Hi Igor,
The two definitions below look almost the same! :)
+NSK_SHARE_JDI_INCLUDES := \
+-I$(VM_TESTBASE_DIR)/nsk/share/native \
+-I$(VM_TESTBASE_DIR)/nsk/share/jdi
+
+NSK_SHARE_JNI_INCLUDES := \
+-I$(VM_TESTBASE_DIR)/nsk/share/native \
+-I$(VM_TESTBAS
Hi Yasumasa,
I just noticed that GcTest01.java and GcCauseTest03.java have also
failed for this reason. I see 9 total failures between the 3 tests so far.
On 5/4/18 6:17 AM, Yasumasa Suenaga wrote:
Hi Chris,
Thank you for your evaluation !
It is very helpful for me.
(I've not reproduced this
Hi Christoph,
It looks like for bsd this code was added to fix JDK-7152800. In that CR
I see the following:
"The attach framework will verify that the file has the same effective
owner and group as the currently running process. This will be true on
linux, since files are created with the ef
Hello,
from the description below it sounds it would also be possible to remove the
Group check. Would this not be an Option which more flexible allows sgid to be
used as intended? (Not that I can imagine anybody Setting a sgid on /tmp?!)
What is the Purpose of validating the Group ownership?
Build change looks good.
/Erik
On 2018-05-03 21:14, Igor Ignatyev wrote:
http://cr.openjdk.java.net/~iignatyev/8199382/webrev.00/index.html
577169 lines changed: 577169 ins; 0 del; 0 mod;
Hi all,
could you please review the patch which open sources JDI tests from vm
testbase? These tests c
Hi,
please review a change for correctly setting the group for the attach listener
file:
Webrev: http://cr.openjdk.java.net/~clanger/webrevs/8202650.0/
Bug: https://bugs.openjdk.java.net/browse/JDK-8202650
The attach listener file, usually /tmp/.java_pid is created from the
hotspot JVM process
Hi Chris,
Thank you for your evaluation !
It is very helpful for me.
(I've not reproduced this issue on linux-x64 !)
It's just very intermittent.
I do not yet decided how do we fix this problem.
IMHO we can add fallback code to get raw PerfCounter data through `jcmd
PerfCounter.print`.
But
