Checking carrier privileges for UIDs with lots of shared apps can incur
a significant performance hit. For UIDs that are fixed and trusted
(system and phone), skip the permission check and always allow.
Bug: 160971853
Test: manual verification -- observed lower rate of cache misses for
getPackageInfo from com.android.phone.
Change-Id: I1399cab579308479d7cf191b8795441cbcd3ff65
Merged-In: I1399cab579308479d7cf191b8795441cbcd3ff65
This used to be done in Make, but with bp-based dist now supporting
tags we can convert these java_library { hostdex: true } libraries
to dist their classes.jar rather than dex jar.
Bug: 161214753
Test: diff results of `m sdk dist`
Merged-In: Id54507e297b5c57d1ce58f14eb347e7661a5328d
Change-Id: Id54507e297b5c57d1ce58f14eb347e7661a5328d
android_robolectric_test is having some implicit deps removed
from it, so add in what we depend on directly.
Test: m RunFrameworksServicesRoboTests
Change-Id: Ice2abeb3f4d031cfe15ce919b750d6de59316c1f
Merged-In: Ice2abeb3f4d031cfe15ce919b750d6de59316c1f
... but not other runtime VINTF info. This may or may
not fix the nullptr exception when accessing getRuntimeInfo().
In either case, this change should improves code health in general:
- it avoids reading things it doesn't need
- it avoids copying the map
Bug: 151092221
Test: m libandroid_runtime -j
Change-Id: I6cb9052eb56fd83ef7ee81751d9791533a36aeed
(cherry picked from commit 77c1f592fa)
Merged-In: I6cb9052eb56fd83ef7ee81751d9791533a36aeed
Two new partitions are created in Android. New notice files are placed
in these new partitions.
Test: none
Bug: 156020364
Change-Id: Iefd5b12b3bbb449b189ade313157e80f36d99c88
From https://github.com/torvalds/linux/blob/master/include/uapi/linux/input-event-codes.h,
KEY_SUBTITLE is near the TV codes. Note that 0x172 == 370.
#define KEY_CHANNEL 0x16b
#define KEY_FAVORITES 0x16c
#define KEY_EPG 0x16d
#define KEY_PVR 0x16e /* Media Select Home */
#define KEY_MHP 0x16f
#define KEY_LANGUAGE 0x170
#define KEY_TITLE 0x171
#define KEY_SUBTITLE 0x172
#define KEY_ANGLE 0x173
#define KEY_FULL_SCREEN 0x174 /* AC View Toggle */
#define KEY_ZOOM KEY_FULL_SCREEN
You can also see on https://www.kernel.org/doc/html/v4.9/media/uapi/rc/rc-tables.html it is listed in the "Movie play control" section:
KEY_SUBTITLE Allow changing the subtitle
Change-Id: I7690269cddd00b787a939f87ba651db38f1160ac
Signed-off-by: Aman Gupta <aman@tmm1.net>
Not specifying any sdk_version defaults the library to building
against the platform internals. We do not want to build the stubs
against the platform internals, as that risks leaking internals
in the stubs.
Follow what the other stubs are doing, which is to build against
no default libs, and compiling with java version 1.8.
Bug: 158850553
Test: m
Change-Id: I310842747422b108ac2a1fcf11afac5c323ed5b1
They can be replaced by the public API provided by ICU4J instead.
No extra layer of caching equivalent to the LocaleData cache is
added because ICU4J has caches instances, e.g. DateFormatSymbols,
etc.
Bug: 160606356
Test: atest FrameworksCoreTests:android.text.format
Change-Id: I07048e0e1a4835d2744b7fce6a5ed79a112e456b
KEY_NR_TIMERS_RESET_IF_NON_ENDC_AND_RRC_IDLE_BOOL:
Whetherdevice reset all of NR timers when device camped on a
network that haven't 5G capability and RRC currently in IDLE state.
Bug: 160365713
Test: build pass.
Change-Id: I9802b8712798054a85307e549fb665bc85166311
Merged-In: I9802b8712798054a85307e549fb665bc85166311
Copy and paste internal setup to AOSP. Not include folders not in AOSP
yet.
Bug: None
Test: repo upload
Change-Id: I08b8c6d1bae7d1954847c207ef8e6d80373b2d05
Merged-In: I08b8c6d1bae7d1954847c207ef8e6d80373b2d05
product_services partition is not used anymore, its name is changed to
system_ext, so following this, change path in LicenseHtml
Test: m
Bug: 134359158
Change-Id: I6907880ab8af456358fb3420b74dcd3e409b41f3
(cherry picked from commit 2f2fe4f72a)
Merged-In: I6907880ab8af456358fb3420b74dcd3e409b41f3
The existence of both fdtrack and the older OpenFdMonitor is confusing
to partners, and not obviously useful because the fdtrack limit is so
much lower than the OpenFdMonitor limit (now that the default
RLIMIT_NOFILE is 32Ki).
Bug: https://issuetracker.google.com/161199049
Test: builds
Change-Id: I86436f287d0e25d1dfae95b915beb78c88dccce2
Merged-In: I86436f287d0e25d1dfae95b915beb78c88dccce2
