Staffan, No it's not to your code. Sorry for not being clean enough.
You cleaned up unix code, but windows code remains bad. Particularly, it has bad unconditional sleep 2 I would propose replace *windows manipulation with CYGWIN/MKS pids* to call to JPS that return windows pid -Dmitry On 2014-01-30 20:53, Staffan Larsen wrote: > > On 30 jan 2014, at 17:35, Dmitry Samersoff <dmitry.samers...@oracle.com> > wrote: > >> Staffan, >> >> Looks good for me except windows part. > > Can you explain? Do any of my changes make things worse on windows? > >> >> Can you replace all this staff with call to ups ? > > I don’t think so. I need to know that the process has started _and_ reached a > certain point in the program. > > /Staffan > >> >> -Dmitry >> >> >> On 2014-01-30 20:11, Staffan Larsen wrote: >>> webrev: http://cr.openjdk.java.net/~sla/8029808/webrev.00/ >>> bug: https://bugs.openjdk.java.net/browse/JDK-8029808 >>> >>> I have tried to improve the stability of this test in a couple of ways. >>> >>> - Timeout set to 120s - it should never need to run that long >>> - When we start the debugee, we wait for the port-file to become available >>> instead of waiting for output from the process. This should be less racy. >>> - stderr output from the debuggee and ShutdownDebuggee programs are >>> captured with the stdout output >>> - Moved “sleep 1” to after the check for the port file so the test is >>> faster if the file already exists >>> >>> I’m not sure any of these are the cause of the timeout, but at least it >>> should be a bit more stable now. >>> >>> /Staffan >>> >> >> >> -- >> Dmitry Samersoff >> Oracle Java development team, Saint Petersburg, Russia >> * I would love to change the world, but they won't give me the sources. > -- Dmitry Samersoff Oracle Java development team, Saint Petersburg, Russia * I would love to change the world, but they won't give me the sources.
