* changes:
Fix a race that could cause GL commands to be executed from the wrong thread.
RefBase subclasses can now decide how they want to be destroyed.
Fix a race in SurfaceFlinger that could cause layers to be leaked forever.
Fix a race-condtion in SurfaceFlinger that could lead to a crash.
LayoutLib: custom styles override the default style instead of replacing it.
Intead of reading either the custom or the default style for a widget, we
read both and use the values from the custom style if it exists, or
from the default style otherwise.
Change-Id: Ibcec2e9b1e8a95295ab26ede145c287ff2f30be4
This adds a destroy() virtual on RefBase which
sublasses can implement. destroy() is called
in lieu of the destructor whenthe last strong
ref goes away.
Bug: 4483050
Change-Id: I8cbf6044a6fd3f01043a45592b5a60fa1e5fade2
The transaction flags were atomically read-and-cleared to determine if
a transaction was needed, in the later case, mStateLock was taken to
keep the current state still during the transaction. This left a small
window open, where a layer could be removed after the transaction flags
were checked but before the transaction was started holding the lock.
In that situation eTraversalNeeded would be set but only seen during the
next transaction cycle; however, because we're handling this transaction
(because of another flag) it will be commited, "loosing" the information
about the layer being removed -- so when the next transaction cycle due
to eTraversalNeeded starts, it won't notice that layers have been removed
and won't populated the ditchedLayers array.
Bug: 4483049
Change-Id: Ibb5989312f871339928ee1aa3f9567770d72969b
Client::mLayers could be accessed from different threads.
On one side from Client::attachLayer() which is currently
called from a binder thread; on the other side from
Client::detachLayer() which is always called from the main
thread.
This could lead to a corruption of Client::mLayers.
We fix this issue by adding an internal lock to Client.
Bug: 4483046
Change-Id: I5262bf1124d9a65ec6f8ffd8e367356fc33a7536
When sending an SMS to an international number in the format
+401234567890, the "+" should be converted to the International
Dialing Prefix (in the US, 011). However, the device drops this
"+" altogether in the outbound data burst message causing the message
to fail or be sent to the wrong address.
Change-Id: If25c092d283f1703b49cf52d0379efa54639f093
Signed-off-by: Soojung Shin <sj46.shin@samsung.com>
- Signal level updates should be updated whenever a signal change
broadcast is received but it is failing to do so.
- Only when state change broadcast is received, signal level was
updated. (Ex: State: Idle -> State: Connected)
Change-Id: I71d86782143b3440a042164a87af64c7bee97ae2
Signed-off-by: TK MUN <tk.mun@samsung.com>
LayoutLib: Fix issue where <include> with no layout params wouldn't display.
The issue is that the layout params from the root element of the included
layout should be used but this failed because loading the layout params
from the <include> tag didn't throw a RuntimeException in our modified
code (BridgeTypedArray).
Because we don't want to throw exception in general we only throw it
when reading the layout params of an include node which is pretty crappy,
but works for now.
Change-Id: I83ccf956e8b476f34dfc9a70aebae2288d53746e
Secure settings are accessible by applications without
permissions. Move to a system-only accessible file
Bug: 4391796
Change-Id: I5448943d26cccce39dd4f32423c68dad4bbf50ad
The native DHCP client renews IP address, but can fail to do so when the device
is in suspend. This has lead to issues where device fails to renew in time and
on networks with lots of devices re-using expired DHCP IP addresses, it leads to
severe issues.
Handle DHCP renewal and wakeup in framework
Bug: 3344732
Change-Id: Ie4062e04a477f4a233946155e40a7c999b337c3f
To avoid blowing past the Binder IPC limit, change the
PackageManagerService to have a DB-like interaction where the client
tells the service the last "row" that it read.
The fact that we use a HashMap instead of a TreeMap makes this
problematic. For now we're just making a new ArrayList for the keys and
then sorting them for each call. This can make the API slower for callers
of this, but it's probably greatly overshadowed by the cost of the data
transfer itself.
Bug: 4064282
Change-Id: Ic370fd148d4c3813ae4f2daffa1a7c28d63d5a09
LayoutLib: Fix issue with rendering mode.
When the layout content is embedded inside a decor layout
to emulate system bar (top or bottom) and title bar (or action bar)
then the code computing the full required size based on the RenderingMode
would fail because the decor layout would prevents the content layout to
take as much room as possible.
There is also an issue with the way we know render dialogs as the
dialogs usually as a frame with some padding and the previous measurements
would not take into account the dialog padding when increasing the
screen size.
This fix makes the code measure the size of the content layout in the
normal rendering, and then separately from the root layout with the proper
MeasureSpec to let the content grown. The difference in size is then added
to the screen size.
Bug: http://code.google.com/p/android/issues/detail?id=15892
Change-Id: Ie9a6c5e254b16785f817dcb9fae755d4936880aa
Add support for encoding and decoding SMS 7 bit user data using the
national language shift tables defined in 3GPP TS 23.038 (GSM/UMTS only),
including the new tables added in Release 9 for Indic languages.
Decoding is always supported, but encoding is only enabled for the
specific language tables added to the new integer array resources
"config_sms_enabled_single_shift_tables" and
"config_sms_enabled_locking_shift_tables" defined in
frameworks/base/core/res/res/values/config.xml. The default empty arrays
should be overridden in an OEM overlay for the specific nationalities where
SMS national language shift table encoding is allowed/mandated (e.g. Turkey).
GsmAlphabet.countGsmSeptets() will try to find the most efficient encoding
among all combinations of enabled locking shift and single shift tables.
If no 7 bit encoding is possible, 16 bit UCS-2 encoding will be used.
This change also fixes a bug in the decoder: when an escape septet
is followed by a septet with no entry in the extension (single shift)
table, TS 23.038 Table 6.2.1.1 states that the MS shall display
the character in the main GSM 7 bit default alphabet table, or the
active national language locking shift table. Previously, we were
decoding this sequence as a space character. Two consecutive escape
septets will continue to decode as a space character, according to
Note 1 of table 6.2.1.1.
Change-Id: I4dab3f0ffe39f3df2064ed93c9c05f26e274d18b
When user uncheck "Data Enabled" check box, WiMAX goes
into "disconnected" state.
Change-Id: I3b9bdbc16cc4ddbf7a1aac0c984cad8994c4e9f2
Signed-off-by: TK MUN <tk.mun@samsung.com>
MarketService.
We use the IMarketBillingService, which is incorrect. It should be
"com.android.vending.billing.MarketBillingService.BIND".
Change-Id: Ie361e58cea8aca6d70bd48bce062ca8d062281ee
Handle the case where the kernel driver is in accessory mode but we failed
to initialize it at the framework level. On disconnnect, check to see if the
accessory kernel driver is enabled rather than checking mCurrentAccessory.
That way we will restore the USB state in the kernel even if mCurrentAccessory
is null.
Change-Id: I2c4f6edb34aae2064f4b62ec0461d1fdd8770541
Signed-off-by: Mike Lockwood <lockwood@android.com>
Herring board exhibits a strong echo even in non speakerphone modes.
To compensate the lack of AEC or AES when not in speakerphone, the mic gain
had been reduced in the ADC. But this has an adverse effect on other VoIP applications
that have their own AEC and are penalized by the weak mic gain.
This workaround enables an acceptable mic gain for other VoIP apps while offering a
SIP call experience which is not worse than it was with the residual echo that was
present even with mic gain reduction.
Change-Id: I33fd37858758e94e42ef5b545d3f0dc233220bf1
eglMakeCurrent() would only deref the previous surfaces if the old and
new contexts were the same. eglTerminate() should not touch TLS.
eglReleaseThread() needs to unbind the current context.
Change-Id: I7f4c090a287ee1e29e4708ae10679fb9d7d8c8c5
Related-Bug: 2964479
If original refuses to tear down, tear down new one. It's better
to have none (which will try to launch them all again) than two.
Really people shouldn't refuse the teardown request.
bug:4183397
Change-Id: I54ea1bf0d2cd2ef16fcf2eafc69895ad2fe33ffd
