79852 Commits

Author SHA1 Message Date
Johan Redestig
3cd42dfd50 Make addAddress locale safe
Using regular string concatenation to avoid unexpected
results in some locales.

Change-Id: I47dd5e174c4a2e88dc18e014002820cdbf63fcad
2012-08-29 08:07:02 +02:00
Daniel Sandler
ad955fe87f Merge "SystemUI: Fix data activity overlay not being removed when data is disabled" 2012-08-28 12:52:49 -07:00
Jean-Baptiste Queru
416362e6d8 Merge "Change StringTokenizer to TextUtils.StringSplitter" 2012-08-28 12:22:39 -07:00
Jean-Baptiste Queru
78ad66b43c Merge "[Audio]Press camera button during Http live Audio playing, then return Browser, DUT display abnormal." 2012-08-28 10:43:10 -07:00
Irfan Sheriff
67ea8c8641 Merge "Added a persistent feature in WiFi Direct." 2012-08-27 17:54:07 -07:00
Dianne Hackborn
32d1fec50b Merge "Fixes relative time formatting issue for dates in future" 2012-08-27 11:39:10 -07:00
Jean-Baptiste Queru
03d16e122a Merge "Make setCaptureRate locale safe" 2012-08-27 10:22:29 -07:00
Jean-Baptiste Queru
eb00f95143 Merge "VibratorService: Fix to ensure actual delay in a vibrate pattern" 2012-08-27 07:55:41 -07:00
Yoshihiko Ikenaga
0879d03f0a Added a persistent feature in WiFi Direct.
Provide a new feature to use persistent group at default.
In the normal connect sequence, if the persistent profile has been
stored, try to use it. Otherwise, a new persistent group is created.
If the persistent profiles are stored over 32, an old profile is
deleted automatically.

Change-Id: Iccb9fa044f1907f0818cd259275e1675f4c3e222
Signed-off-by: Yoshihiko Ikenaga <yoshihiko.ikenaga@jp.sony.com>
2012-08-27 21:54:44 +09:00
Johan Redestig
1c7928e8c6 Make setCaptureRate locale safe
Dont use the default String.format in setCaptureRate as it is both
locale unsafe and less efficient than regular string concatenation
in this case.

Change-Id: I29418ec0352ef8c79967592b6eb8ff66cc9c99f0
2012-08-27 14:24:10 +02:00
Steve Pomeroy
ca3363775e Fixes relative time formatting issue for dates in future
All variations of getRelativeTimeSpanString() now properly handle dates
that are in the future. Prior, the version used by
getRelativeDateTimeString() would occasionally show the time instead of
a date when the future date was the same weekday as the current weekday.
This resulted in the time output being duplicated, eg.: "11:23, 11:23"

Change-Id: If20972a6942cce792fa233437f94dedfb71379f3
Signed-off-by: Steve Pomeroy <steve@staticfree.info>
2012-08-27 02:20:13 -04:00
Wink Saville
13ab2c9fda Merge "Telephony: Add ICC_UNKNOWN intent" 2012-08-23 14:25:23 -07:00
Dianne Hackborn
8bfba7df37 Merge "fix wrong substring" 2012-08-23 14:10:31 -07:00
Alex Yakavenka
ceb9bbc5df Telephony: Add ICC_UNKNOWN intent
Change-Id: I1a666973dd23f510a0cce60ef7db462728717fc0
2012-08-23 14:09:51 -07:00
Jean-Baptiste Queru
322b0a160d Merge "Protecting more views from (bad) MotionEvents" 2012-08-23 14:08:40 -07:00
Jean-Baptiste Queru
8f7100a037 Merge "WifiStateMachine: ignore auth-fail event during WPS connection" 2012-08-23 09:31:03 -07:00
Jean-Baptiste Queru
a4e34ae82f Merge "Properly handle quick switch between Home button and Gallery app" 2012-08-23 08:04:46 -07:00
Johan Rosengren
0dc291eef6 Protecting more views from (bad) MotionEvents
When handling MotionEvents, the method findPointerIndex can return -1
if the current pointer id can't be translated to a pointer index.
Some views are not handling this, which will lead to an out-of-
index crash. In 2585e9b there were checks added, this change adds
some more.

Change-Id: I93ce2420afd83a06b689a1ed35ead7d170cd68f1
2012-08-23 13:46:14 +02:00
Masanori Ogino
08bab56f68 fix wrong substring
The tokenOffsets is the offset from trimed contentLine.
But it is used in substring as the offset from un-trimed contentLine.

How to reproduce bug (ICS and before):
1. Add a contact record with name and a phone number (ex. 090-1111-2222)
2. Search by "0" in Contacts app
3. Contacts shows a wrong number something like "0900111102222"

Change-Id: I1e1cb2f3bd135d98e1573a50aa6ddc021af35b9f
2012-08-23 14:06:45 +09:00
Arik Nemtsov
5770b2dcf6 WifiStateMachine: ignore auth-fail event during WPS connection
Disregard auth failure events during WPS connection. The EAP sequence is
retried several times, and there might be failures (especially for wps pin).
We will get a WPS_XXX event at the end of the sequence anyway.

Without this change, the SupplicantStateTracker class will disable
the WPS network we are connecting to after 2 failed authentication events.
Then, even if WPS succeeds, we will never connect to the selected network.

Change-Id: I57c8e508bfd738405b7f2c2c2325df7838e1b4af
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Vishal Mahaveer <vishalm@ti.com>
2012-08-22 16:29:43 -05:00
Stephen Hines
9f7fa4c417 Merge "DO NOT MERGE - Fix readback lengths" 2012-08-22 14:06:37 -07:00
Irfan Sheriff
e5773274d6 Merge "WifiWatchdog: handle exception from ArpPeer" 2012-08-22 10:57:50 -07:00
Irfan Sheriff
2140d1de47 Merge "Avoid system_server crashing due to mac address is null" 2012-08-22 10:48:23 -07:00
Jean-Baptiste Queru
6ffe259467 Merge "restorecon /data/anr directory." 2012-08-22 07:18:39 -07:00
Chih-Wei Huang
21326d8b77 Avoid system_server crashing due to mac address is null
It's possible that the mac address is null (getMacAddress() of WifiNative.java).
In this case system_server will crash like:

E AndroidRuntime: *** FATAL EXCEPTION IN SYSTEM PROCESS: WifiWatchdogStateMachine
E AndroidRuntime: java.lang.NullPointerException
E AndroidRuntime:        at android.net.arp.ArpPeer.<init>(ArpPeer.java:57)
E AndroidRuntime:        at android.net.wifi.WifiWatchdogStateMachine.doArpTest(WifiWatchdogStateMachine.java:866)

Check the mac address before parsing to avoid crashing.

Change-Id: I5d4205c04d479a3a2837172c6382816ea4bf74d6
2012-08-22 11:26:51 +08:00
Dianne Hackborn
b0b4a70440 Merge "ActivityThread: Avoid creating duplicate handlers and reuse mH" 2012-08-21 15:12:45 -07:00
Jean-Baptiste Queru
b8da225f26 Merge "Tests/RS: Add RS context check for RsBalls setAccel()" 2012-08-21 13:09:19 -07:00
Jean-Baptiste Queru
26f8ce83e0 Merge "To avoid the JNI memory leaks in video editor codes" 2012-08-20 14:29:51 -07:00
Jean-Baptiste Queru
f515c6f4fe Merge "Switch widget: Fix ON/OFF option update error" 2012-08-20 14:28:33 -07:00
Kenny Root
f029c22f82 Merge "Use Libcore.os.umask for setting umask" 2012-08-20 12:32:14 -07:00
Kenny Root
4c74f8c171 Use Libcore.os.umask for setting umask
Libcore.os now provides a umask method call, so just use that since it's
available.

Change-Id: I8826fa40bcab3d61a424ff98a8050d3e44f7ec34
2012-08-20 12:24:05 -07:00
Jean-Baptiste Queru
e00245737c Merge "Release JNI local references as soon as possible." 2012-08-20 10:17:45 -07:00
Amith Yamasani
a3db5297cd Merge "ProviderMap: remove hashmap when entire user specific entries are removed" 2012-08-20 10:16:36 -07:00
rpcraig
ec7ed14c18 restorecon /data/anr directory.
Restore the security contexts of anr directory
when initially created.

Change-Id: Ia731414ccbcdc7369d24be6db0003c53abcf6ef4
Signed-off-by: rpcraig <rpcraig@tycho.ncsc.mil>
2012-08-20 06:08:01 -04:00
Vairavan Srinivasan
e4c56d9367 VibratorService: Fix to ensure actual delay in a vibrate pattern
delay might timeout early as value of duration isn't updated
correctly in the loop, should the wait be interrupted, to reflect
the elapsed time. Fix is to update duration in the loop.

Change-Id: I525b0e97799b288f46ae3a056cff7dcc69701bb0
2012-08-19 12:08:42 -07:00
Vairavan Srinivasan
7335cfd9bb ActivityThread: Avoid creating duplicate handlers and reuse mH
sMainThreadHandler is used for notification of changes in
SharedPreferences and refers to a new instance of Handler instead
of the existing instance, mH, of the ActivityThread. Fix to reuse
existing handlers associated with main looper.

Change-Id: I794a5802a9eacdb188bd0619db5e70e3ae89a07d
2012-08-18 18:36:03 -07:00
Jason Sams
aee28aa6b9 DO NOT MERGE - Fix readback lengths
Change-Id: I4c5f29af03ed3babceef741aa3a982244cfc65d1
2012-08-17 18:24:07 -07:00
Henrik B Andersson
057b086e0e Release JNI local references as soon as possible.
The JNI layer has a fix amount (512) of jni references.
They should be released as soon as possible. In for
and while loops they can quickly reach the limit.

Change-Id: Id984345e1cc4f7aa6eb31a263b796c3da9edf773
2012-08-17 15:59:35 -07:00
Robert Greenwalt
6626398a72 Merge "Fix for updating TCP buffer size when switching network type" 2012-08-17 13:01:49 -07:00
Jean-Baptiste Queru
a688af9c67 Merge "TabHost key handling corrected" 2012-08-17 10:02:50 -07:00
Magnus Strandberg
1968256926 Fix for updating TCP buffer size when switching network type
When PDP connects in GSM network, the TCP buffer window size is
set to max 8760. If the phone swicthes to UMTS, the TCP buffer
size is not updated to UMTS, i.e. GPRS buffer size is still
used with limited data transfer spead as a result.
This fix makes sure the TCP buffer size is updated when
switching network type.

Change-Id: Iaa527c84f56cae108f4c7276bbed10552faaff6c
2012-08-17 11:38:44 +02:00
Joe Onorato
5f61087e26 Merge "Get rid of LOCAL_MODULE_TAGS := user" 2012-08-17 00:40:50 -07:00
Joe Onorato
0e462950d9 Get rid of LOCAL_MODULE_TAGS := user
Change-Id: I1c4e67f935a7c511cda9980f4475dfd0272617a5
2012-08-16 22:37:58 -07:00
bxu10X
96ac720117 [Audio]Press camera button during Http live Audio playing, then return Browser, DUT display abnormal.
When long press camera key, dispatchKeyEvent function will add control surface after the control view hide,
actually, we needn't show control view after we press camera key.

Change-Id: I45175c4eb97b3f13469cd9d8bacb295c9c5803dd
Author: Bin Xu <bxu10X@intel.com>
Signed-off-by: Bin Xu <bxu10X@intel.com>
Singed-off-by: Shuo Gao <shuo.gao@intel.com>
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>
Signed-off-by: Jack Ren <jack.ren@intel.com>
Author-tracking-BZ: 28344
2012-08-17 10:52:33 +08:00
wfeng6
c96a7294a4 To avoid the JNI memory leaks in video editor codes
There is a protential memory leak in video editor JNI codes.
When importing so many video clips, the number of local reference
will exceed 512, which will cause the Dalvik virtual machine
crash. The patch will free each local reference when it's no
longer used to avoid memory leaks.

Change-Id: Iecd986c83f05a1103641746d6d43eef0a9c5ee40
Author: Wei Feng <wei.feng@intel.com>
Signed-off-by: Wei Feng <wei.feng@intel.com>
Singed-off-by: Shuo Gao <shuo.gao@intel.com>
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>
Signed-off-by: Jack Ren <jack.ren@intel.com>
Author-tracking-BZ: 40786
2012-08-17 10:22:57 +08:00
Yong Chen
da1c93ae4d Tests/RS: Add RS context check for RsBalls setAccel()
mRS could be null when lock and unlock screen by pressing power key
several times, it causes RsBalls crash. Add null pointer check to fix.

Change-Id: If37dee8609420daaff17772d5194cad5531f98e6
Author: Yong Chen <yong.a.chen@intel.com>
Signed-off-by: Yong Chen <yong.a.chen@intel.com>
Singed-off-by: Shuo Gao <shuo.gao@intel.com>
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>
Signed-off-by: Jack Ren <jack.ren@intel.com>
Author-tracking-BZ: 31271
2012-08-17 10:04:21 +08:00
Kenny Root
a950daf5c1 Merge changes Ieb566a2a,I953057cd
* changes:
  Use Libcore's stat instead of FileUtils#getFileStatus
  Use Libcore.os.stat instead of FileUtils
2012-08-16 16:06:29 -07:00
Kenny Root
98e15e7893 Use Libcore's stat instead of FileUtils#getFileStatus
Remove the last user of FileUtils#getFileStatus and move it to
Libcore.os.stat instead. Then we can remove the JNI code that does the
equivalent of a stat.

Change-Id: Ieb566a2a8a17c2dd0150724b4eb3ac1cc41c823d
2012-08-16 15:27:55 -07:00
Kenny Root
786cbcacd2 Use Libcore.os.stat instead of FileUtils
PackageManagerService just needed to know the owner for this file, so
just use stat instead so we can remove the old JNI code.

This is the last user of FileUtils#getPermissions so just remove the
FileUtils method as well.

Change-Id: I953057cd6b9de4410f33b6f22e4bddff02fe2988
2012-08-16 15:27:55 -07:00
Jean-Baptiste Queru
d2fb6e99bd Merge "Fix Issue #35489: drm content corruption on GB" 2012-08-16 14:45:12 -07:00