> On 9 Jan 2014, at 18:18, Daniel Fuchs <daniel.fu...@oracle.com> wrote: > >> On 1/8/14 11:37 PM, Chris Hegarty wrote: >> >>> On 8 Jan 2014, at 20:04, Mandy Chung <mandy.ch...@oracle.com> wrote: >>> >>> >>>> On 1/8/2014 11:30 AM, Daniel Fuchs wrote: >>>> >>>> <http://cr.openjdk.java.net/~dfuchs/webrev_8031068-jdk9/webrev.01/> >>> >>> The fix looks fine. The WeakReference and System.gc() call was to help >>> reproduce the test failure. I wonder if it's really needed to be retained >>> in the test (or better to remove it). >> >> Sounds like the most reasonable solution. > > OK - I removed the unnecessary calls. > > http://cr.openjdk.java.net/~dfuchs/webrev_8031068-jdk9/webrev.02/
Thanks. Look good to me. -Chris. > > -- daniel > >> >> -Chris. >> >>> There has been a few test failures due to this eager GC issue running in >>> -server -Xcomp mode (I guess they are from the hotspot nightly testing). >>> It might worth to run all logging tests with -server -Xcomp to find out if >>> all test failures due to this issue are reported and fix them all together. >>> >>> Mandy >