IRC Archive for channel #xwiki on 11 September 2014
Last modified by Vincent Massol on 2014/09/11 23:53
<momomomomo> has joined #xwiki
00:29 <momomomomo> has quit
01:04 <momomomomo> has joined #xwiki
01:21 <abusenius> has quit
02:30 <sburjan> has quit
03:07 <OSIMasson> has quit
03:27 <Emenems> has quit
03:32 <Emenems> has joined #xwiki
03:42 <momomomomo> has quit
03:46 <OSIMasson> has joined #xwiki
03:50 <Denis> has joined #xwiki
03:53 <Emenems> has quit
03:59 <Emenems> has joined #xwiki
04:04 <Denis> has quit
04:07 <OSIMasson> has left #xwiki
04:30 <xwikibot> has joined #xwiki
06:23 <DarkKnightCZ> has joined #xwiki
06:28 <DarkKnightCZ> has quit
07:31 <andreeapopescu> has joined #xwiki
07:48 <woshilapin> has quit
08:05 <mflorea> has joined #xwiki
08:17 <vmassol> has joined #xwiki
08:22 <Denis> has joined #xwiki
08:33 <KermitTheFragger> has joined #xwiki
08:37 <woshilapin> has joined #xwiki
08:47 <cjd> has joined #xwiki
08:53 <tmortagne> has joined #xwiki
08:56 <abusenius> has joined #xwiki
09:02 <DarkKnightCZ> has joined #xwiki
09:02 <abusenius> has quit
09:26 <Enygma`> has joined #xwiki
09:30 <evalica> has joined #xwiki
09:30 <msmeria> has joined #xwiki
09:41 <woshilapin> has quit
09:41 <woshilapin> has joined #xwiki
09:42 <ClemensR> has joined #xwiki
09:57 <cjd> Denis: how do you like my way of getting my code reviewed?
10:01 <Denis> cjd: if you are talking about your checkstyle discussion, it is terribly efficient :)
10:01 <cjd> ;)
10:02 <Denis> I would have much to say as well, but I lack time actually
10:02 <cjd> yeah, vmassol does as well
10:02 <Denis> but checkstyle is not all or nothing
10:02 <cjd> +1 that
10:02 <Denis> you can exclude rules line by lines
10:03 <Denis> it just take time to maintain
10:03 <woshilapin> And pollute code with lines that have nothing to do with code but with checkstyle?
10:03 <cjd> I feel some checks are absolutely necessary, evidenced by the fact that I wrote my own codestyle verifier for cjdns (C)
10:03 <Denis> no, you put that in a separate xml config
10:03 <woshilapin> Ah, OK
10:04 <woshilapin> Seems nicer, indeed
10:04 <cjd> ok that's a bit worse
10:04 <woshilapin> (I didn't know you can do it)
10:04 <cjd> it's like Makefiles with per-file rules, very bad
10:05 <cjd> you need to synchronize them with the state of the files
10:05 <Denis> https://github.com/xwiki/xwiki-commons/blob/master/xwiki-commons-core/xwiki-commons-crypto/xwiki-commons-crypto-common/src/main/checkstyle/checkstyle-suppressions.xml
10:05 <tmortagne> separate xml config has it's drawnback too, it makes it's a bit of a pain to maintain it
10:05 <Denis> this is an example
10:05 <Denis> yes, I fully agree with you Thomas
10:06 <cjd> I use a comment notation in cjdns.. // CHECKFILES_IGNORE it expects a ; here but this is a macro
10:06 <tmortagne> yes I perefer comments system too
10:06 <cjd> that way all information about the code is contained
10:06 <tmortagne> s/perefer/prefer/
10:06 <Denis> this is why my filling, that is only worst it for useful checkstyle check, those that may really reveal poor code that is not obvious
10:06 <tmortagne> when you remove of fix the failure you don't forget to update the excludes
10:07 <cjd> +1
10:07 <Denis> at the same time, comment link you checker with your code, which is not so nice
10:08 <cjd> the code will still compile without the checker
10:08 <cjd> it's not as if I created a magical annotation which needs to be preprocessed away
10:08 <gdelhumeau> has joined #xwiki
10:09 <tmortagne> cjd: note that I think there is some support for things like //CHECKSTYLE:OFF/ON in checkstyle
10:09 <tmortagne> need to get the documentation
10:09 <cjd> it's disabled for us, we had a discussion ages ago about enabling it and sdumitriu1 was not favorable to it at the time
10:10 <cjd> but that was like 2010
10:10 <tmortagne> yep that's it, you need to add a filter for it
10:10 <tmortagne> http://checkstyle.sourceforge.net/config.html
10:11 <tmortagne> <module name="Checker"> ... <module name="SuppressionCommentFilter"/> ... </module>
10:11 <tmortagne> <module name="Checker">
10:11 <tmortagne> ...
10:11 <tmortagne> <module name="SuppressionCommentFilter"/>
10:11 <cjd> except //CHECKSTYLE:OFF begin big function ... means that big function gets *none* of the checks
10:12 <tmortagne> you don't have to put that around the whole method
10:12 <tmortagne> it's around any code as far as I understand
10:12 <cjd> if it's a statement count issue?
10:12 <tmortagne> if it's a statement count issue of course
10:12 <cjd> I suppose statement count would actually be reported at the first line of the function so you could do
10:12 <tmortagne> maybe you can put that around the signature
10:12 <cjd> //CHECKSTYLE:OFF // big function
10:13 <cjd> function (arg, arg2) {
10:13 <cjd> // yeah, that
10:13 <tmortagne> yep
10:24 <mflorea> has quit
10:29 <bouquetf_> has quit
10:37 <bouquetf> has joined #xwiki
10:58 <woshilapin> Hi devs, I have a problem with a 'unknown macro'. This macro is registered since I can access it with Admin user but not with a non-admin user
10:59 <woshilapin> I checked that the macro is registered with a user with programming rights
10:59 <woshilapin> (XWiki.Admin in this case)
11:00 <woshilapin> Something that may be specific to my case, there is one macro A that is using other nested macro B
11:00 <woshilapin> The macro A is well displayed, but the macro B show the 'unknown macro' problem
11:00 <woshilapin> Any idea?
11:02 <woshilapin> (s/I checked that the macro is registered with a user with programming rights/The macros are registered with XWiki.Admin which is in XWikiAdminGroup/)
11:02 <woshilapin> (but I suppose that's equivalent)
11:03 <woshilapin> at least, I'm looking for an idea on how to investigate (where can I find useful log for example)
11:12 <cjd_> has joined #xwiki
11:13 <tmortagne> woshilapin: make sure you macro has the right visibility
11:13 <tmortagne> sounds like your macro has user visibility
11:14 <tmortagne> should be wiki or global
11:14 <woshilapin> "right visibility"? What do you mean?
11:14 <woshilapin> I have only one wiki 'wiki'
11:14 <tmortagne> look at your macro with object editor
11:14 <tmortagne> you have a field in there indicating who should see that macro
11:15 <woshilapin> Indeed :-)
11:15 <tmortagne> maybe you chose user (that should be named "macro author" probably) by mistake
11:15 <tmortagne> user visibility allow a user that does not have programming or admin right to register a wiki macro for himself
11:16 <woshilapin> OK
11:16 <tmortagne> * wiki=admin right
11:16 <tmortagne> * global= programming right
11:16 <woshilapin> But why 'user visibility' is default?
11:16 <cjd> has quit
11:16 <woshilapin> (by the way, it works, thanks)
11:16 <tmortagne> there is no real default, I guess it's just first in the list of available possibility
11:17 <tmortagne> s/possibility/possibilities/
11:17 <woshilapin> Yes
11:17 <cjd_> is now known as <cjd>
11:17 <woshilapin> OK, anyway, thanks for the suppoer
11:18 <woshilapin> support*
11:24 <gdelhumeau> I don't reproduce http://ci.xwiki.org/job/xwiki-enterprise-test-wysiwyg/lastBuild/org.xwiki.enterprise$xwiki-enterprise-test-wysiwyg/testReport/org.xwiki.test.wysiwyg/StandardFeaturesTest/testItalics/ manually
12:02 <cjd> has quit
12:04 <cjd> has joined #xwiki
12:05 <cjd> is now known as <Guest41316>
12:07 <KermitTheFragger> has quit
12:41 <Enygma`> I have my doubts about those tests
12:41 <Enygma`> they were not failing for the 6.2 rc1 release
12:41 <Enygma`> so my take is that they are flukes
12:42 <Enygma`> but we really need to get these flukes into control
12:42 <Enygma`> they are very annoying
12:42 <Enygma`> starting to handle test failures too locall
12:51 <Enygma`> vmassol: any idea about http://ci.xwiki.org/job/xwiki-enterprise-test-ui/lastBuild/org.xwiki.enterprise$xwiki-enterprise-test-ui/testReport/org.apache.maven.surefire.junit4/JUnit4Provider/org_xwiki_test_ui_AllTests/ ?
12:59 <vmassol> Enygma`: what I'd like to do
12:59 <vmassol> modify start_xwiki.sh to stop any running XE process (if a script parameter is passed)
12:59 <vmassol> the first problem here is "Error binding monitor port 8079: java.net.BindException: Cannot assign requested address"
13:00 <vmassol> which could be caused by either a running XE or by an issue with my check in XwikiExecutor
13:00 <vmassol> (which I would remove after I modify start_xwiki.sh since it wouldn't not be needed any longer)
13:00 <vmassol> now
13:00 <vmassol> the NPE is easy to fix, doing it now
13:01 <vmassol> I introduced the problem yesterday actually
13:02 <vmassol> done
13:13 <Denis> has quit
13:28 <Denis> has joined #xwiki
13:33 <Enygma`> not reproducing locally: http://ci.xwiki.org/job/xwiki-enterprise-test-ui-6.2.x/lastBuild/org.xwiki.enterprise$xwiki-enterprise-test-ui/testReport/org.xwiki.test.ui/EditObjectsTest/testPropertyDisplayersForNewObjects/
13:33 <Enygma`> used @Intermittent(repetition=10)
13:43 <vmassol> Hi guys, does anyone know how I can get the process id and continue execution in a shell script?
13:43 <vmassol> I know I can use $! to get it
13:43 <vmassol> currently in our start_xwiki.sh wehave:
13:43 <vmassol> java $XWIKI_OPTS $3 $4 $5 $6 $7 $8 $9 -jar $JETTY_HOME/start.jar ${JETTY_HOME}/etc/jetty.xml ${JETTY_HOME}/etc/jetty-*.xml
13:43 <vmassol> so it's blocking and waiting till the JVM exits
13:43 <vmassol> I'd like to get the pid and save it in a lock file
13:44 <vmassol> but also continue blocking execution till the JVM exits
13:44 <Guest41316> mmm trickey one
13:44 <vmassol> the "fg" command depends on the shell I guess
13:44 <vmassol> so I cannot use & and then fg
13:44 <vmassol> actually maybe
13:44 <Guest41316> don't save the java process, save the shell process
13:45 <Guest41316> echo $$ > pid
13:45 <Guest41316> java .......
13:45 <Guest41316> the script will keep running as long as java is still running
13:45 <vmassol> $$ represents the shell pid?
13:45 <Guest41316> yeap
13:45 <vmassol> cool
13:45 <vmassol> that should work then thanks
13:45 <Guest41316> kill -9 $$
13:45 <vmassol> trying
13:46 <vmassol> I won't try that! :)
13:46 <Guest41316> kill -11 `pidof apache2`
13:46 <Guest41316> (11 is a sigsegv, synthetic segmentation fault)
13:54 <DarkKnightCZ> has left #xwiki
13:58 <vmassol> hmmm
13:58 <vmassol> doesn't work
13:58 <vmassol> kill the shell process doesn't kill the JVM apparently
13:58 <vmassol> s/kill/killing
13:58 <vmassol> (Guest41316)
14:01 <Guest41316> ahh right
14:01 <vmassol> actually I can stop using our stop_xwiki.sh script…
14:02 <Guest41316> reconnecting to get my nick back...
14:03 <Guest41316> has quit
14:03 <cjd> has joined #xwiki
14:04 <Enygma`> vmassol: does the first answer on this question help http://stackoverflow.com/questions/1821968/pid-of-last-started-process-in-bash-script
14:04 <vmassol> ah it was you cjd ;)
14:04 <cjd> yeah, my nick is set to enforce mode so nobody else can use it unless they auth
14:05 <Enygma`> some_program &
14:05 <Enygma`> some_pid=$!
14:05 <cjd> ahh indeed, 'wait'
14:05 <vmassol> actually what I'm doing now is:
14:05 <Enygma`> in your case..
14:05 <Enygma`> java ...
14:05 <Enygma`> xwiki_pid=$!
14:05 <vmassol> run stop_xwiki.sh
14:06 <vmassol> however I need to modify stop_xwiki.sh so that it kills the process if it doesn't stop by itself
14:06 <cjd> +1
14:06 <vmassol> so I still need to store the real JVM pid
14:06 <vmassol> I'll try the wait
14:06 <vmassol> is that standard on all unixes?
14:07 <cjd> hmm
14:07 <cjd> System.getPid() -> pid.file ?
14:08 <cjd> does start_xwiki.sh need to do anything after xwiki shuts down?
14:09 <vmassol> remove the lock file
14:09 <cjd> ok
14:10 <vmassol> I have something strange
14:10 <cjd> yeah, +1 the wait method
14:11 <vmassol> this is what I have:
14:11 <vmassol> https://gist.githubusercontent.com/vmassol/2f1b6b9c3207dd17ed3b/raw/cfa4d26b55198eabe335aabe69eb969ec3a11763/gistfile1.txt
14:11 <cjd> looks like wait is a shell builtin so anywhere there is bash, there is wait
14:11 <vmassol> now when I exeucte the second instance
14:12 <vmassol> the first one stops fine
14:12 <vmassol> but in the console of the second one I get:
14:12 <vmassol> https://gist.githubusercontent.com/vmassol/73707faca38befe362f8/raw/cfc93f2ac2bbb5dcefa0c7750ab7c7d55ff15181/gistfile1.txt
14:12 <vmassol> not sure why I get "Error binding monitor port 8079: java.net.BindException: Address already in use"
14:12 <cjd> cool
14:12 <vmassol> since the stop_xwiki.sh is waiting....
14:13 <vmassol> oh yes
14:13 <vmassol> hmm no
14:13 <vmassol> I don't know
14:15 <vmassol> if the JVM is stopped completely shouldn't it mean that the ports are freed? can they overlast the JVM?
14:16 <vmassol> if so then it's more complex because I'd need to do a loop and check if someone is holding the port
14:16 <vmassol> (in a standard way)
14:16 <cjd> https://ezcrypt.it/j39n#opPHBmCcMRAZoxEJsuWjmRgH
14:16 <cjd> no, they cannot outlast the process
14:17 <cjd> there's always a chance that another process holds the port
14:17 <vmassol> but the JVM doesnet spawn another process AFAIK
14:17 <cjd> oh oops bug
14:19 <cjd> https://ezcrypt.it/k39n#og8g5xXV0cN8FKiopEbUdr1x
14:19 <cjd> that loop might not be 100% right
14:20 <vmassol> ok but that's for later
14:20 <vmassol> ATM I need to understand why I have this "Error binding monitor port 8079: java.net.BindException: Address already in use" error
14:20 <cjd> anyway that's the most resilient way to stop the process
14:20 <vmassol> (and how to fix it)
14:21 <cjd> hmm
14:21 <cjd> race condition?
14:21 <cjd> oh no, it had been up for an hour
14:23 <ClemensR> is it possible that yet another unrelated (or old, defunct) process is still listening at 8079 ?
14:23 <vmassol> with a sleep 2 it still fails but with a sleep 10 it works
14:24 <cjd> ClemensR: we did a `netstat -lnpt` and it didn't show the port as bound
14:24 <cjd> so I'm confused
14:24 <vmassol> cjd: that's different
14:24 <ClemensR> ah, ok, then I join the 'confused' group :/
14:24 <vmassol> cjd is talking about some job running on ci
14:25 <vmassol> but here I'sm experimenting locally
14:25 <vmassol> just running sh start_xwiki.sh twice
14:25 <cjd> mm /nod
14:25 <vmassol> I want the second one to stop the first one
14:25 <cjd> mmm I wouldn't
14:26 <vmassol> it seems that the port is really freed only when the shell exits for the first one
14:26 <vmassol> or something like that
14:26 <cjd> oh
14:26 <cjd> hang on a sec
14:28 <DarkKnightCZ> has joined #xwiki
14:28 <cjd> http://stackoverflow.com/questions/14388706/socket-options-so-reuseaddr-and-so-reuseport-how-do-they-differ-do-they-mean-t
14:29 <cjd> ``The question is, how does the system treat a socket in state TIME_WAIT? If SO_REUSEADDR is not set, a socket in state TIME_WAIT is considered to still be bound to the source address and port and any attempt to bind a new socket to the same address and port will fail until the socket has really been closed, which may take as long as the configured Linger Time.''
14:30 <cjd> Write once, refactor everywhere
14:30 <vmassol> ok I think I know
14:31 <vmassol> it seems the stop returns immediately
14:31 <vmassol> I'm printing the date before and after the actual stop, ie the java execution to stop jetty
14:31 <cjd> mhm
14:31 <vmassol> and it seems too fast to be true
14:31 <vmassol> before stop: Thu Sep 11 14:30:24 CEST 2014
14:31 <vmassol> after stop: Thu Sep 11 14:30:24 CEST 2014
14:31 <vmassol> not even 1 second…
14:32 <vmassol> to start a JVM , run the stop and stop the JVM
14:32 <cjd> doesn't it take 1 second to start the JVM?
14:32 <cjd> <trollface>
14:32 <vmassol> I can't know since it's blocking :)
14:33 <cjd> anyway making stop_xwiki.sh work well I'm in favor of, making start_xwiki.sh automatically call stop_xwiki.sh is too magical IMO
14:33 <vmassol> so I think the stop_xwiki.sh finishes before it's really stopped
14:34 <vmassol> s/before it's/before jetty/
14:34 <cjd> should not be possible to finish before the lock is released
14:34 <cjd> *file deleted
14:34 <vmassol> yes but deleting the file take a split second
14:34 <vmassol> that'snot a problem
14:34 <cjd> but it only happens after xwiki is actually stopped
14:35 <vmassol> apparently not
14:35 <vmassol> otherwise it would work :)
14:35 <cjd> are you using my script or yours?
14:35 <vmassol> mine
14:35 <cjd> I'm not sure of the stuff in that loop
14:35 <cjd> hard for me to follow
14:36 <cjd> try mine :)
14:36 <vmassol> I don't want to do any kill ATM
14:36 <vmassol> that's a fallback only
14:36 <vmassol> if it fails to stop but we need to give it a chance to stop properly first
14:36 <cjd> ok then comment out the kill lines
14:36 <vmassol> you changed the loop?
14:36 <cjd> yeah
14:37 <cjd> https://ezcrypt.it/k39n#og8g5xXV0cN8FKiopEbUdr1x
14:37 <vmassol> ok I'll try but I don't see why it would make a difference
14:38 <vmassol> same
14:38 <vmassol> which I expected
14:38 <cjd> hrm
14:39 <cjd> sec
14:39 <vmassol> there's no reason at all
14:39 <vmassol> it would change anything
14:39 <vmassol> it's not about the deletion of the lock file
14:40 <cjd> echo ">>$XWIKI_LOCK_FILE<< === >>`cat $XWIKI_LOCK_FILE`<<"
14:41 <vmassol> wdym?
14:41 <vmassol> the echo works fine
14:41 <ClemensR> I cannot reproduce locally
14:42 <vmassol> ClemensR: here are my full start_xwiki.sh and stop.xwiki.sh
14:42 <cjd> making sure the lock is indeed there
14:43 <vmassol> start_xwiki.sh: https://gist.githubusercontent.com/vmassol/b10524449403c7eaa09d/raw/bbbe39063b191c5eb6e328a208509e9e100275a6/gistfile1.txt
14:45 <vmassol> stop_xwiki.sh: https://gist.githubusercontent.com/vmassol/32a6a2bcdf4a5849c87c/raw/9b081ecad833deb7140058799c0627e97d544e3d/gistfile1.txt
14:45 <vmassol> what I do:
14:45 <vmassol> in a shell run: sh start_xwiki.sh
14:45 <vmassol> wait till it's fully started
14:45 <vmassol> thenin a second shell in the same directory
14:45 <vmassol> run: sh start_xwiki.sh
14:46 <ClemensR> ah, lock file is in /var/tmp/ ... maybe that makes a difference (though it shouldn't)
14:46 <vmassol> I'm pretty sure the real proboem is that when the java process exits
14:46 <vmassol> the port are not fully freed yet on my OS
14:46 <vmassol> (Mac)
14:46 <vmassol> it may be os-dependent
14:47 <cjd> while [ "$timer" -lt 100 -a -e XWIKI_LOCK_FILE ]; do
14:47 <cjd> while [ "$timer" -lt 100 -a -e $XWIKI_LOCK_FILE ]; do
14:47 <vmassol> yes
14:47 <vmassol> that's just a typo in my paste
14:47 <cjd> and it's not in your file?
14:47 <vmassol> I reverted your loop to my first loop before pasting
14:48 <vmassol> (and I made a typo)
14:49 <vmassol> when stop_xwiki.sh is called from the second eeuction on start_xwiki.sh
14:49 <ClemensR> I dont have that jetty/etc/jetty-*.xml, but I guess that is no problem
14:49 <cjd> [ ! -e $XWIKI_LOCK_FILE ] && echo 'lock missing' && exit 0;
14:49 <cjd> java $XWIKI_OPTS -jar $JETTY_HOME/start.jar --stop
14:49 <vmassol> I can see that my first shell is still active about 4-5 seconds after
14:49 <vmassol> (I have a spinning symbol on my shell command tab)
14:50 <vmassol> first shell is still active == first shell is still executing something
14:51 <ClemensR> now I can reproduce, one :)
14:52 <vmassol> ok good
14:52 <vmassol> wdym by "one"?
14:53 <cjd> ls /var/tmp
14:53 <cjd> is the lock there?
14:53 <vmassol> yes that works fine
14:53 <vmassol> again
14:53 <cjd> ok
14:54 <vmassol> I don't see how the lock file can be an issue
14:54 <vmassol> it's not about the lock file
14:54 <vmassol> but about the port being freed
14:55 <cjd> sec
14:55 <ClemensR> the interesting thing is that the script from the jetty-hsqldb is working w/o problems
14:55 <vmassol> you mean you can start 2 instances of jetty on the same port?
14:55 <vmassol> :)
14:56 <cjd> user@blk1:~/xwiki-enterprise-jetty-hsqldb-6.2-SNAPSHOT$ time ./_stop_xwiki.sh && ps -ef | grep java && netstat -lnpt
14:56 <cjd> lock missing
14:57 <cjd> lock [/var/tmp/xwiki-.lck] missing
14:57 <ClemensR> I meant, when the four "stop when running" lines are pasted into the script from teh original distro
14:57 <cjd> user@blk1:~/xwiki-enterprise-jetty-hsqldb-6.2-SNAPSHOT$ cat ./_stop_xwiki.sh | grep JETTY_PORT
14:57 <cjd> XWIKI_LOCK_FILE="${XWIKI_LOCK_DIR}/xwiki-${JETTY_PORT}.lck"
14:57 <cjd> user@blk1:~/xwiki-enterprise-jetty-hsqldb-6.2-SNAPSHOT$
14:58 <cjd> JETTY_PORT == ""
14:58 <ClemensR> for me the stop-script exits a second before the old process really quits
14:58 <ClemensR> ah, thats it, JETTY_PORT is not set from $1 :)
14:59 <ClemensR> script actually expects JETTY_STOP_PORT as $1
15:00 <vmassol> oh yes
15:00 <vmassol> good catch
15:00 <vmassol> my bad
15:00 <vmassol> fixing that and trying again
15:00 <ClemensR> using XWIKI_LOCK_FILE="${XWIKI_LOCK_DIR}/xwiki-${JETTY_STOP_PORT}.lck" in both files works better
15:01 <vmassol> I need: /bin/sh stop_xwiki.sh $JETTY_STOP_PORT
15:01 <ClemensR> uh, gotta go, meeting at 1500 here/ well, issue found in time :)
15:01 <vmassol> thanks!
15:01 <ClemensR> has left #xwiki
15:02 <vmassol> hmmm still the same probem :(
15:02 <cjd> https://ezcrypt.it/l39n#EKadpZ0QB3LbsfpBGtpX3BYN
15:03 <vmassol> that's normal cjd
15:03 <vmassol> the 3 processes execute in parallel
15:03 <vmassol> since the stop takes the most time
15:03 <cjd> mmm no
15:03 <vmassol> it's normal that netsat would show it still bound
15:03 <cjd> oh
15:03 <vmassol> && means execute in // no?
15:04 <cjd> no, execute after
15:04 <cjd> but only if the previous one succeeds
15:04 <vmassol> ok then indeed it's what I said
15:04 <cjd> like if (something() && somethingElse())
15:04 <vmassol> when the jvm exits the port are not freed yet
15:05 <cjd> actually I think the timer just timed out
15:05 <cjd> because it was 12 seconds
15:05 <cjd> and the next time I ran it, it worked
15:05 <cjd> better make it 20 before trying to kill it
15:06 <cjd> https://ezcrypt.it/m39n#QrSGrJtW0NSTh4LRIlO2upRx
15:06 <cjd> second time works
15:06 <cjd> hey
15:06 <cjd> who installed cryptocurrency in my vm ?!
15:10 <vmassol> hmm it worked for me when I tried:
15:10 <vmassol> time sh stop_xwiki.sh && ps -ef | grep java && lsof -iTCP:8080 -sTCP:LISTEN
15:11 <vmassol> but it's strange
15:11 <vmassol> ps still shows the process
15:12 <vmassol> but the lsof doesn't show the port as bound for 8080
15:12 <vmassol> checking 8079 now
15:12 <OSIMasson> has joined #xwiki
15:12 <vmassol> ok for 8079 it's still boind
15:12 <cjd> increase the timout
15:13 <vmassol> what timeout?
15:13 <vmassol> in stop_xwiki.sh?
15:13 <cjd> yeap
15:13 <cjd> or put an echo below it saying "timed out"
15:13 <cjd> so you know if it times out
15:14 <vmassol> same
15:14 <vmassol> not sure what the timeout would change
15:14 <cjd> user@blk1:~/xwiki-enterprise-jetty-hsqldb-6.2-SNAPSHOT$ ./_start_xwiki.sh
15:14 <cjd> An XWiki instance is already running on port 8080, stopping it!
15:14 <cjd> java.net.ConnectException: Connection refused
15:14 <cjd> this is a bad idea
15:14 <cjd> you should not add that part to start_xwiki.sh
15:15 <cjd> you should check if the process is running and if it's not, clear out the stale lock
15:15 <vmassol> actually I read up a lot on that
15:15 <vmassol> and the consensus is that it's bad to do what you say
15:15 <vmassol> and instead the good practice is to save teh pid
15:16 <vmassol> because doing a ps might not always work as expected
15:16 <gdelhumeau> do we still have the "undelete" right in XWiki?
15:16 <gdelhumeau> I don't see it in the <select> tag
15:16 <gdelhumeau> on a XWikiGlobalRights object
15:16 <vmassol> so I wanted to avoid doing a ps
15:17 <vmassol> in any case the big problem is that the port is not freed when stop_xwiki.sh exits
15:17 <vmassol> so the current stop_xwiki.sh doesn't work
15:17 <vmassol> the wait at the end doesn't do much in practice apparently
15:18 <vmassol> I'm surprised that the port outlives the JVM
15:18 <vmassol> s/JVM/JVM process/
15:19 <vmassol> a simple sleep 10 at the end of stop_Xwiki.sh would work but it's not very nice
15:19 <cjd> 15:15 < vmassol> and instead the good practice is to save teh pid <-- saving the pid is brittle but it's standard practice
15:20 <cjd> what I don't like is that if xwiki is already running, you try to stop it
15:20 <cjd> when the admin didn't ask you to stop anything
15:20 <vmassol> that's not my point
15:20 <cjd> that will lead to angry admins, I'm willing to bet on it
15:20 <vmassol> you're talking details
15:20 <vmassol> and that's absoutely not my goal
15:20 <vmassol> as I mentioned twice already above
15:20 <vmassol> :)
15:20 <vmassol> my goal is to pass a param to the start script to control that behavior
15:21 <vmassol> I just haven't implemented it yet because I don't care at this stage
15:21 <vmassol> ofc we shouldn't stop a running XE if the user has not asked for it!!
15:21 <vmassol> I'd be angry if someone did that to me :)
15:22 <cjd> ahh ok
15:22 <octo> Hello :) qq, XWikiDocument#getXObject() is supposed to return the BaseObject of the document, right?
15:22 <vmassol> octo: the first xobject actually
15:22 <vmassol> there can be a lot...
15:23 <vmassol> depends what param you pass actually
15:24 <vmassol> if you use it without params then it's the first one
15:24 <octo> vmassol: getXObject() calls getXObject(DocumentReference)
15:24 <tmortagne> octo: XWikiDocument#getXObject() is pretty random (unless there is only one object of course), better indicate at least the type of object you want of you know it
15:25 <vmassol> yes but check getXObject(DocumentReference classReference)
15:25 <tmortagne> s/of/if/
15:25 <vmassol> you'll see it iterate on all objects
15:25 <octo> Okay, maybe take a step back.
15:25 <vmassol> and stops on the frist one
15:25 <octo> What I want to get at is the Guid or the document
15:25 <tmortagne> there is no such thing as "Guid or the document"
15:25 <octo> As far as I see the BaseObject is the only class exposign the getGuid() method
15:25 <vmassol> getId
15:25 <vmassol> ?
15:25 <vmassol> getId()?
15:26 <vmassol> XWikiDocument.getId()
15:26 <vmassol> not sure why you'd need this though....
15:26 <tmortagne> be carefull that getId() is local
15:26 <tmortagne> can't you use the document reference ?
15:27 <octo> I'd like to have an identifier of a wiki page that is final, i.e. doesn't change when the page is renamed
15:28 <tmortagne> there is nothing like that
15:28 <tmortagne> what you could use is the value of some field of a custom object you put in those pages
15:28 <octo> Ugh, that will make this complicated :\
15:29 <tmortagne> that way that object field value won't change if you rename the document and you can search the document based on it
15:29 <tmortagne> it's like introducing yourself a uid to a document
15:29 <octo> That sounds like a plan
15:30 <octo> tmortagne: Can I listen for DocumentCreated events and just add such an object automatically in that handler?
15:31 <vmassol> cjd: ok I'll go the ps way then…
15:31 <vmassol> and do some loop till the process is no longer there
15:31 <vmassol> I don't see any other option....
15:32 <tmortagne> octo: yes you can automatically ad such object when a document is created with DocumentCreatedEvent
15:32 <tmortagne> actually would be better to listener to DocumentCreatingEvent
15:32 <tmortagne> and modify the XWikiDocument you get with the event
15:32 <vmassol> note that adding objects to all doc might slow your wiki a bit since loading a doc will also load the object ie do more queries on the db
15:32 <tmortagne> if you work with DocumentCreatedEvent you will have 2 save, the normal save plus you adding your object
15:32 <andreeapopescu> has quit
15:33 <cjd> vmassol: https://ezcrypt.it/q39n#hOvgyeNwO2AirknllVNmgyjC
15:35 <octo> tmortagne: Thank you!
15:35 <tmortagne> octo: you're welcome :)
15:37 <vmassol> cjd: what does this change? BTW your waitTwenty() is not correct as $1 is the jetty stop port
15:38 <vmassol> scratch the last part
15:38 <vmassol> (it's the method param)
15:38 <vmassol> but again I really don't understand what you're trying to do
15:38 <vmassol> why do you keep thinking that the lock file is a problem?
15:38 <cjd> I think it's the same as what you're trying to do
15:38 <cjd> robust stop script
15:39 <vmassol> it's not
15:39 <vmassol> well
15:39 <vmassol> it's going to always kill the process without letting it a chance to stop correctly
15:39 <cjd> not when I tested it
15:39 <ClemensR> has joined #xwiki
15:40 <vmassol> could I talk to you on skype?
15:40 <vmassol> does your skype setup allows this? :)
15:40 <cjd> meaning w/ audio?
15:40 <cjd> heh
15:40 <vmassol> yup
15:41 <cjd> nope, sorry
15:41 <vmassol> ok too bad
15:41 <vmassol> ok so let's do the slow ay
15:41 <vmassol> *way
15:41 <cjd> I think slow so it wouldn't help much to talk live
15:41 <vmassol> the stop_xwiki.sh | ps | netstat
15:41 <vmassol> shows
15:41 <vmassol> that the when the java command returns jetty is not fully stopped
15:41 <vmassol> so in your script
15:42 <vmassol> you get to the line that says:
15:42 <gdelhumeau> vmassol: cjd : use https://www.gruveo.com/ (WebRTC)
15:42 <vmassol> echo 'Failed to stop XWiki, attempting kill';
15:42 <vmassol> and then you execute
15:42 <gdelhumeau> tested during the seminar, and I was happy with it
15:42 <vmassol> kill `cat $XWIKI_LOCK_FILE`
15:42 <vmassol> so even though there's no probem stopping xwiki it'll always end doing a kill
15:43 <cjd> 15:41 < vmassol> that the when the java command returns jetty is not fully stopped <-- I tested a bunch of times and it did not happen
15:43 <vmassol> and the kill will happen when jetty is being stopped so not even sure how jetty will like it....
15:43 <vmassol> cjd: on mac it does all the time
15:43 <cjd> what I saw was the result of the timeout being too short and not telling me it was a timeout
15:44 <vmassol> I don't see the relationshiup between the lock file/timeout and the port beeing freed
15:44 <cjd> try my scripts and tell me if it says it's attempting a kill (you can comment the kill if you'd rather like)
15:44 <vmassol> ok
15:44 <cjd> s/like//
15:45 <cjd> time ./_stop_xwiki.sh && ps -ef | grep java && netstat -lnpt
15:47 <vmassol> note: you changed from #!/bin/sh to bash...
15:47 <vmassol> in the start script
15:47 <cjd> indeed I did
15:47 <cjd> `wait` is standard in bash, 'sh' is not standardized at all
15:47 <OSIMasson> has left #xwiki
15:47 <cjd> meant to mention that
15:48 <vmassol> is bash avaialble on all unixes?
15:48 <cjd> all sane ones :)
15:49 <cjd> might not be on an openwrt router but that's a non-point
15:49 <vmassol> ok
15:49 <vmassol> exactly same problem
15:49 <cjd> paste?
15:49 <vmassol> actually no
15:49 <vmassol> even worse
15:49 <vmassol> :)
15:50 <vmassol> when I start the second instance I get:
15:50 <vmassol> https://gist.githubusercontent.com/vmassol/8a01a48eeabac6fcb07e/raw/1929af3995d3137d1b24a3af994301dea37d9bdb/gistfile1.txt
15:50 <vmassol> ah wait
15:50 <vmassol> checking something
15:51 <cjd> time ./_stop_xwiki.sh && ps -ef | grep java && netstat -lnpt
15:51 <cjd> I want to know how that behaves (w/ 1 instance running)
15:51 <vmassol> I forgot to change stop_xwiki.sh to stop_xwiki2.sh
15:51 <momomomomo> has joined #xwiki
15:52 <cjd> well since my version of the script nolonger calls it, it's kind of a nonpoint
15:52 <vmassol> ah yes
15:52 <vmassol> so no it doesn't work....
15:52 <cjd> time ./_stop_xwiki.sh && ps -ef | grep java && netstat -lnpt
15:52 <vmassol> wow /prioc
15:52 <cjd> proc is in all unics
15:52 <vmassol> that's non standard I believe
15:52 <vmassol> I don't have it on mac
15:52 <cjd> o_O
15:53 <vmassol> I don't have any /proc/ directory
15:53 <vmassol> I've reviewed a lot of ways to check for a process
15:53 <vmassol> yesterday
15:53 <vmassol> and there are only 2 that can work
15:54 <cjd> ok I stand corrected
15:54 <vmassol> best one is with the pid
15:54 <vmassol> and second best is ps
15:54 <cjd> I thought proc was an iron standard
15:54 <vmassol> but you need to be careful with the grep
15:54 <vmassol> ps aux should be standard I think
15:54 <vmassol> and the grep needs to be precise enough to not target other processes
15:55 <vmassol> FTR I was reading this thread http://stackoverflow.com/questions/9117507/linux-unix-command-to-determine-if-process-is-running
15:55 <cjd> anyway
15:55 <cjd> time ./_stop_xwiki.sh && ps -ef | grep java && netstat -lnpt
15:56 <cjd> if we can shut the thing down and know after it stops, we're ok
15:56 <cjd> the rest is details
15:56 <vmassol> netsat won't work
15:56 <vmassol> I pasted what works above
15:56 <vmassol> *netstat
15:56 <vmassol> finding it again
15:57 <cjd> that line is just for testing whether we can stop xwiki w/o a race condition, not to be put into a script
15:57 <vmassol> I know
15:58 <vmassol> found it
15:58 <vmassol> time sh stop_xwiki.sh && ps -ef | grep java && lsof -iTCP:8080 -sTCP:LISTEN
15:58 <vmassol> s/8080/8079
15:58 <DarkKnightCZ> has left #xwiki
15:59 <vmassol> but since stop.sh won't work I don't see what you expect to see
15:59 <vmassol> it doesn't stop XE
15:59 <cjd> it does on my machine
15:59 <vmassol> that's why your script is even worse than the one I had
15:59 <vmassol> which at least stopped XE
16:00 <cjd> what's the backscroll when you run that?
16:01 <vmassol> backscroll?
16:01 <cjd> what's printed to the screen..
16:01 <vmassol> I pasted it above
16:01 <vmassol> :)
16:01 <cjd> https://gist.githubusercontent.com/vmassol/8a01a48eeabac6fcb07e/raw/1929af3995d3137d1b24a3af994301dea37d9bdb/gistfile1.txt <-- ?
16:01 <vmassol> https://gist.githubusercontent.com/vmassol/8a01a48eeabac6fcb07e/raw/1929af3995d3137d1b24a3af994301dea37d9bdb/gistfile1.txt
16:01 <vmassol> yup
16:02 <cjd> $ sh start_xwiki2.sh <-- wrong command
16:02 <cjd> time sh stop_xwiki.sh && ps -ef | grep java && lsof -iTCP:8080 -sTCP:LISTEN
16:02 <cjd> I want to know what that one does
16:02 <vmassol> then you don't understand what I'm trying to do
16:02 <vmassol> :)
16:02 <vmassol> ok
16:02 <vmassol> but it won't be what I 'm trying to achieve
16:02 <cjd> I was under the impression that the stop script returned before stopping xwiki
16:02 <cjd> therefor it would send the kill signal
16:03 <cjd> so I wanted to see what was happening to debug it
16:03 <vmassol> ok trying
16:09 <xwikibot> has joined #xwiki
16:12 <vmassol> I don't need ps anymore
16:12 <vmassol> now that we can stop xwiki for sure
16:13 <cjd> durr
16:13 <cjd> kill -0 $XWIKI_PID
16:13 <cjd> that is universal
16:13 <cjd> not sure how I managed to forget that
16:14 <vmassol> yes they mention it on http://stackoverflow.com/questions/9117507/linux-unix-command-to-determine-if-process-is-running
16:14 <vmassol> but I don't think I even need this
16:14 <vmassol> hmm no same
16:15 <cjd> doesn't work for you?
16:15 <vmassol> so the issue is really triggere by calling stop from start
16:15 <vmassol> *triggered
16:15 <vmassol> stop works because it ends the shell process started
16:15 <vmassol> for some reason
16:16 <cjd> that's odd
16:16 <vmassol> grrr
16:16 <vmassol> we're back to where I was
16:18 <vmassol> wait
16:19 <vmassol> made a mistake
16:21 <vmassol> it works fine :)
16:21 <vmassol> great
16:21 <vmassol> thanks for your help cjd, I'll take over from there
16:22 <cjd> cool, yw
16:22 <vmassol> hmm
16:22 <vmassol> just did a ctrl-c and it didn't stop it cleanly as it did before
16:22 <cjd> ahh of course
16:22 <cjd> because of the start script using & and wait
16:23 <vmassol> btw your script needs to be improve when there's a lock file but no process
16:23 <vmassol> which happened when I did the ctrl-c
16:23 <vmassol> trying to start after that
16:23 <vmassol> leads to
16:24 <cjd> http://alvinalexander.com/java/jwarehouse/jetty-6.1.9/bin/jetty.sh.shtml
16:24 <cjd> JETTY_PID ?
16:24 <vmassol> https://gist.githubusercontent.com/vmassol/ffe0fb87dde12fe8ce6e/raw/1f85d62c08d415289f950dac00c5b906f3437d40/gistfile1.txt
16:24 <Denis> has quit
16:25 <cjd> ahh yea, we should indeed add the check to the stop script as well
16:25 <vmassol> interesting script
16:25 <vmassol> pretty complex too...
16:26 <cjd> kill -0 $XWIKI_PID || ( echo 'XWiki was not stopped cleanly, removing pid file' && rm $XWIKI_PID_FILE
16:26 <cjd> )
16:26 <vmassol> well we need to fix the ctrl-C issue
16:26 <vmassol> it was great
16:26 <Denis> has joined #xwiki
16:26 <cjd> yeah
16:27 <cjd> JETTY_PID=/var/tmp/xwiki-8080.lck
16:27 <vmassol> so wait doesn't respond to ctrl-c
16:27 <cjd> or whatever
16:27 <cjd> well.. technically it does, it stops waiting :)
16:27 <vmassol> I mean the signal is not caught anymore by the JVM for some reason
16:27 <cjd> set JETTY_PID and then let jetty create the actual file
16:27 <cjd> yeah, because the JVM is forked into the background so it won't get signals from the console
16:28 <vmassol> JETTY_PID is not standard
16:28 <vmassol> they use some other script: "start-stop-daemon"
16:29 <cjd> yup
16:30 <cjd> start-stop-deamon is a nice one because it forks the process, (thus getting the pid) then it writes the pid file, then it execve's the program you want to run, therefor not changing pid
16:30 <momomomomo> has quit
16:30 <cjd> but you need to use C to be able to do that
16:33 <vmassol> cjd: why not use "fg" simply?
16:33 <vmassol> process &
16:33 <vmassol> fg
16:33 <cjd> I don't see any reason why not
16:33 <vmassol> trying
16:33 <cjd> never seen it used before so there might be a dragon sleeping under it
16:33 <cjd> *shrug*
16:35 <momomomomo> has joined #xwiki
16:35 <vmassol> hmm doesnet work
16:35 <vmassol> start_xwiki2.sh: line 129: fg: no job control
16:36 <cjd> does your mac have start-stop-dameon ?
16:37 <vmassol> no
16:37 <vmassol> http://apple.stackexchange.com/questions/99888/is-there-any-port-of-start-stop-daemon-for-os-x
16:42 <cjd> https://support.hyperic.com/display/SIGAR/Home
16:42 <cjd> kind of neat
16:42 <vmassol> bbiab
16:54 <sdumitriu1> cjd: No need for inline suppression comments
16:54 <cjd> out of band in an xml file is quite bad IMO
16:55 <sdumitriu1> Why?
16:55 <cjd> refactor the code, don't change the xml file
16:55 <cjd> 2 places to change
16:55 <cjd> coupling
16:55 <sdumitriu1> Yep
16:55 <sdumitriu1> There are two contradicting rules
16:55 <sdumitriu1> There's a rule about not mixing multiple languages in a source file
16:56 <cjd> I disagree about multiple languages in a source file
16:56 <cjd> not sure where that rule came from but I think it needs to justify itself
17:00 <cjd> @Annotations(are not strictly java, they're bolted on to the spec but they're a declaritive thing unto themselves)
17:01 <sdumitriu1> The problem with checkstyle:off is that it's all or nothing
17:01 <sdumitriu1> And maintaining the checkstyle exclusion file isn't that wrong
17:02 <cjd> meh, it's the old makefile mess again
17:02 <sdumitriu1> If you refactor the code so that the exception doesn't apply anymore (rename a function), but the rule is still broken, the build will fail
17:02 <sdumitriu1> And you'll know you have to update one or two lines in the XML
17:02 <cjd> but you refactor the code so the rule is nolonger broken, then you get no feedback
17:03 <cjd> and you have a file full of stuff that nobody's sure which of it is necessary
17:03 <sdumitriu1> In theory that shouldn't be that often required, since exclusions are supposed to be exceptional cases
17:03 <cjd> like #includes which nobody wants to spend the time to try deleting and see if the build fails as a result
17:03 <sdumitriu1> Not a frequent practice
17:03 <sdumitriu1> If we have a larger exclusion config than the java code itself, then you have a bigger problem
17:03 <cjd> well we're going to have to kill off some rules then :)
17:04 <sdumitriu1> Ironically, I'm working on enabling a few more rules right now
17:04 <cjd> smart ones or dumb ones?
17:04 <sdumitriu1> (including the RequireThis rule, which you're right, isn't active yet)
17:04 <cjd> <loaded question>
17:04 <cjd> RequireThis +1
17:04 <cjd> smart, low cost of compliance, tells the reader a lot
17:05 <sdumitriu1> RequireThis is broken now :( https://github.com/checkstyle/checkstyle/pull/274
17:05 <cjd> but don't make it harder to contribute !
17:05 <cjd> if you're going to make it harder, we need to throw some of these things overboard so it's at least a balanced budget
17:06 <sdumitriu1> Is there any codestyle rule that doesn't make it harder to _write_ code?
17:06 <sdumitriu1> All codestyle rules are supposed to make it easier to _read_ code
17:07 <sdumitriu1> At various costs at write time
17:07 <cjd> well some of the bad javadoc ones make it harder to read
17:07 <cjd> and "you must split this up *somewhere*"
17:07 <sdumitriu1> But code, like most data, is more often read than written
17:08 <cjd> duplicate strings and private/package-private javadoc requirements are burning me the most atm
17:08 <sdumitriu1> The problem is that most devs don't realize just how much more code reading they do
17:08 <sdumitriu1> Send a vote for those
17:08 <sdumitriu1> We'll let the community decide
17:08 <cjd> I want to keep the discussion open a bit longer because some people (vmassol) want to discuss it but have not yet had the time
17:09 <cjd> rushing a vote kind of forces a premature -1
17:09 <sdumitriu1> Sure
17:09 <sdumitriu1> But I'd say the veto rule shouldn't apply here
17:10 <sdumitriu1> It's not a design decision, it's a "community feeling good" decision, and no member should dictate how the community is supposed to feel good
17:10 <cjd> that's an interesting point
17:10 <msmeria> has quit
17:10 <cjd> still we shouldn't go forward with something "scary" unless we all feel like we're together on it
17:10 <sdumitriu1> "You'll write clean* code, and you'll enjoy it!" (*clean as defined by the BDFL in this 142 pages document)
17:11 <cjd> well yeah, but we're talking about repealing a law, bearuacracies are scared to death of repealing laws
17:11 <cjd> There will be anarchy in the streets!
17:12 <sdumitriu1> If we're not following the rules because we believe in them, then we're doing it wrong
17:12 <cjd> good point as well
17:13 <cjd> I'm hugely +1 for RequireThis
17:13 <sdumitriu1> You can collect my aphorisms and make a book, we'll split the revenue fairly
17:13 <cjd> heh
17:14 <cjd> but I think it needs to balance with some relaxation of rules
17:14 <sdumitriu1> Sure
17:15 <cjd> and I can squeeze most of the in-package private stuff into private/package-private so I can probably sneek by without even the detection of /internal/ package
17:16 <cjd> just change class/method javadoc rule to be >=protected
17:16 <DarkKnightCZ> has joined #xwiki
17:18 <Slashman> has joined #xwiki
17:18 <cjd> sdumitriu1: RequireThis for both fields and methods or only for fields?
17:19 <sdumitriu1> Only fields
17:19 <cjd> my feeling is only fields since java is -braindead- err not able to store function pointers
17:20 <sdumitriu1> The default for checkstyle is only fields
17:20 <sdumitriu1> And they usually have good defaults
17:20 <sdumitriu1> Not always
17:20 <cjd> do you know a way to quickly bring the whole project into compliance?
17:20 <Enygma`> vmassol: any idea why jobs just stop executing?
17:20 <Enygma`> http://ci.xwiki.org/job/xwiki-enterprise-test-extension/10/console
17:20 <Enygma`> http://ci.xwiki.org/job/xwiki-enterprise-test-pageobjects-6.2.x/2/console (did not even start well)
17:22 <sdumitriu1> cjd: In Eclipse: Alt+Shift+S - F followed by Alt+Shift+S - U
17:22 <sdumitriu1> On the src/main/java directory
17:22 <sdumitriu1> That fixes most of the style issues, but not the deeper issues
17:23 <cjd> I'll need some help with that, I don't have eclipse..
17:24 <cjd> although I could actually run checkstyle and parse the output into a script to fix it :)
17:31 <sdumitriu1> Misspoke, RequireThis applies by default to methods, we must explicitly disable that
17:32 <DarkKnightCZ> has left #xwiki
17:35 <ClemensR> has left #xwiki
18:01 <sdumitriu1> I wonder if anybody would complain if I break the build...
18:01 <cjd> has quit
18:02 <evalica> has quit
18:04 <Enygma`> has quit
18:06 <tmortagne> sdumitriu1: these days you could get lucky and have jenkins blamed for it before you fix it
18:14 <momomomomo> has quit
18:31 <momomomomo> has joined #xwiki
18:34 <Denis> has quit
18:36 <tmortagne> has quit
18:42 <sdumitriu1> nexus unavailable?
18:43 <vmassol> yes momentarily
18:43 <vmassol> I think GF restarted the machine
18:44 <vmassol> pinging him
18:51 <gdelhumeau> has quit
18:53 <vmassol> it's back now
19:18 <Denis> has joined #xwiki
19:19 <momomomomo> has quit
19:36 <woshilapin> has quit
19:43 <sburjan> has joined #xwiki
19:44 <mflorea> has joined #xwiki
19:48 <abusenius> has joined #xwiki
19:52 <Denis> has quit
20:01 <OSIMasson> has joined #xwiki
20:03 <vmassol> has quit
20:13 <OSIMasson> has left #xwiki
20:28 <abusenius_> has joined #xwiki
20:30 <abusenius> has quit
20:39 <momomomomo> has joined #xwiki
22:34 <momomomomo> has quit
22:39 <momomomomo> has joined #xwiki
23:21 <vmassol> has joined #xwiki
23:36 <D-Spair> has quit
23:36 <D-Spair> has joined #xwiki
23:37 <mflorea> has quit
23:46 <xwikiorg_guest_8> has joined #xwiki
23:47 <xwikiorg_guest_8> Hey! >:-) Is the maven repo down?
23:47 <vmassol> yes
23:47 <vmassol> we're haviing some problems ATM
23:47 <vmassol> and the infra guys are sleeping right now… sorry….
23:47 <xwikiorg_guest_8> I am doing my first xwiki install... and couldn't figure out what was happening!
23:48 <vmassol> yeah bad timing
23:48 <xwikiorg_guest_8> Bummer
23:48 <vmassol> it usually doesn't happen...
23:48 <xwikiorg_guest_8> Well, I guess I can try again tomorrow!
23:48 <vmassol> yeah that would be better, again sorry about that
23:49 <xwikiorg_guest_8> Unless there is anywhere else to get the xwiki-enterprise-ui-mainwiki-all-6.1.xar?
23:49 <xwikiorg_guest_8> I could load it through the admin panel with superadmin...
23:49 <xwikiorg_guest_8> No worries ;-)
23:49 <vmassol> well you could get it from OW2 but it's best to use the Distribution Wizard
23:50 <xwikiorg_guest_8> That sounds like a better plan.
23:50 <xwikiorg_guest_8> (Waiting, that is)
23:50 <xwikiorg_guest_8> I have a different question ;-)
23:50 <vmassol> hehe
23:51 <xwikiorg_guest_8> So I am hoping to use basic auth to log in users...
23:51 <xwikiorg_guest_8> I am setting us xwiki as an internal tool and want it behind out production environment... and I would like to use our existing auth.
23:51 <xwikiorg_guest_8> (php session based)
23:52 <xwikiorg_guest_8> Ever heard of somethgin like this?
23:52 <xwikiorg_guest_8> Ideally, I could log users into xwiki with their existing php session.
23:52 <vmassol> xwiki supports lots of auths. Don't know about that php session based on
23:52 <vmassol> but worst case you can write your own auth
23:53 <xwikiorg_guest_8> I guess I could write a custom authenticator to hit a remote database?
23:53 <vmassol> some info here: http://platform.xwiki.org/xwiki/bin/view/AdminGuide/Authentication
23:53 <xwikiorg_guest_8> Alright, I'll git it a shot
23:53 <xwikiorg_guest_8> Thanks!
23:53 <abusenius_> has quit
00:29 <momomomomo> has quit
01:04 <momomomomo> has joined #xwiki
01:21 <abusenius> has quit
02:30 <sburjan> has quit
03:07 <OSIMasson> has quit
03:27 <Emenems> has quit
03:32 <Emenems> has joined #xwiki
03:42 <momomomomo> has quit
03:46 <OSIMasson> has joined #xwiki
03:50 <Denis> has joined #xwiki
03:53 <Emenems> has quit
03:59 <Emenems> has joined #xwiki
04:04 <Denis> has quit
04:07 <OSIMasson> has left #xwiki
04:30 <xwikibot> has joined #xwiki
06:23 <DarkKnightCZ> has joined #xwiki
06:28 <DarkKnightCZ> has quit
07:31 <andreeapopescu> has joined #xwiki
07:48 <woshilapin> has quit
08:05 <mflorea> has joined #xwiki
08:17 <vmassol> has joined #xwiki
08:22 <Denis> has joined #xwiki
08:33 <KermitTheFragger> has joined #xwiki
08:37 <woshilapin> has joined #xwiki
08:47 <cjd> has joined #xwiki
08:53 <tmortagne> has joined #xwiki
08:56 <abusenius> has joined #xwiki
09:02 <DarkKnightCZ> has joined #xwiki
09:02 <abusenius> has quit
09:26 <Enygma`> has joined #xwiki
09:30 <evalica> has joined #xwiki
09:30 <msmeria> has joined #xwiki
09:41 <woshilapin> has quit
09:41 <woshilapin> has joined #xwiki
09:42 <ClemensR> has joined #xwiki
09:57 <cjd> Denis: how do you like my way of getting my code reviewed?
10:01 <Denis> cjd: if you are talking about your checkstyle discussion, it is terribly efficient :)
10:01 <cjd> ;)
10:02 <Denis> I would have much to say as well, but I lack time actually
10:02 <cjd> yeah, vmassol does as well
10:02 <Denis> but checkstyle is not all or nothing
10:02 <cjd> +1 that
10:02 <Denis> you can exclude rules line by lines
10:03 <Denis> it just take time to maintain
10:03 <woshilapin> And pollute code with lines that have nothing to do with code but with checkstyle?
10:03 <cjd> I feel some checks are absolutely necessary, evidenced by the fact that I wrote my own codestyle verifier for cjdns (C)
10:03 <Denis> no, you put that in a separate xml config
10:03 <woshilapin> Ah, OK
10:04 <woshilapin> Seems nicer, indeed
10:04 <cjd> ok that's a bit worse
10:04 <woshilapin> (I didn't know you can do it)
10:04 <cjd> it's like Makefiles with per-file rules, very bad
10:05 <cjd> you need to synchronize them with the state of the files
10:05 <Denis> https://github.com/xwiki/xwiki-commons/blob/master/xwiki-commons-core/xwiki-commons-crypto/xwiki-commons-crypto-common/src/main/checkstyle/checkstyle-suppressions.xml
10:05 <tmortagne> separate xml config has it's drawnback too, it makes it's a bit of a pain to maintain it
10:05 <Denis> this is an example
10:05 <Denis> yes, I fully agree with you Thomas
10:06 <cjd> I use a comment notation in cjdns.. // CHECKFILES_IGNORE it expects a ; here but this is a macro
10:06 <tmortagne> yes I perefer comments system too
10:06 <cjd> that way all information about the code is contained
10:06 <tmortagne> s/perefer/prefer/
10:06 <Denis> this is why my filling, that is only worst it for useful checkstyle check, those that may really reveal poor code that is not obvious
10:06 <tmortagne> when you remove of fix the failure you don't forget to update the excludes
10:07 <cjd> +1
10:07 <Denis> at the same time, comment link you checker with your code, which is not so nice
10:08 <cjd> the code will still compile without the checker
10:08 <cjd> it's not as if I created a magical annotation which needs to be preprocessed away
10:08 <gdelhumeau> has joined #xwiki
10:09 <tmortagne> cjd: note that I think there is some support for things like //CHECKSTYLE:OFF/ON in checkstyle
10:09 <tmortagne> need to get the documentation
10:09 <cjd> it's disabled for us, we had a discussion ages ago about enabling it and sdumitriu1 was not favorable to it at the time
10:10 <cjd> but that was like 2010
10:10 <tmortagne> yep that's it, you need to add a filter for it
10:10 <tmortagne> http://checkstyle.sourceforge.net/config.html
10:11 <tmortagne> <module name="Checker"> ... <module name="SuppressionCommentFilter"/> ... </module>
10:11 <tmortagne> <module name="Checker">
10:11 <tmortagne> ...
10:11 <tmortagne> <module name="SuppressionCommentFilter"/>
10:11 <cjd> except //CHECKSTYLE:OFF begin big function ... means that big function gets *none* of the checks
10:12 <tmortagne> you don't have to put that around the whole method
10:12 <tmortagne> it's around any code as far as I understand
10:12 <cjd> if it's a statement count issue?
10:12 <tmortagne> if it's a statement count issue of course
10:12 <cjd> I suppose statement count would actually be reported at the first line of the function so you could do
10:12 <tmortagne> maybe you can put that around the signature
10:12 <cjd> //CHECKSTYLE:OFF // big function
10:13 <cjd> function (arg, arg2) {
10:13 <cjd> // yeah, that
10:13 <tmortagne> yep
10:24 <mflorea> has quit
10:29 <bouquetf_> has quit
10:37 <bouquetf> has joined #xwiki
10:58 <woshilapin> Hi devs, I have a problem with a 'unknown macro'. This macro is registered since I can access it with Admin user but not with a non-admin user
10:59 <woshilapin> I checked that the macro is registered with a user with programming rights
10:59 <woshilapin> (XWiki.Admin in this case)
11:00 <woshilapin> Something that may be specific to my case, there is one macro A that is using other nested macro B
11:00 <woshilapin> The macro A is well displayed, but the macro B show the 'unknown macro' problem
11:00 <woshilapin> Any idea?
11:02 <woshilapin> (s/I checked that the macro is registered with a user with programming rights/The macros are registered with XWiki.Admin which is in XWikiAdminGroup/)
11:02 <woshilapin> (but I suppose that's equivalent)
11:03 <woshilapin> at least, I'm looking for an idea on how to investigate (where can I find useful log for example)
11:12 <cjd_> has joined #xwiki
11:13 <tmortagne> woshilapin: make sure you macro has the right visibility
11:13 <tmortagne> sounds like your macro has user visibility
11:14 <tmortagne> should be wiki or global
11:14 <woshilapin> "right visibility"? What do you mean?
11:14 <woshilapin> I have only one wiki 'wiki'
11:14 <tmortagne> look at your macro with object editor
11:14 <tmortagne> you have a field in there indicating who should see that macro
11:15 <woshilapin> Indeed :-)
11:15 <tmortagne> maybe you chose user (that should be named "macro author" probably) by mistake
11:15 <tmortagne> user visibility allow a user that does not have programming or admin right to register a wiki macro for himself
11:16 <woshilapin> OK
11:16 <tmortagne> * wiki=admin right
11:16 <tmortagne> * global= programming right
11:16 <woshilapin> But why 'user visibility' is default?
11:16 <cjd> has quit
11:16 <woshilapin> (by the way, it works, thanks)
11:16 <tmortagne> there is no real default, I guess it's just first in the list of available possibility
11:17 <tmortagne> s/possibility/possibilities/
11:17 <woshilapin> Yes
11:17 <cjd_> is now known as <cjd>
11:17 <woshilapin> OK, anyway, thanks for the suppoer
11:18 <woshilapin> support*
11:24 <gdelhumeau> I don't reproduce http://ci.xwiki.org/job/xwiki-enterprise-test-wysiwyg/lastBuild/org.xwiki.enterprise$xwiki-enterprise-test-wysiwyg/testReport/org.xwiki.test.wysiwyg/StandardFeaturesTest/testItalics/ manually
12:02 <cjd> has quit
12:04 <cjd> has joined #xwiki
12:05 <cjd> is now known as <Guest41316>
12:07 <KermitTheFragger> has quit
12:41 <Enygma`> I have my doubts about those tests
12:41 <Enygma`> they were not failing for the 6.2 rc1 release
12:41 <Enygma`> so my take is that they are flukes
12:42 <Enygma`> but we really need to get these flukes into control
12:42 <Enygma`> they are very annoying
12:42 <Enygma`> starting to handle test failures too locall
12:51 <Enygma`> vmassol: any idea about http://ci.xwiki.org/job/xwiki-enterprise-test-ui/lastBuild/org.xwiki.enterprise$xwiki-enterprise-test-ui/testReport/org.apache.maven.surefire.junit4/JUnit4Provider/org_xwiki_test_ui_AllTests/ ?
12:59 <vmassol> Enygma`: what I'd like to do
12:59 <vmassol> modify start_xwiki.sh to stop any running XE process (if a script parameter is passed)
12:59 <vmassol> the first problem here is "Error binding monitor port 8079: java.net.BindException: Cannot assign requested address"
13:00 <vmassol> which could be caused by either a running XE or by an issue with my check in XwikiExecutor
13:00 <vmassol> (which I would remove after I modify start_xwiki.sh since it wouldn't not be needed any longer)
13:00 <vmassol> now
13:00 <vmassol> the NPE is easy to fix, doing it now
13:01 <vmassol> I introduced the problem yesterday actually
13:02 <vmassol> done
13:13 <Denis> has quit
13:28 <Denis> has joined #xwiki
13:33 <Enygma`> not reproducing locally: http://ci.xwiki.org/job/xwiki-enterprise-test-ui-6.2.x/lastBuild/org.xwiki.enterprise$xwiki-enterprise-test-ui/testReport/org.xwiki.test.ui/EditObjectsTest/testPropertyDisplayersForNewObjects/
13:33 <Enygma`> used @Intermittent(repetition=10)
13:43 <vmassol> Hi guys, does anyone know how I can get the process id and continue execution in a shell script?
13:43 <vmassol> I know I can use $! to get it
13:43 <vmassol> currently in our start_xwiki.sh wehave:
13:43 <vmassol> java $XWIKI_OPTS $3 $4 $5 $6 $7 $8 $9 -jar $JETTY_HOME/start.jar ${JETTY_HOME}/etc/jetty.xml ${JETTY_HOME}/etc/jetty-*.xml
13:43 <vmassol> so it's blocking and waiting till the JVM exits
13:43 <vmassol> I'd like to get the pid and save it in a lock file
13:44 <vmassol> but also continue blocking execution till the JVM exits
13:44 <Guest41316> mmm trickey one
13:44 <vmassol> the "fg" command depends on the shell I guess
13:44 <vmassol> so I cannot use & and then fg
13:44 <vmassol> actually maybe
13:44 <Guest41316> don't save the java process, save the shell process
13:45 <Guest41316> echo $$ > pid
13:45 <Guest41316> java .......
13:45 <Guest41316> the script will keep running as long as java is still running
13:45 <vmassol> $$ represents the shell pid?
13:45 <Guest41316> yeap
13:45 <vmassol> cool
13:45 <vmassol> that should work then thanks
13:45 <Guest41316> kill -9 $$
13:45 <vmassol> trying
13:46 <vmassol> I won't try that! :)
13:46 <Guest41316> kill -11 `pidof apache2`
13:46 <Guest41316> (11 is a sigsegv, synthetic segmentation fault)
13:54 <DarkKnightCZ> has left #xwiki
13:58 <vmassol> hmmm
13:58 <vmassol> doesn't work
13:58 <vmassol> kill the shell process doesn't kill the JVM apparently
13:58 <vmassol> s/kill/killing
13:58 <vmassol> (Guest41316)
14:01 <Guest41316> ahh right
14:01 <vmassol> actually I can stop using our stop_xwiki.sh script…
14:02 <Guest41316> reconnecting to get my nick back...
14:03 <Guest41316> has quit
14:03 <cjd> has joined #xwiki
14:04 <Enygma`> vmassol: does the first answer on this question help http://stackoverflow.com/questions/1821968/pid-of-last-started-process-in-bash-script
14:04 <vmassol> ah it was you cjd ;)
14:04 <cjd> yeah, my nick is set to enforce mode so nobody else can use it unless they auth
14:05 <Enygma`> some_program &
14:05 <Enygma`> some_pid=$!
14:05 <cjd> ahh indeed, 'wait'
14:05 <vmassol> actually what I'm doing now is:
14:05 <Enygma`> in your case..
14:05 <Enygma`> java ...
14:05 <Enygma`> xwiki_pid=$!
14:05 <vmassol> run stop_xwiki.sh
14:06 <vmassol> however I need to modify stop_xwiki.sh so that it kills the process if it doesn't stop by itself
14:06 <cjd> +1
14:06 <vmassol> so I still need to store the real JVM pid
14:06 <vmassol> I'll try the wait
14:06 <vmassol> is that standard on all unixes?
14:07 <cjd> hmm
14:07 <cjd> System.getPid() -> pid.file ?
14:08 <cjd> does start_xwiki.sh need to do anything after xwiki shuts down?
14:09 <vmassol> remove the lock file
14:09 <cjd> ok
14:10 <vmassol> I have something strange
14:10 <cjd> yeah, +1 the wait method
14:11 <vmassol> this is what I have:
14:11 <vmassol> https://gist.githubusercontent.com/vmassol/2f1b6b9c3207dd17ed3b/raw/cfa4d26b55198eabe335aabe69eb969ec3a11763/gistfile1.txt
14:11 <cjd> looks like wait is a shell builtin so anywhere there is bash, there is wait
14:11 <vmassol> now when I exeucte the second instance
14:12 <vmassol> the first one stops fine
14:12 <vmassol> but in the console of the second one I get:
14:12 <vmassol> https://gist.githubusercontent.com/vmassol/73707faca38befe362f8/raw/cfc93f2ac2bbb5dcefa0c7750ab7c7d55ff15181/gistfile1.txt
14:12 <vmassol> not sure why I get "Error binding monitor port 8079: java.net.BindException: Address already in use"
14:12 <cjd> cool
14:12 <vmassol> since the stop_xwiki.sh is waiting....
14:13 <vmassol> oh yes
14:13 <vmassol> hmm no
14:13 <vmassol> I don't know
14:15 <vmassol> if the JVM is stopped completely shouldn't it mean that the ports are freed? can they overlast the JVM?
14:16 <vmassol> if so then it's more complex because I'd need to do a loop and check if someone is holding the port
14:16 <vmassol> (in a standard way)
14:16 <cjd> https://ezcrypt.it/j39n#opPHBmCcMRAZoxEJsuWjmRgH
14:16 <cjd> no, they cannot outlast the process
14:17 <cjd> there's always a chance that another process holds the port
14:17 <vmassol> but the JVM doesnet spawn another process AFAIK
14:17 <cjd> oh oops bug
14:19 <cjd> https://ezcrypt.it/k39n#og8g5xXV0cN8FKiopEbUdr1x
14:19 <cjd> that loop might not be 100% right
14:20 <vmassol> ok but that's for later
14:20 <vmassol> ATM I need to understand why I have this "Error binding monitor port 8079: java.net.BindException: Address already in use" error
14:20 <cjd> anyway that's the most resilient way to stop the process
14:20 <vmassol> (and how to fix it)
14:21 <cjd> hmm
14:21 <cjd> race condition?
14:21 <cjd> oh no, it had been up for an hour
14:23 <ClemensR> is it possible that yet another unrelated (or old, defunct) process is still listening at 8079 ?
14:23 <vmassol> with a sleep 2 it still fails but with a sleep 10 it works
14:24 <cjd> ClemensR: we did a `netstat -lnpt` and it didn't show the port as bound
14:24 <cjd> so I'm confused
14:24 <vmassol> cjd: that's different
14:24 <ClemensR> ah, ok, then I join the 'confused' group :/
14:24 <vmassol> cjd is talking about some job running on ci
14:25 <vmassol> but here I'sm experimenting locally
14:25 <vmassol> just running sh start_xwiki.sh twice
14:25 <cjd> mm /nod
14:25 <vmassol> I want the second one to stop the first one
14:25 <cjd> mmm I wouldn't
14:26 <vmassol> it seems that the port is really freed only when the shell exits for the first one
14:26 <vmassol> or something like that
14:26 <cjd> oh
14:26 <cjd> hang on a sec
14:28 <DarkKnightCZ> has joined #xwiki
14:28 <cjd> http://stackoverflow.com/questions/14388706/socket-options-so-reuseaddr-and-so-reuseport-how-do-they-differ-do-they-mean-t
14:29 <cjd> ``The question is, how does the system treat a socket in state TIME_WAIT? If SO_REUSEADDR is not set, a socket in state TIME_WAIT is considered to still be bound to the source address and port and any attempt to bind a new socket to the same address and port will fail until the socket has really been closed, which may take as long as the configured Linger Time.''
14:30 <cjd> Write once, refactor everywhere
14:30 <vmassol> ok I think I know
14:31 <vmassol> it seems the stop returns immediately
14:31 <vmassol> I'm printing the date before and after the actual stop, ie the java execution to stop jetty
14:31 <cjd> mhm
14:31 <vmassol> and it seems too fast to be true
14:31 <vmassol> before stop: Thu Sep 11 14:30:24 CEST 2014
14:31 <vmassol> after stop: Thu Sep 11 14:30:24 CEST 2014
14:31 <vmassol> not even 1 second…
14:32 <vmassol> to start a JVM , run the stop and stop the JVM
14:32 <cjd> doesn't it take 1 second to start the JVM?
14:32 <cjd> <trollface>
14:32 <vmassol> I can't know since it's blocking :)
14:33 <cjd> anyway making stop_xwiki.sh work well I'm in favor of, making start_xwiki.sh automatically call stop_xwiki.sh is too magical IMO
14:33 <vmassol> so I think the stop_xwiki.sh finishes before it's really stopped
14:34 <vmassol> s/before it's/before jetty/
14:34 <cjd> should not be possible to finish before the lock is released
14:34 <cjd> *file deleted
14:34 <vmassol> yes but deleting the file take a split second
14:34 <vmassol> that'snot a problem
14:34 <cjd> but it only happens after xwiki is actually stopped
14:35 <vmassol> apparently not
14:35 <vmassol> otherwise it would work :)
14:35 <cjd> are you using my script or yours?
14:35 <vmassol> mine
14:35 <cjd> I'm not sure of the stuff in that loop
14:35 <cjd> hard for me to follow
14:36 <cjd> try mine :)
14:36 <vmassol> I don't want to do any kill ATM
14:36 <vmassol> that's a fallback only
14:36 <vmassol> if it fails to stop but we need to give it a chance to stop properly first
14:36 <cjd> ok then comment out the kill lines
14:36 <vmassol> you changed the loop?
14:36 <cjd> yeah
14:37 <cjd> https://ezcrypt.it/k39n#og8g5xXV0cN8FKiopEbUdr1x
14:37 <vmassol> ok I'll try but I don't see why it would make a difference
14:38 <vmassol> same
14:38 <vmassol> which I expected
14:38 <cjd> hrm
14:39 <cjd> sec
14:39 <vmassol> there's no reason at all
14:39 <vmassol> it would change anything
14:39 <vmassol> it's not about the deletion of the lock file
14:40 <cjd> echo ">>$XWIKI_LOCK_FILE<< === >>`cat $XWIKI_LOCK_FILE`<<"
14:41 <vmassol> wdym?
14:41 <vmassol> the echo works fine
14:41 <ClemensR> I cannot reproduce locally
14:42 <vmassol> ClemensR: here are my full start_xwiki.sh and stop.xwiki.sh
14:42 <cjd> making sure the lock is indeed there
14:43 <vmassol> start_xwiki.sh: https://gist.githubusercontent.com/vmassol/b10524449403c7eaa09d/raw/bbbe39063b191c5eb6e328a208509e9e100275a6/gistfile1.txt
14:45 <vmassol> stop_xwiki.sh: https://gist.githubusercontent.com/vmassol/32a6a2bcdf4a5849c87c/raw/9b081ecad833deb7140058799c0627e97d544e3d/gistfile1.txt
14:45 <vmassol> what I do:
14:45 <vmassol> in a shell run: sh start_xwiki.sh
14:45 <vmassol> wait till it's fully started
14:45 <vmassol> thenin a second shell in the same directory
14:45 <vmassol> run: sh start_xwiki.sh
14:46 <ClemensR> ah, lock file is in /var/tmp/ ... maybe that makes a difference (though it shouldn't)
14:46 <vmassol> I'm pretty sure the real proboem is that when the java process exits
14:46 <vmassol> the port are not fully freed yet on my OS
14:46 <vmassol> (Mac)
14:46 <vmassol> it may be os-dependent
14:47 <cjd> while [ "$timer" -lt 100 -a -e XWIKI_LOCK_FILE ]; do
14:47 <cjd> while [ "$timer" -lt 100 -a -e $XWIKI_LOCK_FILE ]; do
14:47 <vmassol> yes
14:47 <vmassol> that's just a typo in my paste
14:47 <cjd> and it's not in your file?
14:47 <vmassol> I reverted your loop to my first loop before pasting
14:48 <vmassol> (and I made a typo)
14:49 <vmassol> when stop_xwiki.sh is called from the second eeuction on start_xwiki.sh
14:49 <ClemensR> I dont have that jetty/etc/jetty-*.xml, but I guess that is no problem
14:49 <cjd> [ ! -e $XWIKI_LOCK_FILE ] && echo 'lock missing' && exit 0;
14:49 <cjd> java $XWIKI_OPTS -jar $JETTY_HOME/start.jar --stop
14:49 <vmassol> I can see that my first shell is still active about 4-5 seconds after
14:49 <vmassol> (I have a spinning symbol on my shell command tab)
14:50 <vmassol> first shell is still active == first shell is still executing something
14:51 <ClemensR> now I can reproduce, one :)
14:52 <vmassol> ok good
14:52 <vmassol> wdym by "one"?
14:53 <cjd> ls /var/tmp
14:53 <cjd> is the lock there?
14:53 <vmassol> yes that works fine
14:53 <vmassol> again
14:53 <cjd> ok
14:54 <vmassol> I don't see how the lock file can be an issue
14:54 <vmassol> it's not about the lock file
14:54 <vmassol> but about the port being freed
14:55 <cjd> sec
14:55 <ClemensR> the interesting thing is that the script from the jetty-hsqldb is working w/o problems
14:55 <vmassol> you mean you can start 2 instances of jetty on the same port?
14:55 <vmassol> :)
14:56 <cjd> user@blk1:~/xwiki-enterprise-jetty-hsqldb-6.2-SNAPSHOT$ time ./_stop_xwiki.sh && ps -ef | grep java && netstat -lnpt
14:56 <cjd> lock missing
14:57 <cjd> lock [/var/tmp/xwiki-.lck] missing
14:57 <ClemensR> I meant, when the four "stop when running" lines are pasted into the script from teh original distro
14:57 <cjd> user@blk1:~/xwiki-enterprise-jetty-hsqldb-6.2-SNAPSHOT$ cat ./_stop_xwiki.sh | grep JETTY_PORT
14:57 <cjd> XWIKI_LOCK_FILE="${XWIKI_LOCK_DIR}/xwiki-${JETTY_PORT}.lck"
14:57 <cjd> user@blk1:~/xwiki-enterprise-jetty-hsqldb-6.2-SNAPSHOT$
14:58 <cjd> JETTY_PORT == ""
14:58 <ClemensR> for me the stop-script exits a second before the old process really quits
14:58 <ClemensR> ah, thats it, JETTY_PORT is not set from $1 :)
14:59 <ClemensR> script actually expects JETTY_STOP_PORT as $1
15:00 <vmassol> oh yes
15:00 <vmassol> good catch
15:00 <vmassol> my bad
15:00 <vmassol> fixing that and trying again
15:00 <ClemensR> using XWIKI_LOCK_FILE="${XWIKI_LOCK_DIR}/xwiki-${JETTY_STOP_PORT}.lck" in both files works better
15:01 <vmassol> I need: /bin/sh stop_xwiki.sh $JETTY_STOP_PORT
15:01 <ClemensR> uh, gotta go, meeting at 1500 here/ well, issue found in time :)
15:01 <vmassol> thanks!
15:01 <ClemensR> has left #xwiki
15:02 <vmassol> hmmm still the same probem :(
15:02 <cjd> https://ezcrypt.it/l39n#EKadpZ0QB3LbsfpBGtpX3BYN
15:03 <vmassol> that's normal cjd
15:03 <vmassol> the 3 processes execute in parallel
15:03 <vmassol> since the stop takes the most time
15:03 <cjd> mmm no
15:03 <vmassol> it's normal that netsat would show it still bound
15:03 <cjd> oh
15:03 <vmassol> && means execute in // no?
15:04 <cjd> no, execute after
15:04 <cjd> but only if the previous one succeeds
15:04 <vmassol> ok then indeed it's what I said
15:04 <cjd> like if (something() && somethingElse())
15:04 <vmassol> when the jvm exits the port are not freed yet
15:05 <cjd> actually I think the timer just timed out
15:05 <cjd> because it was 12 seconds
15:05 <cjd> and the next time I ran it, it worked
15:05 <cjd> better make it 20 before trying to kill it
15:06 <cjd> https://ezcrypt.it/m39n#QrSGrJtW0NSTh4LRIlO2upRx
15:06 <cjd> second time works
15:06 <cjd> hey
15:06 <cjd> who installed cryptocurrency in my vm ?!
15:10 <vmassol> hmm it worked for me when I tried:
15:10 <vmassol> time sh stop_xwiki.sh && ps -ef | grep java && lsof -iTCP:8080 -sTCP:LISTEN
15:11 <vmassol> but it's strange
15:11 <vmassol> ps still shows the process
15:12 <vmassol> but the lsof doesn't show the port as bound for 8080
15:12 <vmassol> checking 8079 now
15:12 <OSIMasson> has joined #xwiki
15:12 <vmassol> ok for 8079 it's still boind
15:12 <cjd> increase the timout
15:13 <vmassol> what timeout?
15:13 <vmassol> in stop_xwiki.sh?
15:13 <cjd> yeap
15:13 <cjd> or put an echo below it saying "timed out"
15:13 <cjd> so you know if it times out
15:14 <vmassol> same
15:14 <vmassol> not sure what the timeout would change
15:14 <cjd> user@blk1:~/xwiki-enterprise-jetty-hsqldb-6.2-SNAPSHOT$ ./_start_xwiki.sh
15:14 <cjd> An XWiki instance is already running on port 8080, stopping it!
15:14 <cjd> java.net.ConnectException: Connection refused
15:14 <cjd> this is a bad idea
15:14 <cjd> you should not add that part to start_xwiki.sh
15:15 <cjd> you should check if the process is running and if it's not, clear out the stale lock
15:15 <vmassol> actually I read up a lot on that
15:15 <vmassol> and the consensus is that it's bad to do what you say
15:15 <vmassol> and instead the good practice is to save teh pid
15:16 <vmassol> because doing a ps might not always work as expected
15:16 <gdelhumeau> do we still have the "undelete" right in XWiki?
15:16 <gdelhumeau> I don't see it in the <select> tag
15:16 <gdelhumeau> on a XWikiGlobalRights object
15:16 <vmassol> so I wanted to avoid doing a ps
15:17 <vmassol> in any case the big problem is that the port is not freed when stop_xwiki.sh exits
15:17 <vmassol> so the current stop_xwiki.sh doesn't work
15:17 <vmassol> the wait at the end doesn't do much in practice apparently
15:18 <vmassol> I'm surprised that the port outlives the JVM
15:18 <vmassol> s/JVM/JVM process/
15:19 <vmassol> a simple sleep 10 at the end of stop_Xwiki.sh would work but it's not very nice
15:19 <cjd> 15:15 < vmassol> and instead the good practice is to save teh pid <-- saving the pid is brittle but it's standard practice
15:20 <cjd> what I don't like is that if xwiki is already running, you try to stop it
15:20 <cjd> when the admin didn't ask you to stop anything
15:20 <vmassol> that's not my point
15:20 <cjd> that will lead to angry admins, I'm willing to bet on it
15:20 <vmassol> you're talking details
15:20 <vmassol> and that's absoutely not my goal
15:20 <vmassol> as I mentioned twice already above
15:20 <vmassol> :)
15:20 <vmassol> my goal is to pass a param to the start script to control that behavior
15:21 <vmassol> I just haven't implemented it yet because I don't care at this stage
15:21 <vmassol> ofc we shouldn't stop a running XE if the user has not asked for it!!
15:21 <vmassol> I'd be angry if someone did that to me :)
15:22 <cjd> ahh ok
15:22 <octo> Hello :) qq, XWikiDocument#getXObject() is supposed to return the BaseObject of the document, right?
15:22 <vmassol> octo: the first xobject actually
15:22 <vmassol> there can be a lot...
15:23 <vmassol> depends what param you pass actually
15:24 <vmassol> if you use it without params then it's the first one
15:24 <octo> vmassol: getXObject() calls getXObject(DocumentReference)
15:24 <tmortagne> octo: XWikiDocument#getXObject() is pretty random (unless there is only one object of course), better indicate at least the type of object you want of you know it
15:25 <vmassol> yes but check getXObject(DocumentReference classReference)
15:25 <tmortagne> s/of/if/
15:25 <vmassol> you'll see it iterate on all objects
15:25 <octo> Okay, maybe take a step back.
15:25 <vmassol> and stops on the frist one
15:25 <octo> What I want to get at is the Guid or the document
15:25 <tmortagne> there is no such thing as "Guid or the document"
15:25 <octo> As far as I see the BaseObject is the only class exposign the getGuid() method
15:25 <vmassol> getId
15:25 <vmassol> ?
15:25 <vmassol> getId()?
15:26 <vmassol> XWikiDocument.getId()
15:26 <vmassol> not sure why you'd need this though....
15:26 <tmortagne> be carefull that getId() is local
15:26 <tmortagne> can't you use the document reference ?
15:27 <octo> I'd like to have an identifier of a wiki page that is final, i.e. doesn't change when the page is renamed
15:28 <tmortagne> there is nothing like that
15:28 <tmortagne> what you could use is the value of some field of a custom object you put in those pages
15:28 <octo> Ugh, that will make this complicated :\
15:29 <tmortagne> that way that object field value won't change if you rename the document and you can search the document based on it
15:29 <tmortagne> it's like introducing yourself a uid to a document
15:29 <octo> That sounds like a plan
15:30 <octo> tmortagne: Can I listen for DocumentCreated events and just add such an object automatically in that handler?
15:31 <vmassol> cjd: ok I'll go the ps way then…
15:31 <vmassol> and do some loop till the process is no longer there
15:31 <vmassol> I don't see any other option....
15:32 <tmortagne> octo: yes you can automatically ad such object when a document is created with DocumentCreatedEvent
15:32 <tmortagne> actually would be better to listener to DocumentCreatingEvent
15:32 <tmortagne> and modify the XWikiDocument you get with the event
15:32 <vmassol> note that adding objects to all doc might slow your wiki a bit since loading a doc will also load the object ie do more queries on the db
15:32 <tmortagne> if you work with DocumentCreatedEvent you will have 2 save, the normal save plus you adding your object
15:32 <andreeapopescu> has quit
15:33 <cjd> vmassol: https://ezcrypt.it/q39n#hOvgyeNwO2AirknllVNmgyjC
15:35 <octo> tmortagne: Thank you!
15:35 <tmortagne> octo: you're welcome :)
15:37 <vmassol> cjd: what does this change? BTW your waitTwenty() is not correct as $1 is the jetty stop port
15:38 <vmassol> scratch the last part
15:38 <vmassol> (it's the method param)
15:38 <vmassol> but again I really don't understand what you're trying to do
15:38 <vmassol> why do you keep thinking that the lock file is a problem?
15:38 <cjd> I think it's the same as what you're trying to do
15:38 <cjd> robust stop script
15:39 <vmassol> it's not
15:39 <vmassol> well
15:39 <vmassol> it's going to always kill the process without letting it a chance to stop correctly
15:39 <cjd> not when I tested it
15:39 <ClemensR> has joined #xwiki
15:40 <vmassol> could I talk to you on skype?
15:40 <vmassol> does your skype setup allows this? :)
15:40 <cjd> meaning w/ audio?
15:40 <cjd> heh
15:40 <vmassol> yup
15:41 <cjd> nope, sorry
15:41 <vmassol> ok too bad
15:41 <vmassol> ok so let's do the slow ay
15:41 <vmassol> *way
15:41 <cjd> I think slow so it wouldn't help much to talk live
15:41 <vmassol> the stop_xwiki.sh | ps | netstat
15:41 <vmassol> shows
15:41 <vmassol> that the when the java command returns jetty is not fully stopped
15:41 <vmassol> so in your script
15:42 <vmassol> you get to the line that says:
15:42 <gdelhumeau> vmassol: cjd : use https://www.gruveo.com/ (WebRTC)
15:42 <vmassol> echo 'Failed to stop XWiki, attempting kill';
15:42 <vmassol> and then you execute
15:42 <gdelhumeau> tested during the seminar, and I was happy with it
15:42 <vmassol> kill `cat $XWIKI_LOCK_FILE`
15:42 <vmassol> so even though there's no probem stopping xwiki it'll always end doing a kill
15:43 <cjd> 15:41 < vmassol> that the when the java command returns jetty is not fully stopped <-- I tested a bunch of times and it did not happen
15:43 <vmassol> and the kill will happen when jetty is being stopped so not even sure how jetty will like it....
15:43 <vmassol> cjd: on mac it does all the time
15:43 <cjd> what I saw was the result of the timeout being too short and not telling me it was a timeout
15:44 <vmassol> I don't see the relationshiup between the lock file/timeout and the port beeing freed
15:44 <cjd> try my scripts and tell me if it says it's attempting a kill (you can comment the kill if you'd rather like)
15:44 <vmassol> ok
15:44 <cjd> s/like//
15:45 <cjd> time ./_stop_xwiki.sh && ps -ef | grep java && netstat -lnpt
15:47 <vmassol> note: you changed from #!/bin/sh to bash...
15:47 <vmassol> in the start script
15:47 <cjd> indeed I did
15:47 <cjd> `wait` is standard in bash, 'sh' is not standardized at all
15:47 <OSIMasson> has left #xwiki
15:47 <cjd> meant to mention that
15:48 <vmassol> is bash avaialble on all unixes?
15:48 <cjd> all sane ones :)
15:49 <cjd> might not be on an openwrt router but that's a non-point
15:49 <vmassol> ok
15:49 <vmassol> exactly same problem
15:49 <cjd> paste?
15:49 <vmassol> actually no
15:49 <vmassol> even worse
15:49 <vmassol> :)
15:50 <vmassol> when I start the second instance I get:
15:50 <vmassol> https://gist.githubusercontent.com/vmassol/8a01a48eeabac6fcb07e/raw/1929af3995d3137d1b24a3af994301dea37d9bdb/gistfile1.txt
15:50 <vmassol> ah wait
15:50 <vmassol> checking something
15:51 <cjd> time ./_stop_xwiki.sh && ps -ef | grep java && netstat -lnpt
15:51 <cjd> I want to know how that behaves (w/ 1 instance running)
15:51 <vmassol> I forgot to change stop_xwiki.sh to stop_xwiki2.sh
15:51 <momomomomo> has joined #xwiki
15:52 <cjd> well since my version of the script nolonger calls it, it's kind of a nonpoint
15:52 <vmassol> ah yes
15:52 <vmassol> so no it doesn't work....
15:52 <cjd> time ./_stop_xwiki.sh && ps -ef | grep java && netstat -lnpt
15:52 <vmassol> wow /prioc
15:52 <cjd> proc is in all unics
15:52 <vmassol> that's non standard I believe
15:52 <vmassol> I don't have it on mac
15:52 <cjd> o_O
15:53 <vmassol> I don't have any /proc/ directory
15:53 <vmassol> I've reviewed a lot of ways to check for a process
15:53 <vmassol> yesterday
15:53 <vmassol> and there are only 2 that can work
15:54 <cjd> ok I stand corrected
15:54 <vmassol> best one is with the pid
15:54 <vmassol> and second best is ps
15:54 <cjd> I thought proc was an iron standard
15:54 <vmassol> but you need to be careful with the grep
15:54 <vmassol> ps aux should be standard I think
15:54 <vmassol> and the grep needs to be precise enough to not target other processes
15:55 <vmassol> FTR I was reading this thread http://stackoverflow.com/questions/9117507/linux-unix-command-to-determine-if-process-is-running
15:55 <cjd> anyway
15:55 <cjd> time ./_stop_xwiki.sh && ps -ef | grep java && netstat -lnpt
15:56 <cjd> if we can shut the thing down and know after it stops, we're ok
15:56 <cjd> the rest is details
15:56 <vmassol> netsat won't work
15:56 <vmassol> I pasted what works above
15:56 <vmassol> *netstat
15:56 <vmassol> finding it again
15:57 <cjd> that line is just for testing whether we can stop xwiki w/o a race condition, not to be put into a script
15:57 <vmassol> I know
15:58 <vmassol> found it
15:58 <vmassol> time sh stop_xwiki.sh && ps -ef | grep java && lsof -iTCP:8080 -sTCP:LISTEN
15:58 <vmassol> s/8080/8079
15:58 <DarkKnightCZ> has left #xwiki
15:59 <vmassol> but since stop.sh won't work I don't see what you expect to see
15:59 <vmassol> it doesn't stop XE
15:59 <cjd> it does on my machine
15:59 <vmassol> that's why your script is even worse than the one I had
15:59 <vmassol> which at least stopped XE
16:00 <cjd> what's the backscroll when you run that?
16:01 <vmassol> backscroll?
16:01 <cjd> what's printed to the screen..
16:01 <vmassol> I pasted it above
16:01 <vmassol> :)
16:01 <cjd> https://gist.githubusercontent.com/vmassol/8a01a48eeabac6fcb07e/raw/1929af3995d3137d1b24a3af994301dea37d9bdb/gistfile1.txt <-- ?
16:01 <vmassol> https://gist.githubusercontent.com/vmassol/8a01a48eeabac6fcb07e/raw/1929af3995d3137d1b24a3af994301dea37d9bdb/gistfile1.txt
16:01 <vmassol> yup
16:02 <cjd> $ sh start_xwiki2.sh <-- wrong command
16:02 <cjd> time sh stop_xwiki.sh && ps -ef | grep java && lsof -iTCP:8080 -sTCP:LISTEN
16:02 <cjd> I want to know what that one does
16:02 <vmassol> then you don't understand what I'm trying to do
16:02 <vmassol> :)
16:02 <vmassol> ok
16:02 <vmassol> but it won't be what I 'm trying to achieve
16:02 <cjd> I was under the impression that the stop script returned before stopping xwiki
16:02 <cjd> therefor it would send the kill signal
16:03 <cjd> so I wanted to see what was happening to debug it
16:03 <vmassol> ok trying
16:09 <xwikibot> has joined #xwiki
16:12 <vmassol> I don't need ps anymore
16:12 <vmassol> now that we can stop xwiki for sure
16:13 <cjd> durr
16:13 <cjd> kill -0 $XWIKI_PID
16:13 <cjd> that is universal
16:13 <cjd> not sure how I managed to forget that
16:14 <vmassol> yes they mention it on http://stackoverflow.com/questions/9117507/linux-unix-command-to-determine-if-process-is-running
16:14 <vmassol> but I don't think I even need this
16:14 <vmassol> hmm no same
16:15 <cjd> doesn't work for you?
16:15 <vmassol> so the issue is really triggere by calling stop from start
16:15 <vmassol> *triggered
16:15 <vmassol> stop works because it ends the shell process started
16:15 <vmassol> for some reason
16:16 <cjd> that's odd
16:16 <vmassol> grrr
16:16 <vmassol> we're back to where I was
16:18 <vmassol> wait
16:19 <vmassol> made a mistake
16:21 <vmassol> it works fine :)
16:21 <vmassol> great
16:21 <vmassol> thanks for your help cjd, I'll take over from there
16:22 <cjd> cool, yw
16:22 <vmassol> hmm
16:22 <vmassol> just did a ctrl-c and it didn't stop it cleanly as it did before
16:22 <cjd> ahh of course
16:22 <cjd> because of the start script using & and wait
16:23 <vmassol> btw your script needs to be improve when there's a lock file but no process
16:23 <vmassol> which happened when I did the ctrl-c
16:23 <vmassol> trying to start after that
16:23 <vmassol> leads to
16:24 <cjd> http://alvinalexander.com/java/jwarehouse/jetty-6.1.9/bin/jetty.sh.shtml
16:24 <cjd> JETTY_PID ?
16:24 <vmassol> https://gist.githubusercontent.com/vmassol/ffe0fb87dde12fe8ce6e/raw/1f85d62c08d415289f950dac00c5b906f3437d40/gistfile1.txt
16:24 <Denis> has quit
16:25 <cjd> ahh yea, we should indeed add the check to the stop script as well
16:25 <vmassol> interesting script
16:25 <vmassol> pretty complex too...
16:26 <cjd> kill -0 $XWIKI_PID || ( echo 'XWiki was not stopped cleanly, removing pid file' && rm $XWIKI_PID_FILE
16:26 <cjd> )
16:26 <vmassol> well we need to fix the ctrl-C issue
16:26 <vmassol> it was great
16:26 <Denis> has joined #xwiki
16:26 <cjd> yeah
16:27 <cjd> JETTY_PID=/var/tmp/xwiki-8080.lck
16:27 <vmassol> so wait doesn't respond to ctrl-c
16:27 <cjd> or whatever
16:27 <cjd> well.. technically it does, it stops waiting :)
16:27 <vmassol> I mean the signal is not caught anymore by the JVM for some reason
16:27 <cjd> set JETTY_PID and then let jetty create the actual file
16:27 <cjd> yeah, because the JVM is forked into the background so it won't get signals from the console
16:28 <vmassol> JETTY_PID is not standard
16:28 <vmassol> they use some other script: "start-stop-daemon"
16:29 <cjd> yup
16:30 <cjd> start-stop-deamon is a nice one because it forks the process, (thus getting the pid) then it writes the pid file, then it execve's the program you want to run, therefor not changing pid
16:30 <momomomomo> has quit
16:30 <cjd> but you need to use C to be able to do that
16:33 <vmassol> cjd: why not use "fg" simply?
16:33 <vmassol> process &
16:33 <vmassol> fg
16:33 <cjd> I don't see any reason why not
16:33 <vmassol> trying
16:33 <cjd> never seen it used before so there might be a dragon sleeping under it
16:33 <cjd> *shrug*
16:35 <momomomomo> has joined #xwiki
16:35 <vmassol> hmm doesnet work
16:35 <vmassol> start_xwiki2.sh: line 129: fg: no job control
16:36 <cjd> does your mac have start-stop-dameon ?
16:37 <vmassol> no
16:37 <vmassol> http://apple.stackexchange.com/questions/99888/is-there-any-port-of-start-stop-daemon-for-os-x
16:42 <cjd> https://support.hyperic.com/display/SIGAR/Home
16:42 <cjd> kind of neat
16:42 <vmassol> bbiab
16:54 <sdumitriu1> cjd: No need for inline suppression comments
16:54 <cjd> out of band in an xml file is quite bad IMO
16:55 <sdumitriu1> Why?
16:55 <cjd> refactor the code, don't change the xml file
16:55 <cjd> 2 places to change
16:55 <cjd> coupling
16:55 <sdumitriu1> Yep
16:55 <sdumitriu1> There are two contradicting rules
16:55 <sdumitriu1> There's a rule about not mixing multiple languages in a source file
16:56 <cjd> I disagree about multiple languages in a source file
16:56 <cjd> not sure where that rule came from but I think it needs to justify itself
17:00 <cjd> @Annotations(are not strictly java, they're bolted on to the spec but they're a declaritive thing unto themselves)
17:01 <sdumitriu1> The problem with checkstyle:off is that it's all or nothing
17:01 <sdumitriu1> And maintaining the checkstyle exclusion file isn't that wrong
17:02 <cjd> meh, it's the old makefile mess again
17:02 <sdumitriu1> If you refactor the code so that the exception doesn't apply anymore (rename a function), but the rule is still broken, the build will fail
17:02 <sdumitriu1> And you'll know you have to update one or two lines in the XML
17:02 <cjd> but you refactor the code so the rule is nolonger broken, then you get no feedback
17:03 <cjd> and you have a file full of stuff that nobody's sure which of it is necessary
17:03 <sdumitriu1> In theory that shouldn't be that often required, since exclusions are supposed to be exceptional cases
17:03 <cjd> like #includes which nobody wants to spend the time to try deleting and see if the build fails as a result
17:03 <sdumitriu1> Not a frequent practice
17:03 <sdumitriu1> If we have a larger exclusion config than the java code itself, then you have a bigger problem
17:03 <cjd> well we're going to have to kill off some rules then :)
17:04 <sdumitriu1> Ironically, I'm working on enabling a few more rules right now
17:04 <cjd> smart ones or dumb ones?
17:04 <sdumitriu1> (including the RequireThis rule, which you're right, isn't active yet)
17:04 <cjd> <loaded question>
17:04 <cjd> RequireThis +1
17:04 <cjd> smart, low cost of compliance, tells the reader a lot
17:05 <sdumitriu1> RequireThis is broken now :( https://github.com/checkstyle/checkstyle/pull/274
17:05 <cjd> but don't make it harder to contribute !
17:05 <cjd> if you're going to make it harder, we need to throw some of these things overboard so it's at least a balanced budget
17:06 <sdumitriu1> Is there any codestyle rule that doesn't make it harder to _write_ code?
17:06 <sdumitriu1> All codestyle rules are supposed to make it easier to _read_ code
17:07 <sdumitriu1> At various costs at write time
17:07 <cjd> well some of the bad javadoc ones make it harder to read
17:07 <cjd> and "you must split this up *somewhere*"
17:07 <sdumitriu1> But code, like most data, is more often read than written
17:08 <cjd> duplicate strings and private/package-private javadoc requirements are burning me the most atm
17:08 <sdumitriu1> The problem is that most devs don't realize just how much more code reading they do
17:08 <sdumitriu1> Send a vote for those
17:08 <sdumitriu1> We'll let the community decide
17:08 <cjd> I want to keep the discussion open a bit longer because some people (vmassol) want to discuss it but have not yet had the time
17:09 <cjd> rushing a vote kind of forces a premature -1
17:09 <sdumitriu1> Sure
17:09 <sdumitriu1> But I'd say the veto rule shouldn't apply here
17:10 <sdumitriu1> It's not a design decision, it's a "community feeling good" decision, and no member should dictate how the community is supposed to feel good
17:10 <cjd> that's an interesting point
17:10 <msmeria> has quit
17:10 <cjd> still we shouldn't go forward with something "scary" unless we all feel like we're together on it
17:10 <sdumitriu1> "You'll write clean* code, and you'll enjoy it!" (*clean as defined by the BDFL in this 142 pages document)
17:11 <cjd> well yeah, but we're talking about repealing a law, bearuacracies are scared to death of repealing laws
17:11 <cjd> There will be anarchy in the streets!
17:12 <sdumitriu1> If we're not following the rules because we believe in them, then we're doing it wrong
17:12 <cjd> good point as well
17:13 <cjd> I'm hugely +1 for RequireThis
17:13 <sdumitriu1> You can collect my aphorisms and make a book, we'll split the revenue fairly
17:13 <cjd> heh
17:14 <cjd> but I think it needs to balance with some relaxation of rules
17:14 <sdumitriu1> Sure
17:15 <cjd> and I can squeeze most of the in-package private stuff into private/package-private so I can probably sneek by without even the detection of /internal/ package
17:16 <cjd> just change class/method javadoc rule to be >=protected
17:16 <DarkKnightCZ> has joined #xwiki
17:18 <Slashman> has joined #xwiki
17:18 <cjd> sdumitriu1: RequireThis for both fields and methods or only for fields?
17:19 <sdumitriu1> Only fields
17:19 <cjd> my feeling is only fields since java is -braindead- err not able to store function pointers
17:20 <sdumitriu1> The default for checkstyle is only fields
17:20 <sdumitriu1> And they usually have good defaults
17:20 <sdumitriu1> Not always
17:20 <cjd> do you know a way to quickly bring the whole project into compliance?
17:20 <Enygma`> vmassol: any idea why jobs just stop executing?
17:20 <Enygma`> http://ci.xwiki.org/job/xwiki-enterprise-test-extension/10/console
17:20 <Enygma`> http://ci.xwiki.org/job/xwiki-enterprise-test-pageobjects-6.2.x/2/console (did not even start well)
17:22 <sdumitriu1> cjd: In Eclipse: Alt+Shift+S - F followed by Alt+Shift+S - U
17:22 <sdumitriu1> On the src/main/java directory
17:22 <sdumitriu1> That fixes most of the style issues, but not the deeper issues
17:23 <cjd> I'll need some help with that, I don't have eclipse..
17:24 <cjd> although I could actually run checkstyle and parse the output into a script to fix it :)
17:31 <sdumitriu1> Misspoke, RequireThis applies by default to methods, we must explicitly disable that
17:32 <DarkKnightCZ> has left #xwiki
17:35 <ClemensR> has left #xwiki
18:01 <sdumitriu1> I wonder if anybody would complain if I break the build...
18:01 <cjd> has quit
18:02 <evalica> has quit
18:04 <Enygma`> has quit
18:06 <tmortagne> sdumitriu1: these days you could get lucky and have jenkins blamed for it before you fix it
18:14 <momomomomo> has quit
18:31 <momomomomo> has joined #xwiki
18:34 <Denis> has quit
18:36 <tmortagne> has quit
18:42 <sdumitriu1> nexus unavailable?
18:43 <vmassol> yes momentarily
18:43 <vmassol> I think GF restarted the machine
18:44 <vmassol> pinging him
18:51 <gdelhumeau> has quit
18:53 <vmassol> it's back now
19:18 <Denis> has joined #xwiki
19:19 <momomomomo> has quit
19:36 <woshilapin> has quit
19:43 <sburjan> has joined #xwiki
19:44 <mflorea> has joined #xwiki
19:48 <abusenius> has joined #xwiki
19:52 <Denis> has quit
20:01 <OSIMasson> has joined #xwiki
20:03 <vmassol> has quit
20:13 <OSIMasson> has left #xwiki
20:28 <abusenius_> has joined #xwiki
20:30 <abusenius> has quit
20:39 <momomomomo> has joined #xwiki
22:34 <momomomomo> has quit
22:39 <momomomomo> has joined #xwiki
23:21 <vmassol> has joined #xwiki
23:36 <D-Spair> has quit
23:36 <D-Spair> has joined #xwiki
23:37 <mflorea> has quit
23:46 <xwikiorg_guest_8> has joined #xwiki
23:47 <xwikiorg_guest_8> Hey! >:-) Is the maven repo down?
23:47 <vmassol> yes
23:47 <vmassol> we're haviing some problems ATM
23:47 <vmassol> and the infra guys are sleeping right now… sorry….
23:47 <xwikiorg_guest_8> I am doing my first xwiki install... and couldn't figure out what was happening!
23:48 <vmassol> yeah bad timing
23:48 <xwikiorg_guest_8> Bummer
23:48 <vmassol> it usually doesn't happen...
23:48 <xwikiorg_guest_8> Well, I guess I can try again tomorrow!
23:48 <vmassol> yeah that would be better, again sorry about that
23:49 <xwikiorg_guest_8> Unless there is anywhere else to get the xwiki-enterprise-ui-mainwiki-all-6.1.xar?
23:49 <xwikiorg_guest_8> I could load it through the admin panel with superadmin...
23:49 <xwikiorg_guest_8> No worries ;-)
23:49 <vmassol> well you could get it from OW2 but it's best to use the Distribution Wizard
23:50 <xwikiorg_guest_8> That sounds like a better plan.
23:50 <xwikiorg_guest_8> (Waiting, that is)
23:50 <xwikiorg_guest_8> I have a different question ;-)
23:50 <vmassol> hehe
23:51 <xwikiorg_guest_8> So I am hoping to use basic auth to log in users...
23:51 <xwikiorg_guest_8> I am setting us xwiki as an internal tool and want it behind out production environment... and I would like to use our existing auth.
23:51 <xwikiorg_guest_8> (php session based)
23:52 <xwikiorg_guest_8> Ever heard of somethgin like this?
23:52 <xwikiorg_guest_8> Ideally, I could log users into xwiki with their existing php session.
23:52 <vmassol> xwiki supports lots of auths. Don't know about that php session based on
23:52 <vmassol> but worst case you can write your own auth
23:53 <xwikiorg_guest_8> I guess I could write a custom authenticator to hit a remote database?
23:53 <vmassol> some info here: http://platform.xwiki.org/xwiki/bin/view/AdminGuide/Authentication
23:53 <xwikiorg_guest_8> Alright, I'll git it a shot
23:53 <xwikiorg_guest_8> Thanks!
23:53 <abusenius_> has quit
