| 4d8dc721410a0bf120893ded7a9bf684a962de54 |
|
26-Oct-2016 |
Artem Udovichenko <artem.u@samsung.com> |
Throw IllegalThreadStateException when Thread.start is used in wrong way
Test: run libcore.java.lang.ThreadTest
Change-Id: Ie9bb8aaed5aec6bbe22d4d868ea0b0fa4853ff18
|
| b62710c762bf399ee1a45821eb0740b2d9c92f62 |
|
21-Jul-2016 |
Tobias Thierer <tobiast@google.com> |
Ensure apps cannot prevent uncaught exceptions being logged.
Add a new @hide API to set an additional UncaughtExceptionHandler that
is called before dispatching to the regular handler. The framework
uses this to enforce logging.
Test: Ran ThreadTest.
Bug: 29624607
Change-Id: Ieb8d70662f7a37e9bfcdeb2818bf4863bb6c1890
|
| bf34d8378762aa4d97c30c86ef95502e5d85ed3e |
|
27-Jun-2016 |
Przemyslaw Szczepaniak <pszczepaniak@google.com> |
Fix Unsafe#park for large wait-time values.
jsr166.LockSupportTest#testParkNeg_parkUntilBlocker
fails if the delay between the tests
System.currentTimeMillis is called at least 2ms behind
the Unsafe#park. This is caused by underflow to
large positive value when subtracting current time
from the method argument.
This change fixes it by not doing subtraction before
comparison.
parkUntil takes arguments in milliseconds and is implemented
as a call to parkFor, which takes arguments in nanoseconds.
If parkUntil argument will overflow during conversion to
nanoseconds, it's clamped on max value (~292 years).
Bug: 29746125
Change-Id: Ie0b3923ce4567415bf6478552c51eb31ac4aec62
|
| b814d44b43f5735503eb5c64cf57f9e5e758c0e1 |
|
29-Jun-2016 |
Neil Fuller <nfuller@google.com> |
Remove an infinite loop / put a Thread out of its misery
ThreadTest.testGetStackTrace() created a thread that never
terminated due to a for loop with a missing increment.
The loop was unnecessary and has been removed.
Change-Id: If00018b6330fe1340c293585fb9d64206dd89e20
Test: Ran the CTS tests
Bug: 29820565
|
| 2047ef0b5d272a03297dc1cd307a2dcfa42e280b |
|
23-Mar-2016 |
Narayan Kamath <narayan@google.com> |
Thread: Don't set native names twice for attached threads.
ART already does this when a threads are attached. There's no need
to do it another time.
Also add a test that documents the behaviour wrt. to thread naming
for attached threads.
bug: 27748318
(cherry picked from commit 81888044c73382ef87285593afdf80e840288b06)
Change-Id: I8b1ac9407114d0fc399a3bc644d45e2f00ef7200
|
| 81888044c73382ef87285593afdf80e840288b06 |
|
23-Mar-2016 |
Narayan Kamath <narayan@google.com> |
Thread: Don't set native names twice for attached threads.
ART already does this when a threads are attached. There's no need
to do it another time.
Also add a test that documents the behaviour wrt. to thread naming
for attached threads.
bug: 27748318
Change-Id: I9075f0434b0168280662a1f8a4176f23724748a7
|
| b5a43a8ad530a60469bf9244a157079c93c07c8b |
|
11-Feb-2014 |
Brian Carlstrom <bdc@google.com> |
Fix Thread.sleep(0) of an interrupted thread
Bug: 12929305
Change-Id: I3061c3345c68de8bcf438e6935446b4f7cd76001
|
| c3d53e68cbf91fb4c16abb88f8bf8a8e4be4d706 |
|
22-Nov-2011 |
Jesse Wilson <jessewilson@google.com> |
Add tests to check for a valid context class loader.
(cherry picked from commit cac9b2d67f4570af232bfbf7f065a92878cd3350)
Change-Id: I713fc13be90cdbb13a71ea19502a641f3c1360cb
|
| b20185245f53374b3a86705ac152fd88e4bddec5 |
|
15-Jul-2011 |
jeffhao <jeffhao@google.com> |
Changed Thread.sleep from native to java implementation.
The change removes a global monitor that was being used in the native
sleep method. Part of a multi-project commit.
(cherry picked from commit 39409c767f8e15bd81f32808a2d4ec2796a2c3aa)
Change-Id: Ib8afaea8eb35202bdec29be3d23cf2e90a70197f
|
| 0647bfed6eda99ad77b2dfe8e3696e3fabfaf3cc |
|
13-Dec-2011 |
Jesse Wilson <jessewilson@google.com> |
Bring more tests from frameworks/base/tests/CoreTests into libcore.
Bug: http://b/3073226
Change-Id: I66f88f2bfaadd39e53b2977606731e8579d880b6
|
| 1a9cff8f68deffd618c5cba1f22f0fb0e396e067 |
|
17-Nov-2011 |
Jesse Wilson <jessewilson@google.com> |
Always trigger finalization the same way in our tests.
This came up when I was recently writing a ZipFile finalization test
and I needed to copy-paste the finalizer recipe.
Change-Id: Ia67061b3dba1a7011c93c9a81e2a963876b238a1
|
| be013ce620f6d3bd24f7f0b631a36f70197ac3e3 |
|
23-Oct-2010 |
Jesse Wilson <jessewilson@google.com> |
Change ThreadGroup to no longer hold strong references to threads.
This is intended to fix an issue where threads were leaking when
they weren't started.
This changes a lot of the internals of ThreadGroup, taking advantage
of a new utility method dereferenceIterable that turns a collection
of references into the corresponding iterable of referents.
Change-Id: I8d96c79b4ee30c557276854b38149b9131b656ee
http://b/1937448
|