Displaying #maven-dev/2017-03-03.log:

Fri Mar 3 07:12:03 2017  Michael-O:Joined the channel
Fri Mar 3 09:37:37 2017  tibor_:Joined the channel
Fri Mar 3 10:22:00 2017  Michael-O:Joined the channel
Fri Mar 3 10:55:24 2017  rfscholte:Joined the channel
Fri Mar 3 12:29:30 2017  rfscholte:stephenc: I think the Maven commiter school deserves a new publication date, 2012 is pretty old...
Fri Mar 3 12:35:30 2017  stephenc:rfscholte: do you want to take a stab... maybe post it on the maven project site
Fri Mar 3 12:56:53 2017  rfscholte:stephenc: if you can create a logo of the Maven owl with a student cap :)
Fri Mar 3 13:00:55 2017  stephenc:rfscholte: ok
Fri Mar 3 13:14:21 2017  stephenc:rfscholte: https://www.dropbox.com/sh/3dtxpqr8dyvhzg7/AAAnVKhc81eiUndz212Q5mHHa?dl=0 let me know when I can delete it
Fri Mar 3 13:17:07 2017  rfscholte:stephenc: done
Fri Mar 3 15:29:31 2017  rfscholte:Joined the channel
Fri Mar 3 15:37:12 2017  theRealGent:Joined the channel
Fri Mar 3 21:08:32 2017  tibor_:Joined the channel
Fri Mar 3 21:21:32 2017  Michael-O:Joined the channel
Fri Mar 3 21:48:00 2017  tibor_:Michael-O: Hi Michael.
Fri Mar 3 21:48:10 2017  Michael-O:Hi tibor_
Fri Mar 3 21:48:14 2017  tibor_:Are we going to continue?
Fri Mar 3 21:50:24 2017  Michael-O:Yes, sure.
Fri Mar 3 21:50:42 2017  Michael-O:I was busy the last days. We have a several network failure at work.
Fri Mar 3 21:50:55 2017  Michael-O:I was almost 10 h at work and awfully tired.
Fri Mar 3 21:51:03 2017  Michael-O:had**
Fri Mar 3 21:51:44 2017  tibor_:I finished work at 7:30 cca.
Fri Mar 3 21:51:59 2017  tibor_:If you are tired we can postpone it.
Fri Mar 3 21:52:42 2017  Michael-O:I am not tired today
Fri Mar 3 21:52:44 2017  Michael-O:today is ok
Fri Mar 3 21:52:51 2017  Michael-O:yesterday was hard
Fri Mar 3 21:53:27 2017  tibor_:I will add logger and meanwhile we can uncomment parta of main method in ForkedBooter class.
Fri Mar 3 21:53:35 2017  tibor_:I will check yesterdays logs again
Fri Mar 3 21:54:54 2017  Michael-O:This is what I did. main() was empty
Fri Mar 3 22:02:58 2017  tibor_:This is what I see in the logs positive after you commented out the method:
Fri Mar 3 22:03:05 2017  tibor_:System.out.printf("%d %s before CLI execute
Fri Mar 3 22:03:11 2017  tibor_:final Process p = cl.execute();
Fri Mar 3 22:03:18 2017  tibor_:System.out.printf("%d %s after CLI execute
Fri Mar 3 22:03:32 2017  tibor_:new StreamFeeder(systemIn, p.getOutputStream())
Fri Mar 3 22:03:41 2017  tibor_:System.out.printf("%d %s after getOutputStream()
Fri Mar 3 22:03:52 2017  tibor_:So nothing hangs in both processes.
Fri Mar 3 22:04:33 2017  tibor_:After we uncomment the method we do not see System.out.printf("%d %s after CLI execute, within 20 sec.
Fri Mar 3 22:04:49 2017  Michael-O:So, the processes start normally, but streams are blocked?
Fri Mar 3 22:05:54 2017  tibor_:you mean in normal code when the method is not commented out?
Fri Mar 3 22:05:59 2017  Michael-O:Yes
Fri Mar 3 22:06:21 2017  tibor_:Streams should not be blocked in Utils, because cannot because the Process object is not available since
Fri Mar 3 22:06:23 2017  tibor_:final Process p = cl.execute();
Fri Mar 3 22:06:28 2017  tibor_:does not return
Fri Mar 3 22:06:41 2017  tibor_:and streams are called only on Process obj
Fri Mar 3 22:07:11 2017  tibor_:so since Process obj is not in memory Utils is not able to touch streams
Fri Mar 3 22:07:27 2017  tibor_:strange is not it?
Fri Mar 3 22:07:53 2017  tibor_:But here when you have empty main() method, the streams are available
Fri Mar 3 22:08:16 2017  tibor_:So it means something wrong is in ForkedBooter
Fri Mar 3 22:09:06 2017  tibor_:If we divide the main() method always in two part like, logarithmic division, we canfast find the part which caused haning.
Fri Mar 3 22:09:44 2017  Michael-O:likely. Are you sure that the way Utils creates the Process is sound? There is no bug burried?
Fri Mar 3 22:09:57 2017  Michael-O:Because the test repo you gave me does not use Utils
Fri Mar 3 22:10:25 2017  tibor_:We can integrate Utils in the test code if you want to.
Fri Mar 3 22:10:33 2017  tibor_:Should i?
Fri Mar 3 22:10:45 2017  tibor_:Will e more realistic.
Fri Mar 3 22:10:48 2017  tibor_:be
Fri Mar 3 22:10:59 2017  Michael-O:Yes, please.
Fri Mar 3 22:11:06 2017  tibor_:ok, I go for it.
Fri Mar 3 22:11:11 2017  Michael-O:Isolation is good, but not real in our case.
Fri Mar 3 22:11:24 2017  tibor_:yeah, I understand.
Fri Mar 3 22:11:48 2017  tibor_:I will try to have two forks.
Fri Mar 3 22:11:50 2017  tibor_:ok?
Fri Mar 3 22:12:06 2017  Michael-O:Absolutely
Fri Mar 3 22:13:59 2017  tibor_:In Utils we already use ProcessBuilder:
Fri Mar 3 22:14:05 2017  tibor_:ex = (new ProcessBuilder(this.getShellCommandline())).directory(workingDir);
Fri Mar 3 22:14:07 2017  tibor_:ex.environment().putAll(this.envVars);
Fri Mar 3 22:14:16 2017  tibor_:ex.start();
Fri Mar 3 22:14:26 2017  tibor_:Commandline.java
Fri Mar 3 22:14:41 2017  tibor_:Do you see this code correct?
Fri Mar 3 22:14:41 2017  Michael-O:Just checked. Better code than before.
Fri Mar 3 22:14:49 2017  tibor_:ok
Fri Mar 3 22:16:38 2017  tibor_:The same Commandline is used when running one fork which passed the tests. So the principle starts the JVMs same way.
Fri Mar 3 22:16:47 2017  Michael-O:Looks ok to me
Fri Mar 3 22:18:24 2017  tibor_:I think you are going to sleep soon. So i will wait for you on IRC tomorrow. I will be still up.
Fri Mar 3 22:21:13 2017  tibor_:Can I use Utils from this repo?
Fri Mar 3 22:21:43 2017  tibor_: /net/home/osipovmi/.m2/repository/
Fri Mar 3 22:23:47 2017  Michael-O:Joined the channel
Fri Mar 3 22:24:23 2017  Michael-O:my connection dropped
Fri Mar 3 22:41:51 2017  theRealGent:Joined the channel

Comments