Displaying #maven-dev/2017-02-17.log:

Fri Feb 17 01:29:58 2017  hboutemy:Joined the channel
Fri Feb 17 08:17:54 2017  olamy:Joined the channel
Fri Feb 17 09:25:50 2017  Michael-O:Joined the channel
Fri Feb 17 12:14:28 2017  Michael-O:Joined the channel
Fri Feb 17 12:18:16 2017  tibor_:Joined the channel
Fri Feb 17 12:19:34 2017  Michael-O:tibor_: I have the feeling that MCLEAN could also be affected by Maven Shared Utils issue
Fri Feb 17 12:20:20 2017  tibor_:Michael-O: so can you make install plugin locally and trigger a build?
Fri Feb 17 12:20:46 2017  Michael-O:I did this wih MCLEAN-3.0.1-SNAPSHOT, same issue.
Fri Feb 17 12:20:56 2017  Michael-O:I can install any plugin you want from the source repo
Fri Feb 17 12:21:26 2017  tibor_:can you send the root cause in ML. Maybe everybody want's to know it.
Fri Feb 17 12:22:40 2017  Michael-O:Yes, I wanted to prepare our finds as two mails: one for the surefire issues and one with MCLEAN
Fri Feb 17 12:22:52 2017  tibor_:very good
Fri Feb 17 12:23:58 2017  Michael-O:paired with the log files
Fri Feb 17 12:24:38 2017  Michael-O:what are you currently working on? Surefire or Utils?
Fri Feb 17 12:30:39 2017  tibor_:I am in daily job right now
Fri Feb 17 12:30:53 2017  Michael-O:Me too
Fri Feb 17 12:43:05 2017  tibor_:Michael-O: I will work on Utils as soon as I have freedom from my job.
Fri Feb 17 12:49:15 2017  Michael-O:tibor_: no rush, everything is fine. We need to create the tickets first, then fix code
Fri Feb 17 12:59:06 2017  rfscholte:Joined the channel
Fri Feb 17 15:32:24 2017  Michael-O:tibor_: I'll be later online
Fri Feb 17 16:46:07 2017  Spockz:Joined the channel
Fri Feb 17 16:46:41 2017  Spockz:I just asked this in the user channel but perhaps this channel is more suited
Fri Feb 17 16:46:41 2017  Spockz:I'm trying to get a complete project/model of a pom.xml from my own java application. But the DefaultProjectBuilder and such all require the Plexus container to be setup: a) how can I do it without needing the DI from Plexus or b) how do I setup the container properly
Fri Feb 17 17:16:45 2017  rfscholte:Spockz: https://www.eclipse.org/sisu/docs/api/org.eclipse.sisu.plexus/reference/packages.html
Fri Feb 17 17:17:10 2017  rfscholte:combined with https://eclipse.org/sisu/docs/tutorials/index.html
Fri Feb 17 18:17:10 2017  Michael-O:Joined the channel
Fri Feb 17 18:56:05 2017  Spockz:@rfscholte: thx
Fri Feb 17 19:00:50 2017  Spockz:rfscholte: https://gist.github.com/spockz/6bbb08d18345a0d7b40861a51d0d01c0 << that is basically what I had been doing
Fri Feb 17 19:09:29 2017  rfscholte:Spockz: should work, what if you add "default" as hint?
Fri Feb 17 19:39:39 2017  tibor_:Joined the channel
Fri Feb 17 19:42:20 2017  Spockz:rfscholte: no, no score
Fri Feb 17 19:53:31 2017  Spockz:do I need to scan a package?
Fri Feb 17 20:23:46 2017  Michael-O:Joined the channel
Fri Feb 17 20:49:08 2017  Michael-O:tibor_: mail sent to the ML, have a look
Fri Feb 17 20:56:47 2017  tibor_:yes, I see
Fri Feb 17 20:57:54 2017  tibor_:offtopic: I wam looking for a good book of Web Design for usiness Applications. Do you have hints?
Fri Feb 17 20:58:15 2017  tibor_:for Business Applications
Fri Feb 17 20:59:13 2017  tibor_:Uninitialized/too early attributes made daemon threads to kill forked VMs
Fri Feb 17 20:59:37 2017  tibor_:This cannot be resolved so I decided not to use the timeout in Thread name
Fri Feb 17 21:02:13 2017  Michael-O:we can swap lines and we will have the default value, shouldn't that do it?
Fri Feb 17 21:02:28 2017  Michael-O:We can at least file an issue about it
Fri Feb 17 21:12:36 2017  tibor_:Michael-O: swap lines? I don't know what you are talking about. The email?
Fri Feb 17 21:13:06 2017  tibor_:the variables?
Fri Feb 17 21:13:52 2017  Michael-O:jvmTerminator(), the variable is initalized after the static call of jvmTermniator()
Fri Feb 17 21:15:28 2017  tibor_:let me see, because it was not that straightforward when i saw it in the morning
Fri Feb 17 21:15:39 2017  Michael-O:you remember the 0s thread?
Fri Feb 17 21:15:51 2017  Michael-O:it was supposed to be 30s or so
Fri Feb 17 21:15:51 2017  tibor_:yes, i know that problem
Fri Feb 17 21:18:06 2017  tibor_:JVM_TERMINATOR has name like last-ditch-daemon-shutdown-thread-30s
Fri Feb 17 21:18:41 2017  tibor_:but this is wrong since the scheduler occupied
Fri Feb 17 21:18:48 2017  tibor_:two threads
Fri Feb 17 21:18:54 2017  tibor_:one for shutdown
Fri Feb 17 21:19:02 2017  tibor_:next for pinging forked jvm
Fri Feb 17 21:19:10 2017  tibor_:better would be to have two
Fri Feb 17 21:19:17 2017  tibor_:instances
Fri Feb 17 21:19:32 2017  tibor_:each another clear name
Fri Feb 17 21:20:19 2017  tibor_:what is wrong is the capacity of the pool = 1
Fri Feb 17 21:20:44 2017  tibor_:there are two threads, however one starts after another, but it is more messy to me
Fri Feb 17 21:24:06 2017  tibor_:and the Thread "last-ditch-daemon-shutdown-thread-30s" does not need to be timer nothing but a normal Thread running a Runnable just once
Fri Feb 17 21:24:55 2017  tibor_:and then we might be able to have timeout in the name of thread, because the exit() is called at the exit point of application
Fri Feb 17 21:25:24 2017  tibor_:until that time the timeout will be set properly or by default value
Fri Feb 17 21:25:42 2017  tibor_:Michael-O: are you here?
Fri Feb 17 21:26:23 2017  Michael-O:yes, reading and processing..
Fri Feb 17 21:28:39 2017  tibor_:what about to do something with Utils now/
Fri Feb 17 21:28:54 2017  tibor_:Vote of Utils takes time
Fri Feb 17 21:29:18 2017  Michael-O:I need to verify the patch for utils first
Fri Feb 17 21:29:23 2017  Michael-O:if this works, we can cast a vote
Fri Feb 17 21:29:27 2017  tibor_:of course
Fri Feb 17 21:29:43 2017  Michael-O:if everything is fine, we can have the release in less than a week
Fri Feb 17 21:30:07 2017  tibor_:e.g. svn co .. , edit, install, change dependency version in surefire and run a build
Fri Feb 17 21:30:36 2017  tibor_:I think only you can do that because you have the machine
Fri Feb 17 21:30:58 2017  tibor_:and better not to commit unless we know it is fine
Fri Feb 17 21:31:15 2017  tibor_:WDYT?
Fri Feb 17 21:32:04 2017  tibor_:I did not imaging there would be so much work with Surefire, and I tought just one issue , I was blind :)
Fri Feb 17 21:33:43 2017  Michael-O:Surefire is extremely complex, therefore errorprone
Fri Feb 17 21:33:50 2017  tibor_::))
Fri Feb 17 21:33:54 2017  tibor_:I know, I know
Fri Feb 17 21:34:04 2017  Michael-O:You can commit a branch in the Subversion repo for maven-shared-utils with the fix you image to be
Fri Feb 17 21:34:08 2017  tibor_:I am working in Surefire since 2013
Fri Feb 17 21:34:09 2017  Michael-O:I will check out and build locally
Fri Feb 17 21:34:35 2017  tibor_:should I make the branch like this?
Fri Feb 17 21:34:46 2017  tibor_:mvn release:branch -Dbranch=...
Fri Feb 17 21:35:09 2017  Michael-O:Yes, try it.
Fri Feb 17 21:35:10 2017  tibor_:are we doing it always like this in ASF?
Fri Feb 17 21:35:15 2017  tibor_:ok
Fri Feb 17 21:35:19 2017  Michael-O:I do that at work :-d
Fri Feb 17 21:35:41 2017  tibor_:I go for it..
Fri Feb 17 21:35:48 2017  Michael-O:It is not necessary for a short-lived branch, but good for a maintenance one
Fri Feb 17 21:35:58 2017  tibor_:sure
Fri Feb 17 21:43:46 2017  Michael-O:Btw, unmodified Surefire master passes all tests with Maven master on RHEL6
Fri Feb 17 21:47:13 2017  tibor_:but that's Fedora and amd64, or?
Fri Feb 17 21:47:34 2017  Michael-O:Yes, RHEL is based on Fedora, but RHEL6 is stoneage, still 2.6 kernel
Fri Feb 17 21:47:40 2017  Michael-O:on 12-core Xeon in my case
Fri Feb 17 21:47:53 2017  Michael-O:it runs Oracle and other software
Fri Feb 17 21:48:51 2017  tibor_:flushing of the stream might be different, but I still see issue in our code when missed flush(). Since there are buffers in JRE.
Fri Feb 17 21:50:52 2017  tibor_:Anyway we cannot change FreeBSD but we can change our code
Fri Feb 17 21:52:28 2017  Michael-O:And we shouldn't. We should never rely on cyclic flushes from the OS.
Fri Feb 17 22:56:24 2017  Michael-O:tibor_: I am turning. Just ping me if you think you have something testable
Fri Feb 17 22:57:26 2017  tibor_:I am trying release:branch in dry run mode
Fri Feb 17 22:58:00 2017  Michael-O:ok
Fri Feb 17 22:58:11 2017  Michael-O:It should actually be prettry straight forward
Fri Feb 17 22:58:35 2017  tibor_:I am in tags/maven-shared-utils-0.9
Fri Feb 17 22:58:51 2017  tibor_:mvn release:branch -Dusername=tibordigana -Dpassword=Robit001 -DbranchName=maven-shared-utils-0.9.1 -DupdateBranchVersions=true -Dtag=maven-shared-utils-0.9 -DupdateWorkingCopyVersions=false -DdryRun=true
Fri Feb 17 22:58:55 2017  tibor_:ok?
Fri Feb 17 22:59:01 2017  tibor_:Jesus
Fri Feb 17 22:59:03 2017  tibor_:pswd
Fri Feb 17 22:59:07 2017  tibor_::)
Fri Feb 17 22:59:20 2017  tibor_:can be?
Fri Feb 17 23:01:22 2017  Michael-O:username/pass isn't necessary if subversion has already stored it. Tag isn't necessary. branch should be maven-shared-utils-0.9.x
Fri Feb 17 23:02:22 2017  tibor_:mvn release:branch -Dusername=tibordigana -Dpassword=*** -DbranchName=maven-shared-utils-0.9.x -DupdateBranchVersions=true -DupdateWorkingCopyVersions=false -DdryRun=true
Fri Feb 17 23:02:44 2017  Michael-O:Yep
Fri Feb 17 23:02:46 2017  Michael-O:looks good to me
Fri Feb 17 23:02:47 2017  tibor_:i will run it

Comments