Displaying #maven-dev/2017-02-21.log:

Tue Feb 21 12:46:32 2017  Michael-O:Joined the channel
Tue Feb 21 16:43:43 2017  tibor_:Joined the channel
Tue Feb 21 17:39:41 2017  rfscholte:Joined the channel
Tue Feb 21 18:07:05 2017  Michael-O:Joined the channel
Tue Feb 21 18:07:19 2017  Michael-O:tibor_: I have done further testing.
Tue Feb 21 18:07:55 2017  tibor_:I have also done testing with this:
Tue Feb 21 18:08:28 2017  tibor_:FileOutputStream out = new FileOutputStream( FileDescriptor.out );
Tue Feb 21 18:08:35 2017  tibor_:out.write( encodeBytes, 0, encodeBytes.length );
Tue Feb 21 18:08:41 2017  tibor_:out.getFD().sync();
Tue Feb 21 18:08:46 2017  tibor_:out.close();
Tue Feb 21 18:09:14 2017  tibor_:Close() caused java process hanging, but it was yesterday. I can try again
Tue Feb 21 18:09:46 2017  Michael-O:This is because STDOUT is out of your control, in my opinion
Tue Feb 21 18:12:30 2017  Michael-O:Have you even programmed with Servlets?
Tue Feb 21 18:21:49 2017  tibor_:With servlets, sure.
Tue Feb 21 18:23:57 2017  tibor_:ok, I have reverted out.close();
Tue Feb 21 18:26:43 2017  Michael-O:So probably know that the InputStream of the request as well as the OutputStream from the response shall not be closed. They are out of your control, but unter the container's one.
Tue Feb 21 18:27:37 2017  Michael-O:I think stdio is the same case
Tue Feb 21 18:29:37 2017  Michael-O:I have used git bisect to find the offending commit between 2.19.1 and master
Tue Feb 21 18:30:06 2017  Michael-O:There was one problematic commit. Unfortunately, reverting on top of master did not solve the problem.
Tue Feb 21 18:32:39 2017  tibor_:can you tell me the commit hash?
Tue Feb 21 18:34:56 2017  Michael-O:According to git bisect it is 9dd407e83
Tue Feb 21 18:35:16 2017  Michael-O:I tried forkmodeit and had much less failures before this commit
Tue Feb 21 18:35:31 2017  Michael-O:Now, I am doing a manual bisect going through the history.
Tue Feb 21 18:38:36 2017  rfscholte:Joined the channel
Tue Feb 21 18:38:56 2017  Michael-O:9dd4074e83
Tue Feb 21 18:39:08 2017  Michael-O:that's the one. Sorry
Tue Feb 21 18:42:18 2017  tibor_:but testing with 2.19.1, as you did with using 2.19.1 tag, shown us two test to fail with same message The forked VM terminated without properly saying goodbye.
Tue Feb 21 18:43:03 2017  Michael-O:Same message, but less failures
Tue Feb 21 18:43:19 2017  tibor_:the rootcause?
Tue Feb 21 18:44:02 2017  tibor_:did you try to remove clearCapture(); and test it?
Tue Feb 21 18:46:16 2017  tibor_:clearCapture(); closes file stream to which the strings received from forked std/out and err was redirected to.
Tue Feb 21 18:47:49 2017  tibor_:Usually in this multithreaded s/w the only thing which helped me always was plenty of logs on the console with timestamp
Tue Feb 21 18:48:58 2017  Michael-O:Yes, the commit before clearCapture worked for me.
Tue Feb 21 18:49:09 2017  Michael-O:I can bisect again and show you the previous commit git offered me.
Tue Feb 21 18:53:16 2017  tibor_:try with previous commit and upload the log. The point is that this method call was user's fix and we have to find root cause because this fix has a purpose.
Tue Feb 21 18:56:11 2017  Michael-O:I will. I have found another broken commit.
Tue Feb 21 18:56:22 2017  Michael-O:With this commit, ForkModeIT has three failures
Tue Feb 21 18:56:31 2017  Michael-O:The commit before produces only one failure.
Tue Feb 21 18:57:10 2017  Michael-O:e0bcffd05
Tue Feb 21 18:57:19 2017  Michael-O:this one broke ForkModeIT
Tue Feb 21 19:05:04 2017  tibor_:Let's summarize. 9dd4074e83 broke 3 tests in ForkedModeIT and all other hunderds tests passed?
Tue Feb 21 19:05:27 2017  tibor_:e0bcffd05 broke which tests?
Tue Feb 21 19:06:55 2017  Michael-O:Currently, I tried with ForkModeIT only.
Tue Feb 21 19:07:03 2017  Michael-O:I can rerun the entire suite tomorrow
Tue Feb 21 19:08:06 2017  tibor_:If you have logs from the branch latest commit and one IT, pls upload them.
Tue Feb 21 19:09:32 2017  tibor_:Now my parents are here, so I can check a potentional errors later and I can then bring some logging. Should we better switch to log4j and log to file system?
Tue Feb 21 19:09:50 2017  tibor_:or Java looger api
Tue Feb 21 19:11:03 2017  stephenc:rfscholte: Ping!
Tue Feb 21 19:12:24 2017  rfscholte:stephenc: pong?
Tue Feb 21 19:12:48 2017  Michael-O:tibor_: I will prepare all logs
Tue Feb 21 19:12:53 2017  Michael-O:slf4j simple is enough
Tue Feb 21 19:14:14 2017  stephenc:rfscholte: back
Tue Feb 21 19:14:45 2017  stephenc:rfscholte: did you get a chance to look at my fuller property fix
Tue Feb 21 19:15:33 2017  stephenc:The Windows tests are currently failing because infra updated some plugins so the Jenkinsfile builds get a rather long file path
Tue Feb 21 19:15:39 2017  rfscholte:stephenc: no, not yet. Let me do that right now
Tue Feb 21 19:16:12 2017  Michael-O:stephenc: I started bisecting Surefire master. It turns out that several commits broke some tests.
Tue Feb 21 19:16:23 2017  Michael-O:I am currently trying to the identify them for Tibor.
Tue Feb 21 19:16:27 2017  stephenc:The bit that "worries" me is extending Commons CLI's CommandLine
Tue Feb 21 19:17:00 2017  stephenc:Michael-O: I don't want to take away from the great work you are doing there
Tue Feb 21 19:17:33 2017  stephenc:My concern is getting an alpha of core that is *expected* to build projects the same as 3.3.9
Tue Feb 21 19:19:22 2017  stephenc:From that POV that 3.5.0-S and 3.3.9 were not building surefire the same... that needs fixing... but the surefire failures in and of themselves should not block a core release (unless they differ from on 3.3.9)
Tue Feb 21 19:20:03 2017  stephenc:is always on the side of a GAA pitch in the cold rain when on IRC for some reason!
Tue Feb 21 19:21:13 2017  stephenc:Michael-O: do you think there are any concerns with 3.5.0-S after I merge my system property fix?
Tue Feb 21 19:23:30 2017  Michael-O:stephenc: No, I don't think so. There are several other regressions between surefire-2.19.1 and surefire-master
Tue Feb 21 19:23:50 2017  Michael-O:Honestly, I will take several weeks to find all issues.
Tue Feb 21 19:24:02 2017  Michael-O:Surefire's master branch is messy.
Tue Feb 21 19:24:13 2017  Michael-O:Commit, revert, commit, revert. Very hard to read.
Tue Feb 21 19:24:20 2017  stephenc:Yes. The FreeBSD feature is nasty
Tue Feb 21 19:24:46 2017  stephenc:Can one run FreeBSD in docker?
Tue Feb 21 19:24:57 2017  Michael-O:docker is a VM
Tue Feb 21 19:25:00 2017  Michael-O:isn't
Tue Feb 21 19:25:15 2017  Michael-O:not like KVM, bhyve or VirtualBox
Tue Feb 21 19:25:33 2017  Michael-O:Docker is like Jails or Zones but on a more finegrained level
Tue Feb 21 19:25:46 2017  Michael-O:You are running on the same instace of OS
Tue Feb 21 19:26:02 2017  stephenc:Search on google reveals:
Tue Feb 21 19:26:31 2017  stephenc:You can run Linux docker images on freebsd because there is a Linux syscall adapter
Tue Feb 21 19:26:43 2017  stephenc:No shim currently to do the reverse
Tue Feb 21 19:27:06 2017  stephenc:So we'd need a vm
Tue Feb 21 19:27:08 2017  Michael-O:Yes, FreeBSD offers full binary compat for Linux fieles
Tue Feb 21 19:27:21 2017  Michael-O:You are talking about Jenkins, right?
Tue Feb 21 19:27:25 2017  stephenc:Yes
Tue Feb 21 19:27:31 2017  Michael-O:Yes, we need a dedicated FreeBSD VM for testing
Tue Feb 21 19:27:41 2017  Michael-O:At least for big projects
Tue Feb 21 19:27:44 2017  stephenc:Less work to get a docker image
Tue Feb 21 19:27:50 2017  Michael-O:Surefire, SCM, Release, Core
Tue Feb 21 19:28:32 2017  stephenc:Ok training is over... time to bring him home and thaw out
Tue Feb 21 19:28:42 2017  Michael-O:stephenc: how to initiate the VM with INFRA?
Tue Feb 21 19:30:21 2017  stephenc:Chat with infra on HipChat and ask them
Tue Feb 21 19:34:14 2017  Michael-O:Trying now...
Tue Feb 21 19:36:36 2017  tibor_:@stephenc: I am rather thinking about to finish what i planned in December and surefire release. I guess I will prepare extension in <configuration> in 2.20 and JDK9 support. The extension is programmatic and SOCKET PIPEs enabled by default.
Tue Feb 21 19:36:36 2017  rfscholte:stephenc: logic inside MergedCommandLine makes sense, just a shame that it must be solved like this
Tue Feb 21 19:36:53 2017  rfscholte:IIUC we should have commit rights on commons
Tue Feb 21 19:37:28 2017  rfscholte:so maybe spend some time fixing issues there
Tue Feb 21 19:37:39 2017  tibor_:@rfscholte: where is MergedCommandLine? What does it do?
Tue Feb 21 19:38:16 2017  tibor_:ah, you mean mvn core.
Tue Feb 21 19:38:30 2017  rfscholte::) yes
Tue Feb 21 19:38:40 2017  tibor_:i thought Utils
Tue Feb 21 19:39:07 2017  Michael-O:Joined the channel
Tue Feb 21 19:43:03 2017  stephenc:rfscholte: yes can probably fix in commons but I don't want to hold up 3.5.0-alpha-1 waiting for a new commons-CLI release ;-) hence package private class
Tue Feb 21 19:50:22 2017  Michael-O:stephenc: Just chatted with INFRA. Very bad news.
Tue Feb 21 19:50:38 2017  tibor_:listening..
Tue Feb 21 19:50:56 2017  stephenc:You were asking what you'd need to do to get one
Tue Feb 21 19:51:03 2017  Michael-O:They are consolidating all build nodes to Windows and Ubuntu.
Tue Feb 21 19:51:10 2017  stephenc:Not "give us a freebsd"
Tue Feb 21 19:51:11 2017  Michael-O:All other nodes are to be decomissioned.
Tue Feb 21 19:51:20 2017  stephenc:Hmmm
Tue Feb 21 19:51:31 2017  Michael-O:Several projects complained because they need FreeBSD or macOs
Tue Feb 21 19:51:46 2017  Michael-O:Hadoop, OpenOffice, Subversion. Mesos
Tue Feb 21 19:51:58 2017  stephenc:We should raise this as a PMC
Tue Feb 21 19:52:20 2017  stephenc:But from an ops POV I understand the why
Tue Feb 21 19:52:49 2017  Michael-O:Limited OS selected if lousy from a Jenkins point of view
Tue Feb 21 19:53:09 2017  Michael-O:is*
Tue Feb 21 19:53:28 2017  tibor_:FreeBSD is free of charge
Tue Feb 21 19:53:33 2017  Michael-O:You noticed yourself that the build fails on macOS which does not on Ubuntu
Tue Feb 21 19:53:40 2017  Michael-O:I do not call this reliable of portable
Tue Feb 21 19:54:30 2017  stephenc:Michael-O: we may have to set up our own Jenkins if infra cannot help
Tue Feb 21 19:54:52 2017  Michael-O:This is ridiculous.
Tue Feb 21 19:55:08 2017  Michael-O:All of my work machines are behind a VPN. No chance for a slave.
Tue Feb 21 19:55:11 2017  tibor_:@stephenc: send me logs from testing both surefire branches merged together
Tue Feb 21 19:55:24 2017  tibor_:I know there was one IT failed only on Mac
Tue Feb 21 19:55:38 2017  rfscholte:stephenc: I agree on this solution for 3.5.0
Tue Feb 21 19:55:38 2017  stephenc:tibor_: I think that was the system
Tue Feb 21 19:55:44 2017  stephenc:Property problem
Tue Feb 21 19:56:22 2017  tibor_:The question is the test result succesful or not?
Tue Feb 21 19:56:30 2017  Michael-O:stephenc: right, but this did not show up on Windows or Ubuntu.
Tue Feb 21 19:57:26 2017  stephenc:Ok. Though it passed on 3.3.9
Tue Feb 21 19:57:57 2017  tibor_:Thank you.
Tue Feb 21 19:58:01 2017  stephenc:So remains to be seen if it is a real issue after I merge my fix
Tue Feb 21 19:58:12 2017  tibor_:I will have a look in the tests failed on our Jenkins.
Tue Feb 21 19:58:27 2017  Michael-O:Given that Maven and Surefire is crucial for the entire Java community, I'd have expected more help from INFRA?
Tue Feb 21 20:01:58 2017  Michael-O:tibor_: I'd like to go offline now. I do some more bisecting tomorrow and provide you the tarballs.
Tue Feb 21 20:02:17 2017  tibor_:Michael-O: thx, See You.
Tue Feb 21 20:02:33 2017  Michael-O:stephenc: can you write a proposal for infra? I still convinced that we need more OS coverage.
Tue Feb 21 20:08:04 2017  tibor_:@stephenc: I see only old surefire builds in https://builds.apache.org/view/Maven/
Tue Feb 21 20:08:15 2017  tibor_:wasn't there new with Jenkinsfile?
Tue Feb 21 20:09:27 2017  tibor_:@stephenc: Did you write Java code with socket pipes for Win, Linux, Unix? Is the pipe path different on these platforms?
Tue Feb 21 20:22:42 2017  mbenson_:Joined the channel
Tue Feb 21 20:26:37 2017  stephenc:tibor_: no I think we have our wires crossed
Tue Feb 21 20:32:16 2017  tibor_:https://docs.oracle.com/javase/7/docs/api/java/nio/channels/Pipe.html#open%28%29
Tue Feb 21 20:32:17 2017  tibor_:https://docs.oracle.com/javase/7/docs/api/java/nio/channels/spi/SelectorProvider.html
Tue Feb 21 20:32:31 2017  tibor_:I am looking for elegant way with java pipes
Tue Feb 21 20:33:22 2017  stephenc:tibor_: my experience is with writing an NIO socket server/client using SSLEngine
Tue Feb 21 20:33:37 2017  tibor_:right
Tue Feb 21 20:33:54 2017  tibor_:I will google a bit
Tue Feb 21 20:33:57 2017  stephenc:The benefit of my experience is "oh my god do not use SSLEngine unless you have to"
Tue Feb 21 20:34:24 2017  stephenc:But you have no need for SSLEngine... and likely no need for NIO
Tue Feb 21 20:34:33 2017  tibor_:we do not have to, but datagram udp is fine or?
Tue Feb 21 20:34:55 2017  stephenc:What is your requirement?
Tue Feb 21 20:34:59 2017  tibor_:in local network bound to 127.0.0.1
Tue Feb 21 20:35:18 2017  stephenc:TCP is easier bound to local socket
Tue Feb 21 20:35:48 2017  stephenc:Have the protocol start with a random "secret"
Tue Feb 21 20:36:04 2017  tibor_:The ideal situation would be to bind to address: \\\\.pipes\\<hash>
Tue Feb 21 20:36:17 2017  tibor_:without knowing and reserving port
Tue Feb 21 20:36:33 2017  stephenc:You don't need to know or reserve a port
Tue Feb 21 20:36:45 2017  stephenc:Open a server socket on port 0
Tue Feb 21 20:36:54 2017  tibor_:that's my point
Tue Feb 21 20:37:01 2017  tibor_:but 0 is wildcard
Tue Feb 21 20:37:03 2017  stephenc:Then pass the port that gets bound to the forks
Tue Feb 21 20:37:21 2017  stephenc:0 means bind a random port that is free
Tue Feb 21 20:38:11 2017  tibor_:you mean new Socket(0) and observe local port?
Tue Feb 21 20:38:11 2017  stephenc:Then after binding the port you just pass that port to the forked process
Tue Feb 21 20:38:31 2017  tibor_:and pass local port to fork, right?
Tue Feb 21 20:38:32 2017  stephenc:New ServerSocket(0) and then ask what port it got
Tue Feb 21 20:38:38 2017  stephenc:Yep
Tue Feb 21 20:38:44 2017  tibor_:understand
Tue Feb 21 20:39:15 2017  stephenc:You could even pass the "secret" and port over stdin of the child
Tue Feb 21 20:39:52 2017  stephenc:That way the controller is sure each fork is associated with its client socket
Tue Feb 21 20:39:56 2017  tibor_:then we need to have an abstraction over Utils, because I want to keep std/out for backward compatibility and users choice
Tue Feb 21 20:40:32 2017  stephenc:Not quite
Tue Feb 21 20:40:48 2017  stephenc:Change the stdin protocols to allow tcp upgrade
Tue Feb 21 20:41:14 2017  tibor_:we must have bidirectional communication
Tue Feb 21 20:41:35 2017  stephenc:Send a tcp upgrade command to the fork, the tcp upgrade command is: 1. Address; 2. Port; 3. Hash
Tue Feb 21 20:42:17 2017  stephenc:Then once the socket is set up the fork closes stdout/stdin and replaces with the socket streams
Tue Feb 21 20:42:44 2017  tibor_:oh no
Tue Feb 21 20:42:54 2017  stephenc:What?
Tue Feb 21 20:43:10 2017  stephenc:With this way you keep all the stream code paths
Tue Feb 21 20:43:22 2017  stephenc:And you just have the tcp upgrade
Tue Feb 21 20:43:30 2017  tibor_:we pass properties file to target/surefire/ already now which has configuration for the fork
Tue Feb 21 20:43:39 2017  tibor_:thus
Tue Feb 21 20:43:44 2017  tibor_:before fork starts
Tue Feb 21 20:43:54 2017  tibor_:it configures itself
Tue Feb 21 20:43:55 2017  stephenc:Ok you can do the tcp upgrade via the properties as an alternative
Tue Feb 21 20:43:59 2017  tibor_:this is current style
Tue Feb 21 20:44:16 2017  tibor_:we do it with everyting
Tue Feb 21 20:44:21 2017  tibor_:you know
Tue Feb 21 20:44:24 2017  stephenc:I was assuming that it was setup via the streams
Tue Feb 21 20:44:57 2017  tibor_:it is hard to change all the objects and multithreaded staff in ForkStater meanwhile form runs
Tue Feb 21 20:45:11 2017  tibor_:meanwhile fork runs
Tue Feb 21 20:45:42 2017  stephenc:Yes if you have properties file no need for "upgrade" command
Tue Feb 21 20:45:59 2017  stephenc:You can just start in tcp or stream mode
Tue Feb 21 20:58:26 2017  rfscholte:tibor_:
Tue Feb 21 20:58:35 2017  rfscholte:Failed tests: invokeRuntimeException(org.apache.maven.surefire.its.jiras.Surefire141PluggableProvidersIT): expecting non-empty, but it was empty
Tue Feb 21 20:58:35 2017  rfscholte: invokeReporterException(org.apache.maven.surefire.its.jiras.Surefire141PluggableProvidersIT): expecting non-empty, but it was empty
Tue Feb 21 20:58:35 2017  rfscholte: shouldNotBeVerbose(org.apache.maven.surefire.its.CheckTestNgReportTestIT): log pattern does not match nTimes(..)
Tue Feb 21 20:58:35 2017  rfscholte:Tests run: 714, Failures: 3, Errors: 0, Skipped: 136
Tue Feb 21 20:58:59 2017  rfscholte:are these related to m-clean-p?
Tue Feb 21 21:02:36 2017  tibor_:this i was looking for
Tue Feb 21 21:02:48 2017  tibor_:I could not find the link
Tue Feb 21 21:03:15 2017  tibor_:can be related to clean, but i need to see the logs
Tue Feb 21 21:03:31 2017  tibor_:target/surefire-reports folder is deleted
Tue Feb 21 21:04:01 2017  tibor_:this can be the reason why log pattern does not match.. Maybe.
Tue Feb 21 21:04:45 2017  tibor_:@stephenc: Michael sent me a hint for implementing pipes http://carminedimascio.com/2014/01/named-pipes-with-java/
Tue Feb 21 21:04:57 2017  tibor_:I will check it out on Ubuntu
Tue Feb 21 21:06:54 2017  rfscholte:tibor_: it's in your mailbox, zip with these 3 test results
Tue Feb 21 21:07:25 2017  tibor_:yes
Tue Feb 21 21:09:58 2017  tibor_:https://builds.apache.org/view/Maven%20Master%20Release%20Status/job/maven-master-release-status-test-surefire-linux/32/console
Tue Feb 21 21:14:35 2017  rfscholte:[INFO] Results:
Tue Feb 21 21:14:35 2017  rfscholte:[INFO]
Tue Feb 21 21:14:35 2017  rfscholte:[INFO] Tests run: 2, Failures: 0, Errors: 0, Skipped: 0
Tue Feb 21 21:14:35 2017  rfscholte:[INFO]
Tue Feb 21 21:14:35 2017  rfscholte:[INFO] ------------------------------------------------------------------------
Tue Feb 21 21:14:35 2017  rfscholte:[INFO] BUILD FAILURE
Tue Feb 21 21:14:46 2017  rfscholte:?!?! why
Tue Feb 21 21:15:23 2017  tibor_:how can i know?
Tue Feb 21 21:15:58 2017  rfscholte:add debug?
Tue Feb 21 21:16:17 2017  tibor_:any error behind it?
Tue Feb 21 21:16:44 2017  tibor_:e.g.
Tue Feb 21 21:16:45 2017  tibor_:Process Exit Code: 1 ExecutionException The forked VM terminated without properly saying goodbye. VM crash or System.exit called?
Tue Feb 21 21:16:58 2017  rfscholte:[ERROR] There was an error in the forked process
Tue Feb 21 21:16:58 2017  rfscholte:[ERROR] Let's fail with a reporterexception
Tue Feb 21 21:17:28 2017  tibor_:this is expected
Tue Feb 21 21:17:36 2017  tibor_:this test crashes fork
Tue Feb 21 21:17:36 2017  rfscholte:final cause: Caused by: org.apache.maven.plugin.MojoExecutionException: There are test failures.
Tue Feb 21 21:17:39 2017  tibor_:intended
Tue Feb 21 21:17:43 2017  rfscholte:ok
Tue Feb 21 21:18:08 2017  tibor_:sometimes we implement dummy Surefire Runner and try to crash it
Tue Feb 21 21:18:31 2017  tibor_:sometimes tests want to break s/w more drastically
Tue Feb 21 21:20:04 2017  tibor_:I will download Workspace, the main log is usually useless.
Tue Feb 21 21:50:19 2017  tibor_:@stephenc: this is interesting in NIO
Tue Feb 21 21:50:23 2017  tibor_:https://docs.oracle.com/javase/7/docs/api/java/nio/channels/spi/SelectorProvider.html#inheritedChannel%28%29
Tue Feb 21 21:50:39 2017  tibor_:On many operating systems a process, such as a Java virtual machine, can be started in a manner that allows the process to inherit a channel from the entity that created the process.
Tue Feb 21 21:51:49 2017  tibor_:If it is ServerSocketChannel or DatagramChannel ten it is socket based without configuring fork. WDYT?
Tue Feb 21 21:51:56 2017  tibor_:then
Tue Feb 21 22:02:44 2017  tibor_:@rfscholte: what a hell is going on the Jenkins
Tue Feb 21 22:03:52 2017  stephenc:tibor_: if it's a multibranch Jenkinsfile then Windows tests are failing due to 256 character path limit
Tue Feb 21 22:03:58 2017  tibor_:Surefire141PluggableProvidersIT_invokeRuntimeException directory does not exist
Tue Feb 21 22:04:09 2017  tibor_:aha
Tue Feb 21 22:04:24 2017  tibor_:but this is Ubuntu
Tue Feb 21 22:04:37 2017  tibor_:https://builds.apache.org/view/Maven%20Master%20Release%20Status/job/maven-master-release-status-test-surefire-linux/ws/surefire-master/
Tue Feb 21 22:04:43 2017  stephenc:I need to get infra to set a magic property so that Windows builds work again
Tue Feb 21 22:05:01 2017  tibor_:but this link is not Windows
Tue Feb 21 22:05:05 2017  stephenc:tibor_: that is not multibranch
Tue Feb 21 22:05:34 2017  tibor_:this view is not multibranch?
Tue Feb 21 22:05:34 2017  stephenc:So the path length is only affecting the Maven-3.x-jenkinsfile project at present
Tue Feb 21 22:06:04 2017  tibor_:aha ok
Tue Feb 21 22:06:09 2017  stephenc:Windows is only magic
Tue Feb 21 22:06:31 2017  tibor_:so then I will check the next two tests if it is the same issue
Tue Feb 21 22:06:40 2017  stephenc:Some apis have path limit of 253 others 2045
Tue Feb 21 22:07:13 2017  tibor_:yes, I would by myself with my company projects that Windows cutted path
Tue Feb 21 22:07:21 2017  tibor_:found
Tue Feb 21 22:07:26 2017  tibor_:my typos
Tue Feb 21 22:07:36 2017  tibor_:cut
Tue Feb 21 22:07:37 2017  tibor_::)
Tue Feb 21 22:09:26 2017  tibor_:@stephen: you are natively speaking English?
Tue Feb 21 22:16:28 2017  stephenc:Irish people speak English better than the English themselves
Tue Feb 21 22:17:31 2017  stephenc:James Joyce and Oscar Wilde were Irish
Tue Feb 21 22:17:41 2017  stephenc:;-)
Tue Feb 21 22:18:18 2017  stephenc:tibor_: it's: "are you a native English speaker?" Btw
Tue Feb 21 22:18:40 2017  tibor_::)
Tue Feb 21 22:18:51 2017  tibor_:ok, and where are you from?
Tue Feb 21 22:19:03 2017  stephenc:Ireland
Tue Feb 21 22:19:05 2017  tibor_:Ireland
Tue Feb 21 22:19:09 2017  tibor_:ah
Tue Feb 21 22:19:15 2017  tibor_:i tought
Tue Feb 21 22:19:24 2017  tibor_:maybe you told me before
Tue Feb 21 22:19:51 2017  tibor_:but your people don't speak official English, right?
Tue Feb 21 22:20:27 2017  tibor_:some dialect or
Tue Feb 21 22:21:09 2017  stephenc:No dialect. Different accent
Tue Feb 21 22:22:25 2017  stephenc:We spell colour the correct way... not the way those crazy Americans do
Tue Feb 21 22:22:43 2017  stephenc:We also know how to spell aluminium
Tue Feb 21 22:23:37 2017  stephenc:The only main difference is that we have been known to inject a present continuous tense (because Irish has one)
Tue Feb 21 22:24:23 2017  stephenc:tibor_: so you just have to watch out for Irish people using "does be" as the present continuous tense
Tue Feb 21 22:25:30 2017  stephenc:There is also "youse" being the second person plural... but it is frowned upon by civilised soldiery
Tue Feb 21 22:26:57 2017  tibor_:I am just watching a video with Irish speaker comparing his language with official English
Tue Feb 21 22:27:29 2017  tibor_:I would have problem to understand you first time.
Tue Feb 21 22:34:30 2017  stephenc:That is accent. Same words. Same grammar. Try listening to somebody from Scotland or wales or Lancashire
Tue Feb 21 22:34:46 2017  stephenc:Or the scouse accent
Tue Feb 21 22:35:05 2017  stephenc:If you want your mind blown...
Tue Feb 21 22:35:18 2017  stephenc:Look up Rab C Nesbit on YouTube
Tue Feb 21 22:36:17 2017  stephenc:When I first saw that on telly, I thought that the "joke" was that rab speaks unintelligible gibberish while everyone else has a scotch accent
Tue Feb 21 22:37:05 2017  stephenc:After I'd worked with a guy from Glasgow for a while I noticed something. Rab is not speaking gibberish... he actually gets almost all the best lines
Tue Feb 21 22:37:49 2017  stephenc:So accent is something strange like that
Tue Feb 21 22:38:48 2017  stephenc:The next is that there are many accents in Ireland. The there are at least 3 distinct accents in Dublin alone!
Tue Feb 21 22:39:38 2017  stephenc:The accents get stronger as you move away from Dublin... so the Kerry and Northern Irish accents would be very strong
Tue Feb 21 22:41:53 2017  stephenc:For an example of Kerry accent, YouTube Jackie Healy-Rae (an unusual person with controversial opinions)
Tue Feb 21 22:42:47 2017  stephenc:For an example of a Northern Irish accent, YouTube Gerry Adams
Tue Feb 21 22:43:29 2017  tibor_:Rab C Nesbit :) completely different language for me :)
Tue Feb 21 22:43:37 2017  stephenc:Even there there are many different sub-accents, so these are not typical... just giving the range
Tue Feb 21 22:43:49 2017  tibor_:yeah, i see
Tue Feb 21 22:44:11 2017  stephenc:Rab C Nesbit *is speaking English* and I hear it as English too!
Tue Feb 21 22:44:27 2017  stephenc:But just there is a heavy scotch accent
Tue Feb 21 22:44:29 2017  tibor_::) well
Tue Feb 21 22:44:39 2017  stephenc:But I used to think it was gibberish!
Tue Feb 21 22:45:25 2017  tibor_::) gibberish English
Tue Feb 21 22:46:58 2017  stephenc:No at first I didn't even think it was anything being said
Tue Feb 21 22:47:18 2017  stephenc:Took a while for my ear to learn the sounds
Tue Feb 21 22:47:34 2017  stephenc:Once you start to hear the sounds it all works itself out
Tue Feb 21 22:47:37 2017  stephenc:Just nuts
Tue Feb 21 23:20:36 2017  tibor_:@stephenc: There is one problem with sys prop neverless i fixed with the workaround in MavenLauncher
Tue Feb 21 23:20:37 2017  tibor_:shouldNotBeVerbose(org.apache.maven.surefire.its.CheckTestNgReportTestIT)
Tue Feb 21 23:20:47 2017  tibor_:<property name="testng.version" value="5.7"/>
Tue Feb 21 23:20:58 2017  tibor_:should be 5.10
Tue Feb 21 23:23:37 2017  tibor_:my bad - master was not merged from the surefire branch
Tue Feb 21 23:23:43 2017  tibor_:just fine

Comments