Previously the unparceling code relied on addAction(), which would
incorrectly crash if the node was sealed.
Bug: 21135077
Change-Id: I773ac27b2dee501ffea0a0587536a192af402c32
This reverts commit 97c3813042.
This causes issue with the right IME window getting focus.
Bug: 21144633
Change-Id: I4c75b6e7dd87c10f008444d2059164b52a8f4335
Introduce ACTION_BUTTON_PRESS and ACTION_BUTTON_RELEASE as actions to
signal a button press or release. If these actions happen
simulanteously with a DOWN or UP event then they're explicitly
ordered to happen after the DOWN or preceding the UP in order to send
them to the most recently targeted view.
Also, introduce new stylus button constants that differ from the
constants we use for mouse buttons.
Bug: 20704355
Change-Id: I5b75e5c5e692171c1c117ee687dd185a0d9dd15c
- The mechanism to stop windows drawing while window animator was
animating was somehow flaky. It relied on the fact that the client
would call relayout() whenever the animating state changed. This is
mostly the case, but not for lockscreen animations. Instead, we now
use a push model, where window manager tells the app that the state
has changed.
- In addition, it only stopped drawing if that window was animating,
but then only resumed drawing after all windows have finished
animating. Now, we do this per window, so we only stop drawing for
windows that are currently animating.
- We resume the top activity now at the very beginning of the
unlocking sequence. This gives the app a chance to draw a frame
before the user sees anything. If it's to slow, then we just use the
outdated framebuffer.
Bug: 19964562
Change-Id: Ifef8abd189a3146d854b81b9b948861e4d38c155
IMM#mViewTopLeft and IMM#mViewToScreenMatrix are no longer used.
They should have not been removed in
I4a577bfd02b37b9e56c80b8b41bb25afa95dd8ef.
Change-Id: Ia685956c8f7dda1143135345ee9a3873705975ef
Despite the fact that IMM#focusOut() are called from many focus
mangement logics, the reality is that IMM#focusOut() has done
nothing more than 6 years. This would not a big problem as long
as IMM#focusIn() is called immediately after IMM#focusOut().
However, situations where only IMM#focusOut() is called,
following fields continue keeping object references.
- IMM#mServedView
- IMM#mNextServedView
- IMM#mServedInputConnection
- IMM#mServedInputConnectionWrapper
Even worse, if the IME is showing software keyboard, it will not
be dismissed.
With this CL, IMM#focusOut() starts cleaning up the active IME
session when the associated view loses focus.
This CL also removes IMM#mCurRootView because it is indeed
necessary for above change. The problem when only introducing
above change can be understood as follows.
1. IMM#mCurRootView is correctly set from ViewRootImpl.
2. IMM#focusIn() is called. InputConnection is established.
3. IMM#focusOut() is called, which triggers
IMM#finishInputLocked() because of above change, which
internally clears IMM#mCurRootView.
4. IMM#focusIn() is called but IMM#mCurRootView is still null.
Because of this the focus-in event is ignored and the
software keyboard does not show up anymore.
In this CL, we simply check view.hasWindowFocus() instead.
As far as I've tested, this change looks to be working well.
Bug: 20820914
Change-Id: Ib4bd70ce0305a6bde6a929bcc6ad20a2b8402a97
Knowledge of the various modes of a display is now available to apps, and they can request a specific mode for their windows.
b/18241736
Change-Id: I8eb16ff713e878512faca3ca6662254f08a9be7f
(cherry picked from commit 5c7773d864)
A previous CL introduced a new way of encoding view properties for
use by heirarchy viewer. This CL updates all views using the old
@ExportedProperty annotation to use this new method. The older
mechanism will be removed in a subsequent CL.
Change-Id: I6cc23b90cd9da1c6ce89b4caffe54874db203452
Hierarchy Viewer obtains the properties for each view by using
reflection and looking for fields and methods that have the
@ExportedProperty annotation. Using reflection made it quite slow
for large view hierarchies.
This CL adds a new method (encode) to each class that wishes to
export data to hiererachy viewer. Inside this method, the object
can write a sequence of key, value pairs corresponding to the
values it wants exported.
With this change, the dump hierarchy operation that used to take
more than 10 seconds can be performed in a few hundred milliseconds.
Change-Id: I199ac2e7ca3c59ebcfec7e6bd201e134c41fd583
Fix the various view assist related APIs.
Also remove the blockAssist view attribute, and instead use
the window's FLAG_SECURE to drive blocking of the entire
hierarchy (which is semantically correct, and will protect
existing apps that have already indicated they need it).
Change-Id: I6beebc86b202809cba0a356cae9607d8d0fb5e78
Bug 19437530
While the actual Activity Transitions are running,
input is stopped, except for the "back" button.
Change-Id: I112e6252b9de05ece10a6267681fee5487e5ef6b
This CL relands I1e50ee42838a1bf64a612da4904aa93458d44ea4, which was
reverted by I3decaf37198e5864a1763a059df4a36ebc70c5a7 due to the build
breakage in 'layoutlib' target, with a proper fix.
Hereafter the original CL description is repeated.
The auxiliary subtypes should be listed if the input method picker is
opened from NavBar keyboard icon. However there is only
IMM#showInputMethodPicker() API to open input method picker and this is
also used from LockScreen or Settings UI. Auxiliary subtypes should not
be listed there(Id7cf5d122). Thus framework shows auxiliary subtypes
based on IMMS#mInputShown and LockScreen state, but it is not a perfect
solution. If a physical keyboard is connected, the soft input may be
gone. As the result, auxiliary subtypes won't be listed even if it is
opened from NavBar keyboard icon.
To fix this issue, this CL introduces IMM#showInputMethodPicker(boolean)
to be able to decide showing auxiliary subtypes by caller.
Note that IMM#showInputMethodPicker(boolean) is still hidden with @hide.
There is no public API change in this CL.
Bug: 20763994
Change-Id: Id156c85535a221235737ea6dcc15a67f1c4b9f71
CL is breaking the build. Discussed with Seigo and verting until he can take a look at it.
This reverts commit 80ff4ed6bb.
Change-Id: I3decaf37198e5864a1763a059df4a36ebc70c5a7
Bug 20675857
GhostView was synchronously forcing the View's parent to get
the display list when it was created. TextureView's cannot
create a display list until a surface is available and will
crash if it is attempted. This change moves the request to
an invalidate to get it to execute at a later time.
Change-Id: I12afebdfd57eb29c1a35b23a63d045e319547faf
PhoneWindow, PhoneLayoutInflater and PhoneFallbackEventHandler decided
to @hide out over in the android.view package after the policy jar was
disbanded. Give them a more appropriate home over in framework that
doesn't imply that they should be accessed from other internal layers
of abstraction.
Bug 19606548
Change-Id: Id07b791d178fa447010b49b24726b52208838e88
The auxiliary subtypes should be listed if the input method picker is
opened from NavBar keyboard icon. However there is only
IMM#showInputMethodPicker() API to open input method picker and this is
also used from LockScreen or Settings UI. Auxiliary subtypes should not
be listed there(Id7cf5d122). Thus framework shows auxiliary subtypes
based on IMMS#mInputShown and LockScreen state, but it is not a perfect
solution. If a physical keyboard is connected, the soft input may be
gone. As the result, auxiliary subtypes won't be listed even if it is
opened from NavBar keyboard icon.
To fix this issue, this CL introduces IMM#showInputMethodPicker(boolean)
to be able to decide showing auxiliary subtypes by caller.
Note that IMM#showInputMethodPicker(boolean) is still hidden with @hide.
There is no public API change in this CL.
Bug: 20763994
Change-Id: I1e50ee42838a1bf64a612da4904aa93458d44ea4
Also updates default fade duration for scrollbars to match Material
spec and moves around some padding in AlertDialog so that scrolling
text and list items aren't so close to the title.
Bug: 19098033
Change-Id: I40dca6a931480c4c48463e3ea5b8361534cbd8d7
bug:19967854
Separate properties from Caches, into static, RenderThread-only class.
Also rewrites the means for java to set properties to correctly handle
threading, and adds an override for profile bars so that SysUi doesn't clutter
the screen with them.
Change-Id: I6e21a96065f52b9ecc49d1a126244804ba106fa9
We now correctly throw an IllegalAccessException for non-public accesses,
but we were relying on the old incorrect behavior during inflation. This
CL forces constructors to be accessible so that we emulate the old
behavior and don't crash.
Where we were already doing this, the CL moves the setAccessible() call
immediately after the getConstructor() call, since there's no point in
calling it multiple times.
Bug: 20810495
Change-Id: I63dfcb857e445f688080eea05e86ef22760ed49f
