[Releasing] Meeting minutes: release team meeting 23.10.2012

[Tanilkan Sinan]

Hi,

Thanks for the meeting today. Here are the minutes:

_Installer scripts_
The biggest blockers now are that release builds fail on qtbase and qtwebkit. Both problens are getting attention.

Potential problem: "the rpath in the binaries on Linux , as well as the path to the framekorks in the binaries on bin, are absolute". Iikka will investigate tomorrow.

Qt Creator sources will not be included in the installers.

_Library renaming_
Agreement not reached on how to proceed with this. Lars was asked to make a decision.

_QA_
Rafal gave an overview of the testing effort for the Qt 5 release.

_Keeping communications public_
All communication to the releasing team will go to the releasig mailinglist.

_configure.exe_
configure.exe remain included in the source packages for now.

The irc log follows below.

Br,
Sinan

[16:00:02] <SinanTanilkan> lars, thiago, steveire, ZapB, rosch, mauricek, treinio, iieklund, akseli, joaijala, sahumada, tjenssen, Rafal: Ping
[16:00:10] <iikka_> pong
[16:00:10] <mauricek> SinanTanilkan: pong
[16:00:10] <lars> pong
[16:00:11] <ZapB> pong
[16:00:17] <rosch> pong
[16:00:26] <johanna> pong
[16:00:37] <Rafal> pong
[16:00:46] <SinanTanilkan> Any comments to the agenda before we start?
[16:00:46] <steveire> pong
[16:00:54] <steveire> Yes
[16:01:12] <steveire> Can we add to the agenda a discussion of keeping communications public?
[16:01:41] <SinanTanilkan> Good idea.
[16:01:48] <SinanTanilkan> I'll add it to the end.
[16:01:52] <sahumada> pong
[16:02:15] <tijensse> pong
[16:02:58] <SinanTanilkan> Ok. I suggest we start with an update on the installer scripts.
[16:03:09] <iikka_> ok
[16:03:33] <iikka_> the biggest blocker now is that release builds fail on qtbase and qtwebkit
[16:03:51] <iikka_> oswald is working on qtbase specific issues and simon on webkit issue
[16:04:16] <iikka_> and as we run the builds in strict-mode, which fail now, we don't get any installers
[16:04:46] <iikka_> but we have been running the builds in non-strict mode internally to get some installers for testing
[16:04:48] <mauricek> question from my side: how does strict mode and creator correlate?
[16:05:04] <mauricek> because non-strict means we do not have a creator build with that SHA, right?
[16:05:35] <iikka_> hmm, there are two parts actually, first the qt5 release build, then the packaging step
[16:05:57] <iikka_> the packaging step will fail if any component is missing (qt5 src, qt5 bin, creator, docs, etc)
[16:06:23] <iikka_> but the builds will stop if the qt5 release build fails, it won't even try to package anything
[16:06:41] <iikka_> but I don't know how Tim has made the creator build itself?
[16:06:57] <iikka_> maybe tim can comment on the creator build?
[16:07:09] <tijensse> used older builds and bfore that tried with own built qt versions
[16:08:10] <tijensse> at the moment it was working on windows and linux with an older builds - mac has problems
[16:08:53] <iikka_> but as mentioned above, we have tested installers internally, those built with non-strict mode, looking ok to me, but I have not done any exhaustive testing on those
[16:09:35] <kkoehne> iikka_: Do you know whether the Linux rpath issues + Mac LD_LOAD_PATH issues are fixed in these?
[16:09:38] <-- hanne (~linaae at 84.49.38.202) has quit (Ping timeout: 244 seconds)
[16:10:14] <iikka_> kkoehne: do you mean build time issues?
[16:11:05] <kkoehne> iikka_: No, the problem with at least the 2012-10-5 builds is that the rpath in the binaries on Linux , as well as the path to the framekorks in the binaries on bin, are absolute
[16:11:44] <kkoehne> iikka_: That is, we'd need to to some chrpath / whatever-the-mac-equivalent-is-on-mac on the bin folders for the binary qt packages
[16:11:58] <iikka_> kkoehne: hmm, I didn't know that. The RPath patching has been working previously
[16:12:14] <iikka_> I can check that tomorrow morning
[16:12:16] <johanna> yes, seems a new issue to me also
[16:12:16] <kkoehne> iikka_: Okay, we'll see. If you have packages I'm happy to test it :)
[16:12:30] <kkoehne> Great
[16:12:49] <iikka_> weird, I have not touched anything related to rpath patching...obviously something has then changed somewhere
[16:13:09] <kkoehne> iikka_: Could be that the packages we're using are (again) too old
[16:13:17] <tijensse> iikka_: the isBinary method was new (somehow)
[16:13:31] <iikka_> need to check tomorrow
[16:13:52] <johanna> we haven't had new packages for a week now, so they might be old :)
[16:14:15] <iikka_> question to all
[16:14:32] <iikka_> there was a gerrit commit, the Angle
[16:15:03] <iikka_> do we need that for beta2 or can we configure with "-opengl desktop" ?
[16:15:25] <SinanTanilkan> I think there is a strong wish to get it in. Maybe lars can comment.
[16:15:31] <iikka_> my concern is that the Angle will cause other build issues as well, at least this close to beta2
[16:15:43] <lars> yeah, would be great to get some exposure to this feature
[16:15:59] <iikka_> ok, then we need to update the build machines
[16:16:02] <lars> yes, understand that this is a concern
[16:16:03] <kkoehne> iikka_: You'd need a DIrectX SDK + package the d3d dll, I understand
[16:16:22] <iikka_> is there official isntructions how to setup the Angle stuff by the book?
[16:16:38] <lars> iikka_: the advantage is that QML will work on any windows box, and not fail if no decent opengl driver is installed
[16:16:43] <sahumada> I think that everything that we need to add to the release packages have to pass the CI system first .. that mean that ANGLE needs to be tested in the CI system
[16:17:26] <iikka_> but meanwhile we do can prepare the build machines to have the DirectX stuff in place
[16:17:38] <sahumada> yes, that's correct
[16:18:01] <lars> yes
[16:18:06] <iikka_> ok
[16:18:19] <jbarron> iikka_: I can help you if you have any problems.
[16:18:24] <iikka_> ok, thanks
[16:18:52] <iikka_> I don't have anything else specific on mind, just waiting for those qtbase and qtwebkit fixes to get staged
[16:18:56] <SinanTanilkan> My understanding then is that we will proceed with ANGLE, and that we believe it will be fine?
[16:18:59] <johanna> I have on question
[16:19:23] <jbarron> iikka_: As long the DirectX SDK is installed, everything should just work. We are currently troubleshooting an issue on MinGW 32-bit, but we are close.
[16:19:23] <sahumada> SinanTanilkan: only after is enabled in the CI system
[16:19:38] <iikka_> jbarron: ok
[16:19:54] <SinanTanilkan> Ok. Thanks.
[16:20:03] <SinanTanilkan> johanna: You had a question?
[16:20:07] <johanna> what do we do with the creator sources, do we keep them in installer or do we remove it?
[16:20:11] <mauricek> jbarron: that d3d.dll dependency, is that a redistributable?
[16:20:24] <jbarron> mauricek: Yes.
[16:20:26] <mauricek> jbarron: let's take that offline, sorry for spamming this channel
[16:20:36] <SinanTanilkan> :)
[16:20:42] <lars> mauricek: it is redistributable
[16:20:57] <jbarron> mauricek: There's even a CAB file specifically for that DLL.
[16:21:09] <lars> johanna: can the sources be added as a separate selectable thing in an easy way?
[16:21:10] <kkoehne> johanna: I think shipping Qt Creator sources isn't a big risk, but netiher do much good. And since it's yet another thing that can break, I'd vote for not shipping it.
[16:21:33] <lars> kkoehne: johanna: yep, whatever is easier to implement
[16:21:36] <johanna> lars: yes, and they have been there
[16:21:59] <johanna> it's easy to add and easy to remove
[16:22:11] <johanna> if only I know which way to go :)
[16:22:19] <mauricek> jbarron: lars: if there is something like the c-runtime redistributable, that'd be awesome. Then that should be integrated into the package rather than some dll
[16:22:23] -*- tijensse thinks about line endings and different source packages for windows and unix
[16:22:48] <johanna> only thing is that afaik there is no source packages available atm
[16:22:58] <lars> johanna: leave it out then
[16:23:14] <jbarron> mauricek: What does the c-runtime redistributale look like? :)
[16:23:15] <johanna> lars: that's fine with me
[16:23:24] <mauricek> jbarron: will send a mail later on
[16:23:29] <jbarron> mauricek: Cool.
[16:24:22] <SinanTanilkan> Ok. Anything else before we move to the next topic?
[16:24:33] <johanna> I have one more thing
[16:25:03] <johanna> the sources in general, there has been lots of bugs that prevent our builds
[16:25:35] <johanna> is there something we could to to keep it stable, so that we would get installer packages for testing?
[16:26:07] <johanna> should we have a bit older shas in qtsdk or someghing?
[16:26:27] <kkoehne> THen you'll miss the build fixes, too :)
[16:26:43] <kkoehne> THe obvious one is creating a 5.0-beta2 branch, and actively cherry-pick stuff
[16:26:53] <sahumada> I think the proper thing to do is to have it tested in the CI .. as I said before
[16:27:03] <johanna> well, if we had all the fixed then they would compile :)
[16:27:23] <johanna> sahumada: I agree, why it's not right now?
[16:27:52] <sahumada> johanna: because of the change from Brisbane to Digia I'd guess .. people are busy with that
[16:28:07] <sahumada> change of the CI system, that is
[16:28:12] <johanna> sahumada: yeah, unfortunate timing..
[16:28:16] <kkoehne> sahumada: That's qtwebkit only?
[16:28:34] <sahumada> kkoehne: right now only webkit on Windows I'd say
[16:28:45] <sahumada> next is ANGLE on Windows
[16:28:53] <kkoehne> sahumada: Well, that's the one that is broken all the time ;)
[16:28:56] <iikka_> "custom" icu usage revealed one bug in qtbase as well
[16:29:25] <iikka_> but oswald is working on that one at the moment
[16:29:34] <sahumada> we should use the same configuration in both places .. CI system and Release systems
[16:29:47] <steveire> We still have a lot on the agenda. Are we still on 'Installer scripts' ?
[16:29:54] <SinanTanilkan> yes.
[16:30:06] <SinanTanilkan> good point steveire.
[16:30:13] <steveire> The meeting is half-over, I suggest moving on.
[16:30:20] <johanna> yes, that was just my concern, maybe we can think about that and continue offline
[16:30:30] <SinanTanilkan> Ok. Let's do that.
[16:30:38] <SinanTanilkan> Has thiago joined yet?
[16:30:51] <SinanTanilkan> We should discuss the library renaming.
[16:30:59] <SinanTanilkan> But let's wait and see if he joins later.
[16:31:09] <SinanTanilkan> Can we have a short update on Qt Creator?
[16:31:15] <SinanTanilkan> Sorry.
[16:31:27] <SinanTanilkan> Are there other issues that we should be concerned about, before we move to creator?
[16:31:45] <tijensse> yes we can: used old qt build to have some packages, still waiting for new ones
[16:31:57] <tijensse> and src packages are not there at the moment
[16:32:19] <tijensse> I think that is all - kkoehne?
[16:32:52] <kkoehne> Yep. We have 7z files for Linux and Windows, hit a wall on mac though (which we still try to sneak around).
[16:33:36] <SinanTanilkan> I talked with jing, who had built creator on qt 5 from qtsdk.git. Her experience is that mac is looking better than it did for some time.
[16:33:45] <SinanTanilkan> Do you have the same experiences?
[16:34:40] <kkoehne> I haven't seen Qt Creator / 5on Mac for some time, sorry
[16:34:56] <tijensse> as kkoehne said we still have no package here, I think only eike could say something about the current situation, but he is not at the keyboard at the moment
[16:35:06] <SinanTanilkan> Ok.
[16:36:04] <SinanTanilkan> What is your overall feeling for creator on qt 5? If all qt5 blockers are fixed during the next few days. Would creator be close to ready for release?
[16:36:41] <SinanTanilkan> No need for specific dates, just "it's looking good" or "it's looking bad".
[16:37:07] <kkoehne> From my perspective it's looking good on Windows and Mac. I'm working daily with it, without any annoying issues any more.
[16:37:15] <kkoehne> Sorry s/Mac/Linux/ :)
[16:37:21] <SinanTanilkan> Sounds good.
[16:37:46] <SinanTanilkan> Ok. Anything else before we move to the next topic?
[16:37:57] <tijensse> from my side: no
[16:38:09] <SinanTanilkan> Error management. Topi was not able to join.
[16:38:27] <SinanTanilkan> The status is that there are no issues he is very concerned about
[16:38:34] <SinanTanilkan> (except the ones we have discussed).
[16:38:49] <SinanTanilkan> So unless anyone here wants to raise anything. I suggest we move to the next topic.
[16:39:18] <SinanTanilkan> Whic is QA - Rafal.
[16:39:39] <Rafal> hi all
[16:40:02] <Rafal> I was asked for collecting available pieces information about testing and quality, related to Qt5.
[16:40:34] <Rafal> Since the previous point was Err Mgmnt  I want to add something...
[16:40:53] <SinanTanilkan> Please do.
[16:40:55] <Rafal> Qt5 BUGS SITUATION:
[16:40:55] <Rafal> Number of unresolved bugs:  P0:3, P1:58     (and only those reported after August 2012:  P0:2, P1:31).
[16:41:30] <Rafal> From QA perspective this is something to be concerned about.
[16:41:51] <Rafal> ... if we want to release Qt 5 SDK.
[16:42:15] <Rafal> (This is what I found, please feel free to interrupt me anytime you want to add something.
[16:42:15] <Rafal> I'm still learning the way how qt works and is structured, so please let me know when I'm saying something untrue.
[16:42:15] <Rafal> )
[16:42:25] --> thiago (~thiago at kde/thiago) has joined #qt-releases
[16:42:47] <SinanTanilkan> Thanks for raising this. It's important, but i think we can still have a valuable beta release with known bugs.
[16:43:41] <iikka_> SinanTanilkan: Rafal: should be create Jira tickets from those qtbase and qtwebkit bugs I mentioned earlier and link them into beta2 release?
[16:44:00] <SinanTanilkan> iikka_: please do. (unless they are allready fixed)
[16:44:23] <Rafal> I think it may be a question to Topi... I'm not sure.
[16:44:29] <iikka_> not fixed yet, I'll add those tomorrow
[16:44:35] <Rafal> ok
[16:44:37] <SinanTanilkan> iikka_: thanks.
[16:44:40] <SinanTanilkan> Rafal: Please proceed.
[16:44:54] <Rafal> MANUAL TEST:
[16:45:08] <Rafal> After installer package is ready for testing, the smoke test will be performed by QA team in Oslo.
[16:45:16] <Rafal> (one day)
[16:45:23] <Rafal> Then, a sanity test is planned to be executed. Everybody is welcome to help. Developers in Oslo will be asked for help. Mac/Linux/Windows to be covered.
[16:45:53] <Rafal> (In the future: This effort could be organized. Demos and examples could be tested. It would be nice to have an automated release test)
[16:46:26] <Rafal> About autotest:
[16:46:41] <Rafal> After migration to new Jenkins CI system, we lost graphical representation of autotest results. Now, the work on getting it back is ongoing. We're going to use xUnit Jenkins plugin.
[16:47:12] <Rafal> (In the future: we want to be able to collect and show information about the test coverage and think how to evaluate test efficiency.)
[16:47:50] <Rafal> about performance test:
[16:48:11] <Rafal> I'm not aware of if there is any relevant performance test. Do you know something about it?
[16:48:28] <lars> Rafal: nothing really relevant, no.
[16:48:39] <Rafal> (In the future: A testing approach should be decided (load test, performance test, stress test, ...?))
[16:49:10] <Rafal> That's all from my side...
[16:49:31] <lars> Rafal: let's leave perf testing out from here for now. We won't have anything automated in place for 5.0 anyway
[16:49:50] <Rafal> I agree. It is long time perspective.
[16:50:09] <SinanTanilkan> Ok. Any comments to what we have for the qa overview?
[16:50:29] <Rafal> Can we go back to those P1s we have?
[16:50:44] <SinanTanilkan> Go ahead.
[16:50:48] <Rafal> How are we going to handle them?
[16:51:19] <SinanTanilkan> There are quite a few people in the office working on fixing bugs.
[16:51:27] <steveire> Will the autotest results you have ever be publically visible?
[16:51:57] <lars> steveire: that's the goal with the xUnit integration afaik.
[16:52:20] <Rafal> We're thinking about it. First we want to see them by ourselves.
[16:52:31] <steveire> Cool, thanks.
[16:52:40] <lars> steveire: need to get something working at all first though... :)
[16:53:14] <sahumada> steveire: i have a proof of concept here http://qtstaticanalysis.no-ip.org:8080/job/Qt5Base_master_Tests/lastCompletedBuild/testReport/
[16:53:45] <Rafal> sahumada: Many thanks to you  :-)
[16:54:10] <SinanTanilkan> Ok. Anything more on the qa side?
[16:54:47] <SinanTanilkan> Next topic then.
[16:54:50] <SinanTanilkan> Two topics left. Thiago, could we have an update on the library renaming?
[16:55:20] <thiago> SinanTanilkan: yes
[16:55:25] <steveire> sahumada: Thanks, yes, I've seen that before too. I didn't know if it was showing me actual CI runs though.
[16:55:28] <thiago> last proposal that most people agreed with has been posted
[16:55:39] <thiago> there are two people against them, one is ossi
[16:56:01] <thiago> at this point, I'd like to ask lars to give his official blessing one way or another
[16:56:36] <lars> thiago: isn't the only conflict now about documentation?
[16:57:30] <thiago> lars: ossi doesn't like the principle at all
[16:57:51] <lars> thiago: ok, will re-read in detail tonight
[16:58:26] <lars> mauricek: are there any issues left with thiago's latest proposal from the tooling side?
[16:58:28] <thiago> I can get started in the wrapper in my afternoon today so you'll have a prototype by your morning tomorrow
[16:59:03] <mauricek> lars: best is to talk to Andre about it. I think there is still concerncs regarding the other platforms and the impact that might be there
[16:59:33] <lars> thiago: ok, but the wrapper would IMO not be crucial for beta2, right?
[16:59:49] <lars> mauricek: ok, will give him a call tomorrow morning
[17:00:16] <mauricek> lars: ping him first to turn on his phone. He is not that much into this mobile usage thingy ;)
[17:00:26] <lars> mauricek: hehe, will do
[17:00:29] <SinanTanilkan> Do these changes have any implications for the installer scripts (or other packaging reltated work)?
[17:00:46] <thiago> lars: the wrapper itself?
[17:00:52] <thiago> but I'd like to move the tools, including qmake
[17:01:01] <thiago> so people would need to find qmake in libexec...
[17:01:17] <thiago> temporarily, we could have a copy/symlink of qmake
[17:01:27] <kkoehne> thiago: But things stay like they are in the binary SDK , right?
[17:01:33] <iikka_> thiago: can you mail me and joaijala about that change so we can check if it affect packaging?
[17:02:04] <thiago> kkoehne: except for tools moving around, yes
[17:02:17] <thiago> iikka_: yes
[17:02:23] <iikka_> thanks
[17:02:40] <mauricek> thiago: hm... can you check with joerg, how visual studio currently searches for the tools?
[17:02:56] <mauricek> I'm afraid vs projects will fail after that, but he knows best
[17:02:57] <lars> mauricek: most likely in the path
[17:03:23] <mauricek> lars: talking about using visual studio itself, not the command prompt with visual studio environment setup
[17:03:52] <mauricek> lars: thiago: significant amount of customers use that setup, hence we must have something prepared for them at final release latest
[17:04:05] <lars> mauricek: agree
[17:04:27] <SinanTanilkan> Ok. Anything else before we move to the next topic?
[17:04:39] <thiago> mauricek: will do
[17:04:39] <SinanTanilkan> Last topic of the day: Keeping communications public
[17:04:41] <lars> it needs to work for them with as little changes as possible
[17:05:16] <SinanTanilkan> steveire ^
[17:05:25] <lars> SinanTanilkan: that one's simple. I agree with steveire. There's no reason not to post stuff to the ML. Simply a  matter of transparency
[17:05:35] <lars> at least simple from my POV
[17:05:52] <SinanTanilkan> Ok. Will do that.
[17:06:05] <SinanTanilkan> I guess that's all for now then.
[17:06:09] <steveire> Yes, all group communication to the mailing list by default.
[17:06:15] <sahumada> I got one little thing
[17:06:25] <SinanTanilkan> sahumada: Go ahead.
[17:06:25] <sahumada> should we ship configure.exe in our source packages ?
[17:06:37] <steveire> I'd also like to see logs from these meetings posted to the mailing list.
[17:06:40] <lars> thiago: can we talk tonight about the remaining work on the renaming front? I could give you a call...
[17:07:05] <SinanTanilkan> steveire: Ok. I can add them to the bottom of the minutes.
[17:07:22] <lars> sahumada: IMO no, since bootstrapping works fine afaik
[17:07:34] <steveire> SinanTanilkan: Thanks.
[17:07:35] <thiago> lars: sure, I can call you too
[17:07:52] <lars> thiago: cool, thanks. 9pm Oslo time?
[17:08:11] <sahumada> ok .. then we should remove it from the source packages and just let people bootstrap it
[17:08:39] <lars> sahumada: as long as it's documented in the build instructions it's fine IMO
[17:08:45] <kkoehne> sahumada: Does it require perl or something?
[17:09:03] <johanna> sahumada, lars: is this final decision? I think ossi has something to say on that, and this change requires us to update our builds
[17:09:07] <sahumada> kkoehne: no .. it's boostrapped with configure.bat
[17:09:12] <lars> sahumada: my long term goal is that the source packages are just a checkout from git with a certain sha1 :)
[17:09:13] <sahumada> https://codereview.qt-project.org/37906 if people can +2 that change
[17:09:17] <kkoehne> sahumada: Which reads "@perl.exe %~dp0configure %*"
[17:09:17] <thiago> ossi wants configure.exe in the tarballs
[17:09:24] <thiago> I'd prefer not to
[17:09:28] <thiago> he has his reasons, though
[17:09:38] <thiago> especially that "configure -help" starts building before printing the help
[17:09:55] <sahumada> kkoehne: https://qt.gitorious.org/qt/qtbase/blobs/master/configure.bat
[17:10:09] <lars> johanna: if it requires additional work, I'd say let's postpone any changes until after beta2
[17:10:18] <kkoehne> sahumada: Oh, I see
[17:10:35] <lars> thiago: configure -help on linux also runs the tests (ie. builds qmake IIRC)
[17:10:49] -*- kkoehne is envious to give it a try on mingw32 :)
[17:10:52] <johanna> lars: ok, I can see this raises opinions :), just  want to know in which state we start doing anything..
[17:11:12] <-- iikka_ (~iikka at 188.238.161.199) has quit (Quit: L?hd?ss?)
[17:11:12] <sahumada> kkoehne: but syncqt needs perl though :)
[17:11:12] <johanna> if we do change anything for beta2..
[17:11:18] <thiago> lars: no, the help comes from the script. The problem is that the .bat file on windows does nothing, it needs to build configure.exe
[17:11:23] <kkoehne> sahumada: BUt that one is predone?
[17:12:07] <thiago> I'd prefer the tarball to be pristine. If we're modifying it for any reason, we might as well include configure.exe.
[17:12:15] <mauricek> kkoehne: i think that has changed some while ago
[17:12:22] <lars> thiago: actually configure -help shows the defaults afair, which in turn requires running the configure tests, which requires qmake
[17:12:24] <thiago> I haven't looked at mksrc.sh recently to see if it modifies in other ways
[17:12:24] <mauricek> kkoehne: at some point we stopped calling syncqt for the source package
[17:12:47] <lars> thiago: easy to try out with a clean tree though :)
[17:13:07] <thiago> lars: some tests, not all. It shows the defaults after parsing the command-line, but not the tests. The "show defaults" feature is a mis-feature.
[17:13:09] <kkoehne> Alright. I just remember this s***storm from a previous release where people where crying out because we added perl or whatever to the requirements
[17:13:16] <mauricek> ok, are we done? Lots of parallel discussion right now
[17:13:27] <lars> thiago: ok.
[17:13:28] <johanna> mksrc.sh doesn't create the configure.exe now, we have another build job for that
[17:13:29] <sahumada> https://codereview.qt-project.org/37906 https://codereview.qt-project.org/37702 https://codereview.qt-project.org/37704 if people is interested in the configure.exe history :)
[17:14:03] <sahumada> yes, and that's error probe because we need to download the .zip file .. build configure.exe .. put configure.exe into the .zip file .. and upload it again
[17:14:09] <lars> johanna: thiago: so I'd say we leave it in for now. we can still try to get to a fully clean source tree in later releases :)
[17:14:28] <lars> it's not that important right now
[17:14:40] <thiago> agreed
[17:14:46] <johanna> I'm good with that
[17:14:55] <lars> good :)
[17:15:42] <lars> ok, guess we're done.
[17:16:12] <SinanTanilkan> Thanks you all for your contributions.
[17:16:19] <SinanTanilkan> See you next time.
[17:16:24] <mauricek> bye
[17:16:26] <ZapB> bye
[17:16:26] <kkoehne> bye
[17:16:30] <thiago> see ya
[17:16:30] <johanna> thanks, bye
[17:16:32] -*- thiago goes to the office
[17:16:35] <Rafal> bye
[17:16:37] <SinanTanilkan> :) bye
[17:17:09] <lars> bye
[17:20:52] <tijensse> bye