OSDN Git Service

android-x86/frameworks-base.git
12 years agoMerge "Modify installd to set the SELinux security context on package directories."
Kenny Root [Fri, 8 Jun 2012 19:27:41 +0000 (12:27 -0700)]
Merge "Modify installd to set the SELinux security context on package directories."

12 years agoMerge "Add JNI bindings for some of the libselinux interfaces."
Kenny Root [Fri, 8 Jun 2012 19:25:07 +0000 (12:25 -0700)]
Merge "Add JNI bindings for some of the libselinux interfaces."

12 years agoMerge "Added mpga as valid MediaFile extension"
Jean-Baptiste Queru [Fri, 8 Jun 2012 17:20:57 +0000 (10:20 -0700)]
Merge "Added mpga as valid MediaFile extension"

12 years agoMerge "Aligning native Parcel implementation to Java."
Jean-Baptiste Queru [Fri, 8 Jun 2012 15:03:03 +0000 (08:03 -0700)]
Merge "Aligning native Parcel implementation to Java."

12 years agoMerge "Fix "Too many open files" error for aapt built with Mac OS X SDK 10.6."
Jean-Baptiste Queru [Wed, 6 Jun 2012 01:31:02 +0000 (18:31 -0700)]
Merge "Fix "Too many open files" error for aapt built with Mac OS X SDK 10.6."

12 years agoFix "Too many open files" error for aapt built with Mac OS X SDK 10.6.
Ying Wang [Thu, 31 May 2012 01:02:55 +0000 (18:02 -0700)]
Fix "Too many open files" error for aapt built with Mac OS X SDK 10.6.

With SDK 10.6 without _DARWIN_UNLIMITED_STREAMS aapt sometimes fails
right away with the error "Too many open files" when calling fopen().

Change-Id: Ifa7bd8a9530d706aa47f98be8186f1aefe943d76

12 years agoMerge "Tweak default ignore-assets path"
Tor Norbye [Tue, 5 Jun 2012 14:42:43 +0000 (07:42 -0700)]
Merge "Tweak default ignore-assets path"

12 years agoAligning native Parcel implementation to Java.
Magnus Strandberg [Tue, 3 May 2011 13:44:00 +0000 (15:44 +0200)]
Aligning native Parcel implementation to Java.

The Java implementation of writing the RPC response header
calculates the length of the header including the 4 bytes
specifying the header length but the native implementation
excludes the 4 bytes specifying the length from the header
length.
The native implementation has been aligned to the Java impl.

Change-Id: I325bf272a63152d8fded4cf4e51a906b5a9bfe19

12 years agoAdded mpga as valid MediaFile extension
Mathias Franzén [Tue, 28 Sep 2010 11:41:36 +0000 (13:41 +0200)]
Added mpga as valid MediaFile extension

Added mpga as a valid media file extension, it is treated as mp3.

Change-Id: Ie39036d61b6ca50439861fa10be0d824dcc65ed3

12 years agoTweak default ignore-assets path
Tor Norbye [Mon, 4 Jun 2012 17:38:13 +0000 (10:38 -0700)]
Tweak default ignore-assets path

Add .DS_Store to the list of ignored files that are silently
ignored (other dot-files are ignored but aapt emits a "(skipping <x>)"
message.)

Also, add a "!" prefix to the *~ pattern for Emacs/Vim/Gedit backup
files.

Finally, move the !*.scc pattern up in front of the .* pattern, such
that it doesn't match the earlier .* pattern (which is verbose, unlike
!*.scc).

Change-Id: Id3e96490f1802486aea8c58366d43e9d413971b8

12 years agoMerge "aapt: Remove terminal pointer incrementaion"
Jean-Baptiste Queru [Sat, 2 Jun 2012 16:21:36 +0000 (09:21 -0700)]
Merge "aapt: Remove terminal pointer incrementaion"

12 years agoMerge "Merge "Fix length of pattern." DO NOT MERGE."
Raphaël Moll [Wed, 30 May 2012 21:46:34 +0000 (14:46 -0700)]
Merge "Merge "Fix length of pattern." DO NOT MERGE."

12 years agoMerge "Input: Fix some Applications ANR issue in monkey test"
Jean-Baptiste Queru [Tue, 29 May 2012 20:05:35 +0000 (13:05 -0700)]
Merge "Input: Fix some Applications ANR issue in monkey test"

12 years agoMerge "Lid switch state is incorrect"
Jean-Baptiste Queru [Tue, 29 May 2012 20:04:57 +0000 (13:04 -0700)]
Merge "Lid switch state is incorrect"

12 years agoInput: Fix some Applications ANR issue in monkey test
Zhenghua Wang [Mon, 28 May 2012 13:49:43 +0000 (21:49 +0800)]
Input: Fix some Applications ANR issue in monkey test

Issue Description:
When do monkey test on some applications, there is big probability get ANR issue. one example is as following:
    adb shell monkey -p com.google.android.street -v 500000

Root Cause:
situation 1. InputDispatcher Thread calls findTouchedWindowTargetsLocked routine, there is a AMOTION_EVENT_ACTION_DOWN event
             and it can find a splittable touched window for this event , then mTouchState.split set to true.

situation 2. WMS Thread calls setInputWindows routine and all TouchedWindows associated with mTouchState are removed ,
             mTouchState.split status still keep true.

situation 3. InputDispatcher Thread calls findTouchedWindowTargetsLocked routine, there is a AMOTION_EVENT_ACTION_POINTER_DOWN event
             and it can find found touched window, exit window loop in following code :
                  if (windowInfo->visible) {
                    if (! (flags & InputWindowInfo::FLAG_NOT_TOUCHABLE)) {
                        isTouchModal = (flags & (InputWindowInfo::FLAG_NOT_FOCUSABLE
                                | InputWindowInfo::FLAG_NOT_TOUCH_MODAL)) == 0;
                        if (isTouchModal || windowInfo->touchableRegionContainsPoint(x, y)) {
                            if (! screenWasOff
                                    || (flags & InputWindowInfo::FLAG_TOUCHABLE_WHEN_WAKING)) {
                                newTouchedWindowHandle = windowHandle;
                            }
                            break; // found touched window, exit window loop
                        }
                    }

situation 4. The following code will have problem after situation 3 in below conditions:
             newTouchedWindowHandle doesn't support split , isSplit is true ( last time touched window is splittable )
             mTempTouchState.getFirstForegroundWindowHandle will always return NULL due to situation 2

            // Figure out whether splitting will be allowed for this window.
            if (newTouchedWindowHandle != NULL
                    && newTouchedWindowHandle->getInfo()->supportsSplitTouch()) {
                // New window supports splitting.
                isSplit = true;
            } else if (isSplit) {
                // New window does not support splitting but we have already split events.
                // Assign the pointer to the first foreground window we find.
                // (May be NULL which is why we put this code block before the next check.)
                newTouchedWindowHandle = mTempTouchState.getFirstForegroundWindowHandle();
            }

            The window says it does not want to support splitting, so the touches should go to one of the previously split windows except
            that there are none, which result to "goto Unresponsive" in findtouchedWindowTargetsLocked routine.
            This is not the correct behavior because the user did actually touch something.

situation 5. InputDispatcher Thread will call dispatchOnce repeated to try to dispatch this AMOTION_EVENT_ACTION_POINTER_DOWN event,
             then ANR timeout ( 5000ms ) will be exceeded after several times "goto Unresponsive"in findtouchedWindowTargetsLocked routine.

Solution:
In situation 4 we should drop the touch. If the newly touched window was splittable then we wouldn't drop the touch.
It's only when the newly touched window was not splittable where we will drop the touch.

Change-Id: Iab2c06ce0597ac77eb886ccd9d84646c86723bdb
Author: Jeffrey Brown <jeffbrown@android.com>
Author: Erjun Ding <erjunx.ding@intel.com>
Author: Zhenghua Wang <zhenghua.wang@intel.com>
Author: Jack Ren <jack.ren@intel.com>
Author: Bruce Beare <bruce.j.beare@intel.com>

12 years agoMerge "stagefright aacenc: Fix reading out of bounds in pow2_xy"
Jean-Baptiste Queru [Fri, 25 May 2012 17:38:05 +0000 (10:38 -0700)]
Merge "stagefright aacenc: Fix reading out of bounds in pow2_xy"

12 years agostagefright aacenc: Fix reading out of bounds in pow2_xy
Martin Storsjo [Fri, 25 May 2012 16:46:34 +0000 (19:46 +0300)]
stagefright aacenc: Fix reading out of bounds in pow2_xy

This fixes cases where x was a large number, causing fPart to
exceed the 32 bit signed integer range (while fitting in an
unsigned 32 bit integer), making the table index a negative
number.

Change-Id: I674047db65f89148a93d218c138b42cd8305f80e

12 years agoMerge "Wipe the user data out in any case."
Jean-Baptiste Queru [Fri, 25 May 2012 14:31:26 +0000 (07:31 -0700)]
Merge "Wipe the user data out in any case."

12 years agoMerge "Fix length of pattern." DO NOT MERGE.
Ying Wang [Tue, 22 May 2012 18:24:22 +0000 (11:24 -0700)]
Merge "Fix length of pattern." DO NOT MERGE.

After skipping * with "token++", the length should decrease by 1 as
well.

(merged from 996b073e813ba1a22a13282ccdebb664f14ba898)

Change-Id: Ie6232ef603bb31e25e03b926e6c1bb92ac34902d

12 years agoMerge "FileInputStream is not closed in "UserManager.java : readUserList()""
Amith Yamasani [Thu, 24 May 2012 16:42:29 +0000 (09:42 -0700)]
Merge "FileInputStream is not closed in "UserManager.java : readUserList()""

12 years agoMerge "libstagefright/matroska: Fix build with gcc 4.7"
Jean-Baptiste Queru [Thu, 24 May 2012 13:53:51 +0000 (06:53 -0700)]
Merge "libstagefright/matroska: Fix build with gcc 4.7"

12 years agoMerge "libmedia: Fix build with gcc 4.7"
Jean-Baptiste Queru [Thu, 24 May 2012 13:25:54 +0000 (06:25 -0700)]
Merge "libmedia: Fix build with gcc 4.7"

12 years agoMerge "Add a new API on DRM Framework for streaming"
Jean-Baptiste Queru [Wed, 23 May 2012 21:06:54 +0000 (14:06 -0700)]
Merge "Add a new API on DRM Framework for streaming"

12 years agoMerge "String on Toast notification does not aligned to center"
Jean-Baptiste Queru [Tue, 22 May 2012 20:52:59 +0000 (13:52 -0700)]
Merge "String on Toast notification does not aligned to center"

12 years agoString on Toast notification does not aligned to center
Taeho Kim [Tue, 22 May 2012 05:57:01 +0000 (14:57 +0900)]
String on Toast notification does not aligned to center

If length of the string being showed on Toast is short (less than 8 chars),
it is aligned to left rather than to center.
Added attribute android:layout_gravity="center_horizontal" on the TextView where Toast's message are shown to fix issue.

Change-Id: I77669fd8a11d1ec8f6b082348eb818671ce158bb
Signed-off-by: Taeho Kim <jyte82@gmail.com>
12 years agoMerge "stagefright avcenc: Use intptr_t for casting pointers to integers"
Jean-Baptiste Queru [Mon, 21 May 2012 21:12:51 +0000 (14:12 -0700)]
Merge "stagefright avcenc: Use intptr_t for casting pointers to integers"

12 years agoMerge "stagefright avcenc: Switch tmp variables to use the right type, to avoid casting"
Jean-Baptiste Queru [Mon, 21 May 2012 21:12:48 +0000 (14:12 -0700)]
Merge "stagefright avcenc: Switch tmp variables to use the right type, to avoid casting"

12 years agoMerge "Support a new ANDROID_AAPT_IGNORE env var."
Raphael Moll [Fri, 18 May 2012 20:00:31 +0000 (13:00 -0700)]
Merge "Support a new ANDROID_AAPT_IGNORE env var."

12 years agoMerge "Move kxml2 to prebuilts/misc"
Jean-Baptiste Queru [Fri, 18 May 2012 16:51:03 +0000 (09:51 -0700)]
Merge "Move kxml2 to prebuilts/misc"

12 years agoSupport a new ANDROID_AAPT_IGNORE env var.
Raphael Moll [Mon, 7 May 2012 23:16:46 +0000 (16:16 -0700)]
Support a new ANDROID_AAPT_IGNORE env var.

AAPT has a fixed built-in list of files and directories
to ignore when parsing resource files. Over the years we
always had developers requiring specific patterns.
If the env var ANDROID_AAPT_IGNORE is set, it is parsed
to find which file/directory patterns to ignore.
Otherwise a default is used that matches the current behavior.

Added a command-line option for it:
  aapt di --ignore-assets "foo*:*.blah"

SDK Bug: 5343 24067

Change-Id: Ia4caa2a8188c8c1df143f884e459b8182645995f

12 years agoMerge "Fixed x86 GCC 4.6 compilation error about unused variable"
Christopher Tate [Wed, 9 May 2012 21:12:58 +0000 (14:12 -0700)]
Merge "Fixed x86 GCC 4.6 compilation error about unused variable"

12 years agoFixed x86 GCC 4.6 compilation error about unused variable
Andrew Hsieh [Tue, 8 May 2012 15:13:17 +0000 (23:13 +0800)]
Fixed x86 GCC 4.6 compilation error about unused variable

Change-Id: I97a24cc25f71db4e61e9d93cd64a51bb16d2e09a

12 years agoMerge "Revert "Telephony: Register/Unregister Phones with CallManager""
Conley Owens [Sat, 5 May 2012 00:56:13 +0000 (17:56 -0700)]
Merge "Revert "Telephony: Register/Unregister Phones with CallManager""

12 years agoRevert "Telephony: Register/Unregister Phones with CallManager"
Conley Owens [Sat, 5 May 2012 00:41:32 +0000 (17:41 -0700)]
Revert "Telephony: Register/Unregister Phones with CallManager"

This reverts commit cbca38522ca90322cfc9ccbb7872586807bf6c48

12 years agoMerge "Telephony: Register/Unregister Phones with CallManager"
Conley Owens [Sat, 5 May 2012 00:40:55 +0000 (17:40 -0700)]
Merge "Telephony: Register/Unregister Phones with CallManager"

12 years agoMerge changes Ib28636e6,I93ebc433
Jean-Baptiste Queru [Fri, 4 May 2012 18:05:43 +0000 (11:05 -0700)]
Merge changes Ib28636e6,I93ebc433

* changes:
  WallpaperManagerService does not properly propagate setDimensionHints()
  ImageWallpaper : get bitmap width before calc screen offset

12 years agoMerge "Enable chromium http stack for x86"
Jean-Baptiste Queru [Thu, 3 May 2012 16:02:14 +0000 (09:02 -0700)]
Merge "Enable chromium http stack for x86"

12 years agoEnable chromium http stack for x86
Jun Tian [Wed, 2 May 2012 07:07:05 +0000 (15:07 +0800)]
Enable chromium http stack for x86

Use chromium http stack as the online media http stack for x86 platform.

Change-Id: I16a7a8697127a51fab8105833f7463911852f0b3

12 years agoMerge "stagefright: Remove useless declarations"
Jean-Baptiste Queru [Tue, 1 May 2012 21:48:06 +0000 (14:48 -0700)]
Merge "stagefright: Remove useless declarations"

12 years agoLid switch state is incorrect
Masanori Ogino [Wed, 7 Mar 2012 07:25:16 +0000 (16:25 +0900)]
Lid switch state is incorrect

InputManager#getSwitchState() returns AKEY_STATE_UP(0=OPEN)
and AKEY_STATE_DOWN(1=CLOSE).
mLidOpen is opposite from a real lid status.

Change-Id: I40aa4e0defb95d82b6144ff6b7514f721bf9030f

12 years agoMerge "Improve robustness of resource overlay test suite."
Jean-Baptiste Queru [Tue, 1 May 2012 16:20:47 +0000 (09:20 -0700)]
Merge "Improve robustness of resource overlay test suite."

12 years agoMerge "Runtime resource overlay: clean-up."
Jean-Baptiste Queru [Tue, 1 May 2012 16:20:32 +0000 (09:20 -0700)]
Merge "Runtime resource overlay: clean-up."

12 years agoMerge "Build overlay packages just like regular packages."
Jean-Baptiste Queru [Tue, 1 May 2012 15:49:37 +0000 (08:49 -0700)]
Merge "Build overlay packages just like regular packages."

12 years agoMerge "Fixed group and child view caching in SimpleExpandableListAdapter."
Jean-Baptiste Queru [Tue, 1 May 2012 13:43:03 +0000 (06:43 -0700)]
Merge "Fixed group and child view caching in SimpleExpandableListAdapter."

12 years agoMerge "Fix off-by-one error when filtering application UIDs"
Jean-Baptiste Queru [Tue, 1 May 2012 13:42:35 +0000 (06:42 -0700)]
Merge "Fix off-by-one error when filtering application UIDs"

12 years agoMerge "Fixes an issue that occured unexpected exception "pointerIndex out of range"."
Jean-Baptiste Queru [Tue, 1 May 2012 13:42:11 +0000 (06:42 -0700)]
Merge "Fixes an issue that occured unexpected exception "pointerIndex out of range"."

12 years agoMerge "Avoiding horizontal keypad navigation trapping within gallery."
Jean-Baptiste Queru [Tue, 1 May 2012 12:52:22 +0000 (05:52 -0700)]
Merge "Avoiding horizontal keypad navigation trapping within gallery."

12 years agoMerge "Prevent empty overflow popup when there are no overflow items."
Jean-Baptiste Queru [Tue, 1 May 2012 12:52:12 +0000 (05:52 -0700)]
Merge "Prevent empty overflow popup when there are no overflow items."

12 years agoMerge "Possible NullPointerException in restorePanelState"
Jean-Baptiste Queru [Tue, 1 May 2012 12:51:59 +0000 (05:51 -0700)]
Merge "Possible NullPointerException in restorePanelState"

12 years agoMerge "statistics from some cpus don't fit"
Jean-Baptiste Queru [Tue, 1 May 2012 12:51:48 +0000 (05:51 -0700)]
Merge "statistics from some cpus don't fit"

12 years agoMove kxml2 to prebuilts/misc
Jean-Baptiste Queru [Tue, 1 May 2012 00:35:26 +0000 (17:35 -0700)]
Move kxml2 to prebuilts/misc

Change-Id: Ie4b45d2073f26d8313843f9860cf658e288c850b

12 years agoMerge "Modify getState to prefer AKEY_STATE_DOWN and AKEY_STATE_UP over AKEY_STATE_UN...
Jeff Brown [Tue, 1 May 2012 00:18:43 +0000 (17:18 -0700)]
Merge "Modify getState to prefer AKEY_STATE_DOWN and AKEY_STATE_UP over AKEY_STATE_UNKNOWN."

12 years agoMerge "Do not dispatch context selection events to non-visible fragments."
Dianne Hackborn [Wed, 25 Apr 2012 22:41:42 +0000 (15:41 -0700)]
Merge "Do not dispatch context selection events to non-visible fragments."

12 years agoFileInputStream is not closed in "UserManager.java : readUserList()"
Seohyeon.Song [Wed, 25 Apr 2012 09:20:05 +0000 (18:20 +0900)]
FileInputStream is not closed in "UserManager.java : readUserList()"

Change-Id: I50f54c70c8b6bc808fa59ce589b0480780cc811b

12 years agoMerge "Distance max set incorrectly"
Dianne Hackborn [Tue, 24 Apr 2012 21:38:31 +0000 (14:38 -0700)]
Merge "Distance max set incorrectly"

12 years agoDo not dispatch context selection events to non-visible fragments.
Jake Wharton [Tue, 24 Apr 2012 05:02:32 +0000 (22:02 -0700)]
Do not dispatch context selection events to non-visible fragments.

When used in a `ViewPager`, fragments that are present on the adjacent,
cached pages will receive context selection dispatches which, depending
on your fragment contents, can be difficult to determine whether or not
the event truly originated from your view.

By using the visible hint we restrict dispatching to only those fragments
which are marked as being visible. Since the fragment pager adapter
updates this setting properly most implementations will be afforded this
fix without any change required. If the user is implementing their own
adapter they likely already understand the implications of these cached
fragments and the reponsibility for updating the boolean falls to them.

Mirrors support library change Ie6a72c1c82c2784774373670007b6f5948fe16da

Change-Id: I1bc6f74e26660ef4a3c20419c077e37992412e94

12 years agoMerge "Fix potential NPE when saving fragment state."
Dianne Hackborn [Tue, 24 Apr 2012 02:00:47 +0000 (19:00 -0700)]
Merge "Fix potential NPE when saving fragment state."

12 years agoFix potential NPE when saving fragment state.
Jake Wharton [Sun, 22 Apr 2012 21:17:01 +0000 (17:17 -0400)]
Fix potential NPE when saving fragment state.

If a fragment's saved view state is null and the user
visible hint is true then the `result` bundle will have
never been initialized to a value resulting in a
`NullPointerException`.

Mirrors support library change I8ba585bc6b9298841490d64bc22a8219cd261adb.

Change-Id: Iabd5ac293d2ece3771da9ef257479eca0dcd523c

12 years agoDistance max set incorrectly
Andreas Sandblad [Wed, 21 Mar 2012 13:39:57 +0000 (14:39 +0100)]
Distance max set incorrectly

The distance max used in the MotionRange is calculated
incorrectly.

Looks like a copy-and-paste error.

Change-Id: I2b6daab088df0fb69e05682b67ca33524ff35987

12 years agoMerge "Issue: Foreground activity performs [Resume] and [Pause] when any process...
Dianne Hackborn [Sat, 21 Apr 2012 01:36:52 +0000 (18:36 -0700)]
Merge "Issue: Foreground activity performs [Resume] and [Pause] when any process died in sleep mode."

12 years agoMerge "Fix for too many binder calls in packagemanager"
Dianne Hackborn [Sat, 21 Apr 2012 01:34:52 +0000 (18:34 -0700)]
Merge "Fix for too many binder calls in packagemanager"

12 years agoFix for too many binder calls in packagemanager
Henrik Engström [Tue, 21 Feb 2012 08:05:17 +0000 (09:05 +0100)]
Fix for too many binder calls in packagemanager

The packagemanager uses a ParceledListSlice to send back its lists
of installed packages and apps. The list slice has a method append
which, in addition to adding the item to the list, also returns true
if the list has passed a size limit (about 1/4 of the total possible
IPC parcel size) to let the caller know that he should send the
slice. However, when used by the pm, it has an extra ! that makes it
send whenever it ISN'T over this limit instead (and conversely, not
send if it is under). This causes a lot more calls than needed since
it sends tiny one item slices instead of larger ones. This patch
removes the extra ! making it behave correctly.

Change-Id: I8db46d380a25406b55f3214aee1505e81949acc5

12 years agoIssue: Foreground activity performs [Resume] and [Pause] when any process died in...
p13451 [Wed, 18 Apr 2012 02:39:23 +0000 (11:39 +0900)]
Issue: Foreground activity performs [Resume] and [Pause] when any process died in sleep mode.

Step to Reproduce
1) Turn off device’s screen. (Sleep mode)
2) Kill any process.
A. Engineer Version: kill [PID]
B. User Version: am force-stop [Package Name]
3) Foreground activity proceed [Resume] and [Pause] consecutively.

Reason: Since ICS version, activity goes to stopped status when screen turns off.
stopIfSleepingLocked( ) makes activity to stopped status but, pauseIfSleepingLocked( ) was used in GB
and, activity keep paused status and, this problem did not occur.
This change give effect to resuming activity when any process was killed.
Because, resume is proceed without exception for activity status.
The exception only filtered for [ActivityState.PAUSED] in sleep or shutdown mode.
and, resume complete flow when activity status was [ActivityState.STOPPED].

Solution for this issue:
We think that exception’s condition have to change if stopped activity status is intended in sleep mode.
According to activity life cycle, activity can not resume from stop status.

Also check [ActivityState.STOPPING]. :)

Change-Id: Icca3366ac30ffa3b18f6e2393e4d7309089ef26a

12 years agoMerge "Fix build - re-generate current.txt"
Jean-Baptiste Queru [Thu, 19 Apr 2012 21:31:54 +0000 (14:31 -0700)]
Merge "Fix build - re-generate current.txt"

12 years agoFix build - re-generate current.txt
Jean-Baptiste Queru [Thu, 19 Apr 2012 19:29:56 +0000 (12:29 -0700)]
Fix build - re-generate current.txt

Change-Id: Id6a3100e4aaed473172237ab368e5aeeb7ccf577

12 years agoMerge "Remove deprecation on android:enabled attribute."
Dianne Hackborn [Thu, 19 Apr 2012 01:26:31 +0000 (18:26 -0700)]
Merge "Remove deprecation on android:enabled attribute."

12 years agoMerge "use utf8_length() instead of local function, isValidUtf8()"
Dianne Hackborn [Thu, 19 Apr 2012 01:26:01 +0000 (18:26 -0700)]
Merge "use utf8_length() instead of local function, isValidUtf8()"

12 years agoMerge "Volume Overlay Bug When swapping streams"
Dianne Hackborn [Thu, 19 Apr 2012 00:57:45 +0000 (17:57 -0700)]
Merge "Volume Overlay Bug When swapping streams"

12 years agoMerge "Transparent activity orientation problem when previous landsacpe fullscreen...
Dianne Hackborn [Thu, 19 Apr 2012 00:57:21 +0000 (17:57 -0700)]
Merge "Transparent activity orientation problem when previous landsacpe fullscreen activity not yet destroyed."

12 years agoVolume Overlay Bug When swapping streams
Steven Harper [Wed, 18 Apr 2012 09:15:34 +0000 (10:15 +0100)]
Volume Overlay Bug When swapping streams

Merged in http://review.cyanogenmod.com/#change,14109

Fixing a bug that when an app is dismissed or loaded that changes
the current active audio stream, if the volume overlay is visible during the swap,
if you adjust the volume (with hard keys) after the stream has changed while
the panel is still visible, the wrong panel (view) is still visible on-top of the
one that is being adjusted.

A good way to replicate this is to
   Open the Phone APP
   Go to the Keypad Screen
   Adjust the volume (not the icon on the overlay)
   While the volume is visible Hit the HOME hard key
   Before the volume dissapears, adjust the volume with hard keys

You will hear and feel the volume adjusting, but the overlay will not update.

Change-Id: Ied50ed83b153234cff82c282e3fd76ed671b420b

12 years agoTransparent activity orientation problem when previous landsacpe fullscreen activity...
p13451 [Wed, 18 Apr 2012 06:13:26 +0000 (15:13 +0900)]
Transparent activity orientation problem when previous landsacpe fullscreen activity not yet destroyed.

After terminating landsacpe fullscreen activity,
when user launch transparent activity via portrait home app, transparent activity is shown as landscape mode.

At this time AppWindowToken of previous acitivity has not been deleted, because Activity.onDestory() has not been returned yet.
In this case, getOrientationFromAppTokensLocked() returned ActivityInfo.SCREEN_ORIENTATION_LANDSCAPE.

Ignore hidden application is terminated on the top.
See also http://code.google.com/p/android/issues/detail?id=28927

Change-Id: I51239431120ec6ba8f8ff76871efb2347b9810ca

12 years agoAdd header for forward compatibility
Chris Craik [Tue, 17 Apr 2012 17:45:52 +0000 (10:45 -0700)]
Add header for forward compatibility

Change-Id: I6e3da852d50fc14cfb6fb50f7dd867f37daea19d

12 years agostagefright: Remove useless declarations
Martin Storsjo [Mon, 16 Apr 2012 10:44:19 +0000 (13:44 +0300)]
stagefright: Remove useless declarations

The encoder common library doesn't contain any source
code with assembly optimizations (there is nothing
using the defines in the single source file in the library),
and the source directories added to the include path do
not exist.

Change-Id: If8854ea3bac4ad3504615ae4ad022267103a8b65

12 years agoRemove deprecation on android:enabled attribute.
Yuku on yuku4 [Mon, 16 Apr 2012 08:15:06 +0000 (16:15 +0800)]
Remove deprecation on android:enabled attribute.

The android:enabled was never deprecated and replaced by
android:state_enabled as it was mentioned in the documentation.

The attribute android:enabled when applied to a TextView (or its
descendants) still works until now. In fact that's the way to
disable widgets from layout XMLs.

The deprecation made ADT and lint suggest users not to use this
attribute.

Change-Id: I5e7d7060d9e79a04342ebc723d7937b9bc12a018
Signed-off-by: Yuku on yuku4 <yukuku@gmail.com>
12 years agostagefright avcenc: Switch tmp variables to use the right type, to avoid casting
Martin Storsjo [Fri, 13 Apr 2012 11:16:26 +0000 (14:16 +0300)]
stagefright avcenc: Switch tmp variables to use the right type, to avoid casting

This fixes building on platforms where int is smaller than
pointers, e.g. 64 bit platforms.

Change-Id: I3da0459bac10cbffd1db699bf4782fe926cd391f

12 years agostagefright avcenc: Use intptr_t for casting pointers to integers
Martin Storsjo [Fri, 13 Apr 2012 11:15:58 +0000 (14:15 +0300)]
stagefright avcenc: Use intptr_t for casting pointers to integers

This is used for doing integer arithmetic on the pointers, for
checking pointer alignment.

Change-Id: I7a7ff834dc631ac9cce90e46216f64645e1ae6e5

12 years agouse utf8_length() instead of local function, isValidUtf8()
Homin Lee [Thu, 5 Apr 2012 11:11:27 +0000 (20:11 +0900)]
use utf8_length() instead of local function, isValidUtf8()

utf8_length() from libutils returns -1 when source not contains
valid sequence for UTF-8. Fixed to use it and removed the local
function isValidUtf8().

Change-Id: If2834ce1d1ae07fd8526ce8bc5df3fd3f44e85c8
Signed-off-by: Homin Lee <suapapa@insignal.co.kr>
12 years agoUpdate header for forward compatibility
Teng-Hui Zhu [Thu, 12 Apr 2012 16:57:04 +0000 (09:57 -0700)]
Update header for forward compatibility

Change-Id: Idf38142dc073dce14741e8da005e7d2eaf9242de

12 years agoUpdate SurfaceTexture interface for forward compatibility
Teng-Hui Zhu [Mon, 9 Apr 2012 21:51:48 +0000 (14:51 -0700)]
Update SurfaceTexture interface for forward compatibility

Change-Id: I32d224428c0127800714a976b4d7bb68cc7c2464

12 years agoFix hybrid build
Chris Craik [Tue, 3 Apr 2012 20:53:46 +0000 (13:53 -0700)]
Fix hybrid build

Change-Id: Ieef556473e1878d1319f9844fb0bbb974637999e

12 years agoPossible NullPointerException in restorePanelState
Dmitry Tsyganyuk [Mon, 2 Apr 2012 18:10:52 +0000 (13:10 -0500)]
Possible NullPointerException in restorePanelState

curFeatureId is just an index in a cycle and can not be used for getting an object. May get null that will cause NullPointerException. Solution is to use key to get an object from icicles.

Change-Id: I61d5c6bd1c3226a993f60efa74872995ba0e42ac

12 years agoAdd JNI bindings for some of the libselinux interfaces.
Stephen Smalley [Fri, 13 Jan 2012 13:31:39 +0000 (08:31 -0500)]
Add JNI bindings for some of the libselinux interfaces.

Change-Id: Ifcc68cb06f9f56a04f3bc64dd9906a9436fabc88

12 years agoaapt: Remove terminal pointer incrementaion
Nobuhiro Hayashi [Mon, 30 Jan 2012 09:42:20 +0000 (18:42 +0900)]
aapt: Remove terminal pointer incrementaion

When initialization, 'end' pointer refers terminal
character in a string. But it is incremented in
while loop. It occurs the access to non-secure memory.

Issue detail
  http://code.google.com/p/android/issues/detail?id=24896

Change-Id: I751d1154d797738c87a66374a66933528c045be5

12 years agostatistics from some cpus don't fit
Martin Wallgren [Tue, 20 Sep 2011 11:45:48 +0000 (13:45 +0200)]
statistics from some cpus don't fit

Cpu statistics from time_in_state does not fit
in the data structures in ProcessState.java. The
time_in_state file eventually becomes larger than
the buffer of 256 bytes witch cause missing data
in the result or a NoMoreElementsException when
parsing the data.
The numger of cpu speeds can be larger than the
MAX_SPEEDS limit.

Solution
Make sure that the whole file gets read and increase
the MAX_SPEEDS that the system can handle.

Change-Id: I6dfc127b2934988006649cce321d58750a946017

12 years agolibstagefright/matroska: Fix build with gcc 4.7
Bernhard Rosenkraenzer [Thu, 29 Mar 2012 09:41:17 +0000 (11:41 +0200)]
libstagefright/matroska: Fix build with gcc 4.7

gcc 4.7 doesn't like the redefinition of i in the
same scope

Change-Id: Idf1b33af7041f0302f2f2562442a3e11ce9607d9
Signed-off-by: Bernhard Rosenkraenzer <Bernhard.Rosenkranzer@linaro.org>
12 years agolibmedia: Fix build with gcc 4.7
Bernhard Rosenkraenzer [Thu, 29 Mar 2012 09:38:59 +0000 (11:38 +0200)]
libmedia: Fix build with gcc 4.7

gcc 4.7 doesn't like the redefinition of n in the same
scope:
for(size_t n=...;;) {
  size_t n = 2; // <-- duplicate definition
}

Change-Id: Idfc2e768a0eb85a3dacccd3431b4ad44775432a4
Signed-off-by: Bernhard Rosenkraenzer <Bernhard.Rosenkranzer@linaro.org>
12 years agoPrevent empty overflow popup when there are no overflow items.
Jake Wharton [Thu, 29 Mar 2012 08:42:37 +0000 (01:42 -0700)]
Prevent empty overflow popup when there are no overflow items.

On a device with a hardware menu key, an empty overflow window can be displayed
if there are no items in said overflow menu. This leads to the next user touch
being swallowed in the action of dismissing the invisible window.

Change-Id: I9191e4b2e1ed4888e4bc7261e8b3f095a272a9b5

12 years agoFix off-by-one error when filtering application UIDs
Magnus Eriksson [Tue, 27 Mar 2012 13:40:18 +0000 (15:40 +0200)]
Fix off-by-one error when filtering application UIDs

A filtering check in writeLPr() mistakenly includes the first application UID
(10000) with the result that the package with UID 10000 is missing from
packages.list. This patch fix the error.

Change-Id: I3651beb346290db8e09317391b95a77aed1946b6
Signed-off-by: Magnus Eriksson <eriksson.mag@gmail.com>
12 years agoam 093ed383: Merge "DOC CHANGE: IME subtype article" into ics-mr1
Joe Malin [Sun, 25 Mar 2012 20:04:59 +0000 (13:04 -0700)]
am 093ed383: Merge "DOC CHANGE: IME subtype article" into ics-mr1

* commit '093ed3838c3f2847be1893ea49c6b15b90e0afc3':
  DOC CHANGE: IME subtype article

12 years agoMerge "DOC CHANGE: IME subtype article" into ics-mr1
Joe Malin [Sun, 25 Mar 2012 20:02:53 +0000 (13:02 -0700)]
Merge "DOC CHANGE: IME subtype article" into ics-mr1

12 years agoam 20f394df: Docs: Added efficient downloads Android U class
Reto Meier [Fri, 23 Mar 2012 20:30:04 +0000 (13:30 -0700)]
am 20f394df: Docs: Added efficient downloads Android U class

* commit '20f394df4db02c492d7920f4f366b777451c2067':
  Docs: Added efficient downloads Android U class

12 years agoDocs: Added efficient downloads Android U class
Reto Meier [Wed, 14 Mar 2012 16:27:11 +0000 (09:27 -0700)]
Docs: Added efficient downloads Android U class

Change-Id: I2a9484c5a224a7e0688fbd9afb16246db8f3ddf9

12 years agoDOC CHANGE: IME subtype article
Joe Malin [Sat, 10 Mar 2012 00:04:35 +0000 (16:04 -0800)]
DOC CHANGE: IME subtype article

Change-Id: Ie7addcc6155c275b247e15787fe226b43b0dee1a

12 years agoModify getState to prefer AKEY_STATE_DOWN and AKEY_STATE_UP over AKEY_STATE_UNKNOWN.
David Deephanphongs [Mon, 14 Nov 2011 22:50:45 +0000 (14:50 -0800)]
Modify getState to prefer AKEY_STATE_DOWN and AKEY_STATE_UP over AKEY_STATE_UNKNOWN.

For the getState() routines that operate on multiple devices / mappers:
Return AKEY_STATE_DOWN (or AKEY_STATE_VIRTUAL) if any of the devices/mappers
have the key down.
Otherwise, return AKEY_STATE_UP if at least one device/mapper returns AKEY_STATE_UP.
Otherwise, return AKEY_STATE_UNKNOWN.

Change-Id: I45a1f83d751c3c0cbfcc744dd3d6f0137ec8e01a

12 years agoam 3347d659: am 1f9e7499: Added rules to build f/b/libs/utils in 64-bit: lib64utils.a
Andrew Hsieh [Thu, 22 Mar 2012 17:18:59 +0000 (10:18 -0700)]
am 3347d659: am 1f9e7499: Added rules to build f/b/libs/utils in 64-bit: lib64utils.a

* commit '3347d6597bd6214f037825e41be7e58198d9079d':
  Added rules to build f/b/libs/utils in 64-bit: lib64utils.a

12 years agoam 1f9e7499: Added rules to build f/b/libs/utils in 64-bit: lib64utils.a
Andrew Hsieh [Thu, 22 Mar 2012 17:16:42 +0000 (10:16 -0700)]
am 1f9e7499: Added rules to build f/b/libs/utils in 64-bit: lib64utils.a

* commit '1f9e749914c7c636bd759da0209e68d5801407ac':
  Added rules to build f/b/libs/utils in 64-bit: lib64utils.a

12 years agoAdded rules to build f/b/libs/utils in 64-bit: lib64utils.a
Andrew Hsieh [Tue, 13 Mar 2012 07:55:34 +0000 (00:55 -0700)]
Added rules to build f/b/libs/utils in 64-bit: lib64utils.a

It's needed to build four shared libraries in 64-bit for 64-bit
emulator with "-gpu on"
  lib64OpenglRender.so
  lib64EGL_translator.so
  lib64GLES_CM_translator.so
  lib64GLES_V2_translator.so

Change-Id: Ia6c05b23df1e9cd9e7f2e94e4cd5bde4be5d336b

12 years agoam 66f1bafb: Merge "docs: Android SDK r17 (RC7) Release Notes (fix for build feature...
Joe Fernandez [Wed, 21 Mar 2012 23:43:33 +0000 (16:43 -0700)]
am 66f1bafb: Merge "docs: Android SDK r17 (RC7) Release Notes (fix for build feature info)" into ics-mr1

* commit '66f1bafb2326e12814df703a4285373e8651a3ce':
  docs: Android SDK r17 (RC7) Release Notes (fix for build feature info)

12 years agoMerge "docs: Android SDK r17 (RC7) Release Notes (fix for build feature info)" into...
Joe Fernandez [Wed, 21 Mar 2012 23:41:19 +0000 (16:41 -0700)]
Merge "docs: Android SDK r17 (RC7) Release Notes (fix for build feature info)" into ics-mr1

12 years agodocs: Android SDK r17 (RC7) Release Notes (fix for build feature info)
Joe Fernandez [Wed, 21 Mar 2012 23:25:21 +0000 (16:25 -0700)]
docs: Android SDK r17 (RC7) Release Notes (fix for build feature info)

Change-Id: Id10c2d5eec9f81d1c6ce52067b7e25b7f411d91c