OSDN Git Service

android-x86/frameworks-base.git
10 years agoMerge "Fix a JNI local reference leak in MediaDrm"
Jason Sams [Sat, 11 Jan 2014 00:42:19 +0000 (00:42 +0000)]
Merge "Fix a JNI local reference leak in MediaDrm"

10 years agoMerge "Fix visibility of multiple non-fullscreen activities."
Craig Mautner [Fri, 10 Jan 2014 23:55:43 +0000 (23:55 +0000)]
Merge "Fix visibility of multiple non-fullscreen activities."

10 years agoMerge "Cursor leakage due to cancellation signal"
Christopher Tate [Thu, 9 Jan 2014 20:41:26 +0000 (20:41 +0000)]
Merge "Cursor leakage due to cancellation signal"

10 years agoFix visibility of multiple non-fullscreen activities.
riddle_hsu [Thu, 9 Jan 2014 12:24:34 +0000 (20:24 +0800)]
Fix visibility of multiple non-fullscreen activities.

Issue detail:
Assume X, Y are non-fullscreen activities.
 a.Home starts an activity X in task A in application stack.
 b.X starts an activity Y in <task A> or <new task B>
 c.Activity X will be invisible.

How to fix:
Because the function "isActivityOverHome" means an activity is able to see home.
But there may have many non-fullscreen activities between the top non-fullscreen activity and home.
If flag "behindFullscreen" is set, those middle activities will be invisible.
So it should only take care from who is adjacent to home.
Then check two flags frontOfTask(task root) and mOnTopOfHome for constraining the condition.

Change-Id: I60bcea304976414e44835a0a38675aae365e9e19

10 years agoMerge "telephony: Add same named operator configuration for some operators"
Wink Saville [Wed, 8 Jan 2014 19:25:11 +0000 (19:25 +0000)]
Merge "telephony: Add same named operator configuration for some operators"

10 years agoMerge "Added missing USB device descriptor fields needed for intent filters"
Mike Lockwood [Wed, 8 Jan 2014 15:56:13 +0000 (15:56 +0000)]
Merge "Added missing USB device descriptor fields needed for intent filters"

10 years agoMerge "Fix a bug introduced in change 73712"
Narayan Kamath [Wed, 8 Jan 2014 12:47:31 +0000 (12:47 +0000)]
Merge "Fix a bug introduced in change 73712"

10 years agoFix a bug introduced in change 73712
Narayan Kamath [Wed, 8 Jan 2014 12:26:28 +0000 (12:26 +0000)]
Fix a bug introduced in change 73712

"return (bool) true" was incorrectly changed to
"return (jboolean) JNI_FALSE"

Change-Id: I45f62f60c5affbae7970bb8889c55981e26d01ba

10 years agoMerge "AArch64: Use long for pointers in App/Backup"
Narayan Kamath [Wed, 8 Jan 2014 11:54:29 +0000 (11:54 +0000)]
Merge "AArch64: Use long for pointers in App/Backup"

10 years agoAArch64: Use long for pointers in App/Backup
Ashok Bhat [Thu, 2 Jan 2014 16:52:41 +0000 (16:52 +0000)]
AArch64: Use long for pointers in App/Backup

For storing pointers, long is used, as
native pointers can be 64-bit.

In addition, some minor changes have been done
to conform with standard JNI practice (e.g. use
of jint instead of int in JNI function prototypes)

Change-Id: I7aee49dc26cf6c86af8f1d882e9cd1cc145a1977
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
Signed-off-by: Marcus Oakland <marcus.oakland@arm.com>
Signed-off-by: Kévin PETIT <kevin.petit@arm.com>
10 years agoMerge "AArch64: Use long for pointers"
Narayan Kamath [Wed, 8 Jan 2014 11:53:50 +0000 (11:53 +0000)]
Merge "AArch64: Use long for pointers"

10 years agoAArch64: Use long for pointers
Ashok Bhat [Thu, 2 Jan 2014 13:42:56 +0000 (13:42 +0000)]
AArch64: Use long for pointers

For storing pointers, long is used in CursorWindow
and SQLiteConnection classes as native pointers can
be 64-bit.

Change-Id: Ia686006a7b8bdc7b95e5de0d0a294b155034a921
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
Signed-off-by: Marcus Oakland <marcus.oakland@arm.com>
Signed-off-by: Kévin PETIT <kevin.petit@arm.com>
10 years agoMerge "AArch64: Use long for pointers in hardware classes"
Narayan Kamath [Wed, 8 Jan 2014 11:53:11 +0000 (11:53 +0000)]
Merge "AArch64: Use long for pointers in hardware classes"

10 years agoMerge "AArch64: Use long for pointers in services"
Narayan Kamath [Wed, 8 Jan 2014 11:53:00 +0000 (11:53 +0000)]
Merge "AArch64: Use long for pointers in services"

10 years agoAArch64: Use long for pointers in hardware classes
Ashok Bhat [Fri, 3 Jan 2014 14:37:19 +0000 (14:37 +0000)]
AArch64: Use long for pointers in hardware classes

For storing pointers, long is used in hardware classes,
as native pointers can be 64-bit.

In addition, some minor changes have been done
to conform with standard JNI practice (e.g. use
of jint instead of int in JNI function prototypes)

Change-Id: Icdeb67f9273fb2d8f6d88ca68d7f7d0950796fc1
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
Signed-off-by: Marcus Oakland <marcus.oakland@arm.com>
Signed-off-by: Kévin PETIT <kevin.petit@arm.com>
10 years agoAArch64: Use long for pointers in services
Ashok Bhat [Thu, 2 Jan 2014 19:43:30 +0000 (19:43 +0000)]
AArch64: Use long for pointers in services

For storing pointers, long is used, as native pointers
can be 64-bit.

In addition, some minor changes have been done
to conform with standard JNI practice (e.g. use
of jint instead of int in JNI function prototypes)

Change-Id: Ib4435f0794740d545c1e640087849215e6844802
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
Signed-off-by: Marcus Oakland <marcus.oakland@arm.com>
10 years agoMerge "Revert "Do not show roaming icon for Videotron operator when access Canada...
Robert Greenwalt [Wed, 8 Jan 2014 00:38:29 +0000 (00:38 +0000)]
Merge "Revert "Do not show roaming icon for Videotron operator when access Canada network""

10 years agoMerge "Add a call to registerAppInfo for the VMRuntime"
Ian Rogers [Tue, 7 Jan 2014 21:31:17 +0000 (21:31 +0000)]
Merge "Add a call to registerAppInfo for the VMRuntime"

10 years agoMerge "AArch64: Use long for pointers in MTP classes"
Narayan Kamath [Tue, 7 Jan 2014 12:54:11 +0000 (12:54 +0000)]
Merge "AArch64: Use long for pointers in MTP classes"

10 years agoAArch64: Use long for pointers in MTP classes
Ashok Bhat [Tue, 17 Dec 2013 19:04:19 +0000 (19:04 +0000)]
AArch64: Use long for pointers in MTP classes

For storing pointers, long is used in MTP classes,
as native pointers can be 64-bit.

In addition, some minor changes have been done
to conform with standard JNI practice (e.g. use
of jint instead of int in JNI function prototypes)

Change-Id: I67805547251722e7b77611d47d0bb632a64d3e6d
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
Signed-off-by: Marcus Oakland <marcus.oakland@arm.com>
Signed-off-by: Kévin PETIT <kevin.petit@arm.com>
10 years agoMerge "AArch64: Use long for pointers in VideoEditor classes"
Narayan Kamath [Tue, 7 Jan 2014 12:50:07 +0000 (12:50 +0000)]
Merge "AArch64: Use long for pointers in VideoEditor classes"

10 years agoAArch64: Use long for pointers in VideoEditor classes
Ashok Bhat [Tue, 17 Dec 2013 14:13:08 +0000 (14:13 +0000)]
AArch64: Use long for pointers in VideoEditor classes

For storing pointers, long is used in VideoEditor
classes, as native pointers can be 64-bit.

In addition, some minor changes have been done
to conform with standard JNI practice (e.g. use
of jint instead of int in JNI function prototypes)

Change-Id: Ifff3a28f2ab6774ee89d31770ad63451c8726431
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
Signed-off-by: Marcus Oakland <marcus.oakland@arm.com>
Signed-off-by: Kévin PETIT <kevin.petit@arm.com>
10 years agoMerge "AArch64: Use long for pointers in DisplayEventReceiver"
Narayan Kamath [Tue, 7 Jan 2014 12:49:27 +0000 (12:49 +0000)]
Merge "AArch64: Use long for pointers in DisplayEventReceiver"

10 years agoAArch64: Use long for pointers in DisplayEventReceiver
Ashok Bhat [Wed, 18 Dec 2013 18:00:05 +0000 (18:00 +0000)]
AArch64: Use long for pointers in DisplayEventReceiver

For storing pointers, long is used as
native pointers can be 64-bit.

In addition, some minor changes have been done
to conform with standard JNI practice (e.g. use
of jint instead of int in JNI function prototypes)

Change-Id: I3c0855373c0e4bedc172adb82b103586de9219dc
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoMerge "AArch64: Use long for pointers in audio effect classes"
Narayan Kamath [Tue, 7 Jan 2014 12:48:21 +0000 (12:48 +0000)]
Merge "AArch64: Use long for pointers in audio effect classes"

10 years agoAArch64: Use long for pointers in audio effect classes
Ashok Bhat [Tue, 17 Dec 2013 12:42:00 +0000 (12:42 +0000)]
AArch64: Use long for pointers in audio effect classes

For storing pointers, long is used in AudioEffect
and Visualizer classes, as native pointers can be 64-bit.

In addition, some minor changes have been done
to conform with standard JNI practice (e.g. use
of jint instead of int in JNI function prototypes)

Change-Id: Ief49bbe5baef663e268c7f6d70ac33a4df45621c
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
Signed-off-by: Marcus Oakland <marcus.oakland@arm.com>
Signed-off-by: Kévin PETIT <kevin.petit@arm.com>
10 years agoMerge "AArch64: Use long for pointers in AmrInputStream"
Narayan Kamath [Tue, 7 Jan 2014 12:45:06 +0000 (12:45 +0000)]
Merge "AArch64: Use long for pointers in AmrInputStream"

10 years agoAArch64: Use long for pointers in AmrInputStream
Ashok Bhat [Tue, 17 Dec 2013 14:10:25 +0000 (14:10 +0000)]
AArch64: Use long for pointers in AmrInputStream

For storing pointers, long is used in AmrInputStream
class, as native pointers can be 64-bit.

Change-Id: I78fff3eb54fc7101746b0d693cbc4e923c2bad08
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
Signed-off-by: Marcus Oakland <marcus.oakland@arm.com>
Signed-off-by: Kévin PETIT <kevin.petit@arm.com>
10 years agoMerge "XmlUtils#readThisMapXml should handle null keys."
Narayan Kamath [Tue, 7 Jan 2014 10:03:20 +0000 (10:03 +0000)]
Merge "XmlUtils#readThisMapXml should handle null keys."

10 years agoMerge "Fix broken XML parsing idiom."
Narayan Kamath [Tue, 7 Jan 2014 10:02:57 +0000 (10:02 +0000)]
Merge "Fix broken XML parsing idiom."

10 years agoMerge "CompoundButton: Let text can be changed by config"
Conley Owens [Mon, 6 Jan 2014 23:36:02 +0000 (23:36 +0000)]
Merge "CompoundButton: Let text can be changed by config"

10 years agoMerge "Re-initialize error after attach"
Raph Levien [Mon, 6 Jan 2014 17:00:49 +0000 (17:00 +0000)]
Merge "Re-initialize error after attach"

10 years agoFix broken XML parsing idiom.
Narayan Kamath [Mon, 6 Jan 2014 11:31:35 +0000 (11:31 +0000)]
Fix broken XML parsing idiom.

Code that expected a single top level element in an XML file
was doing something like :

while (type != START_TAG) { next(); }

This would loop forever when the XML being parsed was empty,
where each call to XmlPullParser.next() would return END_DOCUMENT.

bug: https://code.google.com/p/android/issues/detail?id=64173
Change-Id: I7543203e976a8999ae471a6c2d629249a87011bb

10 years agoXmlUtils#readThisMapXml should handle null keys.
Narayan Kamath [Wed, 18 Dec 2013 15:19:17 +0000 (15:19 +0000)]
XmlUtils#readThisMapXml should handle null keys.

Both readThisValueXml & writeThisValueXml have been
implemented to handle values with "null" names.

Change readThisMapXml to allow such names as well, and
map them to the null key in the resulting HashMap.
readThisListXml and readThisSetXml already support
null names.

bug: https://code.google.com/p/android/issues/detail?id=63463

Change-Id: I1c93976656e45e7733113eb67d29b6bad5d25f1f

10 years agoMerge "Keydispatching timeout while finish Activity"
Craig Mautner [Thu, 19 Dec 2013 17:48:58 +0000 (17:48 +0000)]
Merge "Keydispatching timeout while finish Activity"

10 years agoMerge "MediaDrm#mNativeContext is a long, not int."
Narayan Kamath [Thu, 19 Dec 2013 12:35:22 +0000 (12:35 +0000)]
Merge "MediaDrm#mNativeContext is a long, not int."

10 years agoKeydispatching timeout while finish Activity
Martin Wallgren [Wed, 31 Aug 2011 10:39:31 +0000 (12:39 +0200)]
Keydispatching timeout while finish Activity

If there is input to be handled during finish activity we can get a
keydispatching timeout ANR. The reason is that finish activity is some
times not possible, and the activity is instead put on a finish queue.
The activity will then be finished sometime in the future. When we add
the activity to the finish queue, key dispatching is paused, and there
is an ANR timer waiting for it to be resumed again. Since it can take a
long time before the activity is actually finished, we need to resume
the key dispatching to avoid the ANR.

Change-Id: Icea4ab3b5ad05c8bfbadf8f5cece1a59ec621469

10 years agoCursor leakage due to cancellation signal
Peter Eliasson [Fri, 28 Jun 2013 11:55:41 +0000 (13:55 +0200)]
Cursor leakage due to cancellation signal

If a query call was cancelled, the cursor adaptor might leak. The
adaptor is now closed if any exception is thrown during query.

Change-Id: Ic4c2edeaf2fcef56b4ef59484a36d3233aa12dbc

10 years agoMerge "Track Looper decoupling from ALooper"
Brian Carlstrom [Thu, 19 Dec 2013 07:13:02 +0000 (07:13 +0000)]
Merge "Track Looper decoupling from ALooper"

10 years agoFix a JNI local reference leak in MediaDrm
Patrik2 Carlsson [Tue, 10 Dec 2013 13:52:43 +0000 (14:52 +0100)]
Fix a JNI local reference leak in MediaDrm

A reference leak was found in the subtitle usecase that also
applies to android_media_MediaDrm.cpp
(Code taken from the similar fix in android_media_MediaPlayer.cpp)

Change-Id: Icc4b25e4adc9a600ff4ac99a7a8478bb55a6e348

10 years agoTrack Looper decoupling from ALooper
Brian Carlstrom [Fri, 13 Dec 2013 07:12:55 +0000 (23:12 -0800)]
Track Looper decoupling from ALooper

Change-Id: I54f4d36f105e60eaaa453ae60f591d634c681fd7

10 years agoMerge "Handle user-defined cache path for RS."
Tim Murray [Thu, 19 Dec 2013 01:42:41 +0000 (01:42 +0000)]
Merge "Handle user-defined cache path for RS."

10 years agoHandle user-defined cache path for RS.
Tim Murray [Fri, 13 Dec 2013 20:57:36 +0000 (12:57 -0800)]
Handle user-defined cache path for RS.

Change-Id: I92e3346dca331ae6a9dfbe4f95292bc604dfd196

10 years agoAdd a call to registerAppInfo for the VMRuntime
Dave Allison [Thu, 14 Nov 2013 01:15:50 +0000 (17:15 -0800)]
Add a call to registerAppInfo for the VMRuntime

This calls into the VMRuntime tells it where the
application directory is located.

Bug: 11539952
Change-Id: I20e0b8c63e459699a1bc9af435e65ae96f1e1e73

10 years agoMerge "DO NOT MERGE - Fix improper declaration/use of pointer."
Stephen Hines [Thu, 19 Dec 2013 00:30:25 +0000 (00:30 +0000)]
Merge "DO NOT MERGE - Fix improper declaration/use of pointer."

10 years agoDO NOT MERGE - Fix improper declaration/use of pointer.
Stephen Hines [Thu, 19 Dec 2013 00:21:30 +0000 (16:21 -0800)]
DO NOT MERGE - Fix improper declaration/use of pointer.

https://code.google.com/p/android/issues/detail?id=61547

This variable is shadowing the outer "ptr" declaration, resulting in a
potential NULL pointer being passed to the message API.

Change-Id: If96bfae8d5e874e12597182678a180ba137b78da

10 years agoMerge "Remove unused field from MediaMuxer"
Narayan Kamath [Wed, 18 Dec 2013 18:08:58 +0000 (18:08 +0000)]
Merge "Remove unused field from MediaMuxer"

10 years agoRemove unused field from MediaMuxer
Narayan Kamath [Wed, 18 Dec 2013 17:51:06 +0000 (17:51 +0000)]
Remove unused field from MediaMuxer

Change-Id: If1ebb2302654b64924084a52425366f1b69c7c30

10 years agoMediaDrm#mNativeContext is a long, not int.
Narayan Kamath [Wed, 18 Dec 2013 16:53:54 +0000 (16:53 +0000)]
MediaDrm#mNativeContext is a long, not int.

Therefore, we must use SetLongField and not SetIntField.

Change-Id: I2a805dd6b7c8e4905e742d8632d941f381a05930

10 years agoMerge "Don't pseudolocalize untranslatable strings."
Narayan Kamath [Wed, 18 Dec 2013 12:46:40 +0000 (12:46 +0000)]
Merge "Don't pseudolocalize untranslatable strings."

10 years agoMerge "Add /dev/rtc backend to SystemClock"
Greg Hackmann [Tue, 17 Dec 2013 18:11:06 +0000 (18:11 +0000)]
Merge "Add /dev/rtc backend to SystemClock"

10 years agoMerge "Add timerfd backend to AlarmManagerService"
Greg Hackmann [Tue, 17 Dec 2013 18:10:49 +0000 (18:10 +0000)]
Merge "Add timerfd backend to AlarmManagerService"

10 years agoMerge "Clean up native AlarmManagerService tabs/spaces and unused parameter warnings"
Greg Hackmann [Tue, 17 Dec 2013 18:10:32 +0000 (18:10 +0000)]
Merge "Clean up native AlarmManagerService tabs/spaces and unused parameter warnings"

10 years agoMerge "Fix to NullPointerException on move back in ActivityStack."
Craig Mautner [Tue, 17 Dec 2013 16:36:44 +0000 (16:36 +0000)]
Merge "Fix to NullPointerException on move back in ActivityStack."

10 years agoFix to NullPointerException on move back in ActivityStack.
Daniel 2 Olofsson [Mon, 16 Dec 2013 12:24:25 +0000 (13:24 +0100)]
Fix to NullPointerException on move back in ActivityStack.

In ActivityTask.moveTaskToBackLocked NullPointerException may occur
when moving back with only current Activity in stack. This due to a
condition that may trigger despite a TaskRecord being null and then
attempt accessing the TaskRecord.mOnTopOfHome variable.

TaskRecord task may be set to null when no resumed activity remain.

Resolved by assuring that flag mOnTopOfHome is instead set to false
for current TaskRecord in case where there are no remaining activities
above home.

The above bug has already been corrected in the following commit,
ada62fca51d314cefe2c5da4e007df5b9abf320d, but it does not set the
cottect value to mTopOfHome for the current taks, see below.

Variable mOnTopOfHome will not be set to false in situations where
stack is of size 1 or less and task is null, perhaps from already
having finished current activity.

To avoid current TaskRecord maintaining value mOnTopOfHome to true
after launching Home this variable is set to false.

Impact should not be major due to correction earlier that makes sure
that there is always a TaskRecord.mOnTopOfHome set to true above Home
activity but if not correctly set for current task still gives a
possibility of bad behavior.

Change-Id: Ie86ad99c188aaa05b0de9d58eaa16c42b6fc4341

10 years agoAdd /dev/rtc backend to SystemClock
Greg Hackmann [Thu, 12 Dec 2013 02:00:34 +0000 (18:00 -0800)]
Add /dev/rtc backend to SystemClock

On devices without /dev/alarm, use the settimeofday() syscall and the
standard RTC_SET_TIME ioctl (which are collectively equivalent to the
ANDROID_ALARM_SET_RTC ioctl).

Change-Id: I3c1d741099e253186e43c9369b62603b214b9c9a
Signed-off-by: Greg Hackmann <ghackmann@google.com>
10 years agoAdd timerfd backend to AlarmManagerService
Greg Hackmann [Tue, 10 Dec 2013 00:56:53 +0000 (16:56 -0800)]
Add timerfd backend to AlarmManagerService

On devices without /dev/alarm, use a new backend based on timerfd.
timerfd has near-equivalent syscalls for the /dev/alarm ioctls we care
about, with two key differences:

1) /dev/alarm uses one fd for all clocks, while timerfd needs one fd per
clock type.

AlarmManagerService addresses this by replacing the fd (int) with an
opaque pointer (long) to the backend-specific state.

2) When the RTC changes, the /dev/alarm WAIT ioctl always returns, while
timerfd cancels (and signals events) only on specially-flagged RTC
timerfds.

The timerfd backend masks this by creating an extraneous RTC timerfd,
specifically so there's always something to signal on RTC changes.

Change-Id: I5aef867748298610347f6e1479dd8bf569495832
Signed-off-by: Greg Hackmann <ghackmann@google.com>
10 years agoMerge "AArch64: Use long for pointers in MediaDrm and MediaMuxer"
Elliott Hughes [Mon, 16 Dec 2013 20:58:01 +0000 (20:58 +0000)]
Merge "AArch64: Use long for pointers in MediaDrm and MediaMuxer"

10 years agoMerge "Fix ResTable_config::CONFIG_* typo (MCC vs MNC)"
Christopher Tate [Mon, 16 Dec 2013 20:44:33 +0000 (20:44 +0000)]
Merge "Fix ResTable_config::CONFIG_* typo (MCC vs MNC)"

10 years agoMerge "[bootanimation] Drop user loadable bootanimation.zip abiltiy."
Nick Kralevich [Mon, 16 Dec 2013 18:28:42 +0000 (18:28 +0000)]
Merge "[bootanimation] Drop user loadable bootanimation.zip abiltiy."

10 years ago[bootanimation] Drop user loadable bootanimation.zip abiltiy.
Robert Craig [Mon, 16 Dec 2013 14:42:51 +0000 (09:42 -0500)]
[bootanimation] Drop user loadable bootanimation.zip abiltiy.

A prior change removed the ability of
the shell to push to /data/local. Because
of this, the bootanimation code no longer
supports users loading custom bootanimation.zip
files. This patch drops the access attempt
from the code.

Change-Id: Ie30dd73699c8296be00dc6fbf14691e08baa6410
Signed-off-by: rpcraig <rpcraig@tycho.ncsc.mil>
10 years agoFix ResTable_config::CONFIG_* typo (MCC vs MNC)
Mårten Kongstad [Tue, 26 Nov 2013 13:48:22 +0000 (14:48 +0100)]
Fix ResTable_config::CONFIG_* typo (MCC vs MNC)

In commit 08d5b8f, the ResTable_config::CONFIG_* enum was incorrectly
changed to assign the same value to both CONFIG_MCC and CONFIG_MNC:
    - CONFIG_MCC = 0x0001,
    - CONFIG_MNC = 0x0002,
    + CONFIG_MCC = ACONFIGURATION_MCC,
    + CONFIG_MNC = ACONFIGURATION_MCC,

This commit assigns CONFIG_MNC the value ACONFIGURATION_MNC, as
intended.

Change-Id: I3b8799914a55debdb3ff30a82ada7ad990c49970

10 years agoMerge "Use StrictJarFile instead of JarFile for cert collection."
Narayan Kamath [Mon, 16 Dec 2013 11:29:45 +0000 (11:29 +0000)]
Merge "Use StrictJarFile instead of JarFile for cert collection."

10 years agoRe-initialize error after attach
Daniel 2 Olofsson [Tue, 13 Aug 2013 08:30:26 +0000 (10:30 +0200)]
Re-initialize error after attach

If flag for error after attach is not reinitialized, it may cause
NullPointerException on subsequent onAttachedWindow if error text
was reset to null.

Change-Id: I8976c28a6ca71017e5c4b3f29d4eeeb2eda48c38

10 years agoRevert "Do not show roaming icon for Videotron operator when access Canada network"
Sungmin Choi [Mon, 16 Dec 2013 07:14:44 +0000 (16:14 +0900)]
Revert "Do not show roaming icon for Videotron operator when access Canada network"

This reverts commit 119e24f98706cb0a1217b41e7e750b94c1a8941a.

Bug: 12163013
Change-Id: I574d1492253eb6e1f570411e441c5bea7376112c

10 years agotelephony: Add same named operator configuration for some operators
Sungmin Choi [Fri, 13 Dec 2013 05:50:40 +0000 (14:50 +0900)]
telephony: Add same named operator configuration for some operators

Add same named operator configuration for 404, 405, and 520.
That is, show roaming icon though same named operators.

Bug: 12121959
Change-Id: I9148bab727170f670fd2679db69b123b0526bb32

10 years agoMerge "Fix incorrect name in eglSwapBuffers log wrapper"
Jesse Hall [Sun, 15 Dec 2013 00:11:21 +0000 (00:11 +0000)]
Merge "Fix incorrect name in eglSwapBuffers log wrapper"

10 years agoMerge "Use java.util.Objects instead on internal API"
Kenny Root [Fri, 13 Dec 2013 22:12:52 +0000 (22:12 +0000)]
Merge "Use java.util.Objects instead on internal API"

10 years agoUse java.util.Objects instead on internal API
Kenny Root [Fri, 13 Dec 2013 20:00:26 +0000 (12:00 -0800)]
Use java.util.Objects instead on internal API

Not needed since java.util.Objects implements all the needed
functionality.

Change-Id: Icd31d49a9801d1705427f028e9ac927d58e7d34c

10 years agoUse StrictJarFile instead of JarFile for cert collection.
Narayan Kamath [Thu, 12 Dec 2013 16:54:33 +0000 (16:54 +0000)]
Use StrictJarFile instead of JarFile for cert collection.

This ensures that we use the same underlying zip
processing code as the runtimes.

bug: 10193060

(cherry picked from commit eb565dc527eda8c0a43df0d1f30132638ca4ba20)

Change-Id: Iaaa26b02678278394619d0a41613d9ceeae3203c

10 years agoMerge "Fix one and hide the other warnings from core/jni"
Kristian Monsen [Fri, 13 Dec 2013 06:32:03 +0000 (06:32 +0000)]
Merge "Fix one and hide the other warnings from core/jni"

10 years agoMerge "Add GCMParameterSpec to preloaded-classes"
Kenny Root [Fri, 13 Dec 2013 01:08:26 +0000 (01:08 +0000)]
Merge "Add GCMParameterSpec to preloaded-classes"

10 years agoAdd GCMParameterSpec to preloaded-classes
Kenny Root [Fri, 13 Dec 2013 00:37:45 +0000 (16:37 -0800)]
Add GCMParameterSpec to preloaded-classes

Needed for Bouncycastle 1.50 upgrade with art.

Change-Id: I901122c1ce21abba437a7537e70331724c3cbc9c

10 years agoClean up native AlarmManagerService tabs/spaces and unused parameter warnings
Greg Hackmann [Fri, 6 Dec 2013 18:52:09 +0000 (10:52 -0800)]
Clean up native AlarmManagerService tabs/spaces and unused parameter warnings

Change-Id: I64da1437dd0ed30957b43450c72b9d3797112a2e
Signed-off-by: Greg Hackmann <ghackmann@google.com>
10 years agoFix one and hide the other warnings from core/jni
Kristian Monsen [Thu, 12 Dec 2013 03:46:41 +0000 (19:46 -0800)]
Fix one and hide the other warnings from core/jni

Most of the warnings are unused parameter from skia functions.

Change-Id: I9df57718d117c26b96163f751ccb056ca86ac072

10 years agoMerge "Don't try to set service.bootanim.exit"
Nick Kralevich [Wed, 11 Dec 2013 21:46:08 +0000 (21:46 +0000)]
Merge "Don't try to set service.bootanim.exit"

10 years agoDon't try to set service.bootanim.exit
Nick Kralevich [Wed, 11 Dec 2013 21:10:55 +0000 (13:10 -0800)]
Don't try to set service.bootanim.exit

init has never allowed the bootanimation code to
set service.bootanim.exit, and has always generated the
dmesg error message

  <3>[   17.644615] init: sys_prop: permission denied uid:1003  name:service.bootanim.exit

Since setting this property has no effect and never worked, delete
the code which tries to do it.

Change-Id: Idacc5467d85479a8cf974702af8895011be585ea

10 years agoMerge "Parse UUID string with Long.parseLong"
Marco Nelissen [Wed, 11 Dec 2013 18:07:50 +0000 (18:07 +0000)]
Merge "Parse UUID string with Long.parseLong"

10 years agoMerge "Add libutils dependency for libandroidfw host build."
Narayan Kamath [Wed, 11 Dec 2013 17:36:20 +0000 (17:36 +0000)]
Merge "Add libutils dependency for libandroidfw host build."

10 years agoAdd libutils dependency for libandroidfw host build.
Narayan Kamath [Wed, 11 Dec 2013 16:48:07 +0000 (16:48 +0000)]
Add libutils dependency for libandroidfw host build.

Change-Id: I7fe0bb466f0b1cdbfef8934da2231ad20f05b648

10 years agoMerge "Reimplement ZipFileRO in terms of libziparchive."
Narayan Kamath [Wed, 11 Dec 2013 11:41:10 +0000 (11:41 +0000)]
Merge "Reimplement ZipFileRO in terms of libziparchive."

10 years agoMerge "fix possible buffer overrun and memory leak"
Elliott Hughes [Wed, 11 Dec 2013 01:29:27 +0000 (01:29 +0000)]
Merge "fix possible buffer overrun and memory leak"

10 years agoFix incorrect name in eglSwapBuffers log wrapper
Brian Harris [Tue, 10 Dec 2013 23:23:51 +0000 (15:23 -0800)]
Fix incorrect name in eglSwapBuffers log wrapper

Correctly identify eglSwapBuffers in EGLLogWrapper, which had been
labeled as "eglInitialize".

Change-Id: I67c61e100e2f84400ceebc28712f036e5a055017

10 years agoAArch64: Use long for pointers in MediaDrm and MediaMuxer
Ashok Bhat [Thu, 28 Nov 2013 10:56:06 +0000 (10:56 +0000)]
AArch64: Use long for pointers in MediaDrm and MediaMuxer

For storing pointers, long is used in MediaDrm and
MediaMuxer classes, as native pointers can be 64-bit.

Change-Id: I77f26cde627baf0dce70b6aa3a4dbd974051d9a6
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoDon't pseudolocalize untranslatable strings.
Narayan Kamath [Tue, 10 Dec 2013 09:49:59 +0000 (09:49 +0000)]
Don't pseudolocalize untranslatable strings.

bug: https://code.google.com/p/android/issues/detail?id=63004
Change-Id: I8e47a53d887144dc867111e42313d0e113b83ea1

10 years agoMerge "Fix inconsistency in LoggingPrintStream."
Narayan Kamath [Tue, 10 Dec 2013 09:30:56 +0000 (09:30 +0000)]
Merge "Fix inconsistency in LoggingPrintStream."

10 years agoMerge "Update Dalvik process state through VMRuntime."
Mathieu Chartier [Tue, 10 Dec 2013 02:13:51 +0000 (02:13 +0000)]
Merge "Update Dalvik process state through VMRuntime."

10 years agoMerge changes Ib4450c39,Ie7ed5655
Kenny Root [Mon, 9 Dec 2013 22:28:59 +0000 (22:28 +0000)]
Merge changes Ib4450c39,Ie7ed5655

* changes:
  preloaded-classes: remove some OpenSSLCipher modes
  Add CertStoreParameters to preloaded-classes

10 years agopreloaded-classes: remove some OpenSSLCipher modes
Kenny Root [Mon, 9 Dec 2013 22:01:54 +0000 (14:01 -0800)]
preloaded-classes: remove some OpenSSLCipher modes

Some cipher modes were removed from OpenSSLCipher. This change removes
those classes from the preloaded classes list.

Change-Id: Ib4450c392513973836684b4d5df7fffb200b6260

10 years agoAdd CertStoreParameters to preloaded-classes
Kenny Root [Mon, 9 Dec 2013 21:49:27 +0000 (13:49 -0800)]
Add CertStoreParameters to preloaded-classes

For image creating, art needs this in the preloaded-classes list.
Otherwise it complains loudly about not being able to load it from a
class initializer in Provider.java.

Change-Id: Ie7ed5655cba038b504767c3de9f70b47a8965cce

10 years agoReimplement ZipFileRO in terms of libziparchive.
Narayan Kamath [Tue, 3 Dec 2013 13:16:03 +0000 (13:16 +0000)]
Reimplement ZipFileRO in terms of libziparchive.

This lets us share zip archive processing code with both
the runtime (Art, dalvik) and critical java code
(StrictJarFile).

This change also moves several utility methods to ZipUtils
and dedups code across several zip inflation methods.

One of the side effects of this change is that several
processing loops are now O(n) instead of O(n^2).

bug: 10193060

Change-Id: I3c7188496837a47246c4f342e45485a70fef3169

10 years agoFix inconsistency in LoggingPrintStream.
Narayan Kamath [Mon, 9 Dec 2013 11:09:21 +0000 (11:09 +0000)]
Fix inconsistency in LoggingPrintStream.

LoggingPrintStream prints "null" for null strings from
print(String) but throws an NPE from println(String).

Fix this by delegating to the underlying builder (and avoiding
the "fast" path when the input String == null)

bug: https://code.google.com/p/android/issues/detail?id=63380
Change-Id: I89c5c12db6ed5af8037d46f67369b38c5b4cbc9f

10 years agoMerge "Add address flags and scope to LinkAddress."
Lorenzo Colitti [Sat, 7 Dec 2013 02:43:02 +0000 (02:43 +0000)]
Merge "Add address flags and scope to LinkAddress."

10 years agoParse UUID string with Long.parseLong
Tom Marshall [Fri, 6 Dec 2013 20:04:25 +0000 (21:04 +0100)]
Parse UUID string with Long.parseLong

Integer.parseInt throws an NPE for values greater than 0x7fffffff.

10 years agoam f7422885: Merge "Augment SELinuxMMAC functionality."
Nick Kralevich [Fri, 6 Dec 2013 16:17:23 +0000 (08:17 -0800)]
am f7422885: Merge "Augment SELinuxMMAC functionality."

* commit 'f7422885a99c5d240f70c2f8227ae44abeea3e5c':
  Augment SELinuxMMAC functionality.

10 years agoMerge "Augment SELinuxMMAC functionality."
Nick Kralevich [Fri, 6 Dec 2013 16:13:58 +0000 (16:13 +0000)]
Merge "Augment SELinuxMMAC functionality."

10 years agoAugment SELinuxMMAC functionality.
Robert Craig [Mon, 2 Dec 2013 15:24:23 +0000 (10:24 -0500)]
Augment SELinuxMMAC functionality.

* No longer support a package name stanza outside of
  a signature tag. Package names, by themselves, have
  no security associated with them in Android and thus we
  should not be allowing or encouraging this
  type of policy.

* Allow for nested package name stanzas inside
  signature stanzas. There are cases where a finer
  distinction needs to be made among apps signed with
  the same cert. New code allows a different seinfo
  tag to be assigned to the listed package names
  signed by the parent cert. When a determination needs
  to be made concerning seinfo assignments, the inner
  seinfo tag takes precedence over the outer seinfo
  labels which are assigned to just the signature.

* Temp structures are now used to parse new policy files
  until the entire xml file is parsed and deemed correct,
  at which time the temp structures are copied over to the
  permanent class structures. This ensures that any structural
  errors with the policy will not result in partial loads.

* Valid stanzas look like the following with the inner
  package piece being optional.

   <signer signature="">
     <seinfo value=""/>
     <package name="">
       <seinfo value=""/>
     </package>
   <signer>

   <default>
     <seinfo value=""/>
   </default>

Change-Id: Ia204d71211776dcf9b2dcc86ad6d77c4ad39dc25

10 years agoAdd address flags and scope to LinkAddress.
Lorenzo Colitti [Fri, 15 Nov 2013 09:43:52 +0000 (18:43 +0900)]
Add address flags and scope to LinkAddress.

This is necessary so that the framework can know whether an IPv6
address is likely to be usable (i.e., if it's global scope and
preferred). Also, it will simplify the address notification
methods in INetworkManagementEventObserver, which currently take
the address, the flags, and the scope as separate arguments.

1. Add flags and scope to the class and update the unit test.
   Use the IFA_F_* and RT_SCOPE_* constants defined by libcore.
   Since most callers don't know about flags and scope, provide
   constructors that default the flags to zero and determine the
   scope from the address. Addresses notified by the kernel will
   have these properly set. Make multicast addresses invalid.
   Update the class documentation.
2. Provide an isSameAddressAs() method that compares only the
   address and prefix information between two LinkAddress
   objects. This is necessary because an interface can't have
   two addresses with the same address/prefix but different
   flags.
3. Update LinkProperties's addLinkAddress and removeLinkAddress
   to identify existing addresses to add/remove using
   isSameAddressAs instead of implicit equals(). Specifically:
   - If addLinkAddress is called with an address that is already
     present, the existing address's flags and scope are updated.
     This allows, for example, an address on an interface to go
     from preferred to deprecated when it expires, without it
     having to be removed and re-added.
   - If removeLinkAddress is called with an address that is
     present but with different flags, it deletes that address
     instead of failing to find a match.
4. Update the INetworkManagementEventObserver address
   notification methods to take just a LinkAddress instead of
   LinkAddress, flags, and scope. While I'm at it, change the
   order of the arguments for consistency with the other
   functions in the interface.

Change-Id: Id8fe0f09a7e8f6bee1ea3b52102178b689a9336e

10 years agoMerge commit 'bac61807d3bcfff957b358cb9ad77850bd373689' into HEAD
The Android Open Source Project [Thu, 5 Dec 2013 21:10:46 +0000 (13:10 -0800)]
Merge commit 'bac61807d3bcfff957b358cb9ad77850bd373689' into HEAD

Change-Id: I29374270c8e0c2f2859efaf1d55af9f73da0f8d7

10 years agoMerge "wifi: Remove obsolete libwpa_client library"
Ying Wang [Thu, 5 Dec 2013 19:13:20 +0000 (19:13 +0000)]
Merge "wifi: Remove obsolete libwpa_client library"