2c0685b74ea5ba368c777d9646b9d3ac24dca9
If the keyguard is already secure (i.e. if the backup screen lock is
already set up), do not show the skip dialog which warns user about
the danger of not having a screen lock.
Bug: 34129157
Test: adb shell am instrument -w com.android.settings.tests.unit
Change-Id: I6f777631487de89ab25a08ea017dd6194dde464d
