Step 2 in refactoring DatabaseIndexingManager.
This step moves the insertion of data into the database
into a new class. This removes the remaining bulk of the
code outside of DIM, but it does not fix the actual issue
with the indexing code.
The indexing code still chains functions together to
insert data into the database at the end of the functions.
It is exceedingly hard to read, and hard to track down bugs.
I would like the converter to eventually return a list of
IndexData objects, which lets us dissociate the database
from the data collection. I.e. we can store the database
in the Search app, and just pass IndexData objects via
IPC.Fixing this requires more of a refactor, and will be
done in a subsquent CL.
Bug: 33577327
Test: make RunSettingsRoboTests
Test: Took a database dump before and after change,
and they were the same. Cool.
Change-Id: Ia9bb815657b76f6cb9163014e746ec5eb6db8c5e
The first step in refactoring the god class,
DatabaseIndexingManager.
The class has one major entry point: indexDatabase
which begins a chain of calls that first collects all the
data from the fragments, and then massages that data into
the SQLite database. Unfortunately, most of the methods
do not return data, and just pass along some mutated
form of the data until it can be insterted.
Reading and testing this class is very difficult.
This first step moves the collection of the indexable data
into a new class which has a few benefits:
- The data can be easily mocked in tests
- Reduces complexity of D.I.M.
- Separates data collection from indexing, which allows the
indexable data to be piped into a new API that unbundled
search can consume.
Bug:33577327
Test: make RunSettingsRoboTests
Test: Grabbed a DB dump before change, compared to DB dump after change
to make sure everything is still indexed.
Change-Id: Ibc91e3d75ff5dcf5274b93b29bf3544f90b2194d
- Title set if defined in the Bundle returned by summaryUri's content
provider.
- summaryUri invoked inline with onBindView call
bug: 62713030
Test: Manually tested, verified that title and summary are retrieved
when settings app is resumed
Change-Id: Id82531eec5ec11ec3492f033fb34ec65a5437a48
- Never return null when querying userDictionaryLocales
- Auto mirror mobile setting icon in RTL to match status bar.
Fixes: 65298627
Fixes: 65361092
Test: robotests
Change-Id: I0f9827f7bc23baf4895712c0f86584aeccfb9c73
Fixes the bug that, on the new WifiDetailPreference page, the "Forget"
button is not disabled properly for wifi configs created and protected
by DevicePolicyManger.
Robolectric test to follow in a separate CL.
Test: manual, by locking down wifi config with TestDPC
Test: make RunSettingsRoboTests
Bug: 64971700
Bug: 65396674
Change-Id: I27740eabd5eb94415e4258c9c80f91df2d9ab476
If the user moves away from the storage fragment and returns, this
allows us to use cached data from the previous calculation. If the data
is > 1 minute old, we consider it stale. Otherwise, we can bypass the
loading screen.
Fixes: 37923463
Test: Settings Robotest
Change-Id: I7650d4d742852f8d447878c077b9190bc0a0bb22
The preset initial state helps eliminating animation jank when first
landing on the page.
Change-Id: Ia7ba83983f18409b1c653cc1ebb0f3aad281358c
Fixes: 64811322
Test: robotests
The symptom observed is that the Bluetooth master switch on the
Connected devices page doesn't properly respond to Bluetooth turning off
via quicksettings - either turning on airplane mode or just toggling
Bluetooth.
The root cause was that MasterSwitchPreference's isChecked method would
not return the true value of whether the switch was checked - if the
control is disabled, it always just returns false. This interacts badly
with code in BluetoothEnabler - we disable the switch when the Bluetooth
state is in transition (eg becomes STATE_TURNING_OFF), and we also
attempt to avoid calling setChecked if the switch is already in the
desired state. So the switch would be checked but disabled, and we'd
avoid ever calling setChecked(false) on it.
A thorough fix would be to remove the code from MasterSwitchPreference's
isChecked method that looks at the enabled state, since enabled and
checked really should be treated as separate concerns. But given the
timeframe of MR1, we're opting for a more conservative fix of directly
accessing the switch and checking it's state, to avoid introducing bugs
in other consumers that might be depending on the current
behavior. We'll then do the thorough fix on the master branch which will
give a lot more time for any unexpected issues to be found (I audited
other usages and none seemed likely to be a problem, but it's better to
be safe than sorry).
Change-Id: I19a6c6b71e74595be3ef32a9718a430b67a89d53
Bug: 64940731
Test: make RunSettingsRoboTests
We had special behavior for it in the past, but this defines new
behavior that is much closer to what the other storage preferences do.
A photo app filter is used and a photos/video files preference exists on
it which intents over to the gallery app.
Fixes: 64147318
Test: Settings robotests
Change-Id: I47284515fe2dfcc924ae61a44bc47051e9f5fda6
In the entity header layout, the action buttons resource is set to null by
default. When we bind the button with the app preference action, we
should also set the drawable to the settings icon as well.
Change-Id: Ic259b4c538f529671ca5a9c67664ef32fbbb25ae
Fixes: 64826061
Test: make RunSettingsRoboTests
For the work profile drilldown, we used to show all apps when the user
drilled down into the categories. This makes it so that the drill down
only shows the work apps when that deep.
Change-Id: I492cd3e9b9b923b87b68645a871dcfb2b91b4f95
Fixes: 62963093
Test: Settings robotest
This fixes the code in ClassScanner for finding all classes in a given
package to not depend on directory entries in the .jar files generated
by the build system. This dependency caused our tests in
CodeInspepectionTest.java to fail when this CL:
https://android-review.googlesource.com/#/c/platform/build/+/456418/
stopped adding directory entries in the .jar files generated by the
build process. Instead of depending on directories being present in the
list of resources provided by the classloader, this CL switches to using
Guava's ClassPath class to enumerate all loadable classes and filter
them to the ones in the package of interest.
Change-Id: I583919096450b61d4816256be280e2f5f1ce2316
Fixes: 64840107
Test: make RunSettingsRoboTests
If app is device or profile app, we disable the background check
toggle. This cl also create an util method for this check and
remove duplicate code
Bug: 64665807
Test: RunSettingsRoboTests
Change-Id: Id8336eadaac8832327bc3653aaa7dfbacde352ac
SearchFragment and SavedQueryLoader share a loader manager
and had been using the same loader id for different
loaders.
Occaisionally (in monkey tests) two loaders with the same
IDs would be started and crash when they finished.
The loaders now have different IDs.
Change-Id: I11e9b7365605fbcef44cf7d2323183415422f5c8
Fixes: 64756515
Test: robotests
Two issues:
1. UID of settings app == system UID. But isSystemUid failed to recognize
Settings in secondary user is also a system uid.
2. For USER drain type, we should launch the detail page in current
user.
Fix: 64506728
Test: make ROBOTEST_FILTER=AdvancedPowerUsageDetailTest -j40 RunSettingsRoboTests
Test: Switch to seconday user. Go to Settings->Battery, tap owner user
battery entry, observe that detail page is shown.
Test: Stay long enough in Settings app, make sure no
either Settings nor Android Framework battery entry.
Change-Id: I8d66ad55f18fcb3d9567b3bf753f737f5b98c609
* Add a developer menu option to allow name-less devices to be shown
when a Bluetooth developer needs it, but hide it for non-developer
users.
* Set BluetoothDevicePreference to invisible when CachedBluetoothDevice
does not have a name besides MAC address and the above developer option
is false.
* This affects BluetoothPairingDetail and DevicePickerFragment, but does
not affect BluetoothSettings. BluetoothSettings will show all paired
devices regardless whether an user friendly name exists.
Bug: 34685932
Test: pair Bluetooth device, send file over Bluetooth, unit tests
Change-Id: Idd7ad4b1671dfdcf3204efb50eddb6dae1065aa5
This cl change util method in bluetooth package to return
drawable instead of resId.
If the bt device has battery level, then method return customized
layerDrawable, otherwise return a simple drawable created from
resId.
Bug: 63393322
Test: RunSettingsRoboTests
Change-Id: Ib21822eafda0e8570212ce5cb070478e4f4876a2
Merged-In: Ib21822eafda0e8570212ce5cb070478e4f4876a2
When we first initialize the dashboard view, and register the condition
listener, it will trigger the condition changed callback immediately.
This results in unnecessary refresh of the dashboard header. Add check
to not do the refresh when we first initialize the view.
Change-Id: If7c69637463734c150b7f5eb7f3c042cf73837fa
Fixes: 64811475
Test: make RunSettingsRoboTests
