Displaying #maven-dev/2017-03-22.log:

Wed Mar 22 09:52:30 2017  tibor_:Joined the channel
Wed Mar 22 10:12:17 2017  olamy:Joined the channel
Wed Mar 22 17:03:13 2017  Michael-O:Joined the channel
Wed Mar 22 18:47:58 2017  rfscholte:Joined the channel
Wed Mar 22 19:21:09 2017  tibor_:Joined the channel
Wed Mar 22 19:35:14 2017  Michael-O:Joined the channel
Wed Mar 22 19:55:19 2017  gboue:Joined the channel
Wed Mar 22 19:58:55 2017  tibor_:gboue: Hi :)
Wed Mar 22 19:59:10 2017  tibor_:Michael-O: Hi Michael. What's up.
Wed Mar 22 19:59:12 2017  gboue:hey!
Wed Mar 22 19:59:20 2017  Michael-O:Hi
Wed Mar 22 19:59:26 2017  Michael-O:Just read gboue's mail.
Wed Mar 22 19:59:29 2017  Michael-O:Good finding.
Wed Mar 22 19:59:30 2017  tibor_:good catch in the mkdirs, thx
Wed Mar 22 20:00:06 2017  tibor_:the fix was pushed
Wed Mar 22 20:00:12 2017  gboue:I'm testing a change in mkdirs: returning true if the parent file exist
Wed Mar 22 20:00:49 2017  gboue:it appears to have fixed it, yep!
Wed Mar 22 20:01:59 2017  tibor_:Michael has noticed that the semaphore was not notified from maven process.
Wed Mar 22 20:03:23 2017  tibor_:gboue: From your side FreeBSD build is ok?
Wed Mar 22 20:03:45 2017  gboue:the Ubuntu build with JDK 8 is OK now, I'm running it a few times, and will move on to FreeBSD
Wed Mar 22 20:03:55 2017  tibor_:ok
Wed Mar 22 20:10:53 2017  Michael-O:let me test on the BSD box
Wed Mar 22 20:13:04 2017  Michael-O:9bc02fd862b6443c604d861b28a539fc69d234bd, correct?
Wed Mar 22 20:13:38 2017  Michael-O:^M are back in DumpUtils
Wed Mar 22 20:23:06 2017  gboue:Yeah there's a quirk with the line endings
Wed Mar 22 20:32:23 2017  tibor_:I see 0x0A which is \n
Wed Mar 22 20:36:08 2017  Michael-O:Running Surefire141PluggableProvidersIT only now
Wed Mar 22 20:36:17 2017  Michael-O:tibor_: this is again Git likely
Wed Mar 22 20:39:56 2017  gboue:I just ran Surefire141PluggableProvidersIT on my FreeBSD box, it's OK with JDK 7 and 8!
Wed Mar 22 20:41:03 2017  gboue:will run it a few times to make sure, but I'm fairly convinced the root issue was fixed here
Wed Mar 22 20:45:27 2017  tibor_:.gitattributes has
Wed Mar 22 20:45:29 2017  tibor_:* text=auto
Wed Mar 22 20:45:40 2017  tibor_:I should change it to LF
Wed Mar 22 20:48:25 2017  gboue:hmm this is what I have also tibor (text=auto), and I don't remember having this issue
Wed Mar 22 20:48:35 2017  gboue:Tomorrow I'll run the whole ITs with FreeBSD, I won't have time this evening
Wed Mar 22 21:00:13 2017  Michael-O:tibor_: Just passed on OpenJDK 8
Wed Mar 22 21:00:18 2017  Michael-O:running OpenJDK 7
Wed Mar 22 21:00:23 2017  tibor_:cool
Wed Mar 22 21:00:42 2017  Michael-O:I did not pass before
Wed Mar 22 21:00:54 2017  Michael-O:failed with a message: non-empty directory
Wed Mar 22 21:05:31 2017  gboue:Michael-O: it is still failing for you with OpenJDK 7 on FreeBSD?
Wed Mar 22 21:19:32 2017  tibor_:I di this configuration. I hope it helps.
Wed Mar 22 21:19:38 2017  tibor_:git config --global core.autocrlf false
Wed Mar 22 21:19:39 2017  tibor_:git config --global core.eol lf
Wed Mar 22 21:23:27 2017  Michael-O:7 passes too now
Wed Mar 22 21:23:45 2017  Michael-O:This fix was really good now
Wed Mar 22 21:23:59 2017  stephenc:autocrlf is evil
Wed Mar 22 21:24:05 2017  Michael-O:I would to test this fix on top of 2.19.2-experimental
Wed Mar 22 21:24:44 2017  Michael-O:stephenc: git itself is evil ;-)
Wed Mar 22 21:24:49 2017  stephenc:no
Wed Mar 22 21:24:53 2017  stephenc:git is good
Wed Mar 22 21:25:05 2017  stephenc:but autocrlf is almost never what you want
Wed Mar 22 21:25:20 2017  stephenc:most people working on windows use a toolchain that handles lf without the cr
Wed Mar 22 21:25:30 2017  stephenc:and will preserve the lf without cr
Wed Mar 22 21:25:49 2017  Michael-O:tibor_: can you reenable PING on the experimental branch and apply this fix too?
Wed Mar 22 21:26:07 2017  stephenc:much better is to just have lf as the convention and use tooling to ban the crlf
Wed Mar 22 21:29:34 2017  tibor_:@stephenc: what tooling you mean?
Wed Mar 22 21:30:26 2017  stephenc:we could ask infra for a pre-commit hook to reject commits with crlf
Wed Mar 22 21:30:49 2017  stephenc:or we could use the enforcer plugin to break the build if any text files have crlf
Wed Mar 22 21:30:59 2017  tibor_:you mean hooks
Wed Mar 22 21:31:24 2017  stephenc:in any case, lf is IMHO the safest line ending
Wed Mar 22 21:31:33 2017  tibor_:yes
Wed Mar 22 21:31:36 2017  stephenc:much better to just force LF only
Wed Mar 22 21:33:23 2017  tibor_:Michael-O: I will apply the patch in experimental branch + PING
Wed Mar 22 21:34:24 2017  Michael-O:+1
Wed Mar 22 21:44:54 2017  gboue:I finished running this specific test with various combinations of Windows / Ubuntu / FreeBSD on OpenJDK 7 / OpenJDK 8 / Oracle JDK 8. It is passing each time, this looks very good!
Wed Mar 22 21:48:08 2017  Michael-O:gboue: sounds great
Wed Mar 22 21:48:25 2017  Michael-O:I will wait for tibors merge and will retest tomorrow on a few machines at work
Wed Mar 22 22:27:37 2017  Michael-O:good night guys
Wed Mar 22 22:32:26 2017  tibor_:Michael-O: done
Wed Mar 22 22:40:50 2017  gboue:tibor_: on Windows, running all the tests, I have UmlautDirIT in error (Oracle JDK 8): "java.io.FileNotFoundException: C:\...\maven-surefire\surefire-integration-tests\target\junit-pathWith▒mlaut\log.txt". I'll see if I can reproduce that tomorrow
Wed Mar 22 22:41:32 2017  gboue:good night!
Wed Mar 22 22:41:47 2017  tibor_:This is encoding issue
Wed Mar 22 22:42:07 2017  gboue:Left the channel

Comments