Hi Phil,
the fix looks fine to me, approved.
Thanks,
Andrew
On 6/19/2013 9:42 PM, Phil Race wrote:
I revamped the test to scrap the unnecessary swing stuff.
No invokeLater needed anymore.
http://cr.openjdk.java.net/~prr/8015334.jdk7.1/
Nothing else has changed.
-phil.
On 6/19/2013 5:28 AM, Andrew Brygin wrote:
Hello Phil,
The fix looks fine to me.
However, the test does not work correctly with jtreg harness:
for some reasons, jtreg reports success immediately after
invokeLater(),
so actual testing is not performed. The option 'othervm' does not
help: the test reports success anyway.
This behavior seems to be a bug in the jtreg harness. It could be
workarounded
by an explicit synchronization between main and EDT threads.
Also, the @test tag needs to be added to the test header.
Thanks,
Andrew
On 6/19/2013 1:49 AM, Phil Race wrote:
Steven & Andrew,
I believe 8015334 : Memory leak in Kerning should be ported to 7u40.
Please sanity check this fix which is the same as the already pushed
JDK 8 fix :
http://cr.openjdk.java.net/~prr/8015334.jdk7/ then I'll see if it'll
get
accepted into the 7u40 stabilisation forest.
-phil.
