On 2013-10-14 09:42, Staffan Larsen wrote:
Looks good!
Thanks, Staffan.
But I wonder why there is a shell script involved. Couldn't this be
just a straightforward java file with several @run tags?
Probably. I'd rather not change that now.
thanks,
StefanK
Thanks,
/Staffan
On 11 okt 2013, at 20:02, Stefan Karlsson <stefan.karls...@oracle.com
<mailto:stefan.karls...@oracle.com>> wrote:
On 10/11/13 7:44 PM, Coleen Phillimore wrote:
Stefan,
I added a check about the count exceeding 1 000 000 because it
causes the test to create an invalid class file. With the shell
script changes, it won't happen but I would be happy if that check
were there so no one ever has to debug this thinking it has a
corrupted constant pool again.
OK. I though that was old debugging left-overs.
New webrev:
http://cr.openjdk.java.net/~stefank/7196801/webrev.01/
thanks,
StefanK
Thanks,
Coleen
On 10/11/2013 1:33 PM, Stefan Karlsson wrote:
Please, review this test fix.
http://cr.openjdk.java.net/~stefank/7196801/webrev.00/
Most of the fix was implemented by Coleen, I've only done some
small tweaks and updated some comments.
The fix depends on these changes:
JDK-8025996 Track metaspace usage when metaspace is expanded
JDK-8024547 MaxMetaspaceSize should limit the committed memory
used by the metaspaces
Both of these changesets have been pushed to hotspot-gc and are
currently undergoing testing. LowMemoryTest2 can only be removed
from the test/ProblemList.txt when both JDK-8025996 and JDK-8024547
have propagated through all necessary repositories.
thanks,
StefanK
