GitLab

lock()
lock(long, long, boolean)
tryLock()
tryLock(long, long, boolean)

Bug: 27186422
(cherry-picked from commit 2eff57cc1164d20f8a748f8e752ebac021476167)
Change-Id: Ia2d67993d61116e83e636e9305e40932e79d2584

We don't support lambda [de]serialization so this shouldn't have been
in the test list to begin with.

Bug: 28505376
Change-Id: I2cd29b16aa9646d5dd1204912e7d9e28ee94267d
(cherry picked from commit b2edc632)

We should be able to generate key pairs when the default locale
is set to farsi (or any other locale with non arabic digits).

bug: 28384942

(cherry picked from commit 5690417e)

Change-Id: Ice3484f31978e2c5aa53b9d76b61a25dd56b10e7

The testAllBlockDevicesAreSecure test is broken. Having
testAllBlockDevicesAreSecure as a device side test would require
granting permissions to the untrusted_app domain, which was
undesirable.

Move this test to host side and add permissions for shell to getattr
on block devices.

bug: 28306036
Change-Id: I36566e0b28b5c8ca69dc99a49e4cf9518b8323e2
Signed-off-by: William Roberts <william.c.roberts@intel.com>

Bug: 27521545
Change-Id: I2a954401ce47604987a437a2741bcab252d32707
(cherry picked from commit f995e958)

LinkerNamespacesTest did not take into consideration wear devices
do not support PackageManager.FEATURE_WEBVIEW. This CL fixes this
by adding check this feature.

Bug: http://b/26894351
Change-Id: I28eceeba35b438f2131b8ae20bfa26b310fdcdf3
(cherry picked from commit baae03a7)

Bug: 23535985

Adapted from Pirama's tests in RsTest, to make it work with the CTS
testing harness.

Refactored so that tests are in seperate groups depending on what
they are testing. In the allocation creation test, also moved the loop
over data types and vector sizes into the script from Java.

Dropped the return values for testing kernels, since it is no longer
a requirement in the final implementation of rsForEach().

Change-Id: Ie5512809412c1229dd975395fd33c18ec643e1dc
(cherry picked from commit 1238e2e8)

Move testAllCharacterDevicesAreSecure() to a host side test.
Currently, this CTS test is unable to enumerate devices on /dev
becuase SE Linux policy is blocking it. Rather then open this
access to untrusted_app domain, move the test host side where
shell domain shall be granted access.

bug: 28306036
Change-Id: Ie75366f0efe88a0d265d62b6cdc8014827e2e069

This improvements helps avoid situation when some of the libraries
expected to reside in /system/lib are moved to /vendor/lib directory.

This is not allowed and the test now checks the limited list of known
libraries to make sure that they are in /system/lib

Bug: http://b/26894351
Change-Id: I4943a8ad808537b217dfcba5e242b037ca2626a6
(cherry picked from commit c93b8585)

* changes:
  Removed unused global variable
  CTS tests for single-source RS rsForEach

Bug: 27874893

(cherry picked from commit 2b9f262c)

Change-Id: I1d99c58cf1bacffc77bd6b0c42072334073e2348

Bug: 23535985

Added CTS tests for the rsForEach API for Single-Source RenderScript.

Adjusted one existing CTS test (get_allocation.rs for GetAllocationTest)
to avoid calling the deprecated rsForEach API with a script argument.

Temporarily added the 0 target API level for all RenderScript tests, since
the new API rsForEach is not officially released yet. This would be
removed once the API is released, and assigned an official API level.

(cherry picked from commit 1b9f8e15)

Change-Id: Ife5fd3e30de8c223dc7b5ea4d8c6c1a7a9e852f5

Bug: 28199918
Change-Id: I7626404617a21b2dd3301db8224b4c7a74362b2d

Device may not be built with libbacktrace_offline shared library,
so use libbacktrace_offline static library instead.

Bug: 28152982
Change-Id: I46fdb53e711dd60902dca7c9d46030c16eeb222a

Bug: 27747093

Relaxed the precision of the new Refocus test a bit more to tolerate
GPU run.

More cleanups
* Avoided sending new picture intent, since CTS runs in StrictMode
 that forbids exposing piture URI.
* Renamed a parameter with a better name
* Comment rewording

Change-Id: I4393f637ac833142a5ad43c7b2b1851cc9394b29
(cherry picked from commit 067aaadc)

Bug: 27747093

Replaced the test image with a smaller one, so that it can be loaded
into memory on an X86 emulator, or a device with small RAM.

Separated the depthmap from the test image itself, to make it easy to
resize both the image and the depthmap in sync. Adjusted DepthImage
and RGBZ classes to handle a separate image and depthmap.

Adjusted the DepthOfFieldOptions class to make the output image
look reasonable. The blur-at-infinity parameter was improperly set
previously.

Use PSNR for image comparison. Updated the ImageCompare class to do
in-place comparision, instead of loading redundant copies of the images
into memory.

Adjusted MediaStoreSaver class to use a different path to write the
output image, in the case where Pictures directory does not exist,
e.g., on an X86 emulator.

Use PNG instaed of JPG test images, to avoid noises from decompression
on different target platforms.

Change-Id: I281c738780d98bb70404d05b59a430735c546715

The system is allowed to load libraries specified
in /vendor/etc/public.libraries.txt from /vendor/lib

Bug: http://b/27073931
Bug: http://b/26894351
Change-Id: I6b7bd077843810f5166e47095750bc8f33c344cf

The tests check the behaviour of FileLock when more than one processe or
thread try to hold the file lock.

The FileLock should return null if the multiple processes try to obtain
the lock, however, it should raise OverlappingFileLockException if the
lock is already held by the same runtime.

Bug: 27186422
(cherry-picked from commit f16ac2dcde20737e2f437d5c8cd5a321c0c38273)
Change-Id: I97184d7735ad05a9b88cf8ee7dbc76177b0210d2

http://b/27886613

Put a runtime check around rsDebug calls to prevent excessive logging,
while simultaneously ensuring that overeager compiler optimizations
don't optimize away writes to kernel inputs.

Change-Id: I5af22c8c60a73615bf61ee61d78d19236d88fe00

Bug: http://b/28035006
Change-Id: I2bef1c2965c32d282e64968c3b1d703090de5b01

Use mockito-api instead of mockito-target-minus-junit4

Change-Id: I8e131b9c54f213565a12f06a075e37a8fed4bdae