By calling WindowManagerService#reAddWindowLocked(), the window object
in DisplayContent.mWindows is rearranged. If this window has children,
the window objects belonging to WindowState.mChildWindows are also
rearranged together. At this time, while DisplayContent.mWindows is
arranged in ascending order of Z-Order, WindowState.mChildWindows is
arranged in descending order. Therefore, if children
(= WindowState.mChildWindows) are added to DisplayContent.mWindows
in the original order, the order of children is switched. I think
that children must be sorted before they are added.
Change-Id: I28a506f17fcd99c801a3faa17cca66fffb291596
Previously a change to a surface would be logged with the old value
and you had to scroll through the logs to see what the new value
was. This change reflects the change to the surface immediately.
Change-Id: I2a6566466287922d08f4ce2329c61aa46d692ee1
The min layer was set to only show the windows that matched
the specified app token. But that meant when dialogs were
launched it only showed the dialogs and not the background
windows.
Added improved debugging.
fixes bug 8502844.
Change-Id: I26b49568b872801ec9aa088df20317aa752dacd6
If LockScreen is enhanced using SurfaceView/GLSurfaceView,
deadlock problem between LockScreen and WindowManagerService
can occur because of IWindow.resized() callback.
And it must lead to watchdog and reset.
IWindow.resized() callback is one-way function so calling resized()
callback of a remote IWindow object is never blocked.
However, calling resized() callback of a local IWindow object
(LockScreen is running on the same system_server process)
is always blocked until resized() callback returns.
Because resized() callback of SurfaceView/GLSurfaceView can lead to
WindowManagerService.relayoutWindow() call, deadlock can occur
between relayoutWindow() and performLayoutAndPlaceSurfacesLockedInner().
(Both functions need locking mWindowMap)
So this patch simulate one-way call when calling resized() callback
of a local IWindow object.
Change-Id: I2a6a5c74ed22d8e6b7a3bea3424ff2879d227105
Signed-off-by: Sangkyu Lee <sk82.lee@lge.com>
Conflicts:
services/java/com/android/server/wm/WindowManagerService.java
If LockScreen is enhanced using SurfaceView/GLSurfaceView,
deadlock problem between LockScreen and WindowManagerService
can occur because of IWindow.resized() callback.
And it must lead to watchdog and reset.
IWindow.resized() callback is one-way function so calling resized()
callback of a remote IWindow object is never blocked.
However, calling resized() callback of a local IWindow object
(LockScreen is running on the same system_server process)
is always blocked until resized() callback returns.
Because resized() callback of SurfaceView/GLSurfaceView can lead to
WindowManagerService.relayoutWindow() call, deadlock can occur
between relayoutWindow() and performLayoutAndPlaceSurfacesLockedInner().
(Both functions need locking mWindowMap)
So this patch simulate one-way call when calling resized() callback
of a local IWindow object.
Change-Id: I2a6a5c74ed22d8e6b7a3bea3424ff2879d227105
Signed-off-by: Sangkyu Lee <sk82.lee@lge.com>
When a tablet rotates, FUL must be stopped and restarted in a new
position. 90 degree rotations cause a configuration change, causing
FUL to be automatically reconstructed in the new location. However,
a 180 degree rotation is not a configuration change, so FUL was not
restarting. A 180 degree rotation happens more often than one might
think. If you set the tablet down and later picked it up in the
opposite orientation, FUL would not work prior to this fix.
This change adds a rotation watcher to KeyguardFaceUnlockView. It
watches for 180 degree rotations and stops and restarts FUL
accordingly.
The rotation watcher callback must be unregistered when
KeyguardFaceUnlockView is recreated (as during 90 degree rotation
changes), otherwise the number of rotation watcher callbacks will keep
growing and they will never go away. This is a problem not just
because there are many callbacks hanging around, but also because the
old callbacks end up trying to access biometric unlock views that no
longer exist, resulting in crashes. So, a simple function was added
to the window manager to unregister a rotation watcher.
Change-Id: Ie1ef20a9a22b8f4e39918987dff2b8ad444fcfd1
Pause and retry capturing a thumbnail if the window hasn't yet
been drawn. Still will show black thumbnails if recents is pressed
before window has anything drawn.
Improves bug 7552304.
Change-Id: I550c5a60510bfc4547bbb44451d57b4bc9f49880
The root cause is:
There is a defect in window manager service: When a new
activity that can be ime target is added into window manager
but the Z order of input method window don't need to be
changed, then the target app token of input method window
would not be updated to new one. This defect may cause that
the layer of input method window is calculated incorrectly.
The solution:
Correct the target app token for input method window.
Change-Id: I008311e3c9b1cf5fc320b614d8675c183c506d50
This is a class representing a window and providing limited
interaction with it, which can be handed across processes.
Change-Id: I22885f2064a9cc8c68d690a5858c2e28bbb6a0f3
Keeping all activity=>task changes in master and removing them
from jb-mr2.
Revert "Update histories simultaneously."
Revert "Add null check to setAppGroupId."
Revert "Fix crashing bug in validator."
Revert "Switch topRunning* and moveTaskTo*"
Revert "Begin switch over to task based history."
Revert "Reset and reuse Iterators and don't new() one."
Revert "Remove AppWindowToken lists."
Revert "Fix build."
Revert "Remove unused App methods."
Revert "Stop using AppToken movement and start using Task."
Revert "Replace access to mAppTokens with AppTokenIterator"
Revert "Refactor setAppOpVisibility implementation."
Revert "Add AppWindowTokens to TaskList."
Revert "Make ActivityStack.mHistory private."
Revert "Migrate AppWindowToken lists into DisplayContent."
Change-Id: I5722c9a4956dccb52864207e2967690bc58e4ebb
The window manager now maintains and reports a new formal
"overscan insets" for each window, much like the existing
content and visible insets. This is used to correctly
position the various UI elements in the various combination
of layout options. In particular, this allows us to have
an activity that is using fitSystemWindows to have the content
of its UI extend out to the visible content part of the screen
while still positioning its fixed UI elements inside the
standard content rect (and the entire window extending all
the way into the overscan area to fill the screen as desired).
Okay, maybe that is not written so clearly. Well, it made
my head hurt too, so suffer!
The key thing is that windows now need to know about three
rectangles: the overall rectangle of the window, the rectangle
inside of the overscan area, and the rectangle inside of the
content area. The FLAG_LAYOUT_IN_OVERSCAN option controls
whether the second rectangle is pushed out to fill the entire
overscan area.
Also did some improvements to debug dumping in the window
manager.
Change-Id: Ib2368c4aff5709d00662c799507c37b6826929fd
Allow fullscreen windows to specify crossfade or jumpcut animations
that override the default rotation animation. Only if the incoming
and outgoing topmost windows are fullscreen and both specify the
same animation to use.
Fixes bug 8182773.
Change-Id: I6b3c0020d7bd2cdfba5c66189e114ec62cd54fcf
In moveTaskToBackLocked calls that used mHistory were being made
between the times that mTaskHistory and mHistory were modified. This
caused an inconsistent state that led to Windows arranged out of
order. Updating both history stacks at the same time fixes this.
Fixes bug 8244261.
Change-Id: I9669762ad39b06ab6d401122702b74969d4dc658
- More of the Activity to Task changeover.
- Fix bug in validateAppTokens().
- Improved validation of changeover.
- Eliminated iterator classes.
Change-Id: I934a208eabfc9a2668e5a6162452e1406f2c8d3a
The window manager now keeps track of the overscan of
each display, with an API to set it. The overscan impacts
how it positions windows in the display. There is a new set
of APIs for windows to say they would like to go into the
overscan region. There is a call into the window manager to
set the overscan region for a display, and it now has a
concept of display settings that it stores presistently.
Also added a new "wm" command, moving the window manager
specific commands from the "am" command to there and adding
a new now to set the overscan region.
Change-Id: Id2c8092db64fd0a982274fedac7658d82f30f9ff
Save time and memory by resuing permanent mTmpXxxIterator for
AllWindowsIterator and AppTokenIterator.
Change-Id: I2e5143364651d9471e9e70eee514d3f45d595468
