This change essentially backs out ag/683646, which added an API to click
on ClickableSpans within a TextView. This API has the flaw, however,
that ClickableSpans are not parcelable, so they are not in general
reported to AccessibilityServices. That means that services will have no
idea what they are activating.
Since the API is not usable end-to-end, I'm backing it out before the
API is final.
Bug: 17726921
Change-Id: I541c6218f2899ff67a6b32a13fd9ae6f3b53b3c4
The root cause of Bug 18920212 is that when AutoCompleteTextView tries
to show completion window upon text input,
InputMethodManager#focusInLocked() can be called with a View that is
temporarily detached, which should be ignored.
Bug: 18920212
Change-Id: Ia79bbd8468f768d546354382b47b39dd31ef7bb5
When using applyOverrideConfiguration, subsequent callbacks to
onConfigurationChanged didn't take into account the locally overridden
configuration. This patches the incoming configuration to match
what is expected by the application.
Bug:27644297
Change-Id: I3a090e9862a56470d999aa0aa4d5bae29f533a11
If the relevant frames for the app don't actually change, there is
no reason to call forceLayout and force a layout pass.
Bug: 27864358
Change-Id: I2add8cc5791023937afc154e6e50bbf7efb35adb
bug:27918183
Fixes an issue where all convex outlines were pushed as empty rects to
RenderNode.
Also adds outline/reveal clip property logging, so such issues are
easier to debug in the future.
Change-Id: Ic4a996ecd09a8ef84cdf8b963bdb4853c7f6d180
Update configuration with WM first and check if the stacks need to be
resized due to the update. If so, let activity manager resize the stacks
inline, instead of letting WM schedule another pass of resizeStack. This
way the configuration will be updated to the latest before ensureActivity-
ConfigurationLocked, and we don't need another relaunch there.
bug: 27834014
Change-Id: Ib761a96cada0c3247b0480f18370670c593159da
Add WebViewDelegate.setDrawGLFunctionDetachedCallback system API that's
used for webview to receive the functor detach callback.
BUG: 27709981
Change-Id: Ie6b5e445c0090a181f94fcd2ec1ea77095c9cb03
In order to fix Bug 18920212, we have to track when a View enters
temporarily detached state and when it exits from that state. To do
that, ListView needs to use View#dispatchStartTemporaryDetach() instead
of directly calling View#onStartTemporaryDetach() because there is no
guarantee that existing applications have internally followed Call-Super
pattern.
With this CL, we are going to expose temporary detach state and its
dispatching methods as public APIs. Major changes are:
1. ListView's indirect children will start receiving temporary
dispatch callbacks. Previously only direct children have received
View#on{Start, Finish}TemporaryDetach() callbacks.
2. TextView can no longer assume that ListView never calls
View#View#dispatchStartTemporaryDetach() but directly call
View#onStartTemporaryDetach() instead. See the commit message
of [1] for details.
This also enables us to do the following fixes, which will be handled in
subsequent CLs.
A. ViewCompat support lib is finally able to rely on temporary
dispatch mechanism without reflection.
B. InputMethodManager is now able to ignore focus-in events from
temporarily detached Views. This will be done in the next CL [2].
[1]: a440b002aa
[2]: Ia79bbd8468f768d546354382b47b39dd31ef7bb5
Bug: 18920212
Bug: 27905921
Change-Id: If8f780f8b71754f7533a65097304113ae1f5cf12
This reverts commit 35e2ea0203.
This patch was based on two different wrong assumptions.
Bug 27822069
Change-Id: I20b1017f91f3fce3c23dd8446459d6f3e3150a48
This is an input consumer similar to the one used when hiding the navbar,
but placed above wallpapers. It might be useful for processing touch
events over "desktop" in freeform MW mode.
Re-landing I9d6d28a624f750ad48fc39f9b149dd1f989cceba after fixing build.
Bug:26688904
Change-Id: I89fdabd9c72cdd4a1d7ca626c33ddc99ddea97f9
This is an input consumer similar to the one used when hiding the navbar,
but placed above wallpapers. It might be useful for processing touch
events over "desktop" in freeform MW mode.
Bug:26688904
Change-Id: I9d6d28a624f750ad48fc39f9b149dd1f989cceba
This reverts commit e9a4f87852.
This revert is to test to see if the gralloc lock problems are caused by this CL.
Change-Id: I395fd3675dd430c44a5d29b050ba248f3f2bf4cd
It turns out that BaseInputConnection has still depended on a private
API named BaseInputConnection#reportFinish(), which was introduced
4 years ago to work around a UI freeze due to an unbalanced batch edit
count [1]. Note that such an unbalanced batch edit count cannot always
be avoidable. It can easily occur in the following situations.
- The current IME crashed during batch edit.
- The user changed the View focus during batch edit.
- The current IME called IMM#switchToNextInputMethod() during batch
edit.
The remaining problem is that #reportFinish() is still an internal API
and only subclasses of BaseInputConnection can implement it, and IMM
calls it when and only when the current InputConnection is
BaseInputConnection or its subclass. InputConnectionWrapper and any
other InputConnection implementations will never receive such a callback
to clean up InputConnection#{begin, end}BatchEdit(), which is considered
to be a major contributor to UI freeze.
To address the above issue, we unhide BaseInputConnection#reportFinish()
as InputConnection#closeConnection() so that application developers can
receive an appropriate callback to clean up internal state including
unfinished batch edit.
[1] I5525d776916f0c42d5e6d4a4282aed590d7f0e9a
9d69ecbf61
Bug: 24688781
Bug: 25332806
Change-Id: I234309c5880c9fe0b299b8bd0f8862796d4dda0d
Also add support for shortcuts with no base character or keycode,
as needed by "Search" being triggered with just Meta.
Bug: 27454997
Change-Id: I75bc9d22c30f9ebcdcbea3fb53f166d942dc7b90
The extras may contain custom Parcelables which aren't in the
system's default classpath, so touching them would end up clobbering
the contents.
Update code to leave the parcelled data untouched inside the bundle
until someone tries reading the extras.
Bug: 27897919
Change-Id: I99381dd50c9a0e8887667076362ea98805f0437c
We need to change the SurfaceView size from the UI thread
so that we can appropriately deliver the SurfaceChanged
callback. We also need to not preserve geometry
in this case, as if we don't update the surface
and layout size together we could get scaling. This still has
some potential for holes, as transactions are not synced with
the parent renderer, but we have other methods to avoid
these in the case of resizing. This fixes the remaining
issues with content sizing and surface view "out of sync".
Bug: 27780983
Bug: 27687126
Bug: 27676101
Change-Id: Idd7864f00e5cf7a4eb32dd66c0b389292a788069
This reverts commit 3127c2a471.
The original CL caused various issues for which we don't have the
time/not willing to take the risk.
Bug: 27864389
Bug: 27451341
View.hasOverlappingRendering() is an important performance tweak that
subclasses can override to do the right thing return false when appropriate
to avoid expensive operations when view is translucent).
But this requires subclassing View to get this behavior.
This new API allows the property to be set from outside, enabling
standard views to have this behavior set. When the new method is called,
the behavior will derive from whatever it was set to. Otherwise, it
will default to the old/overriden behavior.
Issue #16561361 Make hasOverlappingRendering settable from outside/XML
Change-Id: If0fbc8667cdb82b1d85e795e782716a07196f3c0
