When sending a ViewNode to the auto-fill service, it's necessary to sanitize
it so it does not contain PII (Personally Identifiable Information).
A good rule of thumb is to mark as sanitized nodes whose value were statically
set from resources. Ideally, that logic should belong to View, but View does
not know how its contexts are set, so the decision must be made by the
subclasses, and this CL implements that logic on TextView (which is the main
auto-fill type - others will come soon...)
Bug: 33269702
Test: CtsAutoFillServiceTestCases pass
Test: manual verification
Change-Id: I4dcdd8419a35812bd2f15a63a5be85f8cd97bb4e
* LRU cache of recently-used is dead, replaced
disk storage
* ASHMEM size is read from native by the system service,
no longer requires keeping a sizeof() in sync with a
constant in Java
* Supports dumping in proto format by passing --proto
* Rotates logs on a daily basis
* Keeps a history of the most recent 3 days
Bug: 33705836
Test: Manual. Verified log rotating works by setting it up to
rotate every minute instead of day. Confirmed /data/system/graphicsstats
only has the most recent 3 entries after several minutes
Change-Id: Ib84bafb26c58701cc86f123236de4fff01aaa4aa
1. Introduce config_scrollbarSize, default value=4dp
2. Use it to define android:scrollbarSize in themes.
3. Remove the explicit size from the MD scrollbar thumb shape.
Bug: 35432058
Test: android.cts.view.ViewTest#testScrollbarSize
Change-Id: Ia97b8f34ae1b9296ca2afcf62ff5bde55f70a63e
When activity that is moved between displays handles all configuration
changes, it won't be restarted. This CL adds a callback to the client
to notify it about display change. Usually it will be followed by
onConfigurationChanged, except when configuration didn't actually change.
When activity is recreated, it won't receive onMovedToDisplay.
Bug: 34862802
Test: android.server.cts.ActivityManagerDisplayTests
Test: #testOnMovedToDisplayCallback
Change-Id: I9a9501cab788623ada15a31efb53e4b2378639fe
This is a follow up CL to my previous CL [1], which caused an unexpected
side effect that leads the current IME to crash due to a too strict
requiremnt.
It turns out that it was too early for us to start requiring non-null
StartInputToken in InputMethodManagerService#setImeWindowStatus()
because in many places we have assumed that an IME can show/hide its
software keyboard even before an IME target window is associated with
the IME.
There are two major cases that we missed:
- InputMethodManager#showSoftInputFromInputMethod(IBinder, int):
This method does not require that the calling IME is already bound
to an IME target window.
- InputMethodManager#showSoftInputUnchecked(int ,ResultReceiver):
This @hide method allows the caller application to let current IME
show the software keyboard with bypassing all the normal focus
management tasks in InputMethodManager. We should seriously
consider to deprecate this @hide method, but to do that we have to
clean up some internal components and SearchView in the support
library that still rely on this method.
Bug 35395372 is triggered by the second scenario, but until we sort out
all the possible corner cases, we should allow null startInputToken as a
valid request, like we had have done so before the CL [1] introduced
such a validation.
[1]: I9921b381e02106dbffff5e0b3d13f0a1245ce807
6db3bfe33d
Test: Made sure Bug 35395372 is no longer reproducible, that is,
1. Flash a new image and complete the setup wizard on a
direct-boot unaware device.
2. Set a device password and require it upon each device boot.
3. adb reboot
4. Observe the default IME does not crash because of
InvalidParameterException thrown by IMMS.
Test: Made sure IMM#showSoftInputFromInputMethod(IBinder, int) does
not throw an InvalidParameterException even in an extreme case.
1. Rebuild LatinIME with the following code in LatinIME.java
@Override
public AbstractInputMethodImpl onCreateInputMethodInterface() {
return new InputMethodService.InputMethodImpl() {
@Override
public void attachToken(IBinder token) {
super.attachToken(token);
final InputMethodManager imm =
getSystemService(InputMethodManager.class);
final IBinder imeToken =
getWindow().getWindow().getAttributes().token;
imm.showSoftInputFromInputMethod(imeToken, 0);
}
};
}
2. adb install -r LatinIME.apk
3. adb shell ime enable com.android.inputmethod.latin/.LatinIME
4. adb shell ime set com.android.inputmethod.latin/.LatinIME
5. Tap any text field.
4. Observe LatinIME does not crash because of
InvalidParameterException thrown by IMMS.
Bug: 34628091
Bug: 35079353
Fixes: 35395372
Change-Id: Ib9448c551d9a30776a999c27a5ff20f1a095633a
The URL will be typically used by views rendering HTML pages (like WebView),
and used in 2 cases
- Root node (containing the URL of the HTML page)
- Child nodes that represent hyperlinks (contains the hyperlink URL).
Fixes: 20902429
Test: manual verification
Test: android.assist.cts.WebViewTest verifies it
Change-Id: Ic49472f1fca89047369b7187176f54efd277e204
Cluster navigation would get stuck if:
- there was an empty cluster (nothing to focus)
- there were views NOT in a cluster and the global
default-focused view was in a cluster.
This will skip over empty clusters during cluster navigation
and when the default "cluster" is active, it will try to
focus a non-clustered focusable instead of the global default
(and if there aren't any non-clustered focusables, it will
skip over the default "cluster" as it now does with actual
clusters.
Bug: 35273052
Bug: 34361282
Test: Added/Updated CTS tests and checked behavior in a test app.
Change-Id: I0b4d55707203f6a0453f25313f940927d59a9b31
Suppressing whenever the keyguard is locked, not just when
it's showing. This should reduce false triggering.
Bug: 34820720
Test: Verfied that shortcut no longer activates on emergency
dialer or lock screen.
Change-Id: Ic43c11798f04217c2871efe7d0fa21dbf61fd953
This resolves the binary compatibity issue with some
existing apps.
Bug: 35150720
Test: framework builds
Change-Id: I4242186b7103ad35a1e09264c0c9f86bbcf88f49
* Fix a layering issue where auto-fill manager which is in view
depended on activity which is in app
* Moved auto-fill classes to view or service based on their
purpose and removed dependecy on the classes in view to the
classes in service
* Push state to local auto-fill manager whether auto-fill is
enabled to avoid making IPC for every focus transition if
the user did not enable the feature
* Remove unnecessary offload to messages when handling calls
to auto-fill manager service as these are made over a oneway
interface and in general they do almost no work and typically
we do these on the binder thread
* Removed id from data set and fill response as the provider
can embed everything it needs to id them in the auth pending
intent
* Enforce the auth UI to be only an activity as this will work
with multi-window, recents, and back and also does not require
draw on top of other app special permission
* Authentication also no longer requires passing a remotable
callback to the auth activity but the activity handles the
request as if called for a result
* Handling stopping of a user to clean up in-memory state as
well as handling when a user gets unlocked as a provider may
be non-direct boot aware
* User the correct context when creating an auto-fill manager
* Move the receiver that listens for requests to hide system
windows to the manager service as the UI is a singleton and
no need every per-user state to register its own
* Removed extras from dataset as the only case a provider needs
to associate state with a dataset is for auth and the provider
can embed this data in the auth pending intent
Test: manual and CTS
Change-Id: I4bc54c13cf779d7f6fdb3ab894637f9fac73f603
setImeWindowStatus() introduced in the previous CL [1] can help
WindowManagerService (WMS) address Bug 34628091. As a preparation
this CL has InputMethodManagerService propagate relevant information
to WMS WMS for later use.
[1]: I9921b381e02106dbffff5e0b3d13f0a1245ce807
db2afde0e584904ee4ce115ce176a4dfe1773990
Test: Verified the callback in WindowManagerService as follows:
1. Rebuild the system with DEBUG_INPUT_METHOD=true
2. adb logcat -s WindowManager:*
to make sure the visibility passed to
WindowManagerService#updateInputMethodWindowStatus matches
to the actual IME window visibility
Bug: 34628091
Bug: 35079353
Change-Id: I54fb1faf513883a54293a756e4c9d2ae9453d778
This is part of work to introduce historical debugging infrastructure
for Android IME.
In this CL, we will focus on the following two event flows.
A1. IMMS#attachNewInputLocked() queues MSG_(RE)START_INPUT to deliver
new InputConnection/EditorInfo to the current IME
A2. The IME triggers IMS#onStartInput()/IMS#onRestartInput() and
updates the following fields:
- InputMethodService#mStartedInputConnection
- InputMethodService#mInputEditorInfo
B1. IME is expected to call back IMM#setImeWindowStatus() to notify
its window visibility change to IMMS.
B2. IMMS updates the following field if the caller is still the
current IME.
- InputMethodManagerService#mImeWindowVis
What this CL aims to do is to enable IMMS to access A1 state when it
was in B2 state, by considering that for given a B1 the last A2
happened before B1 is the cause of B1 and B2.
To do this, IMMS issues a binder token in A1 and each IME keeps it
so that it can be passed in B1. By using this Binder token as a key,
IMMS can keep tracking state snapshot taken from each A1. Note that
those state snapshots keep alive until the Binder token's proxy in the
IME process loses strong reference from its GC root.
Test: Make sure `adb shell dumpsys input_method | grep mImeWindowVis`
matches to the IME window visibility.
Test: Make sure the current IME is not receiving any
InvalidParameterException from IMMS.
Bug: 35079353
Change-Id: I9921b381e02106dbffff5e0b3d13f0a1245ce807
Currently ACTION_DOWN and ACTION_SCROLL cause
entering touch mode regardless of the event source.
This is especially awkward for mouse.
This should only happen when the event source is
SOURCE_TOUCHSCREEN.
Bug: 35314124
Test: android.cts.view.ViewTest#testTouchMode
Change-Id: I2205d4240be4afcbd9a3ac5558e701f7cac51926
This cl also updates label strings, renames an internal method,
and fixes some bugs in TextClassifierImpl caught by tests.
Test: See: I31ee94177e58c986e61e9b24c6c5816fa4b022e1
Bug: 34778263
Bug: 34922928
Change-Id: Ibf60d345d462dfc8a13dead6f01053f1ae50f35a
For processes with a window of type TYPE_APPLICATION_OVERLAY adjust their
oom importance to PERCEPTIBLE_APP_ADJ to reduce the chance of them getting
killed by the low-memory-killer since they are displaying something that
is perceptible to the user.
Also z-order TYPE_DREAM windows above alerts windows.
Bug: 33256752
Test: cts/.../run-test CtsAppTestCases android.app.cts.AlertWindowsTests
Change-Id: I4c05a9fee6fad61399bf4d10c8647467cc596ca6
Doing this allows us to annotate integer values that should corresponds
to WindowManager.LayoutParams#softInputMode.
Test: There should be no behavior change. RetentionPolicy.SOURCE
annotation should change nothing in production code.
Test: checkbuild
Bug: 32784563
Bug: 35079353
Change-Id: I96300b090edce327d0515c740183abe91ded6bac
A field in WindowManager.LayoutParams softInputMode is something that
definitely needs to be kept tracking in historical debugging
infrastructure across IME-related processes (Bug 35079353) [1]. As a
preparation, this CL enables InputMethodManagerService (IMMS) to
include the last softInputMode specified in IMMS#windowGainedFocus()
in the dumpsys in human readable format.
[1]: As explained in b.android.com/224318, softInputMode misspecified
by app developers is a typical root cause of unexpected behavior
in keyboard visibility. Bugs such as Bug 23168250, Bug 27275709,
and Bug 31770400 fall into this category.
Test: `adb shell dumpsys input_method | grep softInputMode=`
Bug: 35079353
Change-Id: I485ced030def179dad78b4b811c6eb52b5e5c951
