Hi Amy,
yes, this looks fine; no need for the RMI Registry in the test.
Perhaps you can fix the typo: line 54: "Exdecute" =-> "Execute";
no need for another webrev.
Thanks, Roger
On 11/4/2016 9:33 AM, Daniel Fuchs wrote:
Hi Amy,
Looks good to me.
Might be good to have confirmation from Roger or Stuart
that no longer using the registry is OK WRT to the
purpose of this test. If so then this looks like
the ideal solution :-)
best regards,
-- daniel
On 04/11/16 13:27, Amy Lu wrote:
Hi, Daniel
I updated webrev, removed the test dependency on TestLibrary.
(RMI logging still be triggered thus no change to testing purpose.)
Please review again:
http://cr.openjdk.java.net/~amlu/8037278/webrev.02/
Thanks,
Amy
On 11/3/16 9:42 PM, Daniel Fuchs wrote:
On 03/11/16 13:35, Daniel Fuchs wrote:
Hi Amy,
Looks good to me.
Maybe a followup should be logged to try
and no longer call TestLibrary.getUnusedRandomPort(),
as this the method that is causing the rogue
output (and I see that it might print things
to System.err as well).
best regards,
-- daniel
-- daniel
On 03/11/16 13:22, Amy Lu wrote:
Thank you Daniel for your review.
Yes, agree and reverted those three lines.
webrev updated:
http://cr.openjdk.java.net/~amlu/8037278/webrev.01
Thanks,
Amy
On 11/3/16 8:37 PM, Daniel Fuchs wrote:
Hi Amy,
I think it might still be useful to see what the subprocess
prints on System.out in case of failure.
I would leave lines 83, 86 and 87 unchanged.
best regards,
-- daniel
On 03/11/16 11:58, Amy Lu wrote:
Please review the patch for test
sun/rmi/runtime/Log/6409194/NoConsoleOutput.java
bug: https://bugs.openjdk.java.net/browse/JDK-8037278
webrev: http://cr.openjdk.java.net/~amlu/8037278/webrev.00/
This test fails whenever there’s extra output from System.out
(from the
test application DoRMIStuff). But actually this test is for
checking
RMI
logging messages and that are from System.err. Test should
ignore the
output from System.out.
Thanks,
Amy
