Hi Daniel,
The set of affected tests is somewhat limited in that the real cause of
the timeout
is reading the timezone information and the extra code that it throws into
the mix due to -Xcomp. Also, the current sequence of tests may be
hiding some failures since the one-time initialization has already been
performed by some other test.
Thanks, Roger
On 4/27/2015 3:37 PM, Daniel Fuchs wrote:
Hi Roger,
Looks good to me. I'm surprised that only these tests are failing.
IIRC I also had to round up some time comparisons in tests when I
increased the resolution of the SystemClock - and I believe I rounded
them to only 1 second...
Anyway - if they fail we know what to do :-)
best regards,
-- daniel
]
On 4/27/15 8:50 PM, Roger Riggs wrote:
Please review a timing relaxation so tests do not fail on very slow
hardware
with more expensive runtime options. (-Xcomp)
The change will also be backported to jdk 8.
Issue:
8078369: [testbug] java/time/tck/java/time/TCKOffsetTime[now]
fails on slow devices
Webrev:
http://cr.openjdk.java.net/~rriggs/webrev-timeout-8078369/
Thanks, Roger
