Displaying #maven-dev/2017-02-18.log:

Sat Feb 18 06:21:42 2017  hboutemy:Joined the channel
Sat Feb 18 08:08:16 2017  hboutemy:tibor_: Hi
Sat Feb 18 08:08:24 2017  hboutemy:you know, an alpha is an alpha
Sat Feb 18 08:08:53 2017  hboutemy:people reporting for an alpha should be more manageable than people reporting for a final release
Sat Feb 18 08:08:58 2017  tibor_:MCLEAN should be fixed, and Karl made a workaround
Sat Feb 18 08:09:10 2017  tibor_:it seems multithreading issue
Sat Feb 18 08:09:22 2017  hboutemy:oh, MCLEAN and multi-threading
Sat Feb 18 08:09:24 2017  tibor_:if contains() is placed as a workaround
Sat Feb 18 08:09:40 2017  hboutemy:= once again very advanced
Sat Feb 18 08:09:54 2017  tibor_:then what should I think of the logic behind it unless I see Jira issue
Sat Feb 18 08:10:00 2017  hboutemy:and why should it postpone 3.5.0-alpha?
Sat Feb 18 08:10:39 2017  hboutemy:with an alpha, we can even tell we have known issues
Sat Feb 18 08:10:40 2017  tibor_:because useing ConcurrentHashMap instead of Map is simple replacement too
Sat Feb 18 08:10:50 2017  hboutemy:too technical answer
Sat Feb 18 08:11:21 2017  hboutemy:do you fear that there will be other plugins that will be affected?
Sat Feb 18 08:11:21 2017  tibor_:very hard to maintaine the sw having workarounds
Sat Feb 18 08:11:46 2017  hboutemy:I don't understand which sw you're talking about: Maven core or m-clean-p?
Sat Feb 18 08:11:49 2017  tibor_:worse than bad later, you know, too costly later
Sat Feb 18 08:11:58 2017  tibor_:Maven
Sat Feb 18 08:12:14 2017  tibor_:I will send you a link, wait pls
Sat Feb 18 08:12:21 2017  hboutemy:once again, what you tell me is valid for a full release, not an alpha
Sat Feb 18 08:12:35 2017  tibor_:https://git1-us-west.apache.org/repos/asf?p=maven.git;a=commitdiff;h=238f789ad07a18c0dff7f884c9a5d3b47258fc47
Sat Feb 18 08:12:55 2017  tibor_:yes, but one day +/- is not too long
Sat Feb 18 08:13:23 2017  tibor_:how much time you think Karl needs to fix MClean issue?
Sat Feb 18 08:13:40 2017  tibor_:The one Surefire772*IT fails with
Sat Feb 18 08:13:46 2017  hboutemy:why do you switch from Maven core code to m-clean-p ?
Sat Feb 18 08:14:19 2017  tibor_:because the version can be included in default binding
Sat Feb 18 08:14:21 2017  hboutemy:and why MNG-6170 which has been there since 3.1 is now a blocker for 3.5?
Sat Feb 18 08:14:36 2017  hboutemy:ah , default bindings are shit
Sat Feb 18 08:15:17 2017  hboutemy:then people doing advanced use (ie threading) using default bindings should be considered crazy
Sat Feb 18 08:15:21 2017  tibor_:no idea, we blamed surefire as blocker of Maven 3.5
Sat Feb 18 08:15:47 2017  hboutemy:if surefire does not work with 3.5, we'll have many reports
Sat Feb 18 08:16:06 2017  tibor_:itsnot it same with alpha?
Sat Feb 18 08:16:14 2017  hboutemy:I mean existing surefire releases, in conditions that work with previous Maven releases
Sat Feb 18 08:16:19 2017  hboutemy:yes, sure
Sat Feb 18 08:16:54 2017  hboutemy:that's just that clean is easy, surefire is harder since it runs user code and so on
Sat Feb 18 08:17:15 2017  hboutemy:but for sure, if the level of confidence is not too low, an alpha is an alpha
Sat Feb 18 08:17:52 2017  hboutemy:and when we talk about Surefire, are we talking about existing releases, or about latest SNAPSHOT?
Sat Feb 18 08:17:53 2017  tibor_:tell me, are the users requesting a new version?
Sat Feb 18 08:18:01 2017  hboutemy:of Maven?
Sat Feb 18 08:18:06 2017  tibor_:yes Maven
Sat Feb 18 08:18:31 2017  hboutemy:there is no users pressure: just we must prove that we are going forward
Sat Feb 18 08:18:46 2017  hboutemy:and currently, I'm not confident
Sat Feb 18 08:18:58 2017  tibor_:is Maven stable enough to release?
Sat Feb 18 08:19:06 2017  hboutemy:that's the key question
Sat Feb 18 08:19:32 2017  hboutemy:to me, yes, but some say no without explaining me what is the level of issues
Sat Feb 18 08:19:42 2017  hboutemy:"there are issueds"
Sat Feb 18 08:19:46 2017  tibor_:what if we uncover more bugs, we are only few active devs and you are active dev
Sat Feb 18 08:19:56 2017  hboutemy:for sure, there are always issues: nothing will never be perfect
Sat Feb 18 08:20:32 2017  hboutemy:alpha status is the answer to worried people
Sat Feb 18 08:20:48 2017  tibor_:what if the vore has one -1?
Sat Feb 18 08:20:52 2017  hboutemy:I respect worried people (or I would fight against alpha status)
Sat Feb 18 08:21:10 2017  hboutemy:for an alpha, we'll have more +1 than -1
Sat Feb 18 08:21:10 2017  tibor_:is -1 blocker?
Sat Feb 18 08:21:13 2017  hboutemy:no
Sat Feb 18 08:21:34 2017  hboutemy:because I'm confident quality is good: it is usable for most users
Sat Feb 18 08:21:36 2017  tibor_:I will change my mind if it is blocker, do not worry
Sat Feb 18 08:22:14 2017  tibor_:are you here?
Sat Feb 18 08:22:19 2017  hboutemy:even for what would be a blocker in a final release, in an alpha, a blocker...
Sat Feb 18 08:22:32 2017  hboutemy:what do you mean by "here"?
Sat Feb 18 08:22:58 2017  hboutemy:(in fact, I want people to discover colored output and do some buzz)
Sat Feb 18 08:23:17 2017  tibor_:aha, that's good to see
Sat Feb 18 08:23:22 2017  tibor_:i like it
Sat Feb 18 08:23:33 2017  hboutemy:(we need buzz or otehr will do buzz with other tools)
Sat Feb 18 08:24:01 2017  tibor_:sure
Sat Feb 18 08:24:42 2017  hboutemy:then, IMHO, with an alpha in the current state, we're safe
Sat Feb 18 08:25:21 2017  hboutemy:I'm even sure that a release would have beed feasible, and that we are getting too much nervous
Sat Feb 18 08:25:29 2017  hboutemy:we did the big reset
Sat Feb 18 08:26:02 2017  hboutemy:the big reset was serious, because the previous state made me nervous for good causes: I knew anything could happen if we pushed a release
Sat Feb 18 08:26:06 2017  hboutemy:even an alpha
Sat Feb 18 08:26:55 2017  hboutemy:but with current master, carefully chosen commits, I don't see why the sky should fall on our head, even on a final release: for an alpha, even less!!!
Sat Feb 18 08:30:24 2017  tibor_:go to ML and see my email, pls
Sat Feb 18 08:35:12 2017  tibor_:hboutemy: I still hope I am not the bad guy :))
Sat Feb 18 08:39:24 2017  hboutemy:tibor_: thanks, +1 = not not the bad guy ;)à
Sat Feb 18 08:40:17 2017  hboutemy:more seriously: I didn't really track the issues you're facing, and I don't really understand the impact: I'm just feeling the impact is low
Sat Feb 18 08:41:14 2017  hboutemy:and when the blockers are Jira issues that have been proven for bugs that are there wince Maven 3.1, I'm just sure we're missing what is really important
Sat Feb 18 08:41:49 2017  hboutemy:notice on your issues with FreeBSD
Sat Feb 18 08:42:01 2017  hboutemy:IIUC, FreeBSD does more buffering than any other OS?
Sat Feb 18 08:42:35 2017  hboutemy:then you need explicit flush to avoid waiting for buffer that never flush: isn't it?
Sat Feb 18 08:56:41 2017  tibor_:freebsd does not make buffering, that's the issue, the byte stream does not propagate eagerly
Sat Feb 18 08:59:03 2017  tibor_:after reading Javadoc of java.lang.Process, the documentation is free of implicit flushing in jre, so I guess this flushing must be done by our code
Sat Feb 18 09:33:41 2017  Michael-O:Joined the channel
Sat Feb 18 09:36:04 2017  tibor_:Michael-O: Hi Michael. Any good news?
Sat Feb 18 09:36:35 2017  Michael-O:I did not check yet. You did branch, didn't you?
Sat Feb 18 09:36:46 2017  Michael-O:Let me spin up my work laptop to run the build
Sat Feb 18 09:36:47 2017  tibor_:yes, i did
Sat Feb 18 09:36:57 2017  tibor_:branch in Utils
Sat Feb 18 09:45:13 2017  Michael-O_:Joined the channel
Sat Feb 18 09:50:05 2017  Michael-O:Left the channel
Sat Feb 18 09:50:09 2017  Michael-O:Joined the channel
Sat Feb 18 09:53:35 2017  tibor_:Shoul i push SUREFIRE_SYSPROP_DUPLICATES to master and create ticket in Jira?
Sat Feb 18 09:54:18 2017  Michael-O:Let us finish all tests, then create tickets and finally sort commits
Sat Feb 18 09:54:27 2017  tibor_:ok
Sat Feb 18 10:00:06 2017  Michael-O_:Joined the channel
Sat Feb 18 10:03:15 2017  stephenc:hboutemy: yt?
Sat Feb 18 10:03:50 2017  stephenc:notices he actually wired up his phone to his irc relay!
Sat Feb 18 10:04:10 2017  Michael-O:Left the channel
Sat Feb 18 10:04:13 2017  Michael-O:Joined the channel
Sat Feb 18 10:04:27 2017  stephenc:Michael-O: yt?
Sat Feb 18 10:04:42 2017  Michael-O:yt?
Sat Feb 18 10:04:45 2017  Michael-O:youtube?
Sat Feb 18 10:04:50 2017  stephenc:You there
Sat Feb 18 10:04:58 2017  Michael-O:yes
Sat Feb 18 10:05:02 2017  stephenc:I guess that's a yes
Sat Feb 18 10:05:18 2017  stephenc:So how serious are the current issues?
Sat Feb 18 10:05:38 2017  Michael-O:tibor_: can you explain better the severity?
Sat Feb 18 10:05:38 2017  stephenc:It's not DoA on every system as I understand
Sat Feb 18 10:05:52 2017  Michael-O:No, it is not, I guess.
Sat Feb 18 10:06:05 2017  stephenc:Ok, probably not S1
Sat Feb 18 10:06:30 2017  stephenc:On affected systems, does it blow up always?
Sat Feb 18 10:06:50 2017  Michael-O:Yes, always repeatedly
Sat Feb 18 10:06:56 2017  stephenc:Iow can I not do a build at all on FreeBSD
Sat Feb 18 10:07:13 2017  stephenc:Or is it only some builds failing on FreeBSD
Sat Feb 18 10:07:31 2017  Michael-O:Surefire simply does not build due to forking issues. I cannot pass the test. The code compiles fine
Sat Feb 18 10:08:02 2017  stephenc:Ok, if we cannot fork any tests on surefire, that's S1 imho
Sat Feb 18 10:08:21 2017  Michael-O:Bascically, these always fail in tests:
Sat Feb 18 10:08:23 2017  Michael-O:Tests in error:
Sat Feb 18 10:08:25 2017  Michael-O: testForkCountTwoNoReuse(org.apache.maven.surefire.its.ForkModeIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:26 2017  Michael-O: testForkCountTwoReuse(org.apache.maven.surefire.its.ForkModeIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:28 2017  Michael-O: testForkModeOncePerThreadTwoThreads(org.apache.maven.surefire.its.ForkModeIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:29 2017  Michael-O: notReusableForksJUnit47(org.apache.maven.surefire.its.jiras.Surefire1209RerunAndForkCountIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:31 2017  Michael-O: reusableForksJUnit47(org.apache.maven.surefire.its.jiras.Surefire1209RerunAndForkCountIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:33 2017  Michael-O: reusableForksJUnit4(org.apache.maven.surefire.its.jiras.Surefire1209RerunAndForkCountIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:34 2017  Michael-O: notReusableForksJUnit4(org.apache.maven.surefire.its.jiras.Surefire1209RerunAndForkCountIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:36 2017  Michael-O: testForkCountTwoNoReuse(org.apache.maven.surefire.its.ForkModeMultiModuleIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:37 2017  Michael-O: testForkCountTwoReuse(org.apache.maven.surefire.its.ForkModeMultiModuleIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:39 2017  Michael-O: testForkCountTwoReuse(org.apache.maven.surefire.its.ForkModeTestNGIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:41 2017  Michael-O: testForkModeOncePerThreadTwoThreads(org.apache.maven.surefire.its.ForkModeTestNGIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:42 2017  Michael-O: classWithoutCategoryForked(org.apache.maven.surefire.its.jiras.Surefire839TestWithoutCategoriesIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:44 2017  Michael-O: test(org.apache.maven.surefire.its.jiras.Surefire1080ParallelForkDoubleTestIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:45 2017  Michael-O: testRerunFailingErrorTestsTwoRetry(org.apache.maven.surefire.its.JUnit4RerunFailingTestsIT): Exit code was non-zero: 1; command line and log = (..)
Sat Feb 18 10:08:52 2017  stephenc:What about user builds?
Sat Feb 18 10:09:00 2017  Michael-O:tibor_: build started with the patched version of Shared Utils
Sat Feb 18 10:09:06 2017  Michael-O:stephenc: what do you mean exactly?
Sat Feb 18 10:09:25 2017  stephenc:If we have the current to users, would they be affected
Sat Feb 18 10:09:47 2017  stephenc:That surefire tests are finding problems is good
Sat Feb 18 10:10:13 2017  stephenc:But are those problems that most FreeBSD users will find with their builds?
Sat Feb 18 10:10:18 2017  tibor_:I was away
Sat Feb 18 10:10:43 2017  tibor_:any question to me?
Sat Feb 18 10:11:12 2017  stephenc:Eg if I try to build Log4j or Jenkins or Jackson or something else on FreeBSD will the current 3.5.0-Snap fail or succeed or random?
Sat Feb 18 10:11:14 2017  Michael-O:The Core ITs pass flawlessly on FreeBSD. But if you have serious tests running iwth Surefire 2.19.1 and do some forking, they will fail
Sat Feb 18 10:11:32 2017  stephenc:Ok then that is S2 at best
Sat Feb 18 10:11:35 2017  Michael-O:stephenc: it completely depends on the test style
Sat Feb 18 10:11:44 2017  stephenc:Yes so not S1
Sat Feb 18 10:12:06 2017  stephenc:But the feature is dead if you are using the specific pattern
Sat Feb 18 10:12:36 2017  Michael-O:tibor_: can you explain better what surefire case would fail in the user side w/o our fixes?
Sat Feb 18 10:12:53 2017  stephenc:Is there a workaround that users could apply?
Sat Feb 18 10:13:08 2017  stephenc:Or do we just need to fix?
Sat Feb 18 10:13:23 2017  tibor_:1. FreeBSD does not flush std/out => no user can fix it
Sat Feb 18 10:13:39 2017  stephenc:(I am trying to see if this is a blocker for an alpha... it is a blocker for a non-alpha)
Sat Feb 18 10:14:39 2017  Michael-O:stephenc: it issue is partially called by Shared Utils, StreamFeeder. If this is used somewhere in Maven, it could cause issues, but I cannot really assess where.
Sat Feb 18 10:15:31 2017  tibor_:2. our problem that we did not flush() std/out in surefire and maven-shared-utils
Sat Feb 18 10:15:32 2017  stephenc:Ok I think this issue is a blocker.
Sat Feb 18 10:16:01 2017  stephenc:But I think the other issues likely are not blockers for an alpha
Sat Feb 18 10:16:27 2017  stephenc:So let's take the time to fix this right
Sat Feb 18 10:16:31 2017  tibor_:a result is that the JVM exitted with exit code 1
Sat Feb 18 10:16:52 2017  tibor_:looks like test failed but it is not users test failure however it looks so
Sat Feb 18 10:17:16 2017  tibor_:might be confusing to users
Sat Feb 18 10:18:02 2017  stephenc:If a user's test fails will it consistently fail?
Sat Feb 18 10:18:16 2017  tibor_:Michael will launch the build of surefire and we will see, but we will have to proceed thought m-s-u release, takes time
Sat Feb 18 10:18:41 2017  Michael-O:stephenc: I ran the test bed at least 10 times, the problematic cases always failed
Sat Feb 18 10:18:48 2017  stephenc:Ok. Maybe next weekend we can cut an alpha
Sat Feb 18 10:18:51 2017  tibor_:on FreeBSD usually consistently
Sat Feb 18 10:18:58 2017  Michael-O:Yes, exactly.
Sat Feb 18 10:19:06 2017  stephenc:Ok I'm calling this S2
Sat Feb 18 10:19:06 2017  Michael-O:I has to check macOS and Solaris
Sat Feb 18 10:19:41 2017  stephenc:It is not S1 because not every user test will be affected
Sat Feb 18 10:20:35 2017  stephenc:But under specific circumstances- that we could possibly define if we wanted to waste time so doing- it will fail always on at least FreeBSD
Sat Feb 18 10:20:48 2017  stephenc:That is S2
Sat Feb 18 10:21:07 2017  stephenc:is on a phone
Sat Feb 18 10:21:40 2017  stephenc:Are there any other potential S1/S2 issues with 3.5.0 or is this the only one we know of?
Sat Feb 18 10:22:24 2017  tibor_:maybe clean plugin but not sure, Michael wdut?
Sat Feb 18 10:22:43 2017  stephenc:If you mail me how to run the tests I can use my Mac when I finish standing at the side of a GAA pitch
Sat Feb 18 10:23:27 2017  tibor_:I sent local Java changes to Michael, Stephen you dont have them
Sat Feb 18 10:23:28 2017  Michael-O:tibor_: I lest MCLEAN 3.0.0 for now. I do not want to test two issues at once.
Sat Feb 18 10:24:14 2017  stephenc:If neither of you have easy access to osX send them to me and I'll test them
Sat Feb 18 10:25:24 2017  Michael-O:stephenc: check my uploaded log files, you'll see the Maven versions I have used. All builds where on Surefire master with mods from Tibor
Sat Feb 18 10:25:56 2017  stephenc:Do you need help to test on osX?
Sat Feb 18 10:26:02 2017  Michael-O:Joined the channel
Sat Feb 18 10:26:30 2017  stephenc:Iow do you have an osX machine you can test with?
Sat Feb 18 10:27:07 2017  Michael-O:I don't
Sat Feb 18 10:27:18 2017  stephenc:I can see about rescuing my sun ultra t20 from the attic but it is single core :-(
Sat Feb 18 10:27:30 2017  stephenc:tibor_: do you have a Mac?
Sat Feb 18 10:27:53 2017  tibor_:no, sorry, I have Win x64 amd64
Sat Feb 18 10:28:10 2017  stephenc:Ok so you need OS X test coverage
Sat Feb 18 10:28:17 2017  stephenc:I will offer my assistance
Sat Feb 18 10:28:21 2017  Michael-O:stephenc: single core would suffice as long as you have Solaris 10
Sat Feb 18 10:28:24 2017  Michael-O:or 11 running
Sat Feb 18 10:28:27 2017  tibor_:these branches were tested SUREFIRE_STDOUT_FLUSH, SUREFIRE_SYSPROP_DUPLICATES
Sat Feb 18 10:28:29 2017  Michael-O:with recent Java
Sat Feb 18 10:29:12 2017  stephenc:Oh that machine is probably on Solaris 9
Sat Feb 18 10:29:25 2017  stephenc:I will take the osX test case
Sat Feb 18 10:29:48 2017  stephenc:Can you mail me the instructions for an idiot to run the tests
Sat Feb 18 10:29:53 2017  tibor_:https://svn.apache.org/repos/asf/maven/shared/branches/maven-shared-utils-0.9.x
Sat Feb 18 10:29:57 2017  stephenc:Mail
Sat Feb 18 10:30:01 2017  tibor_:ok
Sat Feb 18 10:30:02 2017  stephenc:I am on a phone
Sat Feb 18 10:30:07 2017  Michael-O:stephenc: I will mail it
Sat Feb 18 10:30:19 2017  tibor_:ok, Michael will
Sat Feb 18 10:30:42 2017  stephenc:What to check out and build and then what build step is the critical one
Sat Feb 18 10:31:27 2017  stephenc:My Mac is currently running benchmarks of Log4j and Logback but once it is done I'll run the tests
Sat Feb 18 10:31:47 2017  stephenc:I'll post the results back to the list
Sat Feb 18 10:32:14 2017  tibor_:stephenc: do you have SSD? With ssd and 4 core 3GHz takes 45 min. otehrwise longer
Sat Feb 18 10:32:18 2017  stephenc:Probably better to use a gist for the logs and just mail the gist link...
Sat Feb 18 10:32:49 2017  stephenc:SSD i7 4 cores 8 threads
Sat Feb 18 10:32:56 2017  tibor_::)
Sat Feb 18 10:33:03 2017  stephenc:2.3GHz
Sat Feb 18 10:33:59 2017  stephenc:It's not like that Mac is needed over the weekend
Sat Feb 18 10:34:36 2017  stephenc:I have a non-retina older Mac and my wife's air if I need to do something
Sat Feb 18 10:35:52 2017  stephenc:tibor_: is a gist ok or would you prefer the logs in an email on the DEV list?
Sat Feb 18 10:36:06 2017  stephenc:Michael-O: ^^ same Q
Sat Feb 18 10:37:07 2017  Michael-O:I will rather mail to mailing list for all of us
Sat Feb 18 10:37:17 2017  Michael-O:My logs are on home.apache.org anyway
Sat Feb 18 10:37:51 2017  stephenc:needs to do some cheering on... back in a while
Sat Feb 18 10:46:37 2017  hboutemy:stephenc: Hi Stephen, back there
Sat Feb 18 10:47:11 2017  hboutemy:I'm +1 on S1 vs S2 estimation strategy: that's what I tried to start with Tibor this morning :)
Sat Feb 18 10:48:13 2017  hboutemy:what I still d'ont understand at all about this flush on FreeBSD with Surefire issue is: what could have change in Maven core 3.3.9 vs 3.5.0-S that has an impact?
Sat Feb 18 10:49:04 2017  hboutemy:anyhow, I'm pretty confident for an alpha
Sat Feb 18 10:49:25 2017  Michael-O:hboutemy: there are some differences in test results between 3.3.9 and master
Sat Feb 18 10:49:32 2017  Michael-O:some issues where never reported
Sat Feb 18 10:49:32 2017  hboutemy:I'm using only master to work daily (on Maven) for more than one year
Sat Feb 18 10:49:42 2017  hboutemy:ok
Sat Feb 18 10:50:24 2017  hboutemy:but if this relates to flushing: do you have any idea why flushing has changed between 3.3.9 and master?
Sat Feb 18 10:50:48 2017  Michael-O:hboutemy: changes in core, deps, etc. I did not bisect
Sat Feb 18 10:51:06 2017  hboutemy:even like this, I don't understand
Sat Feb 18 10:51:12 2017  hboutemy:how it is possible :)
Sat Feb 18 10:51:30 2017  Michael-O:hboutemy: did you check the maven-surefire.log in my tarball?
Sat Feb 18 10:51:35 2017  hboutemy:no
Sat Feb 18 10:52:06 2017  hboutemy:but a plugin is quite protected from core, when runnin its internals
Sat Feb 18 10:52:36 2017  hboutemy:IIUC, the issue you're facing is not integration between core and plugin, but plugin when it runs unit tests
Sat Feb 18 10:52:46 2017  hboutemy:= nothing that relys on core
Sat Feb 18 10:52:57 2017  Michael-O:hboutemy: yes, correct.
Sat Feb 18 10:53:11 2017  hboutemy:(the issue with multiple properties order is, for example, an issue at integration level)
Sat Feb 18 10:53:54 2017  rfscholte:Joined the channel
Sat Feb 18 10:54:06 2017  hboutemy:do you confirm that these issues happen only on FreeBSD?
Sat Feb 18 10:54:14 2017  hboutemy:or are there other OSes affected?
Sat Feb 18 10:54:36 2017  Michael-O:hboutemy: this is my observations at the moment. I can't tell anything but our Ubuntu and my work RHEL6 machines
Sat Feb 18 10:54:45 2017  Michael-O:no Solaris, no macOS
Sat Feb 18 10:54:46 2017  hboutemy:(I really don't know Surefire code, then bebugging through your logs won't be easy...)
Sat Feb 18 10:55:14 2017  Michael-O:Tibor and I invested at least 10 hours of work here
Sat Feb 18 10:55:30 2017  hboutemy:yes, I know
Sat Feb 18 10:55:38 2017  hboutemy:that's why I'd like to help :)
Sat Feb 18 10:55:45 2017  tibor_:We added some diagnosis code printed communication between processes
Sat Feb 18 10:56:00 2017  tibor_:the diagnosis code does not exist in Git
Sat Feb 18 10:56:16 2017  tibor_:printed the stream to file
Sat Feb 18 10:56:20 2017  hboutemy:oh yes, now I see: even when you have the OS, since it's about threading and IO, you can't even use a debugger: that's it?
Sat Feb 18 10:56:26 2017  tibor_:did not find the Z, event
Sat Feb 18 10:56:45 2017  Michael-O:hboutemy: yes, debugger immediately changes behavior
Sat Feb 18 10:56:53 2017  Michael-O:we used jstack for threads
Sat Feb 18 10:56:53 2017  hboutemy:arrrrgghhhhh
Sat Feb 18 10:57:08 2017  hboutemy:what a headache....
Sat Feb 18 10:57:15 2017  tibor_:yes, and there are a lot of tests and this use to happen with parallel forks
Sat Feb 18 10:57:40 2017  tibor_:yes and analysed threads dumps
Sat Feb 18 10:58:08 2017  Michael-O:hboutemy: that is the reason why we needed so many nights for this
Sat Feb 18 10:58:15 2017  tibor_:we dumped internal exceptions etc
Sat Feb 18 10:58:23 2017  tibor_:yes
Sat Feb 18 10:58:27 2017  tibor_:not easy realy
Sat Feb 18 10:58:36 2017  hboutemy:from what I see: that does not prevent an alpha, since usual users (not on FreeBSD and not using Surefire master) should not be affected
Sat Feb 18 10:58:39 2017  tibor_:we finished on midnight
Sat Feb 18 10:58:57 2017  hboutemy:but requires to continue work...
Sat Feb 18 11:00:47 2017  tibor_:we continue every day
Sat Feb 18 11:02:36 2017  hboutemy:I mean: doing the alpha release does not mean that we don't care
Sat Feb 18 11:03:06 2017  hboutemy:now, for the rythm of work, you do what you can
Sat Feb 18 12:28:54 2017  rfscholte:hboutemy: got a continuous deployment project :( http://search.maven.org/#search%7Cga%7C1%7Cg%3A%22org.mockito%22%20AND%20a%3A%22mockito-core%22
Sat Feb 18 12:29:40 2017  hboutemy:rfscholte: ouch, yes
Sat Feb 18 12:31:24 2017  hboutemy:looks they started in january 2015 to continuous flow, with 2.0.0
Sat Feb 18 12:31:46 2017  hboutemy:227 versions in 2 years
Sat Feb 18 12:33:07 2017  rfscholte:It was just luck that I noticed this one.
Sat Feb 18 12:33:16 2017  hboutemy:currently, on our central copy (= until 12/2016): 720 MB
Sat Feb 18 12:33:21 2017  rfscholte:would be better to have a complete picture
Sat Feb 18 12:33:38 2017  hboutemy:in fact, we need an algo to guess these cases
Sat Feb 18 12:33:44 2017  rfscholte:+1
Sat Feb 18 12:34:04 2017  hboutemy:perhaps just: when you have more than 100 versions, you're suspect
Sat Feb 18 12:34:12 2017  rfscholte::D
Sat Feb 18 12:36:11 2017  hboutemy:or even simpler: test the size of maven-metadata.xml
Sat Feb 18 12:36:29 2017  hboutemy:currently, in our copy, it's near 7KB
Sat Feb 18 12:38:47 2017  hboutemy:running find /x1/central/data -name maven-metadata.xml -size +4k
Sat Feb 18 12:39:35 2017  hboutemy:cd/go/plugin/go-plugin-api-experimental
Sat Feb 18 12:39:50 2017  rfscholte:what a name....
Sat Feb 18 12:39:50 2017  hboutemy:gov/hhs/cms/bluebutton/datapipeline/bluebutton-data-pipeline
Sat Feb 18 12:40:02 2017  hboutemy:io/fabric8/swagger-model
Sat Feb 18 12:40:16 2017  hboutemy:io/undertow/undertow-servlet
Sat Feb 18 12:41:26 2017  hboutemy:just checked https://repo.maven.apache.org/maven2/io/undertow/undertow-servlet/maven-metadata.xml
Sat Feb 18 12:42:02 2017  hboutemy:perhaps this "maven-metadata.xml site > 4k" is a little too much simple
Sat Feb 18 12:42:25 2017  hboutemy:but give reasonable count
Sat Feb 18 12:43:11 2017  hboutemy:see https://repo.maven.apache.org/maven2/net/openhft/chronicle-bom/maven-metadata.xml
Sat Feb 18 12:47:03 2017  hboutemy:notice: that proves we should work on it
Sat Feb 18 12:47:13 2017  hboutemy:but it's like so many things...
Sat Feb 18 12:47:48 2017  rfscholte::)
Sat Feb 18 12:48:16 2017  hboutemy:and on this one: we need infra's help (not easy) and we're not sure we'll be able to do anything once we have ideas on what to do
Sat Feb 18 12:48:35 2017  hboutemy:in fact, my first issue is: how to report inside the PMC
Sat Feb 18 12:48:49 2017  hboutemy:basically, as a first step
Sat Feb 18 12:49:52 2017  hboutemy:then once we have interesting reports, perhaps this can give people desire to go further
Sat Feb 18 12:50:17 2017  hboutemy:I published my analysis code as pre-first step :)
Sat Feb 18 12:50:32 2017  hboutemy:and the missing part is the reporting channel
Sat Feb 18 12:50:50 2017  hboutemy:mails to private@ proved to be unusable to grab interest
Sat Feb 18 12:50:58 2017  hboutemy:(and I perfectly can understand)
Sat Feb 18 12:51:24 2017  hboutemy:I imagined adding an httpd to make some html report visible
Sat Feb 18 12:51:44 2017  rfscholte:makes sense
Sat Feb 18 12:51:46 2017  hboutemy:(with or without auth, I don't know, that's some details)
Sat Feb 18 12:52:11 2017  hboutemy:even if the sync is stopped for now, this could still be useful
Sat Feb 18 12:52:33 2017  hboutemy:this mirror had good performance for analysis
Sat Feb 18 12:52:54 2017  hboutemy:we really could do a lot of simple things
Sat Feb 18 12:53:16 2017  hboutemy:look at this basic find command that points out to interesting data
Sat Feb 18 12:53:36 2017  hboutemy:no need for much complex things
Sat Feb 18 12:54:33 2017  hboutemy:after thikning at it, for analysis, even without the TBs of jar, the few 100 s of GB of metadata could give excellent informations
Sat Feb 18 12:54:50 2017  hboutemy:(perhaps less than 100 s of GB...)
Sat Feb 18 12:55:25 2017  rfscholte:true, most of the time we're not interested in the content of the jar
Sat Feb 18 12:55:40 2017  hboutemy:I'll try to contact infra to remove some binaries (find some space on our full 5TB), then sync only metadata, then add httpd
Sat Feb 18 12:56:03 2017  rfscholte:great, thanks
Sat Feb 18 12:56:04 2017  hboutemy:then it could give us a good basis to prove value of analysis
Sat Feb 18 12:56:25 2017  hboutemy:since analysis should not cost much (archive costs a lot more)
Sat Feb 18 12:56:44 2017  rfscholte:hboutemy: parallel to M3.5.0 we could work on http://opensourcedesign.net/
Sat Feb 18 12:57:23 2017  hboutemy:another good idea requiring additional time... :)
Sat Feb 18 12:57:32 2017  rfscholte::D yes
Sat Feb 18 12:57:34 2017  hboutemy:you met them at FOSDEM ?
Sat Feb 18 12:58:10 2017  rfscholte:no, but every now ant than I keep thinking about this
Sat Feb 18 12:58:32 2017  hboutemy:lately, I'm trying to go at french meetings to find people interested in involving
Sat Feb 18 12:58:54 2017  hboutemy:in french and face 2 face, since it could be simpler for them, and give us great value
Sat Feb 18 12:59:29 2017  hboutemy:for example, for color output, a guy wrote the parameterization code: I didn't do it, just integrated
Sat Feb 18 12:59:35 2017  rfscholte:I have one dutch guy who is interested in contributing, but somehow it is hard to find a topic to start with
Sat Feb 18 12:59:39 2017  hboutemy:(in maven-shared-util)
Sat Feb 18 13:00:18 2017  hboutemy:if you meet him f2f, such topics like opensourcedesign could be doable
Sat Feb 18 13:00:46 2017  rfscholte:it's not f2f
Sat Feb 18 13:00:54 2017  hboutemy:just need to mentor f2f, since by email, won't be easy: it's quite an open debate to know what to do ont this topic
Sat Feb 18 13:01:03 2017  hboutemy:of at least webcam
Sat Feb 18 13:01:08 2017  hboutemy:something very interactive
Sat Feb 18 13:01:13 2017  rfscholte:yes
Sat Feb 18 13:01:23 2017  hboutemy:even more interactive than IRC
Sat Feb 18 13:02:05 2017  hboutemy:I have a BOF at DevoxxFR this year, like last 2 years
Sat Feb 18 13:02:12 2017  hboutemy:and I try to ask for help
Sat Feb 18 13:02:21 2017  hboutemy:but people say "yes" but don't come back
Sat Feb 18 13:02:45 2017  hboutemy:they don't even imagine that the field of possible is so open
Sat Feb 18 13:03:03 2017  hboutemy:they need to have ideas on what they can do
Sat Feb 18 13:03:15 2017  hboutemy:can+like
Sat Feb 18 13:04:07 2017  hboutemy:this year, I don't have any talk: perhaps I'll take time to prepare slides for the BoF to be more clear on "help wanted" (not only on code, but just on energy to do anything useful)
Sat Feb 18 13:04:21 2017  rfscholte:and it shouldn't be maintenance or refactoring, that alone is too boring
Sat Feb 18 13:05:23 2017  hboutemy:I had contacts with guys who need to change cmplex things in core, or things like that = too much
Sat Feb 18 13:05:54 2017  hboutemy:BTW, how is your new job going?
Sat Feb 18 13:06:09 2017  hboutemy:finding sufficient clients?
Sat Feb 18 13:06:25 2017  hboutemy:lets you sufficient time to still work on Maven?
Sat Feb 18 13:06:52 2017  hboutemy:(this part can go in privatge, if you prefer :) )
Sat Feb 18 13:06:58 2017  rfscholte:just had my first week for a client
Sat Feb 18 13:07:46 2017  rfscholte:4 days, so one day to Maven related stuff and things for other clientss
Sat Feb 18 13:43:36 2017  stephenc:hboutemy: yt?
Sat Feb 18 13:43:57 2017  hboutemy:stephenc: yes
Sat Feb 18 13:44:14 2017  stephenc:So I am running the surefire tests
Sat Feb 18 13:44:43 2017  hboutemy:you mean "building Surefire master with Maven master"?
Sat Feb 18 13:44:46 2017  stephenc:Do you know if the failures are a regressing with 3.3.9->3.5.0 or a surefire regression
Sat Feb 18 13:44:53 2017  stephenc:Yes
Sat Feb 18 13:45:17 2017  hboutemy:no: it's a nighttmare to understand
Sat Feb 18 13:45:31 2017  hboutemy:simply the conditions of what fails
Sat Feb 18 13:45:41 2017  stephenc:Yeah that is what I want to understand too
Sat Feb 18 13:45:46 2017  hboutemy:(with noise everywhere: m-clean-p, for example)
Sat Feb 18 13:46:00 2017  stephenc:We need to get an alpha and start tracking issues
Sat Feb 18 13:46:26 2017  stephenc:Afaik we can have multiple versions in fix-version in JIRA
Sat Feb 18 13:47:17 2017  stephenc:So as long as we can, my proposal is that everything "fixed" for master gets 3.5.0-alpha-1 *added*
Sat Feb 18 13:47:34 2017  stephenc:But we keep 3.5.0 as the fix version
Sat Feb 18 13:48:05 2017  stephenc:So that the release notes for 3.3.9->3.5.0 reflect reality ignoring alphas
Sat Feb 18 13:48:30 2017  stephenc:Wdyt?
Sat Feb 18 13:49:07 2017  stephenc:hboutemy: ^^^
Sat Feb 18 13:49:13 2017  stephenc:rfscholte: ^^^
Sat Feb 18 13:49:48 2017  rfscholte:I agree, all 3.5.0 alpha's should also be marked as fixed for 3.5.0
Sat Feb 18 13:50:25 2017  stephenc:Ok. I'll start teeing up this w/e
Sat Feb 18 13:51:07 2017  stephenc:Once I get a better understanding of where the regressions are: surefire/clean/core then I'll know whether the alpha is blocked
Sat Feb 18 13:52:00 2017  stephenc:rfscholte: do you have a good understanding as to where the regression lies?
Sat Feb 18 13:52:30 2017  rfscholte:I wonder if they are all regressions. Could very well be that we're just extra testing current release and now issues are exposed
Sat Feb 18 13:52:47 2017  stephenc:Well that is also a possibility
Sat Feb 18 13:53:21 2017  hboutemy:(sorry, on phone)
Sat Feb 18 13:53:25 2017  stephenc:I think the surefire FreeBSD issue is a GA release blocker... but I don't know if that is a blocker on core or surefire itself
Sat Feb 18 13:53:42 2017  stephenc:(On phone too... hence slow typing)
Sat Feb 18 13:53:54 2017  tibor_:Discuss it with Michael, because he isolated clean plugin and added 2.6.1 version.
Sat Feb 18 13:53:59 2017  stephenc:waiting for a haircut ;-)
Sat Feb 18 13:55:14 2017  stephenc:tibor_: we need to be very strict and only worry about real regressions for alpha-1
Sat Feb 18 13:55:36 2017  stephenc:Then we can get a list of issues in JIRA, triage and knock down to release
Sat Feb 18 13:56:00 2017  tibor_:if you have a result, please write README and zip the surefire-it/target and we will have a look
Sat Feb 18 13:56:48 2017  stephenc:Well right now I'm just running surefire master against the 4 distorts
Sat Feb 18 13:57:30 2017  stephenc:Then I need to apply the patches to surefire and re-run
Sat Feb 18 13:57:40 2017  tibor_:ok, thx
Sat Feb 18 13:57:59 2017  stephenc:Then I'll check which ones failed, rerun and capture the directory
Sat Feb 18 13:58:24 2017  stephenc:tibor_: can you mail me what to capture for failed builds
Sat Feb 18 13:58:42 2017  hboutemy:stephenc: +1 for the "everything "fixed" for master gets 3.5.0-alpha-1 *added*" strategy
Sat Feb 18 13:58:57 2017  tibor_:there is module surefire-integration-tests/target
Sat Feb 18 13:59:18 2017  tibor_:the log of the build you run as well
Sat Feb 18 13:59:20 2017  stephenc:tibor_: *mail* me... if it's not in a mail I'll forget
Sat Feb 18 13:59:25 2017  tibor_:ok
Sat Feb 18 13:59:34 2017  stephenc:No
Sat Feb 18 13:59:38 2017  stephenc:Problem
Sat Feb 18 14:00:01 2017  stephenc:curses autocorrect for changing NP into No
Sat Feb 18 14:00:35 2017  hboutemy::)
Sat Feb 18 14:00:57 2017  stephenc:tibor_: do you know if it is a regression in core or surfire or a confluence of both?
Sat Feb 18 14:01:23 2017  tibor_:I lost your thread
Sat Feb 18 14:01:40 2017  stephenc:The surefire freebsd issue
Sat Feb 18 14:02:10 2017  tibor_:surefire
Sat Feb 18 14:02:11 2017  stephenc:Is this a regression in maven core, or a regression in surefire or a combination of the two
Sat Feb 18 14:02:39 2017  stephenc:Ok so all is good running the previous release of surefire on FreeBSD
Sat Feb 18 14:03:03 2017  stephenc:S/previous/current/ (as we have not released master)?
Sat Feb 18 14:03:06 2017  tibor_:which version?
Sat Feb 18 14:03:11 2017  stephenc:2.19.x
Sat Feb 18 14:03:21 2017  stephenc:Whatever x is latest
Sat Feb 18 14:04:06 2017  stephenc:Iow does freebsd fail to build surefire-2.19.1 (say) with Maven 3.3.9 or 3.5.0-snap
Sat Feb 18 14:04:19 2017  stephenc:Or do both of those build fine?
Sat Feb 18 14:04:41 2017  stephenc:We need to isolate the problem
Sat Feb 18 14:04:48 2017  tibor_:we did not test with 2.19.1
Sat Feb 18 14:05:11 2017  stephenc:What is the current latest *release* of surefire?
Sat Feb 18 14:05:15 2017  tibor_:We know that one issue was in class ForkedBooter.java
Sat Feb 18 14:05:18 2017  stephenc:on phone
Sat Feb 18 14:05:33 2017  tibor_:latest release is 2.19.1
Sat Feb 18 14:05:37 2017  stephenc:I'd look up my self if not on a phone
Sat Feb 18 14:06:14 2017  tibor_:and ForkedRunListener missed flush() as well
Sat Feb 18 14:06:20 2017  stephenc:Ok. So can somebody check if 2.19.1 will build on freebsd with the different cores
Sat Feb 18 14:06:31 2017  tibor_:maybe somebody deleted a line in code
Sat Feb 18 14:06:54 2017  stephenc:If the answer is those builds fail too, then this is not a blocker for *core* release
Sat Feb 18 14:07:09 2017  tibor_:only Michael has freebsd
Sat Feb 18 14:07:27 2017  tibor_:i understand
Sat Feb 18 14:07:37 2017  stephenc:Yeah, we'll need to see if he can determine where the fault lies
Sat Feb 18 14:07:59 2017  stephenc:I think the system property changes in core smell fishy
Sat Feb 18 14:08:20 2017  stephenc:So they may need to either come out or be fixed in a non-hack way
Sat Feb 18 14:13:14 2017  rfscholte:stephenc: you mean MNG-6172?
Sat Feb 18 14:13:20 2017  tibor_:stephenc: email went out
Sat Feb 18 14:15:40 2017  tibor_:MNG-6172 is not maven issue. Our IT passed default sys props followed by next props written by each test. Reordering the props means that we tested TestNG 5.7 instead of 5.10, but this is not functionality issue - no user issue. I made a fix in ITs.
Sat Feb 18 14:16:03 2017  tibor_:it is the branch
Sat Feb 18 14:16:32 2017  tibor_:this ine SUREFIRE_SYSPROP_DUPLICATES
Sat Feb 18 14:17:26 2017  tibor_:it is hotfix. We can fix each IT but I do not have time now to fix 58 ITs. Later.
Sat Feb 18 15:05:24 2017  tibor_:I checked 2.19.1 ForkingRunListener, ForkedBooter. The flush() was not changed after release 2.19.1. It means the rootcause is same.
Sat Feb 18 16:44:06 2017  rfscholte:hboutemy: you should add release notes for wagon 2.11 too :)
Sat Feb 18 16:51:37 2017  hboutemy:they are empty...
Sat Feb 18 16:52:42 2017  rfscholte:?? weird
Sat Feb 18 17:27:55 2017  rfscholte:Joined the channel
Sat Feb 18 18:51:32 2017  rfscholte:Joined the channel
Sat Feb 18 19:02:21 2017  Michael-O:tibor_: finally back, took longer than expected. Tests finished, still too much failures. New log file has been uploaded. Have a look.
Sat Feb 18 19:07:11 2017  Michael-O:still the missing ping :(
Sat Feb 18 19:07:39 2017  tibor_:I would expect everything to pass
Sat Feb 18 19:08:00 2017  tibor_:I use to run the build in offline mode
Sat Feb 18 19:08:56 2017  tibor_:Do we have changes from both branches?
Sat Feb 18 19:09:35 2017  tibor_:Stephen has got one failure because he did take one branch.
Sat Feb 18 19:10:32 2017  tibor_:<include>org/apache/**/ForkMode*IT.java</include>
Sat Feb 18 19:10:58 2017  tibor_:This would run by seceral tests, not only one
Sat Feb 18 19:11:22 2017  tibor_:several
Sat Feb 18 19:15:48 2017  tibor_:Michael-O: What shall we do in this case?
Sat Feb 18 19:17:40 2017  Michael-O:I have locally patched files, I did not merge the branches
Sat Feb 18 19:17:48 2017  Michael-O:I will rerun ForkMode only now
Sat Feb 18 19:19:16 2017  tibor_:see this
Sat Feb 18 19:19:18 2017  tibor_:[INFO] Total time: 4.441 s
Sat Feb 18 19:19:26 2017  tibor_:[WARNING] 1487413455569 CommandReader NOOP
Sat Feb 18 19:19:35 2017  tibor_:we received ping
Sat Feb 18 19:20:26 2017  Michael-O:why warning then?
Sat Feb 18 19:20:46 2017  tibor_:I don't know
Sat Feb 18 19:21:10 2017  tibor_:but the failure must be because of missing Z event == good buy signal
Sat Feb 18 19:21:18 2017  tibor_:bye
Sat Feb 18 19:21:27 2017  tibor_:so some flush missing in code
Sat Feb 18 19:21:34 2017  Michael-O:I guessed so
Sat Feb 18 19:21:53 2017  Michael-O:we likely need to search for all write()
Sat Feb 18 19:30:26 2017  Michael-O:tibor_: I have updated the tarball for the ForkMode ITs
Sat Feb 18 19:30:36 2017  tibor_:ok
Sat Feb 18 19:32:18 2017  Michael-O:1487445848842 ForkedBooter NOOP timer
Sat Feb 18 19:32:19 2017  Michael-O:1487445848842 ForkedBooter PING timer: plugin did not send me NOOP signal > exit
Sat Feb 18 19:32:41 2017  Michael-O:Quite clear that communication does not work, right?
Sat Feb 18 19:32:46 2017  tibor_:send me a link in IRC
Sat Feb 18 19:32:57 2017  tibor_:is it the same file name?
Sat Feb 18 19:33:27 2017  tibor_:exactly
Sat Feb 18 19:33:40 2017  Michael-O:Yes, the same with 0.9.1-SNAPSHOT
Sat Feb 18 19:33:43 2017  tibor_:ok
Sat Feb 18 19:33:47 2017  Michael-O:but smaller
Sat Feb 18 19:36:04 2017  Michael-O_:Joined the channel
Sat Feb 18 19:36:25 2017  Michael-O_:Got it?
Sat Feb 18 19:41:48 2017  tibor_:[INFO] Tests run: 0, Failures: 0, Errors: 0, Skipped: 0
Sat Feb 18 19:41:59 2017  tibor_:nothing was sent from plugin
Sat Feb 18 19:42:09 2017  tibor_:nothing like run this test
Sat Feb 18 19:42:43 2017  tibor_:we can print all events sent by plugin
Sat Feb 18 19:42:57 2017  tibor_:i will check the code I gave you
Sat Feb 18 19:43:26 2017  Michael-O:ok, I'll wait. This is really, really strange.
Sat Feb 18 19:45:32 2017  Michael-O:stephenc: checked your build results. You were able to reproduce parts of my issues. This is good.
Sat Feb 18 19:47:23 2017  tibor_:I have to think of java bug only.
Sat Feb 18 19:47:47 2017  Michael-O:I will rerun with Java 7
Sat Feb 18 19:47:50 2017  Michael-O:Wait a second
Sat Feb 18 19:48:30 2017  Michael-O:rerunning with 7 Update 111
Sat Feb 18 19:48:37 2017  tibor_:ok
Sat Feb 18 20:14:06 2017  Michael-O_:Joined the channel
Sat Feb 18 20:18:53 2017  Michael-O:tibor_: Java 7 is 7 errors, Java 8 is 9. Not really better
Sat Feb 18 20:19:20 2017  Michael-O:Can we have the stream.out again? We will see the missing U
Sat Feb 18 20:19:21 2017  Michael-O:Z*
Sat Feb 18 20:19:55 2017  tibor_:I am trying to put more logs to see what part of the code did not send bytes to std/out.
Sat Feb 18 20:20:14 2017  Michael-O:OK
Sat Feb 18 20:20:39 2017  Michael-O:If some VMs really exit with 0, it must be missing Z
Sat Feb 18 20:27:54 2017  Michael-O:tibor_: FYI I have resetted master, branched locally to freebsd-master and merge the two experimental branches to have a clean state. But included Shared Utils 0.9.1-SNAPSHOT.
Sat Feb 18 20:31:25 2017  tibor_:i should maybe merge them together into third branch and put logs
Sat Feb 18 20:35:28 2017  Michael-O:this is what I did locally
Sat Feb 18 20:35:50 2017  Michael-O:but publically, I would create a ticket for each of them because these are separate issues
Sat Feb 18 20:36:17 2017  tibor_:and the system.out.println is there as well?
Sat Feb 18 20:49:53 2017  tibor_:I found that we have FlushReceiver which flushed OutputStream after complete byte[] of command was sent to fork vm
Sat Feb 18 20:50:23 2017  tibor_:so the change in m-s-u is not the root cause
Sat Feb 18 20:51:03 2017  tibor_:is there way to dump native std/out?
Sat Feb 18 21:13:55 2017  Michael-O:no, no system out
Sat Feb 18 21:14:14 2017  Michael-O:Native from C?
Sat Feb 18 21:14:19 2017  Michael-O:not without JNI
Sat Feb 18 21:15:20 2017  Michael-O:but you can always use a TeeOutputStream to capture all comm
Sat Feb 18 21:32:20 2017  Michael-O_:Joined the channel
Sat Feb 18 21:49:57 2017  Michael-O:tibor_: do you want me to run anything tonight? If not, I will switch off my laptop.
Sat Feb 18 21:50:14 2017  tibor_:no
Sat Feb 18 21:50:16 2017  tibor_:tomorrow
Sat Feb 18 21:50:25 2017  tibor_:good night
Sat Feb 18 21:51:23 2017  tibor_:I am preparing something and I will run new branch on my PC first of all.
Sat Feb 18 21:52:03 2017  tibor_:all is about the communication, so I need to see data flow, even the binaries
Sat Feb 18 21:53:18 2017  tibor_:i will take care of this and we will have these detail in console
Sat Feb 18 21:53:37 2017  Michael-O:ok
Sat Feb 18 21:53:47 2017  Michael-O:see you tomorrow, thanks for the help!
Sat Feb 18 21:53:56 2017  tibor_:I thank you!

Comments