Refactored ViewRoot, NativeActivity and related classes to tell the
dispatcher whether an input event was actually handled by the application.
This will be used to move more of the global default key processing
into the system server instead of the application.
Change-Id: If06b98b6f45c543e5ac5b1eae2b3baf9371fba28
For possible future changes to android.view.animation.Animation to tie
into StrictMode's animation annotations on violations.
Change-Id: I3a23fa9bd5edb0500586bb0e341f83f998b28ae7
Who knew there were so many Loopers in the system server?
This adds dropbox logging for the activity manager thread, policy
thread, and window manager thread.
The goal's to catch more stuttering.
Change-Id: I999a6ff4a955f0ef549b12a38796b843ade5fe66
IWindowManager now supports two new methods,
freezeRotation() and thawRotation(), that allow a caller to
temporarily stash the device's current rotation as the
default rotation (when no other constraints are present).
The system bar uses this to implement a user-accessible
rotation lock by calling freezeRotation() and then turning
off accelerometer-based display rotation; unless overridden
by an app, the display will continue to appear in the frozen
rotation until the rotation is unlocked by the user (either
via the rotation lock icon in the system bar or by checking
"rotate screen automatically" in Settings).
Bug: 2949639
Change-Id: Icd21c169d1053719590e72401f229424b254622f
Who knew there were so many Loopers in the system server?
This adds dropbox logging for the activity manager thread, policy
thread, and window manager thread.
The goal's to catch more stuttering.
Change-Id: I3ce8518ce183b3c90426750a2992e67200fee5d5
Now, each ViewGroup is tracking which of its child views [which might
themselves be ViewGroups] is currently under the drag point, and when the
drag leaves that child, a DRAG_EXITED is synthesized and dispatched all
the way down to the leaf view previously under the point. ENTERED is
still *not* dispatched down like this; instead, it's calculated and
synthesized directly at each level based on the new LOCATION.
The ViewRoot still tracks the leaf drag target, but solely for the
purpose of reporting changes to the OS after full dispatch of a new
LOCATION -- the entered/exited messaging is no longer initiated at the
ViewRoot level.
Change-Id: I0089cc538b7e33a0440187543fcfd2f8b12e197d
Rewrote interceptKeyBeforeQueueing to make the handling more systematic.
Behavior should be identical except:
- We never pass keys to applications when the screen is off and the keyguard
is not showing (the proximity sensor turned off the screen).
Previously we passed all non-wake keys through in this case which
caused a bug on Crespo where the screen would come back on if a soft key
was held at the time of power off because the resulting key up event
would sneak in just before the keyguard was shown. It would then be
passed through to the dispatcher which would poke user activity and
wake up the screen.
- We propagate the key flags when broadcasting media keys which
ensures that recipients can tell when the key is canceled.
- We ignore endcall or power if canceled (shouldn't happen anyways).
Changed the input dispatcher to not poke user activity for canceled
events since they are synthetic and should not wake the device.
Changed the lock screen so that it does not poke the wake lock when the
grab handle is released. This fixes a bug where the screen would come
back on immediately if the power went off while the user was holding
one of the grab handles because the sliding tab would receive an up
event after screen turned off and release the grab handles.
Fixed a couple of issues where media keys were being handled inconsistently
or not at all, particularly in the case of the new PAUSE, PLAY
and RECORD keys.
Bug: 3144874
Change-Id: Ie630f5fb6f128cfdf94845f9428067045f42892c
Also make sure to properly dispose *both* ends of the drag input
channel when we're done. That was being tidied up after, but was
also making the system work harder and was provoking a bit of
debug logging.
Change-Id: I1ff3100bc4c441f1609c57242864d5744ff1aaa2
Part of issue #3116702: New manifest tags for supported screen sizes
Kind-of.
If you turn your head side-ways.
Change-Id: I446f1e2eadba1ce284c93ff9fb0197bb0e6b0fca
DragEvent.getResult() returns 'true' if the drop was ultimately accepted;
false otherwise. The validity of this datum is only guaranteed when the
DragEvent's action verb is ACTION_DRAG_ENDED.
Also fixes the drag-start timeout handling (though the offending app is
not yet officially declared ANR).
Implements bug 3097807
Change-Id: I6908ac628c72ff7d6193d87060d769a559a78d0e
* Increase the opacity of the drag thumbnail a bit
* Properly present window-local drag locations; this fixes drags into
the status bar
* Pass the current drag touch point in DRAG_STARTED. This may
be reconsidered, but might conceivably be useful to someone.
Change-Id: I77690e82e9831ed6be412990c1c9da92e700a501
Now sets the layer and thumbnail dimensions correctly. Also removed the
leftover bits of the now-replaced thumbnail measurement/drawing mechanism.
Fixes bug 3077339
Change-Id: I02983648e0a7ce2ce5fee10825f744e20a3a5b8d
Bug #3081600
The OpenGL renderer in libhwui uses a single EGL context per process and
thus create it with an RGBA 8888 EGL configuration. To ensure that all
windows are compatible with this configuration, this change modifies
the window manager and SurfaceFlinger.
The window manager now checks the window's flags and if the window is
hardware accelerated, it forces the window's pixel format to be
translucent when creating the surface. The window itself is still
marked as opaque if we know that the window will be opaque on screen.
This keeps existing optimizations in place.
Similarly in SurfaceFlinger, a translucent Surface can now be created
with the Surface.OPAQUE flag, indicating SurfaceFlinger that the surface
does not require blending, despite its RGBA 8888 configuration.
Change-Id: Ic747b6b12564ba064412d842117880fcc199eb7c
Added the concept of a "trusted" event to distinguish between events from
attached input devices or trusted injectors vs. other applications.
This change enables us to move certain policy decisions out of the
dispatcher and into the policy itself where they can be handled more
systematically.
Cherry pick of b931a1b4 from gingerbread into master.
Change-Id: I700a5f07b8b227878cea9437a289a45a245c0424
Merge commit '1da8d4d1697cec871dea8c30dad67e90db10bde6' into gingerbread-plus-aosp
* commit '1da8d4d1697cec871dea8c30dad67e90db10bde6':
Fix an event injection bug when the policy is bypassed.
Added the concept of a "trusted" event to distinguish between events from
attached input devices or trusted injectors vs. other applications.
This change enables us to move certain policy decisions out of the
dispatcher and into the policy itself where they can be handled more
systematically.
Change-Id: I4d56fdcdd31aaa675d452088af39a70c4e039970
* Make View.onDragEvent() public instead of protected.
* No longer @hide View.startDrag()
* Properly manage the boundaries of DRAG_STARTED / DRAG_ENDED notifications
to windows (and as a result don't send bogus empty DRAG_STARTED events or
double-recycle pooled DragEvents)
Change-Id: Ib75f5c1417640c82a5b043c555e02d6e6f4b4d9c
Thumbnail measurement & drawing has been moved out into a delegate
class called View.DragThumbnailBuilder. This consolidates the
thumbnail-related code as well as ensuring that the drag initiator
does not have to know a priori where to place the thumbnail relative
to the touch point *before* the thumbnail measurement step, as was
previously the case.
startDrag() no longer needs to be told where the current touch point
is at the time the drag is launched.
Drag events are now dispatched only to VISIBLE views.
Dispatch shouldn't double-recycle events any more when the target
window is local to the system process.
Change-Id: I49419103765a0cad2e18ddfcdd6dacb94daf1ff1
Merge commit '458fc5f52e25d7a8b4e482010711aed66f0b4999' into gingerbread-plus-aosp
* commit '458fc5f52e25d7a8b4e482010711aed66f0b4999':
Fix issue #3041660: Camera image flips upside down when rotating device
Apps now must explicitly opt in to having their rotation changed while
forcing landscape mode. Also add some new orientation constants for other
things apps may want to do.
Change-Id: If64d84b5ef54793ee717ebda9b4c76408efc9bfd
Merge commit '1fa69735d9e786da4b20c3ab3da3aa66f3cf8116' into gingerbread-plus-aosp
* commit '1fa69735d9e786da4b20c3ab3da3aa66f3cf8116':
Fix issue #3037083: Problem report for Pinball
The app was toggling between hiding and showing the status bar, because
the dialog was large enough to fill the screen but didn't set to hide
the status bar.
It turns out that the code to determine the top fullscreen window is
intrinsically unstable, since it relies on comparing a window's frame
with the screen, yet the window frame can change as a result of changes
such as hiding the status bar.
Instead, we now simplify all of this to just say that the top fillxfill
window gets to control the status bar.
Change-Id: I22913adf7235fe0d52612f6738c7c94351274e38
A View initiates a drag-and-drop operation (hereafter just called a "drag")
by calling its startDrag(ClipData) method. Within the processing of that
call, two callbacks are made into the originating View. The first is to
onMeasureDragThumbnail(). Similarly to the core onMeasure() method, this
callback must respond by calling setDragThumbnailDimension(width, height) to
declare the size of the drag thumbnail image that should be used. Following
this, the View's onDrawDragThumbnail(canvas) method will be invoked to
actually produce the bits of the thumbnail image.
If all goes well, startDrag() will return 'true', and the drag is off and
running. (The other arguments to startDrag() provide reconciliation between
the current finger position and where the thumbnail should be placed on
the screen relative to it.)
Potential receipients of the ClipData behind the drag are notified by a
new dispatch mechanism, roughly parallel to motion event dispatch. The core
routine is the View's onDragEvent(event) callback, with the mechanics of
dispatch itself being routed through dispatchDragEvent(event) -- as in
the case of motion events, the dispatch logic is in ViewGroup, with leaf
View objects not needing to consider the dispatch flow.
Several different event 'actions' are delivered through this dispatch
mechanism:
ACTION_DRAG_STARTED: this event is propagated to every View in every window
(including windows created during the course of a drag). It serves as a
global notification that a drag has started with a payload whose matching
ClipDescription is supplied with the event. A View that is prepared to
consume the data described in this event should return 'true' from their
onDragEvent() method, and ideally will also make some visible on-screen
indication that they are a potential target of the drop.
ACTION_DRAG_ENTERED: this event is sent once when the drag point
enters the View's bounds. It is an opportunity for the View to set up
feedback that they are the one who will see the drop if the finger goes
up now.
ACTION_DRAG_LOCATION: when the drag point is over a given View, that
View will receive a stream of DRAG_LOCATION events, providing an
opportunity for the View to show visual feedback tied to the drag point.
ACTION_DRAG_EXITED: like DRAG_ENTERED, but called when the drag point
leaves the View's bounds. The View should undo any visuals meant to
emphasize their being the hovered-over target.
ACTION_DROP: when the drag ends at a given point, the View under that
point is sent this event, with the full ClipData of the payload.
ACTION_DRAG_ENDED: paralleling the DRAG_STARTED action, this is the global
broadcast that the drag has ended and all Views should return to their
normal visual state. This happens after the DROP event.
Change-Id: Ia8d0fb1516bce8c735d87ffd101af0976d7e84b6
Correction for the error:
When a window is done animating and about to be destroyed,
it is still checked for visibility in the method
updateReportedVisibilityLocked in the WindowManager.
In the case where other windows are ready to be reported as
visible, this, not yet destroyed window which has the state
not visible, will prevent the visibility report message from
being sent.
Change-Id: I9e307a678a204f0883ccd78c2645c373d3f81066
This is a manual cherrypick of:
Change I3a76de15: Fixing font scale in WindowManagerService. There was
a bug that causes WindowManagerService to lose the fontScale setting.
Change-Id: I9b993e11f6a3e75735d8ed05596083893543d732
- New feature to "am monitor" to have it automatically launch
gdbserv for you when a crash/ANR happens, and tell you how to
run the client.
- Update dumpstate to match new location of binder debug logs
- Various commented out logs that are being used to track down
issues.
Change-Id: Ia5dd0cd2df983a1fc6be697642a4590aa02a26a5
- New API for iterating over history that will allow a better implementation
in the future.
- Now do writes asynchronously.
Also improve the documentation for Activity.onRetainNonInstanceState().
Change-Id: Idf67f2796a8868eb62f288bcbb2bad29876c8554
Merge commit '7642c82c291051919aad87c797e2dc1d77ed55e1'
* commit '7642c82c291051919aad87c797e2dc1d77ed55e1':
Make input dispatcher only ANR for foreground windows.
Merge commit 'f4d788c9309bc5480100d980608472e4cb04f309' into gingerbread-plus-aosp
* commit 'f4d788c9309bc5480100d980608472e4cb04f309':
Make input dispatcher only ANR for foreground windows.
Redesigned the input dispatcher's ANR timeout mechanism so it is much
closer to Froyo's policy. ANR is only ever signalled if the dispatcher
is waiting on a window to finish processing its previous event(s) and
there is new pending input.
In the old code, we tracked the dispatch timeout separately for each
input channel. This was somewhat complicated and also resulted in the
situation where applications could ANR long after the user had pushed
them into the background.
Change-Id: I666ecada0952d4b95f1d67b9f733842b745c7f4b
This change is essentially a rewrite of the main input dispatcher loop
with the target identification folded in. Since the input dispatcher now
has all of the window state, it can make better decisions about
when to ANR.
Added a .5 second deadline for processing app switch keys. This behavior
predates Gingerbread but had not previously been ported.
Fixed some timing inaccuracies in the ANR accounting that could cause
applications to ANR sooner than they should have.
Added a mechanism for tracking key and motion events that have been
dispatched to a window so that appropriate cancelation events can be
synthesized when recovering from ANR. This change helps to keep
applications in sync so they don't end up with stuck buttons upon
recovery from ANRs.
Added more comments to describe the tricky parts of PollLoop.
Change-Id: I13dffca27acb436fc383980db536abc4d8b9e6f1
