diff options
| author | Neil Fuller <nfuller@google.com> | 2014-02-07 09:05:34 +0000 |
|---|---|---|
| committer | Neil Fuller <nfuller@google.com> | 2014-02-07 09:05:34 +0000 |
| commit | e37bfbcea35237c5c61f4d94788eafbc5f1f4ee8 (patch) | |
| tree | 61bf6133f079745b9bb97d0437752d201a4e2ba2 /Docs.mk | |
| parent | 0fe3dd19a1d53154b09556e80dd5b2ba69d5f302 (diff) | |
| download | libcore-e37bfbcea35237c5c61f4d94788eafbc5f1f4ee8.zip libcore-e37bfbcea35237c5c61f4d94788eafbc5f1f4ee8.tar.gz libcore-e37bfbcea35237c5c61f4d94788eafbc5f1f4ee8.tar.bz2 | |
Fix CTS test failure under ART
The Thread.interrupted() State from
SelectorTest.testInterrupted() was leaking into
SelectorTest.testManyWakeupCallsTriggerOnlyOneWakeup() causing
that test to fail when the CTS was run with the ART. Under
Dalvik the "interrupted" state was being reset, but it was
causing the test runner to run each subsequent test in its
own Dalvik instance (which probably slowed down the tests but
wasn't causing failures).
It is unclear why a Thread being in "interrupted" after a test
does not cause problems for ART in the same way it causes
problems for Dalvik.
Bug: 12905142
Change-Id: I677c1d46602d10e174ba3b6b8614529b5c902104
Diffstat (limited to 'Docs.mk')
0 files changed, 0 insertions, 0 deletions