IRC Archive for channel #xwiki on 12 September 2014
Last modified by Vincent Massol on 2014/09/12 23:43
<momomomomo> has quit
00:48 <vmassol> has quit
01:02 <Slashman> has quit
01:11 <xwikiorg_guest_8> has left #xwiki
01:16 <jeremiah> has joined #xwiki
01:16 <jeremiah> Anybody have a moment for a quick auth question?
01:17 <jeremiah> I am trying to set up my xwiki so that if a request comes in with basic auth user credentials and the user doesn't exist, that it will automatically create the user.
01:18 <jeremiah> Where can I find the source code for authentication?
01:53 <jeremiah> So I found and cloned the platform-core repo, then created a class that extends XWikiAuthServiceImpl.
01:53 <jeremiah> How do I add this class to my xwiki?
02:16 <sburjan> has quit
03:01 <jeremiah> has quit
07:09 <msmeria> has joined #xwiki
07:54 <DarkKnightCZ> has joined #xwiki
08:04 <mflorea> has joined #xwiki
08:05 <ClemensR> has joined #xwiki
08:08 <Denis> has joined #xwiki
08:10 <vmassol> has joined #xwiki
08:12 <Denis> has quit
08:27 <Denis> has joined #xwiki
08:31 <msmeria> has quit
08:32 <msmeria> has joined #xwiki
08:33 <DarkKnightCZ> has quit
08:40 <cjd> has joined #xwiki
08:43 <evalica> has joined #xwiki
08:44 <Enygma`> has joined #xwiki
08:50 <tmortagne> has joined #xwiki
08:50 <DarkKnightCZ> has joined #xwiki
08:51 <cjd> tmortagne: for (; i < awesome; ++i) ; highfive();
08:55 <woshilapin> has joined #xwiki
09:06 <tmortagne> cjd: :)
09:08 <Enygma`> has quit
09:09 <Enygma`> has joined #xwiki
09:30 <DarkKnightCZ1> has joined #xwiki
09:32 <DarkKnightCZ> has quit
09:35 <KermitTheFragger> has joined #xwiki
10:19 <mflorea> has quit
10:36 <Slashman> has joined #xwiki
10:46 <KermitTheFragger> has quit
11:11 <cjd> Does anyone know if it's possible in velocity to create a reference to a document with a specific langyage?
11:12 <vmassol> I dno't think it is
11:13 <tmortagne> cjd: you mean in wiki syntax ?
11:13 <vmassol> actually it is
11:13 <cjd> $services.model.createReferenceTo....
11:13 <vmassol> public DocumentReference createDocumentReference(DocumentReference reference, Locale locale)
11:13 <vmassol> didn't notice it got added :)
11:13 <vmassol> * @since 5.4RC1
11:14 <cjd> hmm
11:14 <tmortagne> ho you mean a DocumentReference
11:14 <tmortagne> I tough you wanted a link
11:14 <cjd> we should make all of these script services into extensions so I can depend on them rather than waiting for xwiki to be updated...
11:14 <cjd> extensions are really kind of awesome that way
11:14 <vmassol> not really
11:15 <vmassol> that would be pretty bad
11:15 <cjd> why?
11:15 <tmortagne> cjd: you can provide your own script service in your extension if you like
11:15 <vmassol> because when you have tons of core extensions you're into version hell
11:15 <vmassol> and you don't control your stabitlity anymore
11:15 <cjd> that's how debian works
11:15 <vmassol> so it looks appealing but it's almost impossible to do it well
11:15 <tmortagne> it's nice to have such base API by default
11:16 <vmassol> actually it's so hard that you've had companies offering business models based on that!
11:16 <vmassol> curating the repos and finding what versions work with what
11:16 <vmassol> and making distribs
11:17 <tmortagne> technically that API is an en extension, but that extension is sychronized with platform right now like many very standard APIs
11:17 <cjd> mm good base api is nice, it's annoying when the base api is too hard to use and/or doesn't do what people want so you end up with IOUtils and FileUtils and StringUtils etc etc
11:17 <tmortagne> s/an en/in an/
11:17 <tmortagne> having a locale in DocumentReference does not exist since DocumentReference exist
11:18 <cjd> mm true
11:18 <tmortagne> your main issue is that we add things in script service when we need then in XE
11:18 <cjd> it's actually the php model vs. the nodejs model
11:18 <cjd> also the windows model vs the linux model
11:18 <tmortagne> s/then/them/
11:18 <cjd> linux and nodejs are dependency hell, windows and php are unified api
11:19 <tmortagne> you have many extensions in php
11:20 <cjd> it's a bit even the cathedral and the bazaar, the bazaar is the market of extensions all competing to provide the best API
11:20 <cjd> but indeed bazaars are messy
11:21 <vmassol> you need a large community to handle a bazaar too
11:22 <vmassol> (or maybe the community forms because it's a bazaar and it needs sorting out… ;))
11:22 <vmassol> anyway back to my shell scripts now!
11:22 <cjd> I have a hard time imagining a community forming around a cathedral
11:22 <cjd> have fun
11:22 <vmassol> did you see my trick cjd?
11:22 <vmassol> pretty happy about myself
11:23 <cjd> yeah, it's probably going to have problems
11:23 <vmassol> this:
11:23 <vmassol> trap "kill ${XWIKI_PID}; exit 1" INT
11:23 <vmassol> ?
11:23 <cjd> most robust solution is to write the pid from inside of java
11:23 <cjd> /nod
11:23 <vmassol> I don't fully understand it TBH
11:23 <vmassol> but it seems to be working
11:23 <cjd> there is the fact that java has no decent way to get a pid (a mildly disturbing fact btw)
11:24 <cjd> but there's a hacky one which Should Always Work
11:24 <vmassol> so I'll leave it like this for now
11:25 <cjd> user@toshitba:~/wrk/cjdns$ node -e 'require("fs").writeFile("my.pid", process.pid);' && cat ./my.pid
11:25 <cjd> 15725
11:27 <vmassol> guys are you ok that I change the way params are passed to start_xwiki.sh? It's breaking backward compat so I want to know if it's ok
11:27 <cjd> ok for me
11:27 <vmassol> it's only for the standalone version though
11:27 <vmassol> so no big risk for produciton....
11:28 <cjd> oh, you can actually do it without that trick
11:30 <cjd> PID=$$ ; echo $PID > $PID_FILE ; ( wait $PID ; rm $PID_FILE ) & exec java .....
11:30 <cjd> then ctrl+c will go to java directly
11:30 <vmassol> errr
11:30 <tmortagne> vmassol: it's OK IMO
11:30 <vmassol> wait is going to wait on the JVM
11:31 <vmassol> so that JVM won't receive the ctrl+c
11:31 <vmassol> tmortagne: ok thanks
11:31 <cjd> my line switches the processes
11:31 <cjd> it forks the wait/rm pidfile into the background
11:31 <tmortagne> we might need to review some scripts here and there but it's probably not a big deal
11:31 <tmortagne> would be nice to have a clear error when using the old way
11:31 <cjd> then *execs* java rather than spawning it so it reuses the pid number from the shell process
11:32 <cjd> so you already know the pid #
11:32 <vmassol> I don't understand anything! :)
11:32 <vmassol> sounds way more complex
11:32 <cjd> oh a downside of this approach is when you stop java, you probably lose your terminal :)
11:32 <cjd> also fixable
11:32 <cjd> but complex bash is bad
11:32 <cjd> b/c it's different on every machine :D
11:35 <Denis> has quit
11:37 <vmassol> one option would have been to reuse http://dev.eclipse.org/svnroot/rt/org.eclipse.jetty/jetty/trunk/jetty-distribution/src/main/resources/bin/jetty.sh but I find it too complex and I don't want to maintain that!
11:38 <cjd> +1
11:38 <cjd> I fight to keep bash out of the cjdns build because I've been bitten by it before
11:38 <cjd> './do' just downloads nodejs (if necessary) and then runs make.js :)
11:39 <vmassol> I already don't like the complexity I'm bringing but I don't see any other way to make our CI more robust...
11:39 <cjd> /nod
11:40 <cjd> some lines of code are less complex than others..
11:41 <cjd> hm
11:41 <Denis> has joined #xwiki
11:43 <tmortagne> vmassol: if it's really only for the CI we could do that at CI level in a pre build step
11:43 <vmassol> nope
11:43 <vmassol> I thought about it
11:43 <vmassol> ;)
11:43 <vmassol> because a job can start xwiki several times
11:44 <tmortagne> for platform yes
11:44 <vmassol> and in any case it can be useful for the iuser anyway
11:44 <vmassol> *user
11:44 <vmassol> so if we have it we might as well expose it
11:48 <cjd> https://ezcrypt.it/539n#8bQTwqZwlcnogPo83DTMgMsq
11:48 <cjd> that is a nice solution IMO
11:48 <vmassol> I really prefer mine
11:48 <vmassol> I now understand it
11:48 <cjd> /nod
11:48 <cjd> you're trappinga sig and relaying it to the other process
11:48 <cjd> but if it's not a SIGINT then it won't work
11:48 <vmassol> I could add more signals
11:49 <cjd> maybe not a big deal *shrug*
11:49 <vmassol> I just trapped INT
11:49 <vmassol> ie ctrl-c
11:49 <vmassol> for now
11:49 <vmassol> we probably need to add the signals sent by jstack too
11:49 <vmassol> (need to check which one it is)
11:49 <vmassol> for heap dump and memory dump
11:50 <cjd> it will send it to the java process directly
11:50 <cjd> my solution is to put the "remove pid" job in another process and then reuse the pid number from the shell for java
11:50 <vmassol> indeed, so even better
11:51 <cjd> but.. they're both equally good until one of them fails on a weird machine :)
11:51 <cjd> and we can't know which one
12:03 <Denis> has quit
12:15 <evalica> has quit
12:17 <evalica> has joined #xwiki
12:21 <Denis> has joined #xwiki
12:41 <Denis> has quit
13:18 <DarkKnightCZ1> hi, what could cause Caused by: org.artofsolving.jodconverter.office.OfficeException: could not store document: img0.html; errorCode: 2074
13:19 <vmassol> https://code.google.com/p/jodconverter/issues/detail?id=102
13:20 <vmassol> converting to PDF?
13:20 <vmassol> ah it's not just pdf apparently
13:20 <DarkKnightCZ1> nope, just office viewer, but i guess that could be permissions issue (i might be running office server under different user, checking)
13:21 <DarkKnightCZ1> hmm, openoffice server is not storing any files, is it?
13:27 <DarkKnightCZ1> nope, so it's not caused by permissions
13:29 <vmassol> no idea, there could be temporary files
13:30 <vmassol> (that would make sense)
13:35 <DarkKnightCZ1> hmm, cannot find it in filesystem, there might be bug in used version of OO
13:39 <ClemensR> maybe indeed the office-server version: https://code.google.com/p/jodconverter/issues/detail?id=102 "The upgrade to LibreOffice 3.5 solved the problem for us."
13:40 <DarkKnightCZ1> ha, openoffice 4.1.0 is broken, libreoffice 4.3.0 is working
13:41 <DarkKnightCZ1> ok, so not actually working, just not throwing that error (resulting in no slides), going to try 3.x
13:41 <vmassol> what os are you on?
13:42 <DarkKnightCZ1> RHEL
13:42 <ClemensR> happy version hunting :-/ ... seems error code 2074 is a "something went wrong" code from oo
13:42 <DarkKnightCZ1> yeah, they have pretty good error codes :)
13:44 <ClemensR> did the office view work at all before? sometimes this error means modules for spreadsheet etc is missing
13:45 <DarkKnightCZ1> i don't know actually
13:50 <DarkKnightCZ1> hmm, WARN o.x.c.ExecutionContext - Tried to remove non-existing property [officeView.caughtException] from execution context.
13:51 <DarkKnightCZ1> thats when there is no exception (but presentations have 0 slides)
13:51 <Denis> has joined #xwiki
14:05 <Slashman> has quit
14:06 <Slashman> has joined #xwiki
14:12 <msmeria> vmassol: hello
14:14 <msmeria> I've just released File Manager v1.1.7 here: https://github.com/xwiki-contrib/application-filemanager/releases/tag/application-filemanager-1.1.7
14:14 <msmeria> can you promote the release to Nexus ?
14:14 <vmassol> hey why me? :)
14:15 <vmassol> finishgin sth
14:15 <vmassol> you should ask committers in general
14:16 <Denis> has quit
14:19 <msmeria> tmortagne, Enygma`, evalica: can one of you guys do it ?
14:19 <cjd> has quit
14:24 <ClemensR> msmeria: I can give it a try
14:24 <msmeria> ClemensR: sweet! :D
14:25 <cjd> has joined #xwiki
14:26 <msmeria> ClemensR: you have instructions here: http://contrib.xwiki.org/xwiki/bin/view/Main/WebHome#HReleasetheproject
14:26 <msmeria> just before the "Publishing on extensions.xwiki.org" section
14:27 <ClemensR> ah, cool
14:28 <ClemensR> msmeria: ok, its "closed" now. does this need some testing or should I try to publish it right away?
14:29 <vmassol> I'm going to break some jobs but they'll succeed once platform has finished biulding....
14:29 <vmassol> I've had to modify both plarform and enterprise
14:29 <vmassol> and they both need each other to work fine....
14:30 <tmortagne> ClemensR: release it right away
14:30 <msmeria> ClemensR: you could try just installing it and adding an entry just for kicks...maybe I missed something
14:30 <vmassol> well enterprise needs platform
14:30 <msmeria> or not :)
14:31 <tmortagne> msmeria: a test master would miss something ? :)
14:31 <ClemensR> ok, I guess I just publish it, and if it really fails, there can be 1.1.8 ;)
14:32 <msmeria> tmortagne: it's my second release and I just hope I didn't brake anything :D
14:33 <vmassol> msmeria: indeed normally staging is there so that you can try it out before releasing to the final repo :)
14:34 <msmeria> vmassol: yeah, that's what I read in the doc too: "Make sure to test your extension from the closed staging repo first since a released repo cannot be removed!"
14:34 <ClemensR> when testing my problem is that I always point to the wrong repo :p ... anyway, it should be published now
14:34 <vmassol> ah I documented it cool :)
14:34 <msmeria> ClemensR: checking...
14:39 <andreeapopescu> has joined #xwiki
14:45 <Denis> has joined #xwiki
14:51 <msmeria> ClemensR: thanks! I was able to install it and test it
14:51 <ClemensR> ah, ok, cool :)
14:57 <andreeapopescu> has left #xwiki
14:57 <vmassol> hmm my updated scripts don't work on agents apparently
14:57 <vmassol> 14:36:32.471 [Thread-1] INFO o.x.t.i.XWikiLogOutputStream - Starting Jetty on port 8080, please wait...
14:57 <vmassol> 14:36:32.477 [Thread-2] ERROR o.x.t.i.XWikiLogOutputStream - start_xwiki.sh: 161: wait: Illegal number: $
14:57 <vmassol> some wrong shell used apparently
14:58 <msmeria> has quit
14:58 <cjd> start_xwiki.sh: 79: start_xwiki.sh: [[: not found <-- sh vs. bash
14:59 <vmassol> shouldn't the first line ensure that bash is used?
14:59 <cjd> hmm yeah
14:59 <cjd> and bash should have [[
14:59 <cjd> *usually :D
14:59 <tmortagne> only if you execute the script I think
15:00 <cjd> oh yeah, if you run sh ./script
15:00 <tmortagne> if you do "sh start_xwiki.sh" it probably does not care
15:00 <cjd> then it will still be interpreted in sh
15:00 <vmassol> locally I run sh start_xwiki.sh and it workds
15:00 <vmassol> works
15:00 <vmassol> maybe it's specific to mac
15:00 <tmortagne> I guess jenkins is not doing that
15:00 <vmassol> jenkins is not doing anything
15:00 <vmassol> it's our code
15:00 <vmassol> I'll change it to bash then
15:01 <tmortagne> what is our code doing ?
15:01 <vmassol> sh
15:01 <vmassol> sh start_xwiki.sh -p %s -sp %
15:01 <tmortagne> maybe sh is an alias for bash or a special adavanced sh on mac
15:01 <vmassol> I thought it was ok since it works locally
15:01 <vmassol> ah yes good point
15:01 <vmassol> cehcking
15:01 <tmortagne> s/adavanced/advanced/
15:01 <cjd> sh is different on every machine
15:01 <cjd> really
15:02 <cjd> even versions of the same OS
15:02 <tmortagne> anyway batter be explicit in our code since we require bash
15:02 <tmortagne> s/batter/better/
15:03 <Denis> has quit
15:05 <vmassol> let's see how it goes....
15:05 <vmassol> 45 minutes more....
15:05 <vmassol> actually not that long
15:05 <vmassol> just 3-4
15:10 <vmassol> grr still failing
15:10 <vmassol> but different eerror
15:10 <vmassol> 5:10:06.060 [Thread-3] ERROR o.x.t.i.XWikiLogOutputStream - start_xwiki.sh: line 161: wait: `$': not a pid or valid job spec
15:16 <Denis> has joined #xwiki
15:16 <vmassol> oh
15:16 <vmassol> I know
15:16 <vmassol> :)
15:16 <vmassol> (I thinkà)
15:17 <vmassol> ok so
15:17 <vmassol> I believe that the following line:
15:17 <vmassol> XWIKI_PID=$!
15:18 <vmassol> is transformed by velocity to:
15:18 <vmassol> XWIKI_PID=$
15:18 <vmassol> :)
15:18 <ClemensR> ah, if velocity is filtering the resources
15:18 <vmassol> we're using velocity for this line:
15:18 <vmassol> XWIKI_DATA_DIR=${xwikiDataDir}
15:19 <cjd> mm template preprocessors are good until they aren't
15:19 <vmassol> yes it's dangerous to run velocity on a shell script
15:19 <vmassol> very dangerous
15:19 <cjd> yeah, we should use CPP instead :)
15:20 <vmassol> the code is in PackageMojo.expandJettyDistribution() FYI
15:20 <cjd> mhm
15:21 <Denis> has quit
15:21 <cjd> I wrote an excellent little piece of CPP/js which tests if a CPP macro is defined
15:21 <Denis> has joined #xwiki
15:22 <vmassol> it's easy enough to search/replace here
15:22 <Denis> has quit
15:22 <cjd> what do you use velocity for ?
15:23 <cjd> oh right
15:23 <cjd> so do this:
15:23 <cjd> XWIKI_VELOCITY_INJECTED_CONFIGURATION
15:23 <cjd> ## check abcdx.vm for information about that this is here for
15:33 <vmassol> more complex than that, it needs to work for the distribution pom too, I'll try using some escapes
15:35 <cjd> http://imgs.xkcd.com/comics/success.png
15:36 <DarkKnightCZ> has joined #xwiki
15:36 <DarkKnightCZ> has left #xwiki
15:38 <DarkKnightCZ1> has quit
15:50 <cjd> tmortagne: do you know what happens if you install an extension and there is already existing a document which the extension is to override?
15:50 <tmortagne> cjd: conflict
15:50 <vmassol> grrr seems escapes don't work for some reason…. I'd have expected this to work: XWIKI_PID=$\!
15:51 <tmortagne> i.e. "no idea what to do, let the user choose"
15:51 <vmassol> (according to http://velocity.apache.org/engine/devel/user-guide.html#Advanced_Issues:_Escaping_and_ )
15:51 <cjd> ok, Charles did it on the intranet and it seems to have hung
15:52 <tmortagne> what do you mean ?
15:52 <vmassol> ok I'll do it like this:
15:52 <vmassol> #set ($exclamationMark = '!')
15:52 <vmassol> XWIKI_PID=$$exclamationMark
15:52 <vmassol> ;)
15:52 <cjd> he said it refused to proceed
15:53 <cjd> just "trying forever" to continue
15:53 <tmortagne> always the exact same conflict ?
15:54 <tmortagne> never seen something like this
15:54 <tmortagne> maybe he have some old EM css/js in his browser cache
15:55 <cjd> --> Installation manager does not seems to be sur that this extention is installed.
15:55 <cjd> To be continued
15:58 <ClemensR> vmassol: maybe XWIKI_PID=\$! would have worked, but via #set everybody notices velocity has a word to say here first
15:58 <vmassol> no it doesn't work
15:58 <vmassol> right now I'm trying to figure out why it doesn't work with my local variable too
15:58 <vmassol> and yes I agree it's better for clarity
15:59 <vmassol> I'l doing this
15:59 <vmassol> ## Since this file is parsed with Velocity and since Velocity would render '$!' as '$' we need to be tricky...
15:59 <vmassol> XWIKI_PID=$$exclamationMark
15:59 <vmassol> but nothings gets replaced strangely
15:59 <vmassol> in a wiki page it works though
15:59 <vmassol> for ex: http://playground.xwiki.org/xwiki/bin/view/Main/VMA
16:00 <vmassol> but through the assembly plugin with filtering on it doesn't....
16:00 <ClemensR> maybe #set ($exclamationMark = '$!') and XWIKI_PID=$exclamationMark to do less substitution
16:01 <vmassol> I'll try this since it's also nicer
16:02 <ClemensR> if I try this locally, it builds a xwiki-platform-tool-jetty-resources-6.3-SNAPSHOT.zip without any substitutions :/
16:03 <vmassol> same
16:03 <vmassol> yep
16:04 <vmassol> actually it does substitute for ${xwikiDataDir}
16:04 <vmassol> so it's like local variable have no effect
16:05 <vmassol> and using a property variable in the pom.xml would be ugly ;)
16:05 <ClemensR> for me it does not even this ... must have made a mistake somewhere :(
16:06 <vmassol> I rebuild tools/jetty
16:06 <vmassol> and then distribution/jetty/hsqldb
16:06 <vmassol> then in target, I unzip the zip
16:06 <vmassol> and check start_xwiki.sh
16:07 <ClemensR> seems I need to do the second step, too
16:09 <vmassol> maybe there's an option when you run velocity to not define local variables
16:09 <vmassol> oh
16:09 <vmassol> stupid me
16:09 <vmassol> forget all that
16:09 <vmassol> the assembly plugin must not be using Velocity at all
16:09 <vmassol> it just happens to use a similar syntax
16:09 <vmassol> ${var}
16:09 <vmassol> checking that
16:10 <vmassol> yup
16:10 <vmassol> http://maven.apache.org/plugins/maven-assembly-plugin/examples/single/filtering-some-distribution-files.html
16:11 <vmassol> so the isseu with $! is only with the PackageMojo actually
16:11 <vmassol> and I can apply a similar strategy than the assembyl plugin instead
16:11 <ClemensR> ah, great, that might explain something :)
16:13 <ClemensR> actually there is still a XWIKI_PID=$! in the generated xwiki-enterprise-jetty-hsqldb-6.3-SNAPSHOT/start_xwiki.sh anyway, at least I see something
16:14 <Denis> has joined #xwiki
16:15 <ClemensR> ah, PackageMojo is different ...
16:17 <Denis> has quit
16:34 <Pbas> has quit
16:35 <vmassol> let's see how this goes…
16:36 <ClemensR> not sure about, but XWIKI_PID=${!} seems to have worked for the PackageMojo at least ... as different solution in PackageMojo instead
16:36 <ClemensR> ok, maybe nicer to keep it in sync there
16:40 <vmassol> cool it's working
16:40 <vmassol> see http://ci.xwiki.org/job/xwiki-platform/18/console
16:41 <vmassol> actually
16:41 <vmassol> speaking too fat
16:41 <vmassol> fast
16:41 <vmassol> some other problem now....
16:43 <evalica> has quit
16:52 <Denis> has joined #xwiki
17:02 <Denis> has quit
17:09 <vmassol> oh stupid me
17:14 <benoitc> has quit
17:14 <benoitc> has joined #xwiki
17:14 <vmassol> should work this time….
17:14 <vmassol> can't believe how long it all took....
17:15 <cjd> #bash
17:16 <tmortagne> vmassol: you Java developer :)
17:16 <cjd> naw, bash is always like that
17:16 <vmassol> actually I've worked during a few years coding only in shell in my past life :)
17:16 <tmortagne> cjd: was refering to the ;
17:16 <cjd> that's why I avoid it like the plague
17:17 <tmortagne> in https://github.com/xwiki/xwiki-platform/commit/77261c854551d9d3e0690b5583581ff157786d7d
17:17 <cjd> haha no, the ; are all my lines
17:17 <vmassol> I've developed pretty scary long code (full apps) in shell....
17:18 <vmassol> working! :)
17:18 <momomomomo> has joined #xwiki
17:18 <vmassol> (so far)
17:19 <cjd> friday fun: determining whether a given preprocessor macro is defined w/o using an #ifdef https://github.com/cjdelisle/cjdns/blob/master/util/Defined.h#L18
17:37 <momomomomo> has quit
17:38 <ClemensR> has left #xwiki
17:40 <momomomomo> has joined #xwiki
17:52 <vmassol> someone introduced a checkztyle error in rendering
17:52 <vmassol> home/hudsonagent/jenkins_root/workspace/xwiki-rendering/xwiki-rendering-api/src/main/java/org/xwiki/rendering/block/DefinitionDescriptionBlock.java:0: File does not end with a newline.
17:52 <vmassol> sdumitriu: is that you?
17:52 <vmassol> see http://ci.xwiki.org/job/xwiki-rendering/10/console
17:53 <cjd> yeah, I'm pretty sure he enabled that
17:53 <vmassol> actually no
17:53 <vmassol> more likely to be tmortagne
17:53 <vmassol> https://github.com/xwiki/xwiki-rendering/commit/9ebe8bc850538f01c8f04c58dcea4d874cb877d1
17:54 <vmassol> he's the last one to have committed in rendering
17:54 <vmassol> http://ci.xwiki.org/job/xwiki-rendering/8/
17:54 <tmortagne> vmassol: I don't see how this is related
17:54 <vmassol> before the error ahppened :)
17:54 <tmortagne> there was no change in DefinitionDescriptionBlock
17:54 <vmassol> or it's caused by some changes in commons indirectly
17:54 <tmortagne> pretty sure it's a change in checkstyle setup
17:54 <tmortagne> I don't remember ever having this kind of checktyle error
17:55 <vmassol> sdumitriu: could you check that please?
17:55 <tmortagne> and I really don't see how this is important enough to fail the build...
17:55 <Enygma`> has quit
17:55 <cjd> btw I'd like to hear your comments on my most recent checkstyle related mail
17:55 <tmortagne> "File does not end with a newline." seriously
17:55 <vmassol> cjd: no time at all
17:56 <cjd> will you have time over the weekend?
17:56 <vmassol> no
17:56 <vmassol> right now the priority is the biuld
17:56 <vmassol> and the release
17:56 <vmassol> but if you're interested in checkstyle please review my mail about it
17:56 <cjd> from when?
17:56 <vmassol> (the one I sent 2 weeks before yours)
17:57 <vmassol> pointing to that page: http://design.xwiki.org/xwiki/bin/view/Proposal/DefineSonarRules
17:59 <mflorea> has joined #xwiki
18:01 <cjd> I don't see anything that looks related in the past month
18:02 <vmassol> http://xwiki.markmail.org/search/?q=DefineSonarRules
18:03 <cjd> ahh [xwiki-devs] Playing with Sonar Rules -- aug 17
18:03 <vmassol> the next step is replying to this email by listing the checks that in your opinion should be removed
18:03 <vmassol> all violations can be seen on sonar.xwiki.org
18:03 <cjd> a next step
18:04 <vmassol> it's much larger than juste checkstyle
18:05 <cjd> well we would need to think about what it means in terms of possibly weighting the project down with (what I understand is) a proprietary SaaS service
18:10 <vmassol> cjd: it's as proprietary as xwiki is
18:10 <vmassol> actually less
18:11 <vmassol> not less same since it's also LGPL
18:11 <cjd> ahh ok
18:12 <cjd> well see you on monday
18:12 <cjd> has quit
18:14 <vmassol> bye
18:14 <vmassol> sdumitriu: ping
18:33 <sdumitriu1> vmassol: Pong
18:33 <vmassol> sdumitriu: I missed that it was on a branch
18:33 <vmassol> sorry about that
18:33 <vmassol> I still fixed a violation on rendering though
18:33 <vmassol> (master)
18:34 <vmassol> regarding checkstyle, this is a dead tool
18:34 <vmassol> let me reply quickly
18:37 <tmortagne> has quit
18:39 <vmassol> done
18:40 <woshilapin> has quit
18:42 <vmassol> oh I forgot to merge my changes to the 6.x branch
18:42 <vmassol> doing it
18:47 <mflorea> has quit
19:10 <momomomomo> has quit
19:15 <vmassol> sdumitriu: could you please fix all violations caused by your addition of the New line check?
19:16 <vmassol> I fixed all rendering but now all platform is failing
19:16 <vmassol> (see for example http://ci.xwiki.org/job/xwiki-platform/21/console )
19:17 <sdumitriu1> vmassol: Working on that
19:17 <vmassol> k
19:23 <Slashman> has quit
19:25 <sdumitriu1> From yesterday:
19:25 <sdumitriu1> (12:01:18 PM) sdumitriu1: I wonder if anybody would complain if I break the build...
19:25 <sdumitriu1> (12:06:44 PM) tmortagne: sdumitriu1: these days you could get lucky and have jenkins blamed for it before you fix it
19:59 <cjd> has joined #xwiki
20:00 <cjd> is now known as <Guest12167>
20:04 <momomomomo> has joined #xwiki
20:15 <vmassol> sdumitriu: I didn't see it, I would have said no because we're trying to do the opposite
20:15 <vmassol> been at it for more than week already
20:15 <vmassol> opposite = try to stabillize the buikd
20:15 <vmassol> and it's hard
20:15 <vmassol> and painful
20:16 <vmassol> and we're all supposed to be doing other stuff ....
20:16 <vmassol> I'm actually depressed by the state of our build/ci
20:48 <abusenius_> has joined #xwiki
22:39 <vmassol> has quit
23:08 <OSIMasson> has joined #xwiki
23:12 <OSIMasson> has quit
23:25 <OSIMasson> has joined #xwiki
23:33 <sburjan> has joined #xwiki
23:40 <nickless> has joined #xwiki
23:43 <abusenius_> has quit
00:48 <vmassol> has quit
01:02 <Slashman> has quit
01:11 <xwikiorg_guest_8> has left #xwiki
01:16 <jeremiah> has joined #xwiki
01:16 <jeremiah> Anybody have a moment for a quick auth question?
01:17 <jeremiah> I am trying to set up my xwiki so that if a request comes in with basic auth user credentials and the user doesn't exist, that it will automatically create the user.
01:18 <jeremiah> Where can I find the source code for authentication?
01:53 <jeremiah> So I found and cloned the platform-core repo, then created a class that extends XWikiAuthServiceImpl.
01:53 <jeremiah> How do I add this class to my xwiki?
02:16 <sburjan> has quit
03:01 <jeremiah> has quit
07:09 <msmeria> has joined #xwiki
07:54 <DarkKnightCZ> has joined #xwiki
08:04 <mflorea> has joined #xwiki
08:05 <ClemensR> has joined #xwiki
08:08 <Denis> has joined #xwiki
08:10 <vmassol> has joined #xwiki
08:12 <Denis> has quit
08:27 <Denis> has joined #xwiki
08:31 <msmeria> has quit
08:32 <msmeria> has joined #xwiki
08:33 <DarkKnightCZ> has quit
08:40 <cjd> has joined #xwiki
08:43 <evalica> has joined #xwiki
08:44 <Enygma`> has joined #xwiki
08:50 <tmortagne> has joined #xwiki
08:50 <DarkKnightCZ> has joined #xwiki
08:51 <cjd> tmortagne: for (; i < awesome; ++i) ; highfive();
08:55 <woshilapin> has joined #xwiki
09:06 <tmortagne> cjd: :)
09:08 <Enygma`> has quit
09:09 <Enygma`> has joined #xwiki
09:30 <DarkKnightCZ1> has joined #xwiki
09:32 <DarkKnightCZ> has quit
09:35 <KermitTheFragger> has joined #xwiki
10:19 <mflorea> has quit
10:36 <Slashman> has joined #xwiki
10:46 <KermitTheFragger> has quit
11:11 <cjd> Does anyone know if it's possible in velocity to create a reference to a document with a specific langyage?
11:12 <vmassol> I dno't think it is
11:13 <tmortagne> cjd: you mean in wiki syntax ?
11:13 <vmassol> actually it is
11:13 <cjd> $services.model.createReferenceTo....
11:13 <vmassol> public DocumentReference createDocumentReference(DocumentReference reference, Locale locale)
11:13 <vmassol> didn't notice it got added :)
11:13 <vmassol> * @since 5.4RC1
11:14 <cjd> hmm
11:14 <tmortagne> ho you mean a DocumentReference
11:14 <tmortagne> I tough you wanted a link
11:14 <cjd> we should make all of these script services into extensions so I can depend on them rather than waiting for xwiki to be updated...
11:14 <cjd> extensions are really kind of awesome that way
11:14 <vmassol> not really
11:15 <vmassol> that would be pretty bad
11:15 <cjd> why?
11:15 <tmortagne> cjd: you can provide your own script service in your extension if you like
11:15 <vmassol> because when you have tons of core extensions you're into version hell
11:15 <vmassol> and you don't control your stabitlity anymore
11:15 <cjd> that's how debian works
11:15 <vmassol> so it looks appealing but it's almost impossible to do it well
11:15 <tmortagne> it's nice to have such base API by default
11:16 <vmassol> actually it's so hard that you've had companies offering business models based on that!
11:16 <vmassol> curating the repos and finding what versions work with what
11:16 <vmassol> and making distribs
11:17 <tmortagne> technically that API is an en extension, but that extension is sychronized with platform right now like many very standard APIs
11:17 <cjd> mm good base api is nice, it's annoying when the base api is too hard to use and/or doesn't do what people want so you end up with IOUtils and FileUtils and StringUtils etc etc
11:17 <tmortagne> s/an en/in an/
11:17 <tmortagne> having a locale in DocumentReference does not exist since DocumentReference exist
11:18 <cjd> mm true
11:18 <tmortagne> your main issue is that we add things in script service when we need then in XE
11:18 <cjd> it's actually the php model vs. the nodejs model
11:18 <cjd> also the windows model vs the linux model
11:18 <tmortagne> s/then/them/
11:18 <cjd> linux and nodejs are dependency hell, windows and php are unified api
11:19 <tmortagne> you have many extensions in php
11:20 <cjd> it's a bit even the cathedral and the bazaar, the bazaar is the market of extensions all competing to provide the best API
11:20 <cjd> but indeed bazaars are messy
11:21 <vmassol> you need a large community to handle a bazaar too
11:22 <vmassol> (or maybe the community forms because it's a bazaar and it needs sorting out… ;))
11:22 <vmassol> anyway back to my shell scripts now!
11:22 <cjd> I have a hard time imagining a community forming around a cathedral
11:22 <cjd> have fun
11:22 <vmassol> did you see my trick cjd?
11:22 <vmassol> pretty happy about myself
11:23 <cjd> yeah, it's probably going to have problems
11:23 <vmassol> this:
11:23 <vmassol> trap "kill ${XWIKI_PID}; exit 1" INT
11:23 <vmassol> ?
11:23 <cjd> most robust solution is to write the pid from inside of java
11:23 <cjd> /nod
11:23 <vmassol> I don't fully understand it TBH
11:23 <vmassol> but it seems to be working
11:23 <cjd> there is the fact that java has no decent way to get a pid (a mildly disturbing fact btw)
11:24 <cjd> but there's a hacky one which Should Always Work
11:24 <vmassol> so I'll leave it like this for now
11:25 <cjd> user@toshitba:~/wrk/cjdns$ node -e 'require("fs").writeFile("my.pid", process.pid);' && cat ./my.pid
11:25 <cjd> 15725
11:27 <vmassol> guys are you ok that I change the way params are passed to start_xwiki.sh? It's breaking backward compat so I want to know if it's ok
11:27 <cjd> ok for me
11:27 <vmassol> it's only for the standalone version though
11:27 <vmassol> so no big risk for produciton....
11:28 <cjd> oh, you can actually do it without that trick
11:30 <cjd> PID=$$ ; echo $PID > $PID_FILE ; ( wait $PID ; rm $PID_FILE ) & exec java .....
11:30 <cjd> then ctrl+c will go to java directly
11:30 <vmassol> errr
11:30 <tmortagne> vmassol: it's OK IMO
11:30 <vmassol> wait is going to wait on the JVM
11:31 <vmassol> so that JVM won't receive the ctrl+c
11:31 <vmassol> tmortagne: ok thanks
11:31 <cjd> my line switches the processes
11:31 <cjd> it forks the wait/rm pidfile into the background
11:31 <tmortagne> we might need to review some scripts here and there but it's probably not a big deal
11:31 <tmortagne> would be nice to have a clear error when using the old way
11:31 <cjd> then *execs* java rather than spawning it so it reuses the pid number from the shell process
11:32 <cjd> so you already know the pid #
11:32 <vmassol> I don't understand anything! :)
11:32 <vmassol> sounds way more complex
11:32 <cjd> oh a downside of this approach is when you stop java, you probably lose your terminal :)
11:32 <cjd> also fixable
11:32 <cjd> but complex bash is bad
11:32 <cjd> b/c it's different on every machine :D
11:35 <Denis> has quit
11:37 <vmassol> one option would have been to reuse http://dev.eclipse.org/svnroot/rt/org.eclipse.jetty/jetty/trunk/jetty-distribution/src/main/resources/bin/jetty.sh but I find it too complex and I don't want to maintain that!
11:38 <cjd> +1
11:38 <cjd> I fight to keep bash out of the cjdns build because I've been bitten by it before
11:38 <cjd> './do' just downloads nodejs (if necessary) and then runs make.js :)
11:39 <vmassol> I already don't like the complexity I'm bringing but I don't see any other way to make our CI more robust...
11:39 <cjd> /nod
11:40 <cjd> some lines of code are less complex than others..
11:41 <cjd> hm
11:41 <Denis> has joined #xwiki
11:43 <tmortagne> vmassol: if it's really only for the CI we could do that at CI level in a pre build step
11:43 <vmassol> nope
11:43 <vmassol> I thought about it
11:43 <vmassol> ;)
11:43 <vmassol> because a job can start xwiki several times
11:44 <tmortagne> for platform yes
11:44 <vmassol> and in any case it can be useful for the iuser anyway
11:44 <vmassol> *user
11:44 <vmassol> so if we have it we might as well expose it
11:48 <cjd> https://ezcrypt.it/539n#8bQTwqZwlcnogPo83DTMgMsq
11:48 <cjd> that is a nice solution IMO
11:48 <vmassol> I really prefer mine
11:48 <vmassol> I now understand it
11:48 <cjd> /nod
11:48 <cjd> you're trappinga sig and relaying it to the other process
11:48 <cjd> but if it's not a SIGINT then it won't work
11:48 <vmassol> I could add more signals
11:49 <cjd> maybe not a big deal *shrug*
11:49 <vmassol> I just trapped INT
11:49 <vmassol> ie ctrl-c
11:49 <vmassol> for now
11:49 <vmassol> we probably need to add the signals sent by jstack too
11:49 <vmassol> (need to check which one it is)
11:49 <vmassol> for heap dump and memory dump
11:50 <cjd> it will send it to the java process directly
11:50 <cjd> my solution is to put the "remove pid" job in another process and then reuse the pid number from the shell for java
11:50 <vmassol> indeed, so even better
11:51 <cjd> but.. they're both equally good until one of them fails on a weird machine :)
11:51 <cjd> and we can't know which one
12:03 <Denis> has quit
12:15 <evalica> has quit
12:17 <evalica> has joined #xwiki
12:21 <Denis> has joined #xwiki
12:41 <Denis> has quit
13:18 <DarkKnightCZ1> hi, what could cause Caused by: org.artofsolving.jodconverter.office.OfficeException: could not store document: img0.html; errorCode: 2074
13:19 <vmassol> https://code.google.com/p/jodconverter/issues/detail?id=102
13:20 <vmassol> converting to PDF?
13:20 <vmassol> ah it's not just pdf apparently
13:20 <DarkKnightCZ1> nope, just office viewer, but i guess that could be permissions issue (i might be running office server under different user, checking)
13:21 <DarkKnightCZ1> hmm, openoffice server is not storing any files, is it?
13:27 <DarkKnightCZ1> nope, so it's not caused by permissions
13:29 <vmassol> no idea, there could be temporary files
13:30 <vmassol> (that would make sense)
13:35 <DarkKnightCZ1> hmm, cannot find it in filesystem, there might be bug in used version of OO
13:39 <ClemensR> maybe indeed the office-server version: https://code.google.com/p/jodconverter/issues/detail?id=102 "The upgrade to LibreOffice 3.5 solved the problem for us."
13:40 <DarkKnightCZ1> ha, openoffice 4.1.0 is broken, libreoffice 4.3.0 is working
13:41 <DarkKnightCZ1> ok, so not actually working, just not throwing that error (resulting in no slides), going to try 3.x
13:41 <vmassol> what os are you on?
13:42 <DarkKnightCZ1> RHEL
13:42 <ClemensR> happy version hunting :-/ ... seems error code 2074 is a "something went wrong" code from oo
13:42 <DarkKnightCZ1> yeah, they have pretty good error codes :)
13:44 <ClemensR> did the office view work at all before? sometimes this error means modules for spreadsheet etc is missing
13:45 <DarkKnightCZ1> i don't know actually
13:50 <DarkKnightCZ1> hmm, WARN o.x.c.ExecutionContext - Tried to remove non-existing property [officeView.caughtException] from execution context.
13:51 <DarkKnightCZ1> thats when there is no exception (but presentations have 0 slides)
13:51 <Denis> has joined #xwiki
14:05 <Slashman> has quit
14:06 <Slashman> has joined #xwiki
14:12 <msmeria> vmassol: hello
14:14 <msmeria> I've just released File Manager v1.1.7 here: https://github.com/xwiki-contrib/application-filemanager/releases/tag/application-filemanager-1.1.7
14:14 <msmeria> can you promote the release to Nexus ?
14:14 <vmassol> hey why me? :)
14:15 <vmassol> finishgin sth
14:15 <vmassol> you should ask committers in general
14:16 <Denis> has quit
14:19 <msmeria> tmortagne, Enygma`, evalica: can one of you guys do it ?
14:19 <cjd> has quit
14:24 <ClemensR> msmeria: I can give it a try
14:24 <msmeria> ClemensR: sweet! :D
14:25 <cjd> has joined #xwiki
14:26 <msmeria> ClemensR: you have instructions here: http://contrib.xwiki.org/xwiki/bin/view/Main/WebHome#HReleasetheproject
14:26 <msmeria> just before the "Publishing on extensions.xwiki.org" section
14:27 <ClemensR> ah, cool
14:28 <ClemensR> msmeria: ok, its "closed" now. does this need some testing or should I try to publish it right away?
14:29 <vmassol> I'm going to break some jobs but they'll succeed once platform has finished biulding....
14:29 <vmassol> I've had to modify both plarform and enterprise
14:29 <vmassol> and they both need each other to work fine....
14:30 <tmortagne> ClemensR: release it right away
14:30 <msmeria> ClemensR: you could try just installing it and adding an entry just for kicks...maybe I missed something
14:30 <vmassol> well enterprise needs platform
14:30 <msmeria> or not :)
14:31 <tmortagne> msmeria: a test master would miss something ? :)
14:31 <ClemensR> ok, I guess I just publish it, and if it really fails, there can be 1.1.8 ;)
14:32 <msmeria> tmortagne: it's my second release and I just hope I didn't brake anything :D
14:33 <vmassol> msmeria: indeed normally staging is there so that you can try it out before releasing to the final repo :)
14:34 <msmeria> vmassol: yeah, that's what I read in the doc too: "Make sure to test your extension from the closed staging repo first since a released repo cannot be removed!"
14:34 <ClemensR> when testing my problem is that I always point to the wrong repo :p ... anyway, it should be published now
14:34 <vmassol> ah I documented it cool :)
14:34 <msmeria> ClemensR: checking...
14:39 <andreeapopescu> has joined #xwiki
14:45 <Denis> has joined #xwiki
14:51 <msmeria> ClemensR: thanks! I was able to install it and test it
14:51 <ClemensR> ah, ok, cool :)
14:57 <andreeapopescu> has left #xwiki
14:57 <vmassol> hmm my updated scripts don't work on agents apparently
14:57 <vmassol> 14:36:32.471 [Thread-1] INFO o.x.t.i.XWikiLogOutputStream - Starting Jetty on port 8080, please wait...
14:57 <vmassol> 14:36:32.477 [Thread-2] ERROR o.x.t.i.XWikiLogOutputStream - start_xwiki.sh: 161: wait: Illegal number: $
14:57 <vmassol> some wrong shell used apparently
14:58 <msmeria> has quit
14:58 <cjd> start_xwiki.sh: 79: start_xwiki.sh: [[: not found <-- sh vs. bash
14:59 <vmassol> shouldn't the first line ensure that bash is used?
14:59 <cjd> hmm yeah
14:59 <cjd> and bash should have [[
14:59 <cjd> *usually :D
14:59 <tmortagne> only if you execute the script I think
15:00 <cjd> oh yeah, if you run sh ./script
15:00 <tmortagne> if you do "sh start_xwiki.sh" it probably does not care
15:00 <cjd> then it will still be interpreted in sh
15:00 <vmassol> locally I run sh start_xwiki.sh and it workds
15:00 <vmassol> works
15:00 <vmassol> maybe it's specific to mac
15:00 <tmortagne> I guess jenkins is not doing that
15:00 <vmassol> jenkins is not doing anything
15:00 <vmassol> it's our code
15:00 <vmassol> I'll change it to bash then
15:01 <tmortagne> what is our code doing ?
15:01 <vmassol> sh
15:01 <vmassol> sh start_xwiki.sh -p %s -sp %
15:01 <tmortagne> maybe sh is an alias for bash or a special adavanced sh on mac
15:01 <vmassol> I thought it was ok since it works locally
15:01 <vmassol> ah yes good point
15:01 <vmassol> cehcking
15:01 <tmortagne> s/adavanced/advanced/
15:01 <cjd> sh is different on every machine
15:01 <cjd> really
15:02 <cjd> even versions of the same OS
15:02 <tmortagne> anyway batter be explicit in our code since we require bash
15:02 <tmortagne> s/batter/better/
15:03 <Denis> has quit
15:05 <vmassol> let's see how it goes....
15:05 <vmassol> 45 minutes more....
15:05 <vmassol> actually not that long
15:05 <vmassol> just 3-4
15:10 <vmassol> grr still failing
15:10 <vmassol> but different eerror
15:10 <vmassol> 5:10:06.060 [Thread-3] ERROR o.x.t.i.XWikiLogOutputStream - start_xwiki.sh: line 161: wait: `$': not a pid or valid job spec
15:16 <Denis> has joined #xwiki
15:16 <vmassol> oh
15:16 <vmassol> I know
15:16 <vmassol> :)
15:16 <vmassol> (I thinkà)
15:17 <vmassol> ok so
15:17 <vmassol> I believe that the following line:
15:17 <vmassol> XWIKI_PID=$!
15:18 <vmassol> is transformed by velocity to:
15:18 <vmassol> XWIKI_PID=$
15:18 <vmassol> :)
15:18 <ClemensR> ah, if velocity is filtering the resources
15:18 <vmassol> we're using velocity for this line:
15:18 <vmassol> XWIKI_DATA_DIR=${xwikiDataDir}
15:19 <cjd> mm template preprocessors are good until they aren't
15:19 <vmassol> yes it's dangerous to run velocity on a shell script
15:19 <vmassol> very dangerous
15:19 <cjd> yeah, we should use CPP instead :)
15:20 <vmassol> the code is in PackageMojo.expandJettyDistribution() FYI
15:20 <cjd> mhm
15:21 <Denis> has quit
15:21 <cjd> I wrote an excellent little piece of CPP/js which tests if a CPP macro is defined
15:21 <Denis> has joined #xwiki
15:22 <vmassol> it's easy enough to search/replace here
15:22 <Denis> has quit
15:22 <cjd> what do you use velocity for ?
15:23 <cjd> oh right
15:23 <cjd> so do this:
15:23 <cjd> XWIKI_VELOCITY_INJECTED_CONFIGURATION
15:23 <cjd> ## check abcdx.vm for information about that this is here for
15:33 <vmassol> more complex than that, it needs to work for the distribution pom too, I'll try using some escapes
15:35 <cjd> http://imgs.xkcd.com/comics/success.png
15:36 <DarkKnightCZ> has joined #xwiki
15:36 <DarkKnightCZ> has left #xwiki
15:38 <DarkKnightCZ1> has quit
15:50 <cjd> tmortagne: do you know what happens if you install an extension and there is already existing a document which the extension is to override?
15:50 <tmortagne> cjd: conflict
15:50 <vmassol> grrr seems escapes don't work for some reason…. I'd have expected this to work: XWIKI_PID=$\!
15:51 <tmortagne> i.e. "no idea what to do, let the user choose"
15:51 <vmassol> (according to http://velocity.apache.org/engine/devel/user-guide.html#Advanced_Issues:_Escaping_and_ )
15:51 <cjd> ok, Charles did it on the intranet and it seems to have hung
15:52 <tmortagne> what do you mean ?
15:52 <vmassol> ok I'll do it like this:
15:52 <vmassol> #set ($exclamationMark = '!')
15:52 <vmassol> XWIKI_PID=$$exclamationMark
15:52 <vmassol> ;)
15:52 <cjd> he said it refused to proceed
15:53 <cjd> just "trying forever" to continue
15:53 <tmortagne> always the exact same conflict ?
15:54 <tmortagne> never seen something like this
15:54 <tmortagne> maybe he have some old EM css/js in his browser cache
15:55 <cjd> --> Installation manager does not seems to be sur that this extention is installed.
15:55 <cjd> To be continued
15:58 <ClemensR> vmassol: maybe XWIKI_PID=\$! would have worked, but via #set everybody notices velocity has a word to say here first
15:58 <vmassol> no it doesn't work
15:58 <vmassol> right now I'm trying to figure out why it doesn't work with my local variable too
15:58 <vmassol> and yes I agree it's better for clarity
15:59 <vmassol> I'l doing this
15:59 <vmassol> ## Since this file is parsed with Velocity and since Velocity would render '$!' as '$' we need to be tricky...
15:59 <vmassol> XWIKI_PID=$$exclamationMark
15:59 <vmassol> but nothings gets replaced strangely
15:59 <vmassol> in a wiki page it works though
15:59 <vmassol> for ex: http://playground.xwiki.org/xwiki/bin/view/Main/VMA
16:00 <vmassol> but through the assembly plugin with filtering on it doesn't....
16:00 <ClemensR> maybe #set ($exclamationMark = '$!') and XWIKI_PID=$exclamationMark to do less substitution
16:01 <vmassol> I'll try this since it's also nicer
16:02 <ClemensR> if I try this locally, it builds a xwiki-platform-tool-jetty-resources-6.3-SNAPSHOT.zip without any substitutions :/
16:03 <vmassol> same
16:03 <vmassol> yep
16:04 <vmassol> actually it does substitute for ${xwikiDataDir}
16:04 <vmassol> so it's like local variable have no effect
16:05 <vmassol> and using a property variable in the pom.xml would be ugly ;)
16:05 <ClemensR> for me it does not even this ... must have made a mistake somewhere :(
16:06 <vmassol> I rebuild tools/jetty
16:06 <vmassol> and then distribution/jetty/hsqldb
16:06 <vmassol> then in target, I unzip the zip
16:06 <vmassol> and check start_xwiki.sh
16:07 <ClemensR> seems I need to do the second step, too
16:09 <vmassol> maybe there's an option when you run velocity to not define local variables
16:09 <vmassol> oh
16:09 <vmassol> stupid me
16:09 <vmassol> forget all that
16:09 <vmassol> the assembly plugin must not be using Velocity at all
16:09 <vmassol> it just happens to use a similar syntax
16:09 <vmassol> ${var}
16:09 <vmassol> checking that
16:10 <vmassol> yup
16:10 <vmassol> http://maven.apache.org/plugins/maven-assembly-plugin/examples/single/filtering-some-distribution-files.html
16:11 <vmassol> so the isseu with $! is only with the PackageMojo actually
16:11 <vmassol> and I can apply a similar strategy than the assembyl plugin instead
16:11 <ClemensR> ah, great, that might explain something :)
16:13 <ClemensR> actually there is still a XWIKI_PID=$! in the generated xwiki-enterprise-jetty-hsqldb-6.3-SNAPSHOT/start_xwiki.sh anyway, at least I see something
16:14 <Denis> has joined #xwiki
16:15 <ClemensR> ah, PackageMojo is different ...
16:17 <Denis> has quit
16:34 <Pbas> has quit
16:35 <vmassol> let's see how this goes…
16:36 <ClemensR> not sure about, but XWIKI_PID=${!} seems to have worked for the PackageMojo at least ... as different solution in PackageMojo instead
16:36 <ClemensR> ok, maybe nicer to keep it in sync there
16:40 <vmassol> cool it's working
16:40 <vmassol> see http://ci.xwiki.org/job/xwiki-platform/18/console
16:41 <vmassol> actually
16:41 <vmassol> speaking too fat
16:41 <vmassol> fast
16:41 <vmassol> some other problem now....
16:43 <evalica> has quit
16:52 <Denis> has joined #xwiki
17:02 <Denis> has quit
17:09 <vmassol> oh stupid me
17:14 <benoitc> has quit
17:14 <benoitc> has joined #xwiki
17:14 <vmassol> should work this time….
17:14 <vmassol> can't believe how long it all took....
17:15 <cjd> #bash
17:16 <tmortagne> vmassol: you Java developer :)
17:16 <cjd> naw, bash is always like that
17:16 <vmassol> actually I've worked during a few years coding only in shell in my past life :)
17:16 <tmortagne> cjd: was refering to the ;
17:16 <cjd> that's why I avoid it like the plague
17:17 <tmortagne> in https://github.com/xwiki/xwiki-platform/commit/77261c854551d9d3e0690b5583581ff157786d7d
17:17 <cjd> haha no, the ; are all my lines
17:17 <vmassol> I've developed pretty scary long code (full apps) in shell....
17:18 <vmassol> working! :)
17:18 <momomomomo> has joined #xwiki
17:18 <vmassol> (so far)
17:19 <cjd> friday fun: determining whether a given preprocessor macro is defined w/o using an #ifdef https://github.com/cjdelisle/cjdns/blob/master/util/Defined.h#L18
17:37 <momomomomo> has quit
17:38 <ClemensR> has left #xwiki
17:40 <momomomomo> has joined #xwiki
17:52 <vmassol> someone introduced a checkztyle error in rendering
17:52 <vmassol> home/hudsonagent/jenkins_root/workspace/xwiki-rendering/xwiki-rendering-api/src/main/java/org/xwiki/rendering/block/DefinitionDescriptionBlock.java:0: File does not end with a newline.
17:52 <vmassol> sdumitriu: is that you?
17:52 <vmassol> see http://ci.xwiki.org/job/xwiki-rendering/10/console
17:53 <cjd> yeah, I'm pretty sure he enabled that
17:53 <vmassol> actually no
17:53 <vmassol> more likely to be tmortagne
17:53 <vmassol> https://github.com/xwiki/xwiki-rendering/commit/9ebe8bc850538f01c8f04c58dcea4d874cb877d1
17:54 <vmassol> he's the last one to have committed in rendering
17:54 <vmassol> http://ci.xwiki.org/job/xwiki-rendering/8/
17:54 <tmortagne> vmassol: I don't see how this is related
17:54 <vmassol> before the error ahppened :)
17:54 <tmortagne> there was no change in DefinitionDescriptionBlock
17:54 <vmassol> or it's caused by some changes in commons indirectly
17:54 <tmortagne> pretty sure it's a change in checkstyle setup
17:54 <tmortagne> I don't remember ever having this kind of checktyle error
17:55 <vmassol> sdumitriu: could you check that please?
17:55 <tmortagne> and I really don't see how this is important enough to fail the build...
17:55 <Enygma`> has quit
17:55 <cjd> btw I'd like to hear your comments on my most recent checkstyle related mail
17:55 <tmortagne> "File does not end with a newline." seriously
17:55 <vmassol> cjd: no time at all
17:56 <cjd> will you have time over the weekend?
17:56 <vmassol> no
17:56 <vmassol> right now the priority is the biuld
17:56 <vmassol> and the release
17:56 <vmassol> but if you're interested in checkstyle please review my mail about it
17:56 <cjd> from when?
17:56 <vmassol> (the one I sent 2 weeks before yours)
17:57 <vmassol> pointing to that page: http://design.xwiki.org/xwiki/bin/view/Proposal/DefineSonarRules
17:59 <mflorea> has joined #xwiki
18:01 <cjd> I don't see anything that looks related in the past month
18:02 <vmassol> http://xwiki.markmail.org/search/?q=DefineSonarRules
18:03 <cjd> ahh [xwiki-devs] Playing with Sonar Rules -- aug 17
18:03 <vmassol> the next step is replying to this email by listing the checks that in your opinion should be removed
18:03 <vmassol> all violations can be seen on sonar.xwiki.org
18:03 <cjd> a next step
18:04 <vmassol> it's much larger than juste checkstyle
18:05 <cjd> well we would need to think about what it means in terms of possibly weighting the project down with (what I understand is) a proprietary SaaS service
18:10 <vmassol> cjd: it's as proprietary as xwiki is
18:10 <vmassol> actually less
18:11 <vmassol> not less same since it's also LGPL
18:11 <cjd> ahh ok
18:12 <cjd> well see you on monday
18:12 <cjd> has quit
18:14 <vmassol> bye
18:14 <vmassol> sdumitriu: ping
18:33 <sdumitriu1> vmassol: Pong
18:33 <vmassol> sdumitriu: I missed that it was on a branch
18:33 <vmassol> sorry about that
18:33 <vmassol> I still fixed a violation on rendering though
18:33 <vmassol> (master)
18:34 <vmassol> regarding checkstyle, this is a dead tool
18:34 <vmassol> let me reply quickly
18:37 <tmortagne> has quit
18:39 <vmassol> done
18:40 <woshilapin> has quit
18:42 <vmassol> oh I forgot to merge my changes to the 6.x branch
18:42 <vmassol> doing it
18:47 <mflorea> has quit
19:10 <momomomomo> has quit
19:15 <vmassol> sdumitriu: could you please fix all violations caused by your addition of the New line check?
19:16 <vmassol> I fixed all rendering but now all platform is failing
19:16 <vmassol> (see for example http://ci.xwiki.org/job/xwiki-platform/21/console )
19:17 <sdumitriu1> vmassol: Working on that
19:17 <vmassol> k
19:23 <Slashman> has quit
19:25 <sdumitriu1> From yesterday:
19:25 <sdumitriu1> (12:01:18 PM) sdumitriu1: I wonder if anybody would complain if I break the build...
19:25 <sdumitriu1> (12:06:44 PM) tmortagne: sdumitriu1: these days you could get lucky and have jenkins blamed for it before you fix it
19:59 <cjd> has joined #xwiki
20:00 <cjd> is now known as <Guest12167>
20:04 <momomomomo> has joined #xwiki
20:15 <vmassol> sdumitriu: I didn't see it, I would have said no because we're trying to do the opposite
20:15 <vmassol> been at it for more than week already
20:15 <vmassol> opposite = try to stabillize the buikd
20:15 <vmassol> and it's hard
20:15 <vmassol> and painful
20:16 <vmassol> and we're all supposed to be doing other stuff ....
20:16 <vmassol> I'm actually depressed by the state of our build/ci
20:48 <abusenius_> has joined #xwiki
22:39 <vmassol> has quit
23:08 <OSIMasson> has joined #xwiki
23:12 <OSIMasson> has quit
23:25 <OSIMasson> has joined #xwiki
23:33 <sburjan> has joined #xwiki
23:40 <nickless> has joined #xwiki
23:43 <abusenius_> has quit