OSDN Git Service

android-x86/frameworks-base.git
9 years agoMove the remaining ZygoteInit native code over to Java.
Elliott Hughes [Wed, 17 Dec 2014 20:28:29 +0000 (12:28 -0800)]
Move the remaining ZygoteInit native code over to Java.

Bug: 18719692
Change-Id: I5f9834edecdbcc065a33a9921c740a53441636a4

9 years agoMerge "Rewrite selectReadable JNI in Java."
Elliott Hughes [Wed, 17 Dec 2014 18:21:06 +0000 (18:21 +0000)]
Merge "Rewrite selectReadable JNI in Java."

9 years agoMerge "Move frameworks users over to libcore hex encoding API."
Narayan Kamath [Wed, 17 Dec 2014 13:33:59 +0000 (13:33 +0000)]
Merge "Move frameworks users over to libcore hex encoding API."

9 years agoMerge "Fix typo with register native allocation."
Tim Murray [Tue, 16 Dec 2014 21:30:57 +0000 (21:30 +0000)]
Merge "Fix typo with register native allocation."

9 years agoMerge "Cleanup Obsolete LOCAL_PRELINK_MODULE."
Ying Wang [Tue, 16 Dec 2014 20:46:17 +0000 (20:46 +0000)]
Merge "Cleanup Obsolete LOCAL_PRELINK_MODULE."

9 years agoCleanup Obsolete LOCAL_PRELINK_MODULE.
Ying Wang [Tue, 16 Dec 2014 20:43:40 +0000 (12:43 -0800)]
Cleanup Obsolete LOCAL_PRELINK_MODULE.

Bug: 18675947
Change-Id: If4909c5f8495136d381b289a3d55e811313abccc

9 years agoRewrite selectReadable JNI in Java.
Elliott Hughes [Mon, 15 Dec 2014 20:39:51 +0000 (12:39 -0800)]
Rewrite selectReadable JNI in Java.

Bug: 18719692
Change-Id: Id4fa95a31f112c720f7a3ac47a3b6c95c3028438

9 years agoFix typo with register native allocation.
Tim Murray [Tue, 16 Dec 2014 17:59:09 +0000 (09:59 -0800)]
Fix typo with register native allocation.

bug 18579193

Change-Id: I42d30709b79a37d6a4126559a511e98e9c8d4808

9 years agoMove frameworks users over to libcore hex encoding API.
Narayan Kamath [Tue, 16 Dec 2014 12:56:23 +0000 (12:56 +0000)]
Move frameworks users over to libcore hex encoding API.

Removes dependency on apache-http.

bug: 18027885
Change-Id: I508db016747985ecdd68e02d5d4efcd0ddd694df

9 years agoMerge "Turn debug flag off"
Christopher Tate [Tue, 16 Dec 2014 02:53:13 +0000 (02:53 +0000)]
Merge "Turn debug flag off"

9 years agoMerge "Reimplement reopenStdio in Java."
Elliott Hughes [Mon, 15 Dec 2014 19:24:52 +0000 (19:24 +0000)]
Merge "Reimplement reopenStdio in Java."

9 years agoReimplement reopenStdio in Java.
Elliott Hughes [Mon, 15 Dec 2014 19:00:25 +0000 (11:00 -0800)]
Reimplement reopenStdio in Java.

Bug: 18719692
Change-Id: I4ed984cd6148c5cba75d46a60f32a0687a5bcac3

9 years agoTurn debug flag off
Alan Jeon [Sat, 13 Dec 2014 13:44:53 +0000 (22:44 +0900)]
Turn debug flag off

Change-Id: Iced1a24ecfc182eccedd3a749fab9dd32dea41da
Signed-off-by: Alan Jeon <skyisle@gmail.com>
9 years agoMerge "Enable native tracking for RS contexts to improve GC behavior."
Tim Murray [Sat, 13 Dec 2014 00:02:30 +0000 (00:02 +0000)]
Merge "Enable native tracking for RS contexts to improve GC behavior."

9 years agoMerge "Lose some unnecessary JNI from the zygote."
Elliott Hughes [Fri, 12 Dec 2014 23:18:48 +0000 (23:18 +0000)]
Merge "Lose some unnecessary JNI from the zygote."

9 years agoMerge "Fix bugs regarding delay the dispatching of non-wakeup alarms"
Christopher Tate [Fri, 12 Dec 2014 22:11:15 +0000 (22:11 +0000)]
Merge "Fix bugs regarding delay the dispatching of non-wakeup alarms"

9 years agoLose some unnecessary JNI from the zygote.
Elliott Hughes [Fri, 12 Dec 2014 22:07:34 +0000 (14:07 -0800)]
Lose some unnecessary JNI from the zygote.

Bug: 18719692
Change-Id: Ie2d50400f10f3c54cfbf94e54656f6d704eac267

9 years agoEnable native tracking for RS contexts to improve GC behavior.
Tim Murray [Fri, 12 Dec 2014 19:34:48 +0000 (11:34 -0800)]
Enable native tracking for RS contexts to improve GC behavior.

This should prevent apps from leaking RS contexts as easily.

bug 18579193

Change-Id: I2d943ce4443ce7cb90ebdd3dd37d338eda6df3a2

9 years agoMerge "[ActivityManager] Fix NPE when start activity"
Olawale Ogunwale [Fri, 12 Dec 2014 13:49:50 +0000 (13:49 +0000)]
Merge "[ActivityManager] Fix NPE when start activity"

9 years agoFix bugs regarding delay the dispatching of non-wakeup alarms
minho.choo [Fri, 12 Dec 2014 07:13:55 +0000 (16:13 +0900)]
Fix bugs regarding delay the dispatching of non-wakeup alarms

checkAllowNonWakeupDelayLocked() method determines that delay the dispatching of non-wakeup alarms.
(there are no wakeup alarms and the screen is off, it can delay until the future.)

if there is any pending non-wakeup alarms, non-wakeup alarm delivery time is bigger than nowELAPSED.
but, checkAllowNonWakeupDelayLocked() returns false and non-wakeup alarm is not delay until future.

i think it is necessary to reverse the inequality symbol on "mNextNonWakeupDeliveryTime > nowELAPSED"
if it isn't, Could you let me know when we get ¡°stuck in a loop¡± in the comment of checkAllowNonWakeupDelayLocked () method?

Change-Id: I7e972064547f4d0a9b4175dbd4cf735b4837abfd
Signed-off-by: Minho Choo <minho.choo@lge.com>
9 years agoMerge "Adjust format strings to work on both 32-bit and 64-bit builds"
Chih-hung Hsieh [Thu, 11 Dec 2014 20:49:29 +0000 (20:49 +0000)]
Merge "Adjust format strings to work on both 32-bit and 64-bit builds"

9 years agoMerge "Avoid tearing down process groups twice."
Narayan Kamath [Thu, 11 Dec 2014 09:15:05 +0000 (09:15 +0000)]
Merge "Avoid tearing down process groups twice."

9 years agoMerge "Remove some crufty "core" tests."
Narayan Kamath [Thu, 11 Dec 2014 09:14:15 +0000 (09:14 +0000)]
Merge "Remove some crufty "core" tests."

9 years agoRemove some crufty "core" tests.
Narayan Kamath [Wed, 10 Dec 2014 19:49:59 +0000 (19:49 +0000)]
Remove some crufty "core" tests.

nobody runs these, and these things are well tested elsewhere
and a part of CTS.

Change-Id: I1ec00f2a085fc82f902f538ace692ac989ca03df

9 years agoMerge "Only one task within the stack should moveToBack"
Craig Mautner [Wed, 10 Dec 2014 17:51:31 +0000 (17:51 +0000)]
Merge "Only one task within the stack should moveToBack"

9 years agoAvoid tearing down process groups twice.
Narayan Kamath [Wed, 10 Dec 2014 10:00:14 +0000 (10:00 +0000)]
Avoid tearing down process groups twice.

The second removal always triggers a warning.

Change-Id: Ieafa8374dbffdaff0d576de004a688cae6f8ac18

9 years agoMerge "[ActivityManager] fix NPE in ActivityStack.resetTargetTaskIfNeededLocked."
Craig Mautner [Tue, 9 Dec 2014 16:33:25 +0000 (16:33 +0000)]
Merge "[ActivityManager] fix NPE in ActivityStack.resetTargetTaskIfNeededLocked."

9 years agoMerge "Fix typo."
Elliott Hughes [Tue, 9 Dec 2014 16:02:56 +0000 (16:02 +0000)]
Merge "Fix typo."

9 years ago[ActivityManager] fix NPE in ActivityStack.resetTargetTaskIfNeededLocked.
Mark Lu [Tue, 9 Dec 2014 06:47:13 +0000 (14:47 +0800)]
[ActivityManager] fix NPE in ActivityStack.resetTargetTaskIfNeededLocked.

Symptom:
in .ActivityStack.resetTargetTaskIfNeededLocked,  "allowTaskReparenting" if case will call setTask
to remove activities from task, that will caused numActivities in main for-loop not consist with task.mActivities size.
caused NPE will happend in finish activities for-loop due to get null object from activities when clearWhenTaskReset" as true case.

Root Cause:
when clearWhenTaskReset as true, will set "end" variable as numActivities -1, but if calling setTask to remove activities from task,
end value will out of date & not same with task.mActivities size.

Solution:
use activities.size() - 1 to assign end value.

Change-Id: I5d7fe22e1df2fc61738db23402e7c42cf6d8c4cc

9 years agoFix typo.
Elliott Hughes [Tue, 9 Dec 2014 04:47:11 +0000 (20:47 -0800)]
Fix typo.

Bug: https://code.google.com/p/android/issues/detail?id=78422
Change-Id: I0dfbb74334e126062660831a4e01817dde068b56

9 years agoMerge "Add more -d dalvik details to dumpsys"
Mathieu Chartier [Tue, 9 Dec 2014 02:05:44 +0000 (02:05 +0000)]
Merge "Add more -d dalvik details to dumpsys"

9 years agoAdd more -d dalvik details to dumpsys
Mathieu Chartier [Tue, 9 Dec 2014 00:20:26 +0000 (16:20 -0800)]
Add more -d dalvik details to dumpsys

Added .NonMoving for the non moving space.
Added .Zygote for the zygote.
Added .IndirectRef for indirect reference tables (jobjects).

Example:
 Dalvik Details
        .Heap     2264     2264        0        0
         .LOS       80       80        0        0
          .GC      144      144        0        0
      .Zygote      331      196        0     2428
   .NonMoving      540      540        0        0
 .IndirectRef       20       20        0        0

Bug: 18672069
Change-Id: Id14e5bc92ceb15fe7a17d9fb035f18aa2925e50c
(cherry picked from commit 25c5e2b12d5d5d06739e3f70a0da968d97935b46)

9 years agoMerge "Remove unsed check for simulator"
Dmitriy Ivanov [Mon, 8 Dec 2014 21:17:59 +0000 (21:17 +0000)]
Merge "Remove unsed check for simulator"

9 years agoMerge "media: fix isSupportedFormat for integer frame rate"
Lajos Molnar [Mon, 8 Dec 2014 18:02:40 +0000 (18:02 +0000)]
Merge "media: fix isSupportedFormat for integer frame rate"

9 years agoMerge "media: allow full range of Vorbis sampling rates"
Lajos Molnar [Mon, 8 Dec 2014 18:02:17 +0000 (18:02 +0000)]
Merge "media: allow full range of Vorbis sampling rates"

9 years agoMerge "Fix PhoneStateListener constructor javadoc."
Narayan Kamath [Mon, 8 Dec 2014 13:35:45 +0000 (13:35 +0000)]
Merge "Fix PhoneStateListener constructor javadoc."

9 years agoFix PhoneStateListener constructor javadoc.
Karol Wrótniak [Thu, 27 Nov 2014 20:29:15 +0000 (21:29 +0100)]
Fix PhoneStateListener constructor javadoc.

Removed sentence in visible javadoc of #PhoneStateListener() which
was useless because it referred to another hidden constructor.

Change-Id: I14a956ac29881000e48bf31e0090e9e1f93bf6e3

9 years agoAdjust format strings to work on both 32-bit and 64-bit builds
Bernhard Rosenkränzer [Sun, 30 Nov 2014 10:04:10 +0000 (11:04 +0100)]
Adjust format strings to work on both 32-bit and 64-bit builds

Adjust format strings to not produce Clang warnings in both 32-bit and
64-bit builds

Change-Id: I76c29d8d5d0fb4b5e9d9518077652370ffe9e871
Signed-off-by: Bernhard Rosenkränzer <Bernhard.Rosenkranzer@linaro.org>
9 years agoRemove unsed check for simulator
Dmitriy Ivanov [Fri, 5 Dec 2014 22:08:04 +0000 (14:08 -0800)]
Remove unsed check for simulator

Change-Id: I46f134d960db330e7b6e7dbfebc84d602d3a3466

9 years agoMerge "Fix NullPointerException in ListView"
Adam Powell [Fri, 5 Dec 2014 19:26:16 +0000 (19:26 +0000)]
Merge "Fix NullPointerException in ListView"

9 years agomedia: fix isSupportedFormat for integer frame rate
Lajos Molnar [Fri, 21 Nov 2014 00:16:55 +0000 (16:16 -0800)]
media: fix isSupportedFormat for integer frame rate

Bug: 18473065
Change-Id: I670cc043d3cb117c26921cb639ff9eecc8f14b0a

9 years agomedia: allow full range of Vorbis sampling rates
Lajos Molnar [Thu, 4 Dec 2014 17:50:06 +0000 (09:50 -0800)]
media: allow full range of Vorbis sampling rates

Bug: 18615604
Change-Id: I51f0c5c81145b94626f2b7e4505b905b63a572bf

9 years agoMerge "Correct check of is user running"
Amith Yamasani [Thu, 4 Dec 2014 23:09:46 +0000 (23:09 +0000)]
Merge "Correct check of is user running"

9 years agoMerge "Dont call close in in BluetoothA2dp finalize"
Mike Lockwood [Thu, 4 Dec 2014 21:41:12 +0000 (21:41 +0000)]
Merge "Dont call close in in BluetoothA2dp finalize"

9 years agoMerge "Prevent ANR when broadcast receiver is killed"
Craig Mautner [Thu, 4 Dec 2014 01:56:39 +0000 (01:56 +0000)]
Merge "Prevent ANR when broadcast receiver is killed"

9 years agoPrevent ANR when broadcast receiver is killed
Kenji Sugimoto [Thu, 17 Jul 2014 05:50:41 +0000 (14:50 +0900)]
Prevent ANR when broadcast receiver is killed

If the process of a BroacastReceiver is dying at the same time
as the system is trying to send an ordered broadcast to the
receiver, the system will try to start the process again. The
BroadcastQueue will store the BroadcastRecord in mPendingBroadcast
to be able to handle it again when the process is awake. A
timeout Message is posted to the handler of the BroadcastQueue.

As part of the shutdown sequence skipCurrentReceiver is called for
the ProcessRecord. This will check if there is a curReceiver set
for the application and make sure to finish the receiver.

Each of the foreground and background BroadcastQueues have their
own handler for managing broadcast timeouts. If the wrong
BroadcastQueue finishes the receiver, the pending timeout Message
will never be cancelled, leading to an ANR report for a receiver
that has already been finished.

Change-Id: I960c0d8f1a8b739b54a8f09f496b32a3498b9e9a

9 years agoMerge "Add -Wno-unused-parameter to LOCAL_CFLAGS"
Chih-hung Hsieh [Thu, 4 Dec 2014 01:04:09 +0000 (01:04 +0000)]
Merge "Add -Wno-unused-parameter to LOCAL_CFLAGS"

9 years agoMerge "RRO: prevent duplicate system overlays"
Adam Lesinski [Wed, 3 Dec 2014 17:44:17 +0000 (17:44 +0000)]
Merge "RRO: prevent duplicate system overlays"

9 years agoCorrect check of is user running
Sergii Skorokhodov [Tue, 11 Nov 2014 07:22:18 +0000 (08:22 +0100)]
Correct check of is user running

Before the ActivityManagerService sends an intent or
starts an activity it checks if target user is in
mStartedUsers array.

When removing a non-owner user process the
UserStartedState instance will still be in mStartedUsers
array with mState STOPPING or SHUTDOWN.

This should be checked before sending an intent or start
an activity.

isUserRunningLocked(...) will interpret mState STOPPING
and SHUTDOWN as a non running user.

Change-Id: I1b51bcdb62bdd0f6dbe05dab4d529d4ad40d0d44

9 years agoMerge "Fix bug in passing of context flags and version."
Jason Sams [Tue, 2 Dec 2014 23:45:16 +0000 (23:45 +0000)]
Merge "Fix bug in passing of context flags and version."

9 years agoMerge changes If2c7e09f,Ie21f227c
Adam Lesinski [Tue, 2 Dec 2014 22:52:36 +0000 (22:52 +0000)]
Merge changes If2c7e09f,Ie21f227c

* changes:
  RRO idmap: pad with 0xffffffff, not 0x00000000
  RRO: reintroduce lost ResTable insert of assets

9 years agoMerge "Keep deleteIntent when copying Notification attributes"
Chris Wren [Tue, 2 Dec 2014 22:31:06 +0000 (22:31 +0000)]
Merge "Keep deleteIntent when copying Notification attributes"

9 years agoFix bug in passing of context flags and version.
Jason Sams [Tue, 2 Dec 2014 20:36:43 +0000 (12:36 -0800)]
Fix bug in passing of context flags and version.

Change-Id: Id20906f6f47b573d881025453a3b6356bcf6e9c7

9 years agoMerge "Consider RTL layout in DPAD navigation in AbsSeekBar"
Alan Viverette [Tue, 2 Dec 2014 19:21:21 +0000 (19:21 +0000)]
Merge "Consider RTL layout in DPAD navigation in AbsSeekBar"

9 years agoMerge "[ActivityManager] Fix race condition in pss collection"
Craig Mautner [Tue, 2 Dec 2014 18:57:34 +0000 (18:57 +0000)]
Merge "[ActivityManager] Fix race condition in pss collection"

9 years agoMerge "bug fix possible NPE while startProcessLocked()"
Craig Mautner [Tue, 2 Dec 2014 17:57:11 +0000 (17:57 +0000)]
Merge "bug fix possible NPE while startProcessLocked()"

9 years agoMerge "[ActivityManager] Invalidate last screenshot when activity relaunched"
Craig Mautner [Tue, 2 Dec 2014 17:50:30 +0000 (17:50 +0000)]
Merge "[ActivityManager] Invalidate last screenshot when activity relaunched"

9 years ago[ActivityManager] Fix race condition in pss collection
louis_chang [Tue, 2 Dec 2014 04:57:39 +0000 (12:57 +0800)]
[ActivityManager] Fix race condition in pss collection

Symptom:
There has a race condition that two threads are accessing
the mPendingPssProcesses simultaneously. One of the thread
is collecting the process pss by looping the mPendingPssProcesses.
The other thread is requesting to collect pss of all processes,
which clears mPendingPssProcesses and adding processes back.

Solution:
Avoid race condition by adding synchornized protection.

Change-Id: Ifb090eda9c4a1b8e3fd980fe0171e9dd77773b46

9 years ago[ActivityManager] Invalidate last screenshot when activity relaunched
louis_chang [Tue, 2 Dec 2014 04:38:05 +0000 (12:38 +0800)]
[ActivityManager] Invalidate last screenshot when activity relaunched

Symptom:
The task thumbnail is not updated when activity relaunched.

Reproduce Steps:
1. Put device in portrait
2. Launch Calculator
3. Launch Recent App (the Calculator's screenshot is correct)
4. Rotate device to landscape
5. Click Calculator in Recent App to return to Calculator (Calculator has relaunched to landsacpe ui)
6. Launch Recent App again (the Calculator's screenshot is not updated)

Change-Id: I92e951ea2ee215c52ca6e50cf6f9e02deb787bce

9 years agobug fix possible NPE while startProcessLocked()
younghwan1.kim [Tue, 2 Dec 2014 04:04:40 +0000 (13:04 +0900)]
bug fix possible NPE while startProcessLocked()

In case, process doesn't create well while startingProcessLocked().
There is possibility to make NPE.
Setting app's crash handler needs to be assigned after null check routine.

Change-Id: I67fb6427f72d93f79fed36eb44c47d37eafdac31

9 years agoMerge "Avoid overwriting EGL14.EGL_NO_SURFACE"
Jesse Hall [Tue, 2 Dec 2014 00:35:11 +0000 (00:35 +0000)]
Merge "Avoid overwriting EGL14.EGL_NO_SURFACE"

9 years agoMerge "Prevent EntropyMixer from registering with service manager."
dcashman [Mon, 1 Dec 2014 23:49:25 +0000 (23:49 +0000)]
Merge "Prevent EntropyMixer from registering with service manager."

9 years agoMerge "Added Razer USA, Ltd serval"
Michael Wright [Mon, 1 Dec 2014 23:47:37 +0000 (23:47 +0000)]
Merge "Added Razer USA, Ltd serval"

9 years agoMerge "Prevent calling GL functions with an invalid surface"
John Reck [Mon, 1 Dec 2014 21:28:26 +0000 (21:28 +0000)]
Merge "Prevent calling GL functions with an invalid surface"

9 years agoMerge "Avoid capture all black screenshot bitmap"
Craig Mautner [Mon, 1 Dec 2014 16:38:14 +0000 (16:38 +0000)]
Merge "Avoid capture all black screenshot bitmap"

9 years agoMerge "Add tests about annotations and enum."
Yohann Roussel [Mon, 1 Dec 2014 09:30:59 +0000 (09:30 +0000)]
Merge "Add tests about annotations and enum."

9 years agoAvoid capture all black screenshot bitmap
tingna_sung [Tue, 21 Oct 2014 05:22:16 +0000 (13:22 +0800)]
Avoid capture all black screenshot bitmap

Captured screenshot bitmap is all black one if none of windows is
surface shown.

Change-Id: I651b2ade1865926d61f945167d60219535fa8bf8
Signed-off-by: tingna_sung <tingna_sung@htc.com>
9 years agoMerge "Add keylayout for Logitech RedHawk Gamepad"
Michael Wright [Mon, 1 Dec 2014 04:36:23 +0000 (04:36 +0000)]
Merge "Add keylayout for Logitech RedHawk Gamepad"

9 years agoMerge "[ActivityManager] Ensure provider external access count will be released."
Craig Mautner [Mon, 1 Dec 2014 00:20:33 +0000 (00:20 +0000)]
Merge "[ActivityManager] Ensure provider external access count will be released."

9 years agoMerge "[ActivityManager] Prevent provider from been created twice."
Craig Mautner [Mon, 1 Dec 2014 00:16:38 +0000 (00:16 +0000)]
Merge "[ActivityManager] Prevent provider from been created twice."

9 years agoMerge "Fix pointer index out of range exception"
Craig Mautner [Mon, 1 Dec 2014 00:13:27 +0000 (00:13 +0000)]
Merge "Fix pointer index out of range exception"

9 years agoMerge "[ActivityManager] Avoid system server crash due to package removed"
Craig Mautner [Mon, 1 Dec 2014 00:11:51 +0000 (00:11 +0000)]
Merge "[ActivityManager] Avoid system server crash due to package removed"

9 years agoMerge "Reset killedByAm flag at attachApplicationLocked"
Craig Mautner [Sun, 30 Nov 2014 23:55:12 +0000 (23:55 +0000)]
Merge "Reset killedByAm flag at attachApplicationLocked"

9 years agoMerge "Prevent showing immersive confirmation unexpectedly"
Craig Mautner [Sun, 30 Nov 2014 23:54:42 +0000 (23:54 +0000)]
Merge "Prevent showing immersive confirmation unexpectedly"

9 years agoMerge "Mention ProGuard in the JNI documentation."
Elliott Hughes [Thu, 27 Nov 2014 18:01:33 +0000 (18:01 +0000)]
Merge "Mention ProGuard in the JNI documentation."

9 years agoMerge "Stop using ErrorStrings for apache."
Narayan Kamath [Thu, 27 Nov 2014 17:53:04 +0000 (17:53 +0000)]
Merge "Stop using ErrorStrings for apache."

9 years agoStop using ErrorStrings for apache.
Narayan Kamath [Thu, 27 Nov 2014 17:20:21 +0000 (17:20 +0000)]
Stop using ErrorStrings for apache.

These strings only ever end up in logcat (at best), so there's no
point having them translated. Also, rename the ErrorStrings class
and move it android.webkit where the last remaining caller lives.

(congrats webview people, this is now your mess to maintain.)

Change-Id: I04dae37c34191b26a69282970318c1b782af1edf

9 years agoMerge "Delete general sounding "CharArrayBuffers" class."
Narayan Kamath [Thu, 27 Nov 2014 13:57:27 +0000 (13:57 +0000)]
Merge "Delete general sounding "CharArrayBuffers" class."

9 years agoDelete general sounding "CharArrayBuffers" class.
Narayan Kamath [Thu, 27 Nov 2014 12:58:52 +0000 (12:58 +0000)]
Delete general sounding "CharArrayBuffers" class.

Move the code to the only point of use. Preparatory work for
decoupling apache-http from the frameworks.

Change-Id: Ieee54bb725cbac19d0c7513867635df6fbcf2b49

9 years agoMerge "Delete unused class from android.net.http."
Narayan Kamath [Thu, 27 Nov 2014 12:58:32 +0000 (12:58 +0000)]
Merge "Delete unused class from android.net.http."

9 years agoDelete unused class from android.net.http.
Narayan Kamath [Thu, 27 Nov 2014 12:52:48 +0000 (12:52 +0000)]
Delete unused class from android.net.http.

Foundation for moving a lot of these classes out of the frameworks
and @removing these APIs.

Change-Id: I590cdd2867b923533de189ec956b4dd3f82efa46

9 years ago[ActivityManager] Prevent provider from been created twice.
riddle_hsu [Sat, 15 Nov 2014 10:25:34 +0000 (18:25 +0800)]
[ActivityManager] Prevent provider from been created twice.

Root Cause:
If a process which contains provider is starting but not
from accessing provider, e.g. service/activity/receiver,
its providers' will be created during handleBindApplication.
And before the process complete publish provider to
ActivityManagerService, if there is an access call to
process provider before the process completes publishing
providers to ActivityManagerService, then the providers
will be created again.

Solution:
Don't schedule install provider if the provider record
already exist and will be published later.

Change-Id: Ic200ceae042c014c86e7edee0acd013b513af334

9 years agoFix pointer index out of range exception
tingna_sung [Sat, 25 Oct 2014 13:52:47 +0000 (21:52 +0800)]
Fix pointer index out of range exception

Avoid pointer index out of range error if return value is -1 of
findPointerIndex() when there is no data available for that pointer
identifier.

Change-Id: I2a08f04c678164e220852094baa124ba157e2e42
Signed-off-by: tingna_sung <tingna_sung@htc.com>
9 years agoAdd keylayout for Logitech RedHawk Gamepad
Mathieu Meisser [Wed, 26 Nov 2014 23:44:42 +0000 (00:44 +0100)]
Add keylayout for Logitech RedHawk Gamepad

Change-Id: Iad03de3384052b360d6b9d887ca9540de68cd21e

9 years agoMerge "Fix bad exclude region in landscape"
Olawale Ogunwale [Wed, 26 Nov 2014 23:13:59 +0000 (23:13 +0000)]
Merge "Fix bad exclude region in landscape"

9 years agoReset killedByAm flag at attachApplicationLocked
Kazuki Nakayama [Thu, 29 May 2014 07:05:34 +0000 (16:05 +0900)]
Reset killedByAm flag at attachApplicationLocked

ActivityManager reuses a process record object that killed
by him under some situation. That reused process record inherits
a killedByAm flag unexpectedly.
The killedByAm flag must be reset otherwise ActivityManager can't
judge the process can be killed or not.

Change-Id: If95137d91939cc44882ad2813131bcde0edd0c1b

9 years ago[ActivityManager] Fix NPE when start activity
louis_chang [Fri, 1 Aug 2014 02:09:08 +0000 (10:09 +0800)]
[ActivityManager] Fix NPE when start activity

Symptom:
NPE occurs in line 1184 (resultStack.sendActivityResultLocked)
because resultStack is null.

Root cause:
When starting activity with FLAG_ACTIVITY_FORWARD_RESULT flag,
the resultRecord could be updated, but the resultStack is not
updated as well. In that case, the resultStack is still be
null. The exception will occurs if the activity is not
granted to launch due to permission denied.

Solution:
Update resultStack when resultRecord updates.

Change-Id: I91634e4f713c2e8dbd1a71f358a8fd9beed83ec7

9 years agoPrevent EntropyMixer from registering with service manager.
dcashman [Tue, 25 Nov 2014 20:59:12 +0000 (12:59 -0800)]
Prevent EntropyMixer from registering with service manager.

EntropyMixer is designed to be triggered in two ways, as either the
result of a self-set timer or of one of three system broadcasts.  It is not
meant to be triggered in any other way, so exposing it via servicemanager
is undesirable.

Bug: 18106000

Cherry-pick of commit: 9287e0dd272b85b475e33bcbd7d868517a0f98f9

Change-Id: I9aeb35e7ffde75090f4234ea193514fb883b1425

9 years agoMerge "ANR caused by incorrect cleanup in BroadcastQueue."
Craig Mautner [Wed, 26 Nov 2014 21:43:59 +0000 (21:43 +0000)]
Merge "ANR caused by incorrect cleanup in BroadcastQueue."

9 years agoANR caused by incorrect cleanup in BroadcastQueue.
Guobin Zhang [Fri, 7 Mar 2014 09:47:10 +0000 (17:47 +0800)]
ANR caused by incorrect cleanup in BroadcastQueue.

Two broadcasts could be sent to the same app simultaneously:
one foreground, one background. For example, LOCALE_CHANGED
and PACKAGE_CHANGED are delievered to com.android.vending
at the same time.
1. AMS started new vending process to handle LOCALE_CHANGED.
   And set app.curReceiver = LOCALE_CHANGED.
2. Before LOCALE_CHANGED is handled by vending process,
   PACKAGE_CHANGED was delievered to vending process too.
   AMS set app.curReceiver = PACKAGE_CHANGED. Bad!
3. Vending process finished handling LOCALE_CHANGED.
   AMS clear app.curReceiver = NULL. Bad!
   And Vending process  killed itself without handling
   PACKAGE_CHANGED.
4. AMS known vending process has died, but didn't know that
   BgBroadcastQueue was still waiting for finish message
   for PACKAGE_CHANGED.
At last, BgBroadcastQueue reported ANR for PACKAGE_CHANGED.

This patch adds protection before clearing app.curReceiver,
only set to NULL if the finishing receiver = app.curReceiver
So handleAppDied would know that PACKAGE_CHANGED was not
finished yet, it will abort the broadcast and continue.

Change-Id: Ic4f31b35e21823d4a3c27712391ecbede213a494
Signed-off-by: Guobin Zhang <guobin.zhang@intel.com>
9 years agoMention ProGuard in the JNI documentation.
Elliott Hughes [Tue, 25 Nov 2014 19:18:43 +0000 (11:18 -0800)]
Mention ProGuard in the JNI documentation.

Change-Id: I7822ec1eca0290eba1f4b08eca6794b41172d606

9 years agoMerge "APK still has privileged flag after being moved from "/system/priv-app""
Jeff Sharkey [Wed, 26 Nov 2014 18:35:10 +0000 (18:35 +0000)]
Merge "APK still has privileged flag after being moved from "/system/priv-app""

9 years agoMerge "[WebView] Allow the WebView to be compiled against the system SDK."
Ignacio Solla [Wed, 26 Nov 2014 18:31:00 +0000 (18:31 +0000)]
Merge "[WebView] Allow the WebView to be compiled against the system SDK."

9 years agoMerge "RRO tests: fix potential NPE"
Conley Owens [Wed, 26 Nov 2014 17:57:08 +0000 (17:57 +0000)]
Merge "RRO tests: fix potential NPE"

9 years agoMerge "RRO tests: adapt tests to split apks"
Conley Owens [Wed, 26 Nov 2014 17:56:47 +0000 (17:56 +0000)]
Merge "RRO tests: adapt tests to split apks"

9 years agoMerge "Fix Presentation window is removed on stack change"
Craig Mautner [Wed, 26 Nov 2014 17:52:59 +0000 (17:52 +0000)]
Merge "Fix Presentation window is removed on stack change"

9 years agoMerge "Sync focus stack frame and layer with focus app."
Craig Mautner [Wed, 26 Nov 2014 17:43:29 +0000 (17:43 +0000)]
Merge "Sync focus stack frame and layer with focus app."

9 years agoMerge "Not include fullscreen opaque window on screenshot"
Craig Mautner [Wed, 26 Nov 2014 17:37:37 +0000 (17:37 +0000)]
Merge "Not include fullscreen opaque window on screenshot"

9 years agoMerge "Measure absolute value of pointer motion distance"
Craig Mautner [Wed, 26 Nov 2014 17:26:45 +0000 (17:26 +0000)]
Merge "Measure absolute value of pointer motion distance"

9 years agoMerge "Track change to Conscrypt"
Kenny Root [Wed, 26 Nov 2014 17:09:40 +0000 (17:09 +0000)]
Merge "Track change to Conscrypt"