Currently, the padding and the glow was inside the individual
notification. This no longer works if we want to adjust the
padding dynamically whether we are on Keyguard or not. This change
moves the padding outside of the individual notifications, and as
a side effect, removes the glow. The glow wasn't really visible with
the new layout, so it's not a breaking change. We have to discuss
with UX first what the new "glow" solution is going to be.
Change-Id: Iac16892cb3b7bb1de3001954b1428796b07950c1
From its beginning, InputMethodService#onUpdateCursor has
provided the cursor position in local coordinates in the attached
view. However, the local coordinates is not useful for IMEs
to render a floating UI near the cursor because the IME is not
able to know the origin of the attached view.
With this CL, CURSOR_ANCHOR_MONITOR_MODE_CURSOR_RECT also means
that the IME will receive the cursor position in screen
coordinates. Because this is a new constant in the next release,
conditionally changing the coordinates never causes
compatibility issues as long as its behavior is well documented.
BUG: 14323360
Change-Id: I3acf2317ae1d763d11dae5ef73c2a1348b377c71
The scroller could crash due to a race condition when updating
the maxheight of the first view.
Bug: 14295010
Change-Id: I911c724a26c8624e2326118e3b392ee675001bc6
System.gc doesn't always do a GC unless to attempt to run the
finalizers after you call System.gc.
Bug: 14325353
Change-Id: Iebed7fad5576d610cea13a86e7d3d46652c67478
The previous fix did not work when the text was ellipsized, because the
test for whether the line was the last line was incorrect. The new test
handles both the end of the buffer and the case where it is the last
line because of ellipsizing.
So this should be the proper fix for bug 14276128 Clipping at bottom of
TextView when lineSpacingMultiplier < 1
This version of the patch also handles the single-line case (which is
computed in BoringLayout rather than StaticLayout).
Change-Id: I88791acc2aa493cc8c599b374f4d213571260b4b
On the app side, requests are now composed by subclassing
from various types of Request objects.
On the service side, starting a voice interaction session
involves starting another service that will then manage the
session. This leads the service design much more to what
we want, where the long-running main service is very tiny
and all the heavy-weight transient session work is elsewhere
in another process.
Change-Id: I46c074c6fe27b6c1cf2583c6d216aed1de2f1143
Add new "ro.config.wallpaper" and "ro.config.wallpaper_component"
properties which may be defined outside of the bundled framework
resources. Falls back to bundled resources when properties are
undefined.
Also look for boot animation under OEM partition.
Bug: 13340779
Change-Id: Ibdc9935dbdaae3319bf63b40573de0503d82ae67
Otherwise derived fingerprints longer than the maximum system property
size put the device into a crash loop.
Bug: 13983493
Change-Id: I8a88e71b1fd396f1cd63b414e3a62bb25010430c
Some devices leave "ro.build.fingerprint" undefined at build time,
since they need to build it from the components at runtime.
Bug: 13340779
Change-Id: I4d74398817af22079a11c196dc6742b85d2bfb61
We probably don't need to deprecate this constant. It can still be used
by partner devices if they choose to continue to support widgets.
Change-Id: Ie8c33446e0c5c32d96ed0670df1772adedf6a5de
