OSDN Git Service

android-x86/frameworks-base.git
8 years agoRevert "Remove deprecated methods / classes"
Todd Kennedy [Wed, 24 Feb 2016 22:45:00 +0000 (22:45 +0000)]
Revert "Remove deprecated methods / classes"

This reverts commit 9706c01f9b92cee64bb84782138735225c8588e0.

Change-Id: I14c53ab140220e48f6728b1173ea58760bbeb7f9

8 years agoRemove deprecated methods / classes
Todd Kennedy [Fri, 12 Feb 2016 23:17:36 +0000 (15:17 -0800)]
Remove deprecated methods / classes

Now that we've moved the majority of system components to the
PackageInstaller, we can remove the majority of the "installPackage"
methods and the VerificationParams/ContainerEncryptionParams
classes. There are still some lingering references to the two
installPackage() methods left [both internal and external; play
is using these methods via reflection].

Bug: 24542768
Change-Id: Ib65323d842d9f0e7fed60e37e6b5f103c94c859b

8 years agoMerge "Make ServiceWatcher encryption-aware." into nyc-dev
Jeff Sharkey [Tue, 9 Feb 2016 21:59:03 +0000 (21:59 +0000)]
Merge "Make ServiceWatcher encryption-aware." into nyc-dev

8 years agoMerge "Make InputMethodInfo#isEncryptionAware() strict." into nyc-dev
Yohei Yukawa [Tue, 9 Feb 2016 21:48:33 +0000 (21:48 +0000)]
Merge "Make InputMethodInfo#isEncryptionAware() strict." into nyc-dev

8 years agoMake InputMethodInfo#isEncryptionAware() strict.
Yohei Yukawa [Tue, 9 Feb 2016 21:39:50 +0000 (13:39 -0800)]
Make InputMethodInfo#isEncryptionAware() strict.

This is a follow up CL to the previous CL [1], which unintentionally
relied on whether the entire application rather than the
InputMethodService is marked to be encryption-aware.  This wasn't
obvious because right now LatinIME is entirely marked to be
encryption-aware, which probably needs to be reconsidered in subsequent
CLs.

This CL drops "encryptionAware" line from InputMethodInfo#dump() as it
is redundant.  The same information is already dumped in the
'ServiceInfo' section.

 [1] Icf921fe3661eccf4a589b08b616d05decc561356
     69811a98f161a04af8e8ec9978c3a5efe1ea0f29

Bug: 26279466
Change-Id: I6cea5fe5731f4a6a64ea722b854b99c5154a15ae

8 years agoMerge "Delete only the target user" into nyc-dev
Svetoslav Ganov [Tue, 9 Feb 2016 20:54:29 +0000 (20:54 +0000)]
Merge "Delete only the target user" into nyc-dev

8 years agoDelete only the target user
Svetoslav Ganov [Tue, 9 Feb 2016 20:50:49 +0000 (12:50 -0800)]
Delete only the target user

bug:27094971

Change-Id: I5b43d59ff6edeefc0e05a6ec90d04d388e707c29

8 years agoAAPT2: Switch to protobuf for intermediate format
Adam Lesinski [Thu, 4 Feb 2016 23:59:23 +0000 (15:59 -0800)]
AAPT2: Switch to protobuf for intermediate format

Without needing to conform to the runtime data format,
it is much easier to add new features such as debugging symbols
and carrying over product data to link time.

This also simplifies the runtime format parser and serializer,
which will change much less frequently than the protobuf intermediate
format.

Change-Id: I209787bbf087db0a58a534cb8511c51d21133e00

8 years agoMerge "Importance slider logging." into nyc-dev
Julia Reynolds [Tue, 9 Feb 2016 19:58:32 +0000 (19:58 +0000)]
Merge "Importance slider logging." into nyc-dev

8 years agoMerge "Use Context#createConfigurationContext()." into nyc-dev
Yohei Yukawa [Tue, 9 Feb 2016 19:56:58 +0000 (19:56 +0000)]
Merge "Use Context#createConfigurationContext()." into nyc-dev

8 years agoMerge changes Icf921fe3,Ie1a8ee22 into nyc-dev
Yohei Yukawa [Tue, 9 Feb 2016 19:49:19 +0000 (19:49 +0000)]
Merge changes Icf921fe3,Ie1a8ee22 into nyc-dev

* changes:
  Add a utility method to InputMethodInfo for FBE.
  Fix style issues in InputMethodInfo.java.

8 years agoMerge "Add RSN Information Element ID" into mm-wireless-dev
Randy Pan [Tue, 9 Feb 2016 19:32:18 +0000 (19:32 +0000)]
Merge "Add RSN Information Element ID" into mm-wireless-dev
am: 2750b41d0b

* commit '2750b41d0b50eb66c8aada11d62666f934ca950d':
  Add RSN Information Element ID

8 years agoMerge "Fix a crash in SystemUI when using CarNavigationBar." into nyc-dev
Rakesh Iyer [Tue, 9 Feb 2016 19:28:36 +0000 (19:28 +0000)]
Merge "Fix a crash in SystemUI when using CarNavigationBar." into nyc-dev

8 years agoMerge "Removed the notification scroller top stack" into nyc-dev
Selim Cinek [Tue, 9 Feb 2016 19:25:04 +0000 (19:25 +0000)]
Merge "Removed the notification scroller top stack" into nyc-dev

8 years agoMerge "Add RSN Information Element ID" into mm-wireless-dev
Randy Pan [Tue, 9 Feb 2016 19:13:32 +0000 (19:13 +0000)]
Merge "Add RSN Information Element ID" into mm-wireless-dev

8 years agoMerge "Update documentation for offset and size" into nyc-dev
Glenn Kasten [Tue, 9 Feb 2016 19:06:22 +0000 (19:06 +0000)]
Merge "Update documentation for offset and size" into nyc-dev

8 years agoMerge "updates to FrameStatsObserver API" into nyc-dev
Andres Morales [Tue, 9 Feb 2016 19:00:31 +0000 (19:00 +0000)]
Merge "updates to FrameStatsObserver API" into nyc-dev

8 years agoFix build break
Selim Gurun [Tue, 9 Feb 2016 18:55:33 +0000 (10:55 -0800)]
Fix build break

Bug: 26874366
Change-Id: Ic7634cb9d823d4f74d2893f4c46a77a5af02a0a1

8 years agoFix a crash in SystemUI when using CarNavigationBar.
Rakesh Iyer [Tue, 9 Feb 2016 18:51:50 +0000 (10:51 -0800)]
Fix a crash in SystemUI when using CarNavigationBar.

CarNavigationBar overrides getCurrentView() in phone NavigationBarView
but the dump() method in NavigationBarView uses mCurrentView directly,
which may be null in the auto case. Callign getCurrentView() returns
the correct overriden view.

This patch makes the change to dump() to use getCurrentView().

Bug: 27046536
Change-Id: I92e5655325247d27faac1e33d8622783652bdb10

8 years agoUpdate documentation for offset and size
Glenn Kasten [Sat, 31 Oct 2015 01:25:38 +0000 (18:25 -0700)]
Update documentation for offset and size

Change-Id: Id9d5ef62767155cc8597795b43c4422f623574ed

8 years agoupdates to FrameStatsObserver API
Andres Morales [Wed, 3 Feb 2016 00:19:40 +0000 (16:19 -0800)]
updates to FrameStatsObserver API

- Rename to FrameMetrics to avoid collision with existing
  android.view.FrameStats class
- Make FrameMetricsObserver implementation detail,
  exposing FrameMetricsListener interface as public API
  and wrapping in FrameStatsObserver to maintain state
- Remove dropped frame count call, in favor of passing as
  parameter to callback method.
- Move away from raw timestamp access in favor of Metric IDs
  which represent higher-level, more stable stages in a frame
  lifecycle and match the categories exposed in the onscreen
  bars.
- Support many-to-many Window<->FrameMetricsListener relationship

Change-Id: I00e741d664d4c868b1b6d0131a23f8316bd8c5c2

8 years agoRelax mapping rule to make the mapping logic simple.
Daichi Hirono [Sun, 7 Feb 2016 04:20:22 +0000 (13:20 +0900)]
Relax mapping rule to make the mapping logic simple.

MtpDocumentsProvider remembers the mapping between SAF's ID and MTP's
ID. Sometimes we need to do heuristic to restore the mapping when MTP
device is reconnected.

Previously we do the mapping files that shares the same name more
strictly. For example,

1. Found file name "test.txt". Assign document ID "1".
2. MTP device is disconnected and the MTP ID of "1" is lost.
3. Found two files that have same name "test.txt" in the same directory.

Previously we don't reuse existing document ID "1" for neither of two
"test.txt" because it's not 1-to-1 mapping and we cannot determine which
one should be mapped with existing document ID. It means we need the
complete list of files in a directory to remap IDs. It takes long time
to fetch all file names in a directory when a directory has 100+
files. It's rare that a MTP device has the two files sharing the same
name in the same directory. Also the strict rule makes the mapping code
more complex.

The CL relax the rule of mapping, and it allows to reuse existing
document ID even if it is not 1-to-1 mapping. For the previous example,
it assigns "1" for either of "test.txt".

BUG=27053734
Change-Id: I19406fafc21f13ab94ba99411ce5e7f55ce7f658
(cherry picked from commit acdbc6e740ffbd465488b6eb0cf9388d43ae860a)

8 years agoMerge "Introduce a use32bitAbi flag" into nyc-dev
Selim Gurun [Tue, 9 Feb 2016 18:32:27 +0000 (18:32 +0000)]
Merge "Introduce a use32bitAbi flag" into nyc-dev

8 years agoMerge "Update TelecomManager docs regarding PhoneAccounts" into nyc-dev
Brad Ebinger [Tue, 9 Feb 2016 18:26:26 +0000 (18:26 +0000)]
Merge "Update TelecomManager docs regarding PhoneAccounts" into nyc-dev

8 years agoMerge "WifiEnterpriseConfig: allow getting and setting of ca_path" into mm-wireless-dev
Samuel Tan [Tue, 9 Feb 2016 18:23:43 +0000 (18:23 +0000)]
Merge "WifiEnterpriseConfig: allow getting and setting of ca_path" into mm-wireless-dev
am: 8629390701

* commit '8629390701efe0ed74730a06ab2fc868a7ca265d':
  WifiEnterpriseConfig: allow getting and setting of ca_path

8 years agoIntroduce a use32bitAbi flag
Selim Gurun [Sat, 6 Feb 2016 01:50:53 +0000 (17:50 -0800)]
Introduce a use32bitAbi flag

Bug: 26874366

On Multiarch apps, it might be necessary to prioritize 32bit Abi ahead
of 64bit ones. The use32bitAbi flag enables this.

This CL also reverts the public api changes in  I2c1fd1d036efe72b28b5fe996416df69a583959f and Ie3ecea6d84e2cb1522e736a21c3a3a24ac62eb27. Previously
the same functionality was provided using a raw abi string that
utilized cpuabioverride flag.

Change-Id: Idce3cbfedd11ef9079ce8a2901e69d30b1cf9ef4

8 years agoMerge "WifiEnterpriseConfig: allow getting and setting of ca_path" into mm-wireless-dev
Samuel Tan [Tue, 9 Feb 2016 18:18:04 +0000 (18:18 +0000)]
Merge "WifiEnterpriseConfig: allow getting and setting of ca_path" into mm-wireless-dev

8 years agoMake ACTION_RADIO_OFF protected
Pavel Zhamaitsiak [Mon, 1 Feb 2016 20:04:44 +0000 (12:04 -0800)]
Make ACTION_RADIO_OFF protected

Bug: 25113145
Change-Id: Idcb88cb1fd06d4761ef1f634136f8ec697bf0851
(cherry picked from commit ae6688ca5d2b506eff29ba27c7532513f2da7da8)

8 years agoImportance slider logging.
Julia Reynolds [Tue, 9 Feb 2016 17:57:02 +0000 (12:57 -0500)]
Importance slider logging.

Change-Id: I3c4ad73d1460e4d28cb1a10846159128590d2b20

8 years agoMerge "don't allow assistant to block notifications." into nyc-dev
Chris Wren [Tue, 9 Feb 2016 17:49:43 +0000 (17:49 +0000)]
Merge "don't allow assistant to block notifications." into nyc-dev

8 years agoMerge changes I1eda3134,Ia7ad36a3 into nyc-dev
Winson Chung [Tue, 9 Feb 2016 17:45:04 +0000 (17:45 +0000)]
Merge changes I1eda3134,Ia7ad36a3 into nyc-dev

* changes:
  Updating paging animation to spec.
  Adding logging for timeout and other features.

8 years agoMerge "Whitelist camera permission for SetupWizard" into nyc-dev
Steven Ng [Tue, 9 Feb 2016 17:32:00 +0000 (17:32 +0000)]
Merge "Whitelist camera permission for SetupWizard" into nyc-dev

8 years agoMerge "Fixed bad auto-generated method name." into mm-wireless-dev
Jan Nordqvist [Tue, 9 Feb 2016 17:28:33 +0000 (17:28 +0000)]
Merge "Fixed bad auto-generated method name." into mm-wireless-dev
am: cf132a1bbc

* commit 'cf132a1bbc676489cfa757317edf8af90c10a105':
  Fixed bad auto-generated method name.

8 years agoMerge "Fixed bad auto-generated method name." into mm-wireless-dev
Jan Nordqvist [Tue, 9 Feb 2016 17:24:20 +0000 (17:24 +0000)]
Merge "Fixed bad auto-generated method name." into mm-wireless-dev

8 years agoMerge "Clean up UiAutomation working alongside a11y." into nyc-dev
Phil Weaver [Tue, 9 Feb 2016 16:59:25 +0000 (16:59 +0000)]
Merge "Clean up UiAutomation working alongside a11y." into nyc-dev

8 years agodon't allow assistant to block notifications.
Chris Wren [Tue, 9 Feb 2016 16:25:08 +0000 (11:25 -0500)]
don't allow assistant to block notifications.

Bug: 26640716
Change-Id: I5c7b691ff664e8dafd39bf15cb1556731df2b0b9

8 years agoMerge "Fix dumpsys meminfo -c output." into nyc-dev
Martijn Coenen [Tue, 9 Feb 2016 13:33:44 +0000 (13:33 +0000)]
Merge "Fix dumpsys meminfo -c output." into nyc-dev

8 years agoAdd java.util.function APIs
Neil Fuller [Tue, 19 Jan 2016 00:43:38 +0000 (16:43 -0800)]
Add java.util.function APIs

Bug: 26814204
(cherry picked from commit 311a97e7f41c494bc355f5d77adc276e2a400970)

Change-Id: I3fab2b12676bd1460553b4cb45b1ff602f6c09c1

8 years agoCleanup sliding animation between directories.
Tomasz Mikolajewski [Mon, 8 Feb 2016 09:08:18 +0000 (18:08 +0900)]
Cleanup sliding animation between directories.

The shadow is restored + redundant fade removed.

Bug: 26986946
Change-Id: I31dde7d03f60be470e1c5e06dcdce828d3e4eaa0
(cherry picked from commit 3134ddf99d4215041ed749f3b4e682a1910d208e)

8 years agoFix animation for the empty cabinet, also make other animations less jarring.
Tomasz Mikolajewski [Mon, 8 Feb 2016 08:42:51 +0000 (17:42 +0900)]
Fix animation for the empty cabinet, also make other animations less jarring.

Setting android:animateLayoutChanges=true was causing flickering and problems
with sliding the cabinet while a fragment transition was already in progress.

Bug: 27060052
Bug: 26952361
Change-Id: I17e5f8b57d63cd9a586fc70ae43e030c4f232e8d
(cherry picked from commit ed82fd784301c645cb665517edc3006888d6f237)

8 years agoRemove temporary test fixes.
Tomasz Mikolajewski [Mon, 8 Feb 2016 06:05:24 +0000 (15:05 +0900)]
Remove temporary test fixes.

The tests should fail, as the features are broken.

Change-Id: Ib760ba5469164b8ca0d08c57f24273268ae2436f
(cherry picked from commit 94eca73fec7aa1857f5847df355877d523d0c8a6)

8 years agoMerge "Migrate all tests to ActivityTest." into nyc-dev
Tomasz Mikolajewski [Tue, 9 Feb 2016 09:12:13 +0000 (09:12 +0000)]
Merge "Migrate all tests to ActivityTest." into nyc-dev

8 years agoAdd a utility method to InputMethodInfo for FBE.
Yohei Yukawa [Mon, 8 Feb 2016 22:05:27 +0000 (14:05 -0800)]
Add a utility method to InputMethodInfo for FBE.

This is another preparation for Bug 26279466.  Whether an input method
is marked to be encryption-aware or not would become the thing here and
there in the Framework code.  Having a utility method and debug info
in dumpsys should make our lives easier.

Bug: 26279466
Change-Id: Icf921fe3661eccf4a589b08b616d05decc561356

8 years agoFix style issues in InputMethodInfo.java.
Yohei Yukawa [Mon, 8 Feb 2016 21:57:15 +0000 (13:57 -0800)]
Fix style issues in InputMethodInfo.java.

This has nothing to do with Bug 26279466 but might be something nice to
fix just as a preparation.

Bug: 26279466
Change-Id: Ie1a8ee2248081a70728c29e52c657b3a0ed919b1

8 years agoUse Context#createConfigurationContext().
Yohei Yukawa [Tue, 9 Feb 2016 07:21:37 +0000 (23:21 -0800)]
Use Context#createConfigurationContext().

Nothing is changed in the test expectations.
Major changes in this CL are:

  1. To obtain a custom context with the given locale,
     Context#createConfigurationContext() is the way to go.
     Modifying any internal state of the resource that is owned by the
     context is not supported at all.
  2. As a preparation to test cases where multiple locales are specified
     in the system settings, unit tests need to switch to
     Confituration#setLocales(LocaleList).

Bug: 27076327
Change-Id: I6caa76627677c7625a1c42d76a275f67ae65a9b2

8 years agoTIF: Fix the unresolved link to #stopRecording()
Jae Seo [Tue, 9 Feb 2016 03:24:12 +0000 (19:24 -0800)]
TIF: Fix the unresolved link to #stopRecording()

Change-Id: Ifa5aeb7c6f1d8210650a1d276872aa4e22d5ffa3

8 years agoMerge "TIF: Allow the application to indicate a TV program to record" into nyc-dev
Jae Seo [Tue, 9 Feb 2016 02:14:11 +0000 (02:14 +0000)]
Merge "TIF: Allow the application to indicate a TV program to record" into nyc-dev

8 years agoMake ServiceWatcher encryption-aware.
Jeff Sharkey [Tue, 9 Feb 2016 00:49:17 +0000 (17:49 -0700)]
Make ServiceWatcher encryption-aware.

This utility class automates the process of binding to the best
matching service, including rebinding as packages change over time.

This design means it's easy to become encryption-aware: we use the
default PackageManager behavior that filters components based on
their encryption-aware status, and we just kick off another
evaluation pass once the user is unlocked.

This change cleans up some of the internal logic so we only rebind
when the implementation switches, and it fixes several bugs along
the way.  For example, we would never trigger a rebind if a service
was refactored to a different ComponentName.  Another subtle bug was
that we'd never fallback to an older implementation if a higher
version was uninstalled/disabled.  And finally, if all implementations
were uninstalled/disabled, we'd leave the last connection bound.

Bug: 26280056
Change-Id: I259af78e6564d61353a772ac03cf5799a398d535

8 years agoMerge "Put dismiss end target at navigation bar" into nyc-dev
Jorim Jaggi [Tue, 9 Feb 2016 01:56:23 +0000 (01:56 +0000)]
Merge "Put dismiss end target at navigation bar" into nyc-dev

8 years agoMigrate all tests to ActivityTest.
Tomasz Mikolajewski [Thu, 4 Feb 2016 08:46:35 +0000 (17:46 +0900)]
Migrate all tests to ActivityTest.

ActivityInstrumentationTestCase2 takes care of starting and closing
the activity, which prevents flakyness.

Change-Id: I2d3908f6605a67a3c2d11e18a9c4799443828df9
(cherry picked from commit e366886c7d91b62fe1478a057fc4c0a104d12b02)

8 years agoMerge "Enable RTT Responder Role (1/4)." into mm-wireless-dev
Wei Wang [Tue, 9 Feb 2016 01:14:34 +0000 (01:14 +0000)]
Merge "Enable RTT Responder Role (1/4)." into mm-wireless-dev
am: 03a6b18350

* commit '03a6b1835060b6c2082ea231fb1ad539aefd916e':
  Enable RTT Responder Role (1/4).

8 years agoUpdating paging animation to spec.
Winson [Fri, 5 Feb 2016 23:40:29 +0000 (15:40 -0800)]
Updating paging animation to spec.

- To handle the specific animation spec, we just animate the views
  manually instead of animating the stack scroll (like how we do when
  swiping to dismiss)
- Fixing a regression in settings the initial focused index when
  alt-tabbing
- Minor tweak to make the front most task smaller when in the initial
  non-paging mode

Change-Id: Ic5fd54500fd8ce8284c7aaeddb102b2291bcecac

Signed-off-by: Winson <winsonc@google.com>
8 years agoAdding logging for timeout and other features.
Winson [Fri, 5 Feb 2016 18:39:46 +0000 (10:39 -0800)]
Adding logging for timeout and other features.

Change-Id: Ife9fca1a2ced08c6cbd38ab55d9ea878928de439

Signed-off-by: Winson <winsonc@google.com>
8 years agoMerge "Enable RTT Responder Role (1/4)." into mm-wireless-dev
Wei Wang [Tue, 9 Feb 2016 01:07:33 +0000 (01:07 +0000)]
Merge "Enable RTT Responder Role (1/4)." into mm-wireless-dev

8 years agoMerge "Clean up accessibility gestures." into nyc-dev
Phil Weaver [Tue, 9 Feb 2016 01:00:17 +0000 (01:00 +0000)]
Merge "Clean up accessibility gestures." into nyc-dev

8 years agoFix vector "right arrow" drawable to be properly 24dp.
Oren Blasberg [Mon, 8 Feb 2016 23:17:07 +0000 (15:17 -0800)]
Fix vector "right arrow" drawable to be properly 24dp.

This drawable is used for cascading menus. It was 25dp on one dimension
before, for some reason. I verified it looks fine at 24x24.

Change-Id: Iff5d4f00655d77830d9492190b1a74ef6f408f47
(cherry picked from commit c007e91cc0519adeb4ff35e9e8c436231c2be822)

8 years agoPut dismiss end target at navigation bar
Jorim Jaggi [Thu, 4 Feb 2016 06:04:22 +0000 (22:04 -0800)]
Put dismiss end target at navigation bar

This makes the animation when exiting docked mode a bit nicer
when you fling the divider towards to the navigation bar. However,
since the divider ends at the navigation bar, we need to immediately
dismiss it instead of fading out when the divider is fully occluded
by the navigation bar.

Change-Id: Ic5432fd118cb71be36485667b2c537caf5065ce5

8 years agoClean up accessibility gestures.
Phil Weaver [Tue, 2 Feb 2016 22:47:44 +0000 (14:47 -0800)]
Clean up accessibility gestures.

Closing two small holes in the implementation:
1. The gesture was dispatched before the callback was registered. It
was possible for gestures that failed quickly to fail to report any
status.
2. Gestures could be dispatched before the input filter was
installed. Adding a wait to give the filter a chance to install
before reporting a failure.

Also removing an unused method on the input filter.

Change-Id: I77cd80dcd2cec6c72b3761169aba5eaecf62250b
(cherry picked from commit 03465fb874ccf35ead2228b66ec03afc31d7694e)

8 years agoClean up UiAutomation working alongside a11y.
Phil Weaver [Tue, 2 Feb 2016 01:01:44 +0000 (17:01 -0800)]
Clean up UiAutomation working alongside a11y.

Fix two bugs: one where Instrumentation did not properly
handle destroyed UiAutomations, another where UiAutomation
did not properly disable running accessibility services.

Change-Id: I1773dbd373f4d5b0e7c9917ff18d8d9b1a0e07c7
(cherry picked from commit 846522cea043d3113b3b96daa8f3a1b70e3beb5e)

8 years agoRevert "Revert "TaskManager bench""
Chris Craik [Mon, 8 Feb 2016 19:36:46 +0000 (19:36 +0000)]
Revert "Revert "TaskManager bench""

This reverts commit 9640477e3cc075b0f303e817a3ebcc72d7bc878b.

Change-Id: I3aa8f2830b43b9c7b211c5792a311d0bc698c51a

8 years agoMerge "Changed list metered-networks so it returns all networks." into nyc-dev
Felipe Leme [Mon, 8 Feb 2016 23:26:13 +0000 (23:26 +0000)]
Merge "Changed list metered-networks so it returns all networks." into nyc-dev

8 years agoMerge "Add pid on main bugreport notifications." into nyc-dev
Felipe Leme [Mon, 8 Feb 2016 23:25:38 +0000 (23:25 +0000)]
Merge "Add pid on main bugreport notifications." into nyc-dev

8 years agoMerge "Multi packages per APK" into nyc-dev
Svetoslav Ganov [Mon, 8 Feb 2016 23:20:50 +0000 (23:20 +0000)]
Merge "Multi packages per APK" into nyc-dev

8 years agoMerge "Clamp EditText cursor in the drawable boundaries." into nyc-dev
Siyamed Sinir [Mon, 8 Feb 2016 23:16:48 +0000 (23:16 +0000)]
Merge "Clamp EditText cursor in the drawable boundaries." into nyc-dev

8 years agoMulti packages per APK
Svet Ganov [Thu, 17 Dec 2015 19:35:04 +0000 (11:35 -0800)]
Multi packages per APK

This change introduces the ability to have multiple packages per
APK. The feature is currently restricted to privileged apps and
updates to such apps.

In essence the manifest can have multiple child package declarations.
A child package can declare everything an Android package can except
some tags or attributes that are not applicable and instead inherited
from the parent when needed. For example, the target SDK of the parent
applies to all children.

A child package can be updated only through the parent package.
A package with multiple child packages is installed, uninstalled
atomically - no partial installs where some child packages are not
installed.

The remaining work is to ensure broadcasts are also sent for child
packages. This will come in a subsequent change.

Sample app:ag/848432

Design doc: https://docs.google.com/document/d/18nFWtJuZchLxrHf5SBbJW03-Ky9Rh_G0-OVB14b6u78

Change-Id: I6fd021d981bf5786290e0c53502724a14c97358c

8 years agoMerge "Support for Bundle in AudioAttributes" into nyc-dev
Jean-Michel Trivi [Mon, 8 Feb 2016 23:13:19 +0000 (23:13 +0000)]
Merge "Support for Bundle in AudioAttributes" into nyc-dev

8 years agoTIF: Allow the application to indicate a TV program to record
Jae Seo [Sat, 6 Feb 2016 02:11:35 +0000 (11:11 +0900)]
TIF: Allow the application to indicate a TV program to record

Bug: 26962040
Change-Id: I2f062a4621c8a9a6765543826ba2cd607f1468d4

8 years agoMerge "Do not save TextView text when freezesText is false" into nyc-dev
Siyamed Sinir [Mon, 8 Feb 2016 22:51:31 +0000 (22:51 +0000)]
Merge "Do not save TextView text when freezesText is false" into nyc-dev

8 years agoMerge "Fix TextDropShadowCacheTests and glyph_t everywhere" into nyc-dev
Chris Craik [Mon, 8 Feb 2016 22:47:34 +0000 (22:47 +0000)]
Merge "Fix TextDropShadowCacheTests and glyph_t everywhere" into nyc-dev

8 years agoMerge "Use clang to build microbench" into nyc-dev
Chris Craik [Mon, 8 Feb 2016 22:46:52 +0000 (22:46 +0000)]
Merge "Use clang to build microbench" into nyc-dev

8 years agoMerge "Fix drawing cache on sw drawing" into nyc-dev
John Reck [Mon, 8 Feb 2016 22:46:09 +0000 (22:46 +0000)]
Merge "Fix drawing cache on sw drawing" into nyc-dev

8 years agoFix TextDropShadowCacheTests and glyph_t everywhere
Chris Craik [Sat, 6 Feb 2016 04:10:50 +0000 (20:10 -0800)]
Fix TextDropShadowCacheTests and glyph_t everywhere

Change-Id: I943eae4e9408c77bdfba6304ba7ee3e862351a41

8 years agoUse clang to build microbench
Chris Craik [Sat, 6 Feb 2016 02:03:36 +0000 (18:03 -0800)]
Use clang to build microbench

Change-Id: I7ae8049fe05e99fadf905abd18553cb28596ba7f

8 years agoClamp EditText cursor in the drawable boundaries.
Siyamed Sinir [Tue, 2 Feb 2016 02:30:02 +0000 (18:30 -0800)]
Clamp EditText cursor in the drawable boundaries.

EditText tried to draw outside of the padding boundaries because of a
cursor positioning issue in RTL. This CL removes that fix and instead
clamps the cursor position if cursor is outside of the clipped view
boundary.

Bug: 23397961
Change-Id: Id5f1fbe2a0f571100c89b21758fbb81b14d5da57

8 years agoMerge "Hidden permissions to access blocked numbers." into nyc-dev
Abhijith Shastry [Mon, 8 Feb 2016 22:40:35 +0000 (22:40 +0000)]
Merge "Hidden permissions to access blocked numbers." into nyc-dev

8 years agoImport translations. DO NOT MERGE
Geoff Mendal [Mon, 8 Feb 2016 22:33:00 +0000 (22:33 +0000)]
Import translations. DO NOT MERGE
am: ac534a6c82  -s ours

* commit 'ac534a6c8282f2669bbd5e30836f43d2bfdd5b3d':
  Import translations. DO NOT MERGE

8 years agoDo not save TextView text when freezesText is false
Siyamed Sinir [Sat, 6 Feb 2016 00:08:59 +0000 (16:08 -0800)]
Do not save TextView text when freezesText is false

Currently TextView state is saved whenever selection is set even if
freezesText is false. This causes inconsistencies with the described
behavior for the attribute. This CL updates the behavior as:
- Always save the text for EditText
- Always save the selection if there are any.
- Do not save the text for TextView if freezesText is false.
- During onRestoreInstanceState if selection is out of the text
  boundaries, do not restore the selection.

Bug: 22076905
Change-Id: I5e05d343e752a7d106c8881993e4d95ae21a38ce
(cherry picked from commit 8535836673c801effce097c54f55814ef88448e7)

8 years agoMake permission RECEIVE_EMERGENCY_BROADCAST public.
Amit Mahajan [Fri, 5 Feb 2016 02:08:30 +0000 (18:08 -0800)]
Make permission RECEIVE_EMERGENCY_BROADCAST public.

It is required to receive intent SMS_EMERGENCY_CB_RECEIVED_ACTION
which is already public.

Bug: 24914029
Change-Id: I2d5461fcedb2cb347c818ab6f94577008f6af242
(cherry picked from commit 45d9b36fc6a609893022b812084987cda3e3717d)

8 years agoHidden permissions to access blocked numbers.
Abhijith Shastry [Sat, 6 Feb 2016 02:28:15 +0000 (18:28 -0800)]
Hidden permissions to access blocked numbers.

BUG: 26232372
Change-Id: I213622ef4623dadcf56745798fb3452c57fa4bd5

8 years agoFix drawing cache on sw drawing
John Reck [Mon, 8 Feb 2016 22:13:06 +0000 (14:13 -0800)]
Fix drawing cache on sw drawing

Bug: 26644517
Change-Id: Ie087a05e1dbab75040ca39e0c43df985f8afed3e

8 years agoMerge "Add system APIs for A/B update." into nyc-dev
Tao Bao [Mon, 8 Feb 2016 22:07:04 +0000 (22:07 +0000)]
Merge "Add system APIs for A/B update." into nyc-dev

8 years agoMerge "Password security for FBE disk encryption keys" into nyc-dev
Paul Crowley [Mon, 8 Feb 2016 21:45:45 +0000 (21:45 +0000)]
Merge "Password security for FBE disk encryption keys" into nyc-dev

8 years agoFixed issue with drag resize state when animating pinned stack.
Wale Ogunwale [Sat, 6 Feb 2016 06:51:01 +0000 (22:51 -0800)]
Fixed issue with drag resize state when animating pinned stack.

When animating the pinned stack, we set drag resizing on the top
most task in the stack. This has a couple of issues.
- Only the top most task is put in drag sizing mode and all other
task in the stack will not be in resizing mode.
- The top most task of the stack can change during the animation,
so we fail to clear the drag resize flag on the previous top task.

We now track drag sizing at the stack level and have the stack
drag resizing state affect its tasks drag resizing states.

Also added concept of continuing a bounds animation if the same
target called BoundsAnimationController.animateBounds before the
current animation is done. We don't send onAnimationEnd() for the
current animation that is been cancelled and don't send
onAnimationStarted() for the animation that will be replacing it.

Bug: 25672053
Change-Id: I64e89ed09d81e4802dacebc5818dfa1deb0d588f

8 years agoRemoved the notification scroller top stack
Selim Cinek [Thu, 4 Feb 2016 21:32:43 +0000 (13:32 -0800)]
Removed the notification scroller top stack

The top stack is now removed and the notifications
can scroll below the quick settings header.

Bug: 24866646
Change-Id: Ia480a0efb3ed70a072e0596633063cc0cbbdf1fb

8 years agoDo not update Record.visible if not setting visibility
Jorim Jaggi [Fri, 5 Feb 2016 20:41:17 +0000 (12:41 -0800)]
Do not update Record.visible if not setting visibility

In resumeTopActivityInnerLocked we only update the visibility
if not already visible, to avoid a black flicker. However, in the
interaction when docking a task from recents, makeVisible was
called but the conditions weren't met to make it actually visible,
but Record.visible was still set, leading that we didn't call
WM.setAppVisibility in resumeTopActivityInnerLocked

Bug: 27041808
Change-Id: I400470a5b31881542fedd3edb6a97e524666a8a3

8 years agoMulti-window optimizations #3
Jorim Jaggi [Fri, 5 Feb 2016 22:26:16 +0000 (14:26 -0800)]
Multi-window optimizations #3

- Make sure divider is pushing frames only if necessary
- When starting recents while docking, make sure no unnecessary
animations are played to reduce system load.

Change-Id: I536f009aef7282ed85dde79b853aeae62644f3e3

8 years agoMulti-window optimizations #2
Jorim Jaggi [Fri, 5 Feb 2016 03:49:28 +0000 (19:49 -0800)]
Multi-window optimizations #2

- Do not relaunch activities in stacks that are not visible during
resize.

Change-Id: Ia6bd17882c17db045450a466ba5347ce53e9b837

8 years agoMulti-window optimizations #1
Jorim Jaggi [Fri, 5 Feb 2016 03:34:00 +0000 (19:34 -0800)]
Multi-window optimizations #1

- When the window doesn't have a surface, do not add it to
mResizingWindows, so we don't report unnecessary resizes
- computeDragResizing => false when window is not visible, so we
never enter resizing mode even if the window decides to relayout
in the background

Change-Id: I8e6cdef86f1579d128973d4f2f12e87bf9b65524

8 years agoFix disappearing translucent status bar background
Jorim Jaggi [Fri, 5 Feb 2016 02:34:50 +0000 (18:34 -0800)]
Fix disappearing translucent status bar background

When relaunching an activity while preserving the window, the content
view was put on top of the status bar background, making it disappear
in case of translucent status bar backgrounds.

Change-Id: I1edff29c10616bf5386e982d7336327135f0f7a0

8 years agoMerge "Import translations. DO NOT MERGE" into nyc-dev
Geoff Mendal [Mon, 8 Feb 2016 21:14:32 +0000 (21:14 +0000)]
Merge "Import translations. DO NOT MERGE" into nyc-dev

8 years agoImport translations. DO NOT MERGE
Geoff Mendal [Mon, 8 Feb 2016 21:10:39 +0000 (13:10 -0800)]
Import translations. DO NOT MERGE

Change-Id: I477868c67814a6bd353474266dbff375d3b26b4c
Auto-generated-cl: translation import

8 years agoEnsure surfaces stay alive until activity stop.
Robert Carr [Wed, 3 Feb 2016 06:43:27 +0000 (22:43 -0800)]
Ensure surfaces stay alive until activity stop.

Prior to this commit in this case of activity pause, with finishing=true
the activity manager will notify us of app visibility and we will begin
an exit animation. When this exit animation finishes, we will destroy
the application surface (unless its eligible for saving). However there
are two cases where this breaks down:

1. The exit animation finishes before the activity thread handles
the stop transition. Many activities stop rendering on Pause
but many do not and it is totally legal to do so. Sometimes this
results in non fatal dequeue buffer errors and sometimes results in
fatal errors with Pixel Buffers, etc...
2. We may resume the activity shortly after asking the window manager
to pause it. If the window wasn't eligible for animation, we will
immediately destroy it after being told of the visibility change
following PAUSE_FINISHING. It's possible for this to complete
before we process the resume. On the other hand the client
happilly processes the resume and transitions back from PAUSE
and then crashes once it attempts to use it's surface.

In this commit we have the activity manager notify the window manager
when an application has actually finished (or we have timed out
waiting). For windows which have not been explicitly removed by the
client, we defer destruction until we have received both this signal
and the animation has completed.

Bug: 26793431
Change-Id: Ib6ee8fbdd1f03450fbbfd62468a19e97774befab

8 years agoChanged list metered-networks so it returns all networks.
Felipe Leme [Mon, 8 Feb 2016 18:12:01 +0000 (10:12 -0800)]
Changed list metered-networks so it returns all networks.

Prior to this change it was relying on NMPS.getNetworkPolicies(), which
would not return networks that never had any policy attribute (like
metered) set.

This change fix this problem by also querying the WifiManager for the
list of saved wi-fi networks. In fact, the new logic is very similar to
what's done on SettingLib's NetworkPolicyEditor.

BUG: 26685616
Change-Id: I42f37de3a7367718036f45b15c6b5d37b67e7d01

8 years agoFix Fade transition interrupt.
George Mount [Thu, 4 Feb 2016 23:24:30 +0000 (15:24 -0800)]
Fix Fade transition interrupt.

Bug 26963113

When a Fade transition is interrupted and reversed, the
View started the animation from the beginning. This change
captures the previous transitionAlpha and starts the animation
from the previous alpha state.

Change-Id: I801fe9ade6af4cf8446838e231bdc71841668a18
(cherry picked from commit 3cf9fa3db0231d035142c27992818d8f7827ac3e)

8 years agoWhitelist camera permission for SetupWizard
Steven Ng [Wed, 3 Feb 2016 11:53:25 +0000 (11:53 +0000)]
Whitelist camera permission for SetupWizard

+ Camera permission is required for QR provisioning in SetupWizard.

Bug: 26932941
Change-Id: I1573cdd3b8ac708f457a9c7efbbf02c3c8ce6d8d
(cherry picked from commit 950b2dee6648a90bd2e9fedecde137f41f7c4513)

8 years agoMerge "Move graphics and JIT caches to DE storage." into nyc-dev
Jeff Sharkey [Mon, 8 Feb 2016 19:47:20 +0000 (19:47 +0000)]
Merge "Move graphics and JIT caches to DE storage." into nyc-dev

8 years agoMerge "Import translations. DO NOT MERGE" into nyc-dev
Geoff Mendal [Mon, 8 Feb 2016 19:25:33 +0000 (19:25 +0000)]
Merge "Import translations. DO NOT MERGE" into nyc-dev

8 years agoSupport for Bundle in AudioAttributes
Jean-Michel Trivi [Fri, 5 Feb 2016 23:20:35 +0000 (15:20 -0800)]
Support for Bundle in AudioAttributes

Bug 25449464

Change-Id: I99cdfd4589e3161b52620e439ba993ce21d256b4

8 years agoMove graphics and JIT caches to DE storage.
Jeff Sharkey [Mon, 8 Feb 2016 19:16:00 +0000 (12:16 -0700)]
Move graphics and JIT caches to DE storage.

We confirmed with the graphics and JIT teams that no sensitive
user data is written to these caches, so they're safe to point at DE
storage.

Since we don't have control over what is written by the app, we need
to keep the cache environment variable pointing at CE storage.

Fix ensurePrivateDirExists() to always return a path, instead of
returning null which can cause scary bugs.

Change packages.list to no longer canonicalize data paths, since
these fail when CE storage is still locked.

Bug: 27069522
Change-Id: Ifff64a036fa4aa1e61aa0dd98486bc711fbf8f4a

8 years agoImport translations. DO NOT MERGE
Geoff Mendal [Mon, 8 Feb 2016 19:22:17 +0000 (11:22 -0800)]
Import translations. DO NOT MERGE

Change-Id: I14edb5a78192ca001f700de6dcc57c1c55b528fc
Auto-generated-cl: translation import