When native webkit code ask java to provide the visibleRect,
the calculation should not take the overscrolling into concern.
The original change was added for this issue, http://b/issue?id=2496502,
and we don't have overscroll any more, so it should be safe to
revert the original change.
bug:3212547
Change-Id: I57e3353b3277f83302ccfd9916a38aa691fa15c7
This extends the view hierarchy's measure pass to allow view to
propagate up to their parent additional information besides just
their measured size. They can now report that their measured width
and/or height should be larger than the size their parent is
limiting them to (even though by definition they need to contrain
their reported measurements to the limits imposed by the parent).
ViewRoot uses this information to determine if it should remeasure
the window with a larger size limit to try to make it fit.
Change-Id: I90af3b7a8ec45d0a5c003fb009857025209d83eb
If the application scrolls the webview without allowing
WebView.java to receive the motion event, stop any long
press timer in progress. This prevents select text from
appearing in applications that host webviews and do their
own flinging.
If the cursor is moved with the dpad or trackball,
reset the hit test state used by long press. This allows
cts tests to run reliably; they emulate key presses
to move between links.
bug:3250291 (app fling)
bug:3241830 (cts fail)
Change-Id: Ida364051ffa090455ddc6c3c3dd41baea5615293
This allows the user to continue scrolling a layer even after reaching the end
of it. Only pin the fling velocity in the direction of the fling. This is
preparation for flinging layers.
Record the inval region when updating layers so that the new layer will draw.
TODO: Keep track of whether the user started on a layer to prevent scrolling a
layer if the user drags across one.
Bug: 3022562
Change-Id: I25f84497a67dfb1e0fc424207f6c97817715501c
The strings used by the framework for network initiated position
requests are now string resources that can be properly localized.
Change-Id: If1cba89adb1bfcb9c8fdb59b9c5aa1107b11279e
This removes the builder pattern from DrawableHolder setters
to avoid reflection and unnecessary GCs in Animators.
Change-Id: I8e3485823f30f10d768d0ab447a5ee3281a363f0
- Add support in access point parser to support static ip settings
- Add Wi-Fi test with static IP configuration
- Use WifiManager new APIs
- Fix issue: if the expected AP is not in the scan list, retry scanning.
Maximum is 5.
- Fix issue: wait for wifi state change before connecting to AP.
Change-Id: Id09b921c8a6a4b14f3c65cdea0b58c7e6f395520
The last span of the TextLine was defining the FontMetrics, instead of
min/maxing the different values.
Bug 3220698
Change-Id: I7394b1699a15aeee4cc38462d561faf329d3e1f6
Use Vendor ID, Product ID and optionally the Version to
locate keymaps and configuration files for external devices.
Moved virtual key definition parsing to native code so that
EventHub can identify touch screens with virtual keys and load
the appropriate key layout file.
Cleaned up a lot of old code in EventHub.
Fixed a regression in ViewRoot's fallback event handling.
Fixed a minor bug in FileMap that caused it to try to munmap
or close invalid handled when released if the attempt to map
the file failed.
Added a couple of new String8 conveniences for formatting strings.
Modified Tokenizer to fall back to open+read when mmap fails since
we can't mmap sysfs files as needed to open the virtual key
definition files in /sys/board_properties/.
Change-Id: I6ca5e5f9547619fd082ddac47e87ce185da69ee6
- do the reply.writeNoException() only if there are NO exceptions
- before, the code could actually generate an exception when asking for the count or the index,
and then the exception could not be unmaarshalled because we previously calling reply.writeNoException()
Change-Id: I241120878c3fc10fea5fbaeb74f9124b1413a3d4
Refactored to use a custom Callback instance instead.
This instance can be shared by different TextView, which no longer have to
be overloaded.
Change-Id: I4749905d8e2057dab2b3ded62bd7c388d13d4e57
This will allow 3rd party apps to access files on digital cameras via the PTP protocol
BUG: 3210830
Change-Id: I06dbf2842a7eea2e0749e2124bb7b7a65e8901c0
Signed-off-by: Mike Lockwood <lockwood@android.com>
In the old code AssetFileDescriptor.AutoCloseInputStream.read()
called super.read(), which calls read(byte[] buffer, int offset, int count),
which decrements mRemaining. Then read() decremented mRemaining again
after super.read() returned.
I'm extending the CTS tests for AssetFileDescriptor.AutoCloseInputStream
to cover this.
Bug: 3240844
Change-Id: I70b945abc773d3ce5c1317dad23a0d797b31f111
Found reproducible cases of Launcher freezing up due to this change
This reverts commit 62490bd358cd001e00cb643cf3879d736c448788.
Change-Id: I1b10e05ee9f0c43566b53c7a11e61fc168d17a75
Add divider attributes to LinearLayout, plus styles for borderless
buttons. Update text field assets.
Change-Id: I673acab1692cc028a0327e8c154069253a4d52e8
By returning the channel object it self it is unnecessary
to have a token. Also, no current code needed it, if its
needed in the future it can be added back.
Change-Id: Ie1d2a1e885f9cd74e95663711ccefb760811bf16
1.) Change from samples per second (persist.sys.profiler_hz) to
interval between samples (persist.sys.profiler_ms) to match
underlying SamplingProfiler API. This allows samples to be taken
less often than a second, which allows lower overhead for always
on profiling.
2.) Add persist.sys.profiler_depth to control the number of frames
kept. Currently defaults to 4 which is the default hprof depth,
but often 12 is necessary even in benchmarks to get a good idea
where time is being spent.
3.) Moved SNAPSHOT_DIR creation to initialization time instead of
checking it on every sample.
4.) Used ThreadFactory to provide human readable name to writeSnapshot
Executor thread.
5.) Fixed bug where writeZygoteSnapshot was calling wrong variant of
writeSnapshot causing profiling to prevent zygote startup. Renamed
underling private writeSnapshot to writeSnapshotFile to try to
prevent future confusion.
Change-Id: Ifcfc343816b19f13a6eef2cbf25cde334d8adc3b
o getSupportedVideoSizes() allows us to retrieve the list of the supported
video sizes from the camera
o getPreferredPreviewSizeForVideoRecording() allows us to retrieve the
preferred/recommended preview size for camcorder applications
bug - 3237021
Change-Id: I52fc9938d389bb411406320b0309775b6f44f4dc
- save Wi-Fi configuration and restore it after the test, other tests
may need data connection
- Fix issue: reset dhcp after parsing a static IP setting.
bug: 3240713
Change-Id: I2158cab6fb6751414893b6a950d9e5bb20021645
When calling startDrag(), the app can now supply an Object to be passed
along in every DragEvent that the app winds up receiving itself. This
object is *not* passed to any other applications; it's strictly app-
local. The purpose is to allow state tracking/management to be done
directly through the drag mechanism rather than requiring out-of-band
code.
An example of the utility here might be TextEdit widgets. A drag that
starts in one TextEdit but ends in a different one should be treated as
a copy/paste operation, where the originating TextEdit is not altered.
However, a drag that starts and ends in the *same* TextEdit is a 'move'
operation within that TextEdit; the text is removed from its original
position and inserted at the drop point. To support this easily, the
drag/drop code in TextEdit can now pass a pointer to the originating
view as the local state object. Then, the drop recipient could tell
whether the drag started within the same TextEdit without needing to
implement any other out-of-band state tracking.
This CL (and its accompanying CLs in a few other packages where the
startDrag() API is being used) adds the new local-state parameter to
the API, but does not actually change the behavior of any existing
clients.
Change-Id: Icba73b2ab4a650b7a94485a19633065b0ef9058c
