Open in new window / Try shogun cloud
--- Log opened Tue Aug 20 00:00:23 2013
-!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has left #shogun []00:02
@sonney2kdumm die dumm00:12
@wikingwazza00:12
@sonney2kwiking, I think we should rent some core i7 based system from gsoc money for arbitrary tortures00:17
@sonney2kas in it could host the demo and do the buildbot00:17
@wikingsonney2k: hetzner.de00:17
@wiking;)00:17
@sonney2kand we could all have logins there00:17
@wikingindeed00:17
@wikingok big commit is coming up00:18
@sonney2kwiking, or https://www.ovh.de/dedicated_server/isgenug.xml the 24G / i7 ~35 EUR/month00:18
@wikingoooh00:19
@wikingthis seems to be even better00:19
@wikingalthough00:20
@wiking50 eur/month for 32 gigs ram ;P00:20
@wikingwhich could really be useful when we have like 3-4 buildbots in qemu/kvm00:21
@wikinghttp://www.hetzner.de/hosting/produkte_rootserver/ex4000:21
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun00:21
shogun-notifier-shogun: Viktor Gal :develop * d43dd66 / / (22 files): https://github.com/shogun-toolbox/shogun/commit/d43dd6649255b38cc6b655b7670c59829ab0ac0100:21
shogun-notifier-shogun: Add BaggingMachine00:21
shogun-notifier-shogun: Add basic label combination rule implementations for ensemble learning00:21
shogun-notifier-shogun: Add MockMachine, MockFeature, MockLabels00:21
@wikingthere weee goooo ;) baggingmachine ;)00:22
@wikingand ensemble learning00:22
@wikingsonney2k: i think we should clone an osx nightly00:22
-!- iglesiasg [~iglesias@89-77-4-38.dynamic.chello.pl] has joined #shogun00:22
-!- mode/#shogun [+o iglesiasg] by ChanServ00:22
@wikingsonney2k: because it's really not needed to create packages after each and every commit00:22
@iglesiasghi guys00:23
* wiking wonders if this is going to compile on travis 00:23
@wikingooooh i forgot to upgrade modshogun.i :(00:24
@wikingsonney2k: when do we start with the D-classing?00:24
@wikingiglesiasg: yo00:24
@sonney2kwiking, well just do it00:24
@iglesiasgwiking, I just saw you pushed new stuff, nice!00:24
@wikingsonney2k: ah ok00:25
@wikingsonney2k: i guess it's going to be some scripty shit :)00:25
@wikingiglesiasg: heheh yeah it's the first attempt00:25
@sonney2kno hand work00:25
@wikingiglesiasg: haven't tested it at all... :P and i really want to see if i enable openmp how it will crash to pieces :D00:26
@sonney2kwe really have to think hard per class what we want to expose and what not00:26
@iglesiasgwiking, hehe I see00:26
@wikingHeikoS: can u plz check on BaggingMachine why i clone_equal fails on it with segfault (you need to remove it from the ignores list in clone_unittest.cc.py script to get the error)00:27
@wikingiglesiasg: should work with a simple thread00:27
@wikingiglesiasg: i should add an ipython notebook for it :)00:27
@wikingsonney2k: speaking of which... can u fire up that ipython notebook generator script?00:28
@iglesiasghushell, hey! I see that there a few comments in the PR you have not addressed yet, is it still work in progress? Just wondering since you told me to check if it was good to merge :)00:28
@iglesiasgwiking, I will get definitely into ipython notebook this week00:28
@wikingiglesiasg: yeah that's a cool stuff ;)00:29
@wikingiglesiasg: make sure you have 1.0 though...00:29
@iglesiasgwiking, ipython 1.0 at least? why is it so?00:29
@wikingiglesiasg: well that's really the stuff... i mean 0.13 is cool as well but our ipython generator will be 1.000:30
@wikingso to avoid surprises :P00:30
@iglesiasgwiking, it makes sense00:31
@iglesiasgI guess it is easy to get 1.0 in ubuntu in any case00:31
@sonney2kwiking, well can you help getting the static interfaces back to work?00:31
@wikingsonney2k: yeah i'll do it00:31
@wikingsonney2k: there's a FindMatlab.cmake by default in 2.8 cmake00:31
@wikingso at least we dont have to care about that too much00:31
@sonney2kwiking, and btw have you seen HeikoS' last commit? We should drop that file and increase the required CMake version number00:32
@sonney2klisitsyn, btw have you updated INSTALL already? I mean at least a bit?00:32
@wikingsonney2k: ooh noooo... i told him to solve it w/o that function00:33
@sonney2kwiking, this file is at least in .9 but not in .4 (as heiko said)00:33
@wikingsonney2k: i'll change the superLu detection00:33
@wikingsonney2k: hence we wont need this macro00:33
@sonney2kwiking, well it doesn't really matter - we can require a newer cmake version00:33
@wikingof course apt-get is now failing on traviw... yeee \o/00:34
@sonney2kafter all this is a big switch...00:34
@wikingsonney2k: should we support an EnsembleMachine?00:35
@wikingsonney2k: which receives a DynObjArray of machines + weights00:35
@wikingand basically does the learning...00:35
@wikingsonney2k: oh and i see that you started to switch the python scripts to simply just import everything from modshogun00:38
@wikingsonney2k: imo it's a good idea... are going to make this default?00:38
@wiking*are we00:38
@wikingjust that whether i should create an Ensemble/__init__.py or not at all...00:39
@wikingas we are really dropping that00:39
gsomix_wiking, hey. I'll fix problems with csv and parses. it seems it's related to locales and reals representations in files.00:42
gsomix_and thanks for report00:42
@wikinggsomix_: heehhe locale00:43
@wikingwtf00:43
@wikingjesus... great that you've managed to find out the underlying bug00:43
@sonney2kwiking, the only worrying thing about the hetzner server is the fixed traffic limit. with ovh you have 'only' 100MB/s but guaranteed00:45
@sonney2kwiking, yes lets only ever use modshogun00:45
@sonney2kgsomix_, yes sure that is why we always set to C locale in ascii file00:47
@wikingsonney2k: mmm i have like 20TB with hetzner... and i never managed to get over it00:51
@wikingsonney2k: and even then you r only shaped to 10mbit00:51
@sonney2kwiking, well we are talking buildslave right?00:51
@sonney2kwiking, but permanently then00:51
@wikingyep00:51
@wikingsonney2k: no just for the month... or?00:51
@sonney2kwiking, no00:51
@sonney2khttp://wiki.hetzner.de/index.php/Traffic00:51
@sonney2kahh german00:51
@wikingaaah00:51
@sonney2kit says permanent if you have no good excuse for this one time thing00:51
@sonney2khttp://wiki.hetzner.de/index.php/Traffic/en00:51
@sonney2kenglish00:51
@wikingWe will permanently restrict the connection speed if more than 20 TB/month are used (the basis for calculation is for outgoing traffic only. Incoming and internal traffic is not calculated).00:51
@wikingwould we have a lot of outgoing?00:51
@sonney2ktrue00:51
@sonney2kwould just be incoming00:51
@wikingas buildslave is really just d/l-ing00:51
@sonney2kthen yes00:51
@wikingsonney2k: we just need some good virtualisation... kvm is good enough? because i had really troubles with qemu :S00:51
@wikingor we should just go with Xen?00:51
@sonney2kwiking, we don't even need virtualization00:51
@sonney2kwiking, chroot would be good enough00:51
@wikingsonney2k: how do we do the different os/distro thingy?00:51
@wikingsonney2k: chroot a fedora within an ubuntu...?00:51
@sonney2kwiking, just install the os's in a couple of chroot's00:51
@wikingah ok00:51
@sonney2kwiking, why not?00:51
@wikingif that's good enough00:51
-!- shogun-notifier- [~irker@7nn.de] has quit [Ping timeout: 256 seconds]00:51
@wikingsonney2k: i guess we just gonna run into trouble with funky os00:51
@wikinglike netbsd or openbsd00:51
@wiking;P00:51
@sonney2kwell if we needed security on this scratch machine then not ...00:51
@wikingor?00:51
@sonney2ktrue00:51
@wikingbut that we can Xen00:51
@sonney2kbut hey we have a working setup00:52
@sonney2kso I would only move over debian00:52
@wikingsonney2k: fedora, centos and an ubuntu slave would be great in that00:53
@wikingand maybe a funky bsd ;P00:53
gsomix_wiking, arch linux? :]00:55
@sonney2kwiking, well ubuntu/redhat + centos/debian derivatives are 90% of the linux distributions00:57
@sonney2kI would assume bsd people can compile their own code :)00:57
@wikingsonney2k: heheh yeah true... but it'd be great to see at least that the stable centos,fedora, ubuntu comiples/runs fine ;)00:59
@sonney2kwiking, ok I am too lousy with cmake01:08
@sonney2kwiking, when I do01:08
@sonney2krm -rf build; mkdir build; cd build; cmake -DPythonModular=ON -DCMAKE_PREFIX_PATH=/usr/local/stow/shogun-2.2.0 ..01:09
@sonney2kcmake overwrites shogun/Makefile01:09
@wikingsonney2k: that's rather interesting :S01:10
@sonney2kwiking, try your own!01:10
@sonney2kurgs01:11
@sonney2know it is doing this already with cmake ..01:11
@wikingyou mean the <shogun root>/Makefile ?01:12
shogun-buildbotbuild #1559 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1559  blamelist: Viktor Gal <viktor.gal@maeth.com>01:12
@sonney2kwiking, yes01:12
@wikingwoah01:12
@wikingi wonder01:12
shogun-buildbotbuild #1441 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1441  blamelist: Viktor Gal <viktor.gal@maeth.com>01:12
shogun-buildbotbuild #1472 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1472  blamelist: Viktor Gal <viktor.gal@maeth.com>01:17
@sonney2kwiking, can one build with python3 and python2 installed now?01:18
* sonney2k tries01:18
-!- iglesiasg [~iglesias@89-77-4-38.dynamic.chello.pl] has quit [Quit: Ex-Chat]01:18
@sonney2kman what slows down shogun compile time *THAT MUCH*01:18
@sonney2kit is ridiculously slow now01:19
@wikingsonney2k: did u set -DCMAKE_BUILD_TYPE=Debug01:20
@wiking?01:20
@wikingotherwise you are building release01:20
@wikingand that's optimizing shogun's mother as well...01:20
gsomix_sonney2k, why dense labels are float64_t?01:20
@sonney2kgsomix_, well common type for regression,classification etc01:21
@sonney2kwiking, yeah but still we have massively slowed down01:21
gsomix_sonney2k, so labels not only represents discrete classes, right?01:21
@sonney2kgsomix_, indeed real valued numbers!01:21
@sonney2kgsomix_, btw maybe it is a good idea to also be more tolerant when reading in integer's - I mean use the float parsers and then convert to ints01:22
@sonney2kgsomix_, otherwise things like 1e+000 cannot be parsed01:22
gsomix_sonney2k, makes sense01:23
shogun-buildbotbuild #8 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/8  blamelist: Viktor Gal <viktor.gal@maeth.com>01:24
@sonney2keven building the examples is slooow01:26
@sonney2k95%01:26
@wikingsonney2k: yeash because there  forgot to remove the optimization flags when building in release mode01:28
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun01:28
shogun-notifier-shogun: Soeren Sonnenburg :develop * cc383c2 / examples/undocumented/libshogun/parameter_iterate_float64.cpp: https://github.com/shogun-toolbox/shogun/commit/cc383c27fc7614e659c7d9754712bb5e33b3fa7d01:28
shogun-notifier-shogun: make n a local variable to fix name clash01:28
@sonney2kwiking, no but that is ok. should be optimized01:28
@wikingsonney2k: really?01:29
@sonney2kwiking, I am just wondering why it is slow - basically the examples use very little headers and just link against shogun01:29
@sonney2kdone01:29
@sonney2kohh no01:29
@sonney2kI did commit in the meantime01:29
@sonney2know examples are linked again01:29
@sonney2kdone01:30
gsomix_3:30 AM, chickens started screaming in my window01:31
@wikinghehehehe01:33
gsomix_I'd say roosters, not chicks. but chicks are better of course.01:34
@sonney2kgsomix_, time to get your well deserved eggs01:34
* sonney2k finally has modshogun compiled and working for python01:36
@wiking\o/01:36
@sonney2kpython ~/ipynb_run_save.py svm_binary_classification.ipynb01:37
@sonney2krunning svm_binary_classification.ipynb01:37
@sonney2ktesting:01:37
@sonney2k8 code cells from notebook Untitled001:37
@sonney2kwrote svm_binary_classification_run_saved.ipynb01:37
@sonney2kyes works01:37
@sonney2kalright then night!01:38
gsomix_wake me up if I fell asleep01:38
shogun-buildbotbuild #1442 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1442  blamelist: Soeren Sonnenburg <sonne@debian.org>01:38
@wiking\o/01:40
shogun-buildbotbuild #1473 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1473  blamelist: Soeren Sonnenburg <sonne@debian.org>01:42
shogun-buildbotbuild #9 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/9  blamelist: Soeren Sonnenburg <sonne@debian.org>01:50
shogun-buildbotbuild #1263 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1263  blamelist: Viktor Gal <viktor.gal@maeth.com>01:54
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has quit [Ping timeout: 256 seconds]02:14
shogun-buildbotbuild #1560 of deb3 - modular_interfaces is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/156002:28
shogun-buildbotbuild #1264 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1264  blamelist: Soeren Sonnenburg <sonne@debian.org>02:35
-!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has joined #shogun03:03
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]04:28
gsomix_good morning?07:11
@sonney2kgsomix_, good morning!08:25
@sonney2kgsomix_, did you have your eggs for breakfast?08:26
gsomix_sonney2k, I haven't had breakfast yet.08:27
@sonney2kgsomix_, I did!08:38
gsomix_but yes, little chicken coop gives us tasty eggs. little farm, huh.08:39
gsomix_sonney2k, cool. :)08:39
@sonney2kwiking, I guess you are still asleep - anyways nightly_default should compile for all interfaces we support :)08:48
@sonney2kwiking, it also has to install to appropriate paths since the doc generator runs things like sg('help') to generate the doc on-the-fly08:49
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]08:55
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun08:55
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Client Quit]08:59
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun08:59
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Client Quit]08:59
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun08:59
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has quit [Quit: Leaving.]09:18
-!- gsomix_ [~gsomix@178.45.77.251] has quit [Ping timeout: 264 seconds]09:32
-!- gsomix_ [~gsomix@178.45.44.239] has joined #shogun09:37
shogun-buildbotbuild #1265 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/126509:42
@wikingsonney2k:  okey i'm on it09:48
@wikingHeikoS: ping09:48
-!- gsomix_ is now known as gsomix10:06
-!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun10:22
-!- gsomix [~gsomix@178.45.44.239] has quit [Ping timeout: 245 seconds]10:48
lisitsynsonney2k: yes I updated INSTALL a bit10:48
-!- gsomix [~gsomix@178.45.44.239] has joined #shogun11:08
@sonney2kwiking, http://nbviewer.ipython.org/url/shogun-toolbox.org/static/notebook/svm_binary_classification.ipynb11:14
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]11:16
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun11:16
@sonney2kwiking, we have another issue on cygwin - the tests are looking for cygshogun-14.dll11:17
@sonney2kshouldn't it be just libshogun*.dll ?11:18
thoralfCould merge my PR https://github.com/shogun-toolbox/shogun/pull/1444?11:25
thoralfsomebody11:26
-!- van51 [~van51@ppp-2-87-0-126.home.otenet.gr] has joined #shogun11:36
van51hello11:37
@HeikoSthoralf: there you go11:37
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun11:37
shogun-notifier-shogun: Thoralf Klein :develop * dc1ed1f / src/shogun/io/BinaryFile.cpp: https://github.com/shogun-toolbox/shogun/commit/dc1ed1f3081c0d7b726e9988987ff875deef6a5d11:37
shogun-notifier-shogun: Fixed that BinaryFile::get_sparse_matrix() did not set num_feat properly.11:37
shogun-notifier-shogun: Heiko Strathmann :develop * 0f20584 / src/shogun/io/BinaryFile.cpp: https://github.com/shogun-toolbox/shogun/commit/0f20584bcc268837f21a770b1376eafb64d8aaec11:37
shogun-notifier-shogun: Merge pull request #1444 from tklein23/num_features_in_binaryfile11:37
shogun-notifier-shogun:11:37
shogun-notifier-shogun: Fixed that BinaryFile::get_sparse_matrix() did not set num_feat properly11:37
thoralfHeikoS: Thanks.  I'll rebase the other PR and tell you when travis goes green.11:38
@HeikoSthoralf: cool, thanks a lot!11:38
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]11:39
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun11:39
@sonney2kHeikoS, http://nbviewer.ipython.org/url/shogun-toolbox.org/static/notebook/mmd_two_sample_testing.ipynb11:43
@HeikoSsonney2k: autogenerated?11:43
@sonney2kyes11:43
@HeikoSsonney2k: sweet!11:43
@sonney2khttp://shogun-toolbox.org/buildbot/builders/nightly_default/builds/49811:44
@sonney2knotebooks updated11:44
@HeikoSso now we only need to structure them a bit and have a website with a list11:44
@sonney2kand fix the issues11:44
@sonney2k8 cells failed to complete11:44
@sonney2k...11:44
@HeikoSsonney2k: will do11:44
@HeikoSsonney2k: is there a list of notebooks somewhere?11:44
@HeikoShttp://nbviewer.ipython.org/url/shogun-toolbox.org/static/notebook doesnt work11:44
@HeikoSah sorry11:45
@sonney2kohh it is just a path thing11:45
@HeikoSthats the viewer11:45
@HeikoSbut is there a way to see all notebooks?11:45
@HeikoSwe also should have small abstracts for them and then generate a website with abstract and link11:45
@sonney2kHeikoS, well we have to think of one11:45
@HeikoSI will do this in a while11:45
@HeikoSits cool that the notebooks are still generated when cells fail11:46
@sonney2kmaybe we could automatically parse the notebooks folder and then display links11:46
@sonney2ksome image snapshots would be cool11:46
@HeikoSsonney2k: yeah but I would like to have something nicer11:46
@sonney2kHeikoS, well then they are semi-broken though11:46
@HeikoSabstract11:46
@HeikoSgenerated as the documented examples11:46
shogun-buildbotbuild #1443 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1443  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>11:48
shogun-buildbotbuild #1474 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1474  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>11:52
@wikingHeikoS: yo11:57
@HeikoSwiking: yo!11:57
@wikingHeikoS: ideas about clone ?11:58
@HeikoSwiking: sorry I am not up to date, whats happening?11:58
@wikingHeikoS: yesterday i've commited BaggingMachine and a bunch of other stuff11:58
@wikingHeikoS: and i had to remove the BaggingMachine from the clone_unittest.cc.py, i.e. add it to the ignores as for some reason the clone method segfaulted11:59
@HeikoSwiking: I see11:59
@HeikoSwiking: would be good to have these more modular tests11:59
@HeikoSwiking: first generate empty instance11:59
@HeikoSthen get name11:59
@HeikoSthen clone it self11:59
@HeikoSthen clone and equals11:59
@HeikoSbut I can have a look ...11:59
shogun-buildbotbuild #10 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/10  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>12:00
@HeikoSwiking: compiling12:00
@wikingHeikoS: don't forget to remove the BaggingMachine from the .py script12:00
@HeikoSwiking: I will create a small program to test12:01
@HeikoSeasier12:01
lisitsynI shoulda get up to that pace you are pulling shogun to, wiking12:02
@wikingHeikoS: i'll add soon a small dummy example12:02
@wikinglisitsyn: :D12:02
@HeikoSwiking: no need, just give me the classname (BaggingMachine)12:03
@HeikoSso done from your part, I wll check later today12:03
@wikinglisitsyn: well u know, u were up and running in jan->july12:03
@wikingso it's all cool12:03
@wikingHeikoS: CBaggingMachine ;)12:03
lisitsynwiking: I am bothered with the begging machine12:03
lisitsyn:D12:03
@wikingheheh that is still shit :)12:03
lisitsynwhat bothers me is that we will have a looot classes12:03
@wikinglisitsyn: indeed12:04
lisitsynlike niggarule yourmotherrule12:04
@HeikoSdrop half of them!12:04
lisitsynwhatever12:04
@wikinglisitsyn: becoming like java API :)12:04
lisitsynyes and it is bad for swig12:04
@wikinglisitsyn: D-class ;P12:04
@HeikoShalf of the classes are not workin anyway , just stealing compile time12:04
@HeikoSmark stable classes that are tested/documented12:04
lisitsynwhat I could suggest12:04
@HeikoSand use this to produce a stable branch12:04
lisitsynis to hide these classes somehow12:04
lisitsynlike12:04
lisitsynCombinationRule <- some enum or whatever -> CombinationRuleImpl -> various implementations12:05
lisitsynthis is not that OOP but it could be better swig wise I think12:05
@wikingCombinationRuleFactory :D12:05
@wikinglisitsyn: http://zchydem.enume.net/2010/01/19/qt-howto-private-classes-and-d-pointers/12:06
@wikinglisitsyn: it's like interface in java :P12:06
lisitsynwiking: forward declare and use a pointer?12:06
@wikingyep12:07
lisitsynyes that's a good way for sue12:07
lisitsynsure*12:07
lisitsynthe only thing is that we need to change implementation in interfaces12:07
lisitsynsomething that links X and these impls12:07
lisitsynHeikoS: I wanted to ask you whether you made some plans on our SF trip12:08
@HeikoSlisitsyn: not yet12:08
@wikinglisitsyn: but aaaanyyyhoooooow12:08
lisitsynHeikoS: I am thinking of arriving at least a day before to walk around12:08
@wikinglisitsyn: ideas for having views on CFeatures? :) that is actually thread safe :P12:08
@HeikoSlisitsyn: yeah totally12:09
lisitsynbut I don't think it makes sense to re-book hotel12:09
lisitsynwiking: what kind of views?12:09
@wikinglisitsyn: it's like add_subset(index_Vector)12:10
@wikingbut with add_subset the problem is that if i run it with 5 threads12:10
@wikingit's going to be either a mess12:10
@wikingor if i add mutex around it then it's not going to be parallel at all12:10
lisitsynHeikoS: my visa appointment is on 6th sept. but I think it should be ok12:10
lisitsynwiking: oops :D12:11
@wikinglisitsyn: https://github.com/shogun-toolbox/shogun/blob/develop/src/shogun/machine/BaggingMachine.cpp#L10112:12
@wikinglisitsyn: see there...12:12
@wikinglisitsyn: if i enable there OpenMP... w/o a condition it's not going to work12:12
@wikingif i add a condition there its basically not parallel...12:12
lisitsynHeikoS: so we can stay somewhere near or so (but I'll arrive just for a day before I think)12:12
@HeikoSlisitsyn: cool12:13
@HeikoSyeah I will check some airbnb place12:13
@wikingHeikoS: yeah go with that...12:13
@HeikoSthat is cheap and nice and we dont have to go to a hotel12:13
@wikingHeikoS: that was the cheapest option for me back in april12:13
lisitsynwiking: yeaah I see what's the problem12:13
@wikingand in SF it's like airbnb heaven :P12:14
@wikinglisitsyn: so that's why i say that i need views on features/labels12:14
lisitsynwiking: yeah I understand now12:14
@wikinglisitsyn: afaik this would be needed anyways if we start supporting multi-core12:15
lisitsynwiking: yes true12:15
shogun-buildbotbuild #1266 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/126612:21
shogun-notifier-shogun: van51 :develop * cd3608c / / (3 files): https://github.com/shogun-toolbox/shogun/commit/cd3608c1bf7cae5da710735fb02518dae4d2652c12:30
shogun-notifier-shogun: Fixed LBPPyrDotFeature12:30
shogun-notifier-shogun: Soeren Sonnenburg :develop * 29b45bb / / (3 files): https://github.com/shogun-toolbox/shogun/commit/29b45bb04ddf8efbab3e758cf436b3d9548301b912:30
shogun-notifier-shogun: Merge pull request #1445 from van51/feature/lbppyr12:30
shogun-notifier-shogun:12:30
shogun-notifier-shogun: Fixed LBPPyrDotFeatures12:30
shogun-buildbotbuild #1562 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1562  blamelist: Thoralf Klein <thoralf.klein@zib.de>12:32
shogun-buildbotbuild #1444 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1444  blamelist: Thoralf Klein <thoralf.klein@zib.de>12:33
shogun-buildbotbuild #1475 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1475  blamelist: Thoralf Klein <thoralf.klein@zib.de>12:36
shogun-buildbotbuild #1836 of deb1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1836  blamelist: van51 <vangelis_51@hotmail.com>12:37
@wikingdamn12:43
@wikingwhy ensure valid is just doing SG_ERROR instead of returning a bool?12:43
shogun-buildbotbuild #11 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/11  blamelist: Thoralf Klein <thoralf.klein@zib.de>12:44
@wikingany objections to change the behaviour of CLabels::ensure_valid from void to bool?12:45
shogun-buildbotbuild #1267 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1267  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>13:02
thoralfHeikoS: Now it's time for my segfault PR: https://github.com/shogun-toolbox/shogun/pull/144413:09
lisitsynthoralf: added new segfault?13:12
lisitsyn;)13:12
thoralflisitsyn: Yes, as usual!13:12
lisitsynah fixed :(13:12
thoralflisitsyn: Nice idea, though. ;)13:12
lisitsynthoralf: yes, "Added new segfaulting code to X"13:13
lisitsyn"Removed necessary check for NULL"13:13
thoralflisitsyn: But we have some constraints: Should not be detected by unit-tests or examples.13:14
lisitsynyes!13:14
lisitsyn"currently unit-tests don't detect the error"13:15
* thoralf is thinking of random access to arrays of unknown length. :)13:15
lisitsynthoralf: random reference counter13:15
lisitsynrussian roulette13:15
lisitsynsometimes it is 0 and you die13:15
thoralflisitsyn: initialization to 1+some_constant*rand()/number_of_threads()13:16
thoralfI think its the main difference between C/C++ and Java: "Trust the programmer." vs. "Don't even think about trusting him.  Keep him tied."13:18
-!- van51 [~van51@ppp-2-87-0-126.home.otenet.gr] has quit [Quit: Leaving.]13:18
lisitsynthoralf: why?13:20
thoralflisitsyn: I didn't get your question.13:21
lisitsynthoralf: why do you think that's the difference ;)13:21
shogun-buildbotbuild #1837 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/183713:22
shogun-buildbotbuild #1445 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1445  blamelist: Soeren Sonnenburg <sonne@debian.org>13:22
lisitsynI admit I don't get it13:22
shogun-buildbotbuild #1476 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1476  blamelist: Soeren Sonnenburg <sonne@debian.org>13:23
thoralflisitsyn: In C/C++ it's much harder to write sane code than in Java.  So you have to trust the skills of your programmer more.13:23
thoralflisitsyn: In Java you're hiding things that are too complicated/error-prone for most programmers.13:24
thoralf(Which is good.)13:24
* wiking created a bagging machine the make worse results than a simple machine \o/13:24
thoralfwiking: Cool :)13:25
lisitsynwiking: begging machine13:25
lisitsynplease give me some skills I can't recognize it13:25
lisitsynpleasse man give me some knowledge13:25
@wiking:D13:26
lisitsyncome on don't be that greedy13:26
@wikingi hate fucking SGMatrix13:26
thoralflisitsyn: One point should be enough :)13:26
thoralflisitsyn: Now try to generalize.13:26
lisitsynI need a point13:27
lisitsynpoint is that I need13:27
lisitsynif I share with you my story would share a point with me?13:27
thoralflisitsyn: Sharing a point in 2D: I keep x, you're getting y?13:28
lisitsynhah or semi-supervised begging13:28
lisitsynjust give me some point no matter what label it is of13:29
shogun-notifier-shogun: Thoralf Klein :develop * 4fae70c / src/shogun/features/ (2 files): https://github.com/shogun-toolbox/shogun/commit/4fae70ce6a0e4be06ec43a9eaf622d746f6d698f13:32
shogun-notifier-shogun: Fixing segfault.  Off-by-one when setting number of dimensions in StreamingSparseFeatures.13:32
shogun-notifier-shogun: Thoralf Klein :develop * 3ac91b7 / src/shogun/features/streaming/StreamingSparseFeatures.cpp: https://github.com/shogun-toolbox/shogun/commit/3ac91b712d688828d6ef2795a345846c2dd52ff813:32
shogun-notifier-shogun: Simple refactoring: Reuse get_dimension() from SparseVector to reduce code duplication.13:32
shogun-notifier-shogun: Thoralf Klein :develop * bc4bfa0 / src/shogun/features/streaming/StreamingSparseFeatures.cpp: https://github.com/shogun-toolbox/shogun/commit/bc4bfa0c983b7208124bdd375978d5beedd57e6e13:32
shogun-notifier-shogun: Simple refactoring: Reuse dense_dot()/sparse_dot() from SparseVector to reduce code duplication.13:32
shogun-notifier-shogun: Thoralf Klein :develop * 0288090 / / (2 files): https://github.com/shogun-toolbox/shogun/commit/0288090a80c5ad122b654e05e1f22a08bae94a1213:32
shogun-notifier-shogun: * Fixed broken unit test (number of non-zero-entries != number of dimensions).13:32
shogun-notifier-shogun: * Reformatting and fixing of debug output.13:32
shogun-notifier-shogun: Thoralf Klein :develop * 21e82a8 / src/shogun/features/ (2 files): https://github.com/shogun-toolbox/shogun/commit/21e82a83da8c8f43d7d3cab62074f2736061681a13:32
shogun-notifier-shogun: Thanks to HeikoS: Disabling reference counting for temporary sparse vectors.13:32
shogun-notifier-shogun: Sergey Lisitsyn :develop * 92abec1 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/92abec12d3aad4abe350f0abdc77a07c23db5f5a13:32
shogun-notifier-shogun: Merge pull request #1443 from tklein23/fix_streaming_sparse13:33
shogun-notifier-shogun:13:33
shogun-notifier-shogun: Fixing segfault. Off-by-one when setting number of dimensions in StreamingSparseFeatures.13:33
shogun-buildbotbuild #12 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/12  blamelist: Soeren Sonnenburg <sonne@debian.org>13:34
-!- gsomix [~gsomix@178.45.44.239] has quit [Ping timeout: 260 seconds]13:43
shogun-buildbotbuild #1268 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1268  blamelist: Thoralf Klein <thoralf.klein@zib.de>13:44
shogun-buildbotbuild #1563 of deb3 - modular_interfaces is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/156314:08
shogun-notifier-shogun: Viktor Gal :develop * e1ed907 / src/interfaces/modular/modshogun.i: https://github.com/shogun-toolbox/shogun/commit/e1ed907f998f43cd93e4249210645b29c451871c14:09
shogun-notifier-shogun: Add Ensemble.i and Ensemble_includes.i to modshogun.i14:09
shogun-notifier-shogun: Viktor Gal :develop * b4cdd86 / / (2 files): https://github.com/shogun-toolbox/shogun/commit/b4cdd86afc138d29533f1f86e290de778ec465d214:09
shogun-notifier-shogun: Fix CBaggingMachine::apply_get_outputs14:09
shogun-notifier-shogun: Add subset for the labels as well in CBaggingMachine::train_machine14:09
-!- gsomix [~gsomix@178.45.90.93] has joined #shogun14:13
@wikinggsomix: how r the chickens? :)14:13
@wikingsonney2k: ping14:13
gsomixwiking, fine, lol.14:14
shogun-buildbotbuild #1446 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1446  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Thoralf Klein <thoralf.klein@zib.de>14:15
shogun-buildbotbuild #1477 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1477  blamelist: Thoralf Klein <thoralf.klein@zib.de>14:16
-!- twthewizard [~chatzilla@141.0.149.74] has joined #shogun14:19
shogun-buildbotbuild #1838 of deb1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1838  blamelist: Thoralf Klein <thoralf.klein@zib.de>14:20
@wikingtwthewizard: yo :)14:20
shogun-buildbotbuild #1447 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1447  blamelist: Viktor Gal <viktor.gal@maeth.com>14:20
shogun-buildbotbuild #1478 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1478  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Thoralf Klein <thoralf.klein@zib.de>, Viktor Gal <viktor.gal@maeth.com>14:20
twthewizardyo wiking14:21
twthewizard:)14:21
-!- gsomix [~gsomix@178.45.90.93] has quit [Ping timeout: 264 seconds]14:21
-!- twthewizard [~chatzilla@141.0.149.74] has left #shogun []14:22
shogun-buildbotbuild #13 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/13  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Thoralf Klein <thoralf.klein@zib.de>14:24
@wikingooh man that person who created the ruby .debs for the stable debian should be hanged14:24
@wiking:)14:24
@HeikoSwiking: you should not be so extreme ...14:26
shogun-buildbotbuild #1269 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1269  blamelist: Soeren Sonnenburg <sonne@debian.org>14:26
@wikingman it's impossible that they are putting packages of ruby to some random locations14:26
@wikinggod bless digitalocean :)14:29
@HeikoSwiking: ok now to the clone bug14:34
@HeikoSwiking can be ask for a certain cmake version14:36
@HeikoSwiking: the error I had yesterday came from an old one14:36
@wikingHeikoS: yes14:36
@HeikoSsee commit14:36
@HeikoSwiking: but sonney2k didnt like adding the file14:36
@wikingHeikoS: you fixed it an awkward way14:36
@wikingi'll fix it ...14:36
@HeikoSwiking: ok, remove the file14:36
@HeikoSwiking: I will install cmake locally instead14:36
shogun-buildbotbuild #14 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/14  blamelist: Viktor Gal <viktor.gal@maeth.com>14:36
@HeikoSwiking: how to make cmake print all options, like debug etc14:38
@wikingHeikoS: google14:39
@wikingi'm not cmake manual sorry man14:39
lisitsynHeikoS: ccmake14:39
@wikingit's just that i'm getting cmake FAQs all the time here at the irc14:39
@HeikoSwiking: hehe, sorry ;)14:39
lisitsynuse that14:39
lisitsynHeikoS: ask me next time to lower the wiking's load ;)14:39
@HeikoSlisitsyn: ccmake is too complicated14:41
@HeikoSI just want a list14:41
@HeikoSwith these D.... flags14:41
@HeikoSsince I cannot remember them14:41
lisitsynHeikoS: ?? what's complicated in ccmake?14:41
lisitsynthat's exactly what it does aint it?14:42
@HeikoStoo much information I dont want14:42
lisitsynthat's -D flags14:42
shogun-notifier-shogun: Viktor Gal :develop * cdbca67 / cmake/FindRubyNArray.cmake: https://github.com/shogun-toolbox/shogun/commit/cdbca671451cfdaa89b24096c412b8e0238afce914:42
shogun-notifier-shogun: Fix RubyNArray find path issue #143814:42
shogun-notifier-shogun: Add archdir among the search paths, as on debian squeeze that's where14:42
shogun-notifier-shogun: narray files reside14:42
lisitsynHeikoS: you may use cmake -i but it is the same14:42
@HeikoSlisitsyn: ah in fact, it helped :)14:43
@HeikoSthanks14:43
lisitsynHeikoS: did you confuse ccmake with something else? :)14:44
@HeikoSlisitsyn: confusion is a good word for most of my actions ;)14:45
lisitsynI just realized it sounds like a wrong word14:45
lisitsyn:D14:46
lisitsynanyway HeikoS you got what I mean right?14:46
@HeikoSdont know :)14:46
@HeikoSbut currently compiling14:46
lisitsynHeikoS: s/confuse/mix up/ or whatever14:47
lisitsynHeikoS: so ccmake is ok?14:47
@HeikoSyes it worked14:47
@HeikoSlisitsyn:14:51
@HeikoS/home/heiko/Desktop/shogun/shogun/src/shogun/lib/RefCount.h: In constructor 'shogun::RefCount::RefCount(int32_t)':14:51
@HeikoS/home/heiko/Desktop/shogun/shogun/src/shogun/lib/RefCount.h:23:46: error: no matching function for call to 'std::atomic<int>::atomic(int32_t&)'14:51
@HeikoS/usr/include/c++/4.7/bits/atomic_base.h:329:7: error: 'constexpr' does not name a type14:51
@HeikoSand many more14:51
@wikingHeikoS: mmm c++11 :)14:52
lisitsynyeah looks like C++1114:52
@HeikoSthis is my custom eclipse project for debugging single programs14:52
@HeikoSworked before :/14:53
@wikingwell i guess your eclipse compiler is not defining -std=c++1114:53
@wikingwhile cmake realised that your compiler supports it14:53
@HeikoSI see14:53
@HeikoSwiking: what do you use to compile /test single programs?14:54
@wikingHeikoS: i would highly suggest to use cmake generated eclipse project14:54
@HeikoSsuch as examples?14:54
@wikingHeikoS: make :)14:54
@HeikoSlet me try14:54
@wikingand then gdb14:54
@HeikoSdoes eclipse even create cmake projects?14:54
@wikingHeikoS: dude14:55
@wikingHeikoS: cmake -h14:55
@wikingon the end of that you'll see a list of generators14:55
lisitsynHeikoS: cmake creates eclipse projects14:55
@HeikoSwiking: dude, well I never used it14:55
lisitsynHeikoS: use -G "Eclipse blablabal"14:55
@wikingand i suppose that you are going to have something like14:55
@wikingEclipse CDT4 - Unix Makefiles14:55
@wikingin that list14:55
lisitsynyes14:55
@wikingif so14:55
lisitsynthis one14:55
@HeikoSso far I used the wizzard in eclipse14:55
lisitsyn:D14:55
@wikingthen just run14:55
@wikingcmake -G "Eclipse CDT4 - Unix Makefiles" <all the other cmake flags u use> ..14:56
@wiking(in the build directory of course14:56
@wikingand you'll have a nice and shiny eclipse project14:56
@wikingyou problem of course is going to be that the src files are not grouped together14:56
@wikinglike on the filessystem14:56
@wikingthis could be fixed in cmake14:56
@wikingbut seems nobody is so interested in this14:56
@wiking(anyhow now you'll have the src files in one big list)14:57
lisitsynwiking: I thought it is solved14:57
lisitsynain't you found some flag or what14:57
@wikinglisitsyn: i dont seen anywhere a source_group in src/shogun/CMakeLists.txt :P14:57
@wikinglisitsyn: one needs to define the groups...14:57
lisitsynahh14:57
@wikinglisitsyn: it's a bunch of cmake regexp lines14:58
@wikingnothing serious..14:58
@wikingi mean to create the right source_groups14:58
@wikingok so about the static fucking interfaces14:58
@wikingis it like14:58
@wikingcompile the files in the static interface dir14:58
@wikingand link it together with libshogun.so?14:59
lisitsynwiking: looks like14:59
@wikingi'm just asking since i have no clue at all14:59
@wikingand i'm too lazy to check in the repo the Makefile.template file15:00
@HeikoSwiking:15:04
@HeikoSsvn: E155004: Working copy '/home/heiko/Desktop/shogun/shogun/third_party/GoogleMock' locked.15:04
@HeikoSsvn: E155004: '/home/heiko/Desktop/shogun/shogun/third_party/GoogleMock' is already locked.15:04
@HeikoSsvn: run 'svn cleanup' to remove locks (type 'svn help cleanup' for details)15:04
@wikingHeikoS: hahahahaha15:05
@wikingHeikoS: they are using the same fs?15:05
@HeikoSwiking: I am currently just trying to compile on my machine15:05
@HeikoSand to run a program to debug the clone stuff15:05
@wikingHeikoS: ok then you left it in a bad state15:05
@wikingHeikoS: rm -rf third_party/GoogleMock15:06
@HeikoSi did a make clean15:06
@wikingand it'll check it out again15:06
@wikingno it's not about that15:06
@HeikoSok15:06
@HeikoSso again about the eclipse thing15:06
@wikingyes?15:06
@HeikoSwhat I had before was I imported shogun as a makefile project, all fine15:06
@HeikoSand then I had a seperate project for my testing stuff15:07
@wikingyeah well now u can use cmake generated15:07
@wikingeclipse15:07
@HeikoSwiking: ok I did15:07
@wikingall shiny and stuff :)15:07
@HeikoSI though have all this crap in the folder structury15:07
@wikingwhere?15:07
@HeikoScxx11* ...15:07
@wikingone can get rid of it15:07
@wikingcmake just needs some love for that one ;)15:08
@HeikoSok15:08
@HeikoSbut importing the build folder is the way to go I guess15:08
@wikingmmm i wouldn't do that15:08
@wikingif i were you15:08
shogun-buildbotbuild #1270 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1270  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Thoralf Klein <thoralf.klein@zib.de>, Viktor Gal <viktor.gal@maeth.com>15:08
@HeikoShow then?15:08
@HeikoSimport src folder? (I did that before=15:09
@wikingHeikoS: use the cmake generated eclipse proj15:09
@wikinglisitsyn: u know what puzzles me: https://travis-ci.org/shogun-toolbox/shogun/jobs/10376781#L41815:09
@wikinglisitsyn: while having a dummy project with almost the same cmake detectors: https://travis-ci.org/vigsterkr/subgrep/jobs/10233626#L267215:10
@wikinglisitsyn: i cannot get around it15:10
@wikingHeikoS: that -G stuff i've written you should actually generate you an eclipse project that you can just open with eclipse15:10
@wikingas far as i know15:10
@HeikoSwiking: this google mock issue apears whenever I cances a build btw15:10
@wikingHeikoS: heh... amaaazing15:11
@wikingHeikoS: cmake/eclipse bug15:11
@wikingor even svn :)15:11
@wikingdunno why the lock is not removed...15:11
@HeikoSuh15:12
@HeikoSeclipse crash15:15
@HeikoSaaah15:15
@wiking:D15:15
@HeikoSman 30 min since I tried to compile this example15:15
@wikingwhy no just do command line?15:15
@HeikoSwiking: because I dont like it15:15
@HeikoSfor that15:15
@wikingok then start fixing cmake15:15
@HeikoSwiking: btw this project file, I cannot find it15:15
@HeikoSit should be put in build15:15
@wikingok look15:16
@wikingi this this now15:16
@wikingThis is not supported well by Eclipse.  It is strongly recommended to use a build directory which is a sibling of the source directory.15:17
@wikinghence15:17
@wikingi did this15:17
@wikingin /home/wiking/15:17
@wikingi have /home/wiking/shogun and /home/wiking/shogun_build_eclipse15:17
@wikingcd /home/wiking/shogun_build_eclipse15:17
@wikingcmake -G "Eclipse CDT4 - Unix Makefiles" -DCMAKE_PREFIX_PATH=/opt/local -DENABLE_TESTING=ON -DENABLE_CCACHE=ON -DCMAKE_BUILD_TYPE=Debug -DBUILD_STATIC=ON ../shogun/15:18
@wikingi've got15:18
@wiking.project and .cproject in /home/wiking/shogun_build_eclipse15:18
@wikingas well as a big mfucking Makefile ;)15:19
@wikingand now i guess i should somehow import/open this in eclipse15:19
@HeikoSwiking: ah I see thanks, will try that15:19
@wikingi aint got no eclipse on my machine15:19
@HeikoSwiking: does this create another build of the same source I guess?15:21
@HeikoSwiking: this having another build of the shogun source is weird15:24
@HeikoSthen I have to recompile two things on editing15:24
thoralfHeikoS: Were you able to "open" the created project or did you create one with existing code?15:25
@HeikoSyep15:25
@HeikoSopened worked15:25
@HeikoSI am thinking of ditching eclipse actually15:26
thoralfMe too. :(15:26
thoralfHeikoS: Didn't find an enable open button. ;)15:26
@HeikoSworked very nicely with configure but this cmake stuff seems to be annoying with it15:26
lisitsynwe should convert to IDE developing team15:26
@HeikoSthoralf: its import existing project15:26
@HeikoSlisitsyn: maybe you have advice for me15:27
lisitsynHeikoS: what's up?15:27
@HeikoSI just want something that works without too much hassle and hacking15:27
@HeikoSall I need to do is to edit shogun and compile, brwoser classes etc15:27
@HeikoSand another project where I can write shogun programs and link them against the shogun compiled version15:27
lisitsynHeikoS: is it general question or you mean eclipse?15:27
@HeikoSmy point is that I dont like spending my time on this15:28
@HeikoSso something that works easily15:28
@HeikoSor something that you guys know how to use15:28
lisitsynI am in search of ideal IDE too hah15:28
lisitsynwell15:28
@HeikoSbecause I dont know anything about this, I just make things work somehow and then I immediately forget how I did that15:28
thoralfHeikoS: Does not work - always gives "invalid project description"15:28
lisitsynHeikoS: what problems do you have with eclipse?15:28
@HeikoSso currently, I would really like to compile shogun but it doesnt work15:29
lisitsynHeikoS: http://www.lighttable.com/ we need this for C++ :)15:29
thoralflisitsyn: Woah.15:31
shogun-buildbotbuild #1448 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1448  blamelist: Viktor Gal <viktor.gal@maeth.com>15:32
lisitsynI feel really bad when I think about C++ IDEs15:33
lisitsynsomething's gotta change15:33
thoralflisitsyn: Yeah.  Even PHP is better. :(15:33
thoralfbetter supported15:34
lisitsynHeikoS: I am going to try cmaking in my virtual machine now15:34
lisitsynthoralf: I have some thoughts on that but I am afraid I won't be able to create an IDE :D15:34
thoralflisitsyn: First of all its a GUI.  Scientists are not good in building GUIs. ;)15:36
shogun-buildbotbuild #1479 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1479  blamelist: Viktor Gal <viktor.gal@maeth.com>15:36
thoralfwiking: Is there a way of giving travis more cpus/machines to speed up things?15:38
-!- HeikoS [~heiko@nat-167-65.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]15:42
-!- HeikoS [~heiko@nat-167-65.internal.eduroam.ucl.ac.uk] has joined #shogun15:42
-!- mode/#shogun [+o HeikoS] by ChanServ15:42
shogun-buildbotbuild #15 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/15  blamelist: Viktor Gal <viktor.gal@maeth.com>15:44
@HeikoSlisitsyn: uninstall?15:48
lisitsynHeikoS: what?15:49
@HeikoSshogun-15:49
@HeikoSfinallyyy!!15:56
@HeikoSI ran my program again ;)15:57
lisitsynHeikoS: I am struggling with cmake now15:59
lisitsynand eclipse15:59
lisitsyn;D15:59
-!- foulwall [~user@2001:da8:215:c252:7846:786c:36da:abfc] has joined #shogun16:01
@HeikoSwiking: new instance of bagging machine returns NULL16:05
@HeikoSwiking: would be good if we had incremental tests for that generated for all classes16:05
@HeikoSthat new empty instance is not NULL16:05
@HeikoSsecond that clone works16:05
@HeikoSthird that clone and equals returns true16:05
@wikingHeikoS: ive just sent u a nice cmake + eclipse tutorial16:06
@HeikoSwiking: I have seen it nice!16:07
@HeikoSwiking: what do you use for hacking shogun?16:07
@wikingthoralf: travis is for free...16:07
@HeikoSwiking: new_sgserializable("CBaggingMachine", PT_NOT_GENERIC); returns NULL16:07
thoralfwiking: No way of adding own machines to the pool?16:08
@wikingthoralf: you can subscribe, i.e. pay and then i guess you get better service... but then again this is really a question for google, right? i'm not a travis employee16:08
@wikingthoralf: /j #travis16:08
@wikingthoralf: there's some quite knowledgeable people there16:08
@wikingabout travis16:08
@wikingi bet that even travis employees hang around there16:08
thoralfwiking: Hey man, I was about to offer a resources.16:09
thoralfDon't rtfm me all the time.16:09
@wikingthoralf: /j #travis16:09
@wikingthoralf: but these type of questions there's no way of me knowing anything about16:09
@wikingthoralf: i just know that travis has a subscription service16:10
@wikingthoralf: and that the build time is soft limited to 20 mins16:10
@wikingso really 'rtfm' :)16:10
@wikingand join #travis16:10
@wikingHeikoS: mmmm16:10
@HeikoSwiking: ah sorry no C16:10
@HeikoSwe should have the tests anyway16:10
@wikingHeikoS: well i edit in ST3 and compile/debug in normal console16:10
@HeikoSwiking: the error is caused since you have the bags member on stack and register it differently16:12
@HeikoSSG_ADD((CSGObject**)&m_bags, "bags", "Bags array", MS_NOT_AVAILABLE);16:12
@HeikoSyou cannot do this with things on the stack16:12
@HeikoSIll replace it to be on heap that should fix things16:12
thoralfwiking: You seemed involved in travis-stuff, so the chances of getting a quick answer are better than reading the internet+more.16:13
thoralfwiking: And "don't know" is an acceptable answer. ;)16:13
lisitsynthese guys talk about bags and members16:13
@HeikoSwiking: #define SG_UNREF(x) { if (x) { if ((x)->unref()==0) (x)=NULL; } }16:13
@HeikoSso it sets things to NULL16:13
@wikingHeikoS: ah cool i've remembered about if(x) but not the x=null16:14
shogun-buildbotbuild #1271 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1271  blamelist: Viktor Gal <viktor.gal@maeth.com>16:14
@HeikoSwiking: any reason you store the number of bags?16:14
@HeikoSone can get them from the array #define SG_UNREF(x) { if (x) { if ((x)->unref()==0) (x)=NULL; } }16:14
@HeikoSm_bags16:14
@HeikoSsorry16:14
@wikingHeikoS: for classifying16:15
@HeikoSok16:15
@HeikoSissue fixed :)16:15
@HeikoSthat was good, clone test detected a nasty problem thats hard to find in other context16:15
@wikingHeikoS: btw wanted to ask: what's the issue with TreeMachine16:16
@wikingwhy cannot it support clone?16:16
@HeikoSsome generic stuff that avoids it16:16
@HeikoSthere are three classes where it cannot work16:16
@HeikoSor I mean where the automated tests cannot work16:16
@wikingHeikoS: as i'd like to have .clone for the treemachine obviously :P16:17
@wikingso i was just wondering if u could tell me why currently it's not 'working' and then i could maybe fix it somehow16:17
@wikingthoralf: Travis CI GmbH Prinzessinnenstr. 20 10969 Berlin, Germany16:18
@wiking;)16:18
thoralfwiking: I know. ;)16:18
thoralfThanks. ;)16:18
@HeikoSwiking: I forgot actually,16:18
@HeikoSis it blacklisted in the jinja stuff?16:18
@wikingyeps16:19
@wikingit's in ignores16:19
@HeikoSok then the sg_new serialisable causes problems16:19
@wikingbtw we should start maybe considering to use mktemp or such in unittests :)16:22
@wikingthoralf: the deal with our slow builds is that we have much too much build jobs at once... the default upper limit is 5 concurent jobs, and as soon as u start using more you'll have your build times at least doubled16:23
@wikingor at least this is what i've gathered from travis blogs16:23
thoralfwiking: I could just go there and ask why.  It's 15 minutes by foot. ;)16:25
@wikingthoralf: do that ! :)16:26
@wikingthoralf: maybe we'll get prio :)16:26
thoralflol16:26
@wikingthoralf: but i'm sure if u wanna donate build server16:26
@wikingthey will not refuse16:26
thoralfwiking: It's more like donating VMs within our infrastructure.16:27
@wikingthoralf: yeah ... they have their own build VM images and that's all u need actually imho16:27
thoralfwiking: But it seems not to be distributed16:27
@wikingthoralf: yeah i couldn't download it either16:28
@wikingbut16:28
@wikingyou can build it yourself16:28
@wikingthe whole vm buildscript stuff is available somewhere within their github repos16:29
-!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun16:29
pickle27HeikoS: around?16:46
-!- van51 [~van51@ppp-2-87-0-126.home.otenet.gr] has joined #shogun17:00
@HeikoSpickle27: hi!17:10
pickle27HeikoS: any idea about why my clone fails? I just found that the object name was wrong could that be it?17:11
pickle27also did unit tests get moved in a recent pr? I can't find them even though Im pretty sure I am building them17:11
@HeikoSpickle27: there are various reasons: get_name, uninitialised parameters, wrongly registered parameter are the most recent ones17:12
@HeikoSpickle27: for that, you will have to ask wiking :)17:12
pickle27HeikoS: okay because get_name was wrong until now17:12
@HeikoSI am still messing with cmake, so dont know17:12
pickle27HeikoS: is travis up?17:12
@HeikoSpickle27: dont know, I just sent a PR, but did not trigger yet17:13
pickle27HeikoS: okay it hasn't triggered for mine yet either, Im trying to build the tests locally to see if I fixed it17:14
@HeikoSok17:14
@HeikoSpickle, otherwise run this program with valgrind, helps a lot:17:14
@wikingHeikoS: pickle27 http://status.travis-ci.com/17:15
pickle27well17:15
@HeikoSpickle27: https://gist.github.com/karlnapf/628279317:15
pickle27HeikoS: thanks I'll try that if this doesn't work17:15
@HeikoSpickle27: obviously with your classname17:16
pickle27wiking: did anything change with unit tests and cmake?17:16
pickle27they used to be built in /tests under my build dir but now they aren't17:16
@wikingpickle27: indeed :)17:16
pickle27Im also building with python modular17:16
@wikingpickle27: in you cmake build dir just type ctest --output-on-failure17:17
@wikingpickle27: but dont forget to give -DENABLE_TESTING=ON for the cmake17:17
pickle27okay, Im am building right now I;ll try that when its done17:17
shogun-notifier-shogun: Heiko Strathmann :develop * a194c7d / src/shogun/machine/BaggingMachine.cpp,src/shogun/machine/BaggingMachine.h: https://github.com/shogun-toolbox/shogun/commit/a194c7d26e2835c7b3b74b4ae5d95837ebd7004017:17
shogun-notifier-shogun: fixed clone error via putting CDynamicObjectArray on heap rather than stack, which caused memory errors in parameter registration17:17
shogun-notifier-shogun: Heiko Strathmann :develop * 28eecbb / src/shogun/machine/BaggingMachine.cpp: https://github.com/shogun-toolbox/shogun/commit/28eecbb219592814325fb7cae8762b36ba45fb7317:17
shogun-notifier-shogun: removed unset generic, not needed17:17
shogun-notifier-shogun: Heiko Strathmann :develop * 0937bcc / src/shogun/machine/BaggingMachine.cpp,src/shogun/machine/BaggingMachine.h: https://github.com/shogun-toolbox/shogun/commit/0937bcc695f8a23ee01eca07855feba6e9bfb0dc17:17
shogun-notifier-shogun: Merge pull request #1446 from karlnapf/develop17:17
@HeikoSwiking: fixed your clone17:17
shogun-notifier-shogun:17:17
shogun-notifier-shogun: Fix clone for BaggingMachine17:17
@HeikoSwiking: ^17:17
@wikingHeikoS: \o/17:17
@HeikoSwiking: now I got 100 free more stupid questions :)17:18
@HeikoSwiking: never put CSGObjects on the stack btw17:18
@HeikoSah I am dreaming of a shogun hacking guide17:19
@wiking:D17:19
@wikingHeikoS: btw what's with this: https://github.com/shogun-toolbox/shogun/issues/134817:21
@wikingand i need this as well: https://github.com/shogun-toolbox/shogun/issues/135517:22
@wikingHeikoS: any update on that one?17:22
@HeikoSwiking: the obtain from generic is done, someone just has to do the automated generation, but I have put up examples for that already17:23
@HeikoSwiking: the multiple subsets, ah well17:23
@HeikoSI am for views17:23
@HeikoSas "name", "rows", "cols"17:23
shogun-buildbotbuild #1449 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1449  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>17:24
@HeikoSor for features just "indices"17:24
@HeikoSwiking: I can do that17:24
@wikinghttps://gist.github.com/karlnapf/617627717:24
@wikingu want diz?17:24
@HeikoSwiking: yes, that should be easy to generate for all classes rigtht?17:25
@wikingand want it to be autogenerated? :)17:25
@wikinglike with jinja2?17:25
@HeikoSyes17:25
@wikingok the only question here will be17:25
@HeikoSand then put to17:25
@HeikoS#include <shogun/base/obtain_from_generic.h>17:25
@wikingi mean 'concern'17:25
@wikingthat we'll require jinja2 just to build libshogun17:25
@wiking;)17:25
@wikingi mean we already have python as a requirement17:25
@HeikoSit can be optional17:25
@wikingHeikoS: you mean a -DAutomagicObtainFromGeneric=ON17:26
@HeikoSbut that might be annoying17:26
@wiking?: )17:26
@HeikoSyes17:26
@wikingwell17:26
@wikingthen we'll have to have like17:26
@HeikoSbut then we have to use more +ifdefs17:26
@wikingmillions of #ifdef ...17:26
@wikingyeps17:26
@HeikoSI am not against dependencies17:26
@HeikoScan we cmake the jinja?17:26
shogun-buildbotbuild #1480 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1480  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>17:27
@wikingHeikoS: we could :)17:28
@HeikoSthen it would be fine right?17:28
@wikingHeikoS: i've just put there the big question.... just put there what u think17:28
@wikingand then let's decide about this fast17:28
@HeikoSok voted17:29
@HeikoSwiking: and views will come some time this or next month17:29
@HeikoSI first need to run shogun on a cluster17:29
@wikingHeikoS: btw i would do feature_object.add_subset(int32_t fold_id, SGVector<index_t> subset)17:31
@wikingi mean17:31
@wikingfeature_object.add_subset(SGVector<index_t> subset, int32_t fold_id=0)17:31
@HeikoSfold_id?17:31
@wikingcoming from https://github.com/shogun-toolbox/shogun/issues/135517:31
@HeikoSyeah the add_subset will just get an identifier with it, preferably a string17:31
@wikingor you totally changed your mind about that?17:31
@HeikoSfor code readability17:31
@HeikoSand the current calls will be mapped to "default" or whatever17:32
@HeikoSthen its not touching existing code17:32
@HeikoSadd_subset(const char* identifier, SGVector subset)17:32
@wikingthe only reason i'm suggesting to switch the func args is that we dont have to rewrite code where subset is being used17:32
@HeikoSwiking: yes, agreed17:32
@HeikoSor no17:32
@wikingand just give a default value if it's not set17:32
@HeikoSactually, we can just have an inline wrapper17:32
@HeikoSdoesnt really matter17:33
@HeikoSand then we need a serialisable map class17:33
@HeikoSwhich we dont have I think17:33
@wikingadd_subset(SGVector subset, const char* identifier="defaultshit")17:33
@HeikoSwiking, yes17:33
@wikingHeikoS: what ya miss for the clustering btw17:34
@HeikoSwiking: its all basic stuff17:34
@wikingi mean having a cluster where arch/os is different is pretty scary though17:34
@HeikoSI need a script that locally installs shogun including python bindings17:35
@wikingthat's really not how a cluster should look like ;)17:35
@HeikoSwiking: yes for now, its ok to assume common arch17:35
@HeikoSbut later, no17:35
shogun-buildbotbuild #16 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/16  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>17:35
@HeikoSI am currently just writing a bash to install everything17:35
@HeikoSIn file included from /usr/include/hdf5.h:24:0,17:38
@HeikoS                 from /nfs/data3/ucabhst/environments/barlow/shogun/src/shogun/io/HDF5File.cpp:17:17:38
@HeikoS/usr/include/H5public.h:57:20: fatal error: mpi.h: No such file or directory17:38
@HeikoScompilation terminated.17:38
@HeikoS[ 21%] Building CXX object src/shogun/CMakeFiles/shogun.dir/io/SerializableJsonReader00.cpp.o17:38
@HeikoScompile error on cluster17:38
@HeikoSwiking: ^ my fault?17:38
@wikingHeikoS: checking17:39
@wikingmpi? :)17:39
@wikingwhere did mpi came from :)))17:39
@HeikoS/usr/include/H5public.h:57:20:17:39
@wikingmmm so it's hdf5 right?17:40
@HeikoSyes17:40
@HeikoSinclude <mpi.h>17:40
@wikingwtf :)17:40
@HeikoSis it possible to deactivate HD5 file stuff?17:41
@wikingHeikoS: hehehe17:41
@wikingalways have something that is not yet ready17:41
@wiking:)17:41
@HeikoSmmh17:41
@HeikoSjust need to compile17:42
@wikingbut i really wonder how's that possible that you have hdf5 with parallel installed17:42
@wikingand no mpi.h17:42
@wikingthis is scary17:42
@HeikoSnot my machine17:42
@wikingyeah i supposed17:42
@HeikoSthis is other peoples machines17:42
@wiking'other peoples sucks' :D17:42
@HeikoSthe problem is that I cannot change the configs everywhere17:43
@wikingyeah i supposed17:43
@HeikoSbtw is cmake easy to install locally?17:43
@HeikoSbecause this one is so old that shogun only cmakes with my addition yesterday17:43
@wikingHeikoS: yeap17:43
@HeikoSwiking: btw will you remove that and require cmake >2.8.517:43
@wikingHeikoS: you dont even need to install it17:43
@wikingHeikoS: wget (latest cmake tar.gz) && untar && cd && ./configure && make17:44
@HeikoSok cool17:44
@wikingand you'llhave a ./bin/cmake exec17:44
pickle27wiking: I got No tests were found when running ctests17:44
@wikingthat u can use .... no need to install it at all17:44
@wikingpickle27: have u read the otehr line? :)17:44
@wikingpickle27: -DENABLE_TESTING=ON with cmake17:44
pickle27yeah I built with -DENABLE_TESTING=ON17:44
pickle27wait17:44
@wikingand u reran make17:44
@wiking?17:44
pickle27how do you separate cmake args17:44
@wikingspace17:45
pickle27ahh17:45
@wikingas any other command line args17:45
pickle27I used a comma?17:45
@wikingpickle27: dunno why17:45
-!- iglesiasg [~iglesias@89-77-4-38.dynamic.chello.pl] has joined #shogun17:45
-!- mode/#shogun [+o iglesiasg] by ChanServ17:45
@wikingHeikoS: no it's no need fo rthat17:45
@wikingHeikoS: u just made me now a little bit more worx17:45
pickle27wiking: okay now it is building tests17:45
@wikingpickle27: gr817:45
@HeikoSwiking: what?17:45
@wikingHeikoS: well u would have needed to find a more simple FindSuperLU.cmake file17:46
@wikinginstead u backported something17:46
@wiking...17:46
@wikingand now i have to git rm17:46
@HeikoSwiking: I know sorry17:46
@wikinganyhow googling17:46
@wikinganyways what the fuck are we using superlu for?17:46
@HeikoSno idea17:47
@wikinglisitsyn: do we need superlu still?17:47
@wikingHeikoS: it's in tapkee apartently17:47
@wikinglisitsyn: i remember u writing something about superlu is not needed by tapkee anymore...17:48
@HeikoSwiking: well17:48
@HeikoShaha :)17:48
@wikingHeikoS: no really.... lisitsyn wrote i think yesterday that he doesnt need superlu anymore17:48
@wikinglet's see if so17:49
@wikingbecause if really not then i'm just throwing out the whole stuff17:49
@HeikoSyep good17:49
@HeikoSwiking: btw I have a basic question about bash scripts17:50
@HeikoSis it ok to cd around?17:50
@HeikoSor is there a command to run something in a certain dir?17:50
@wikingn17:51
@wikingo17:51
@wikingu can move whereever u want to17:51
@wikingjust make sure that the directory actually exits :P17:51
@HeikoSsure :)17:51
@wikingother than that u do whatever17:51
@wikingHeikoS: i mean u can even do cd / && rm -rf .17:51
@wiking;)17:51
@HeikoSwiking: cool let me try17:52
@wikingheheheh17:52
@HeikoS**************17:52
@wikingHeikoS: root removed :S17:52
@wiking?17:52
@wikingno ping from HeikoS :D17:52
@HeikoShehe :)17:52
thoralfHeikoS: "sudo rm -rf ." is cool17:52
thoralfJust removes the current directory.17:52
thoralf:D17:52
@HeikoSno root rights here17:52
@HeikoSwhat does cmake's configure do?17:53
thoralfHeikoS: Just joking.  It recursively deletes everything :)17:53
@HeikoSrunning tests of socket stuff???17:53
@HeikoStests failing17:53
@wikingHeikoS: lol17:53
@wiking:D17:53
@wikingi have no idea man17:53
@HeikoSconfigure takes longer than compile17:54
@HeikoSwiking: my current solution for cluster is to locally install everything to $HOME/$HOSTNAME17:55
@wikingHeikoS: sounds reasonable... as u said that there's a high variance among your cluster nodes arch/os wise17:57
shogun-buildbotbuild #1450 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1450  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:05
shogun-buildbotbuild #1272 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1272  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:06
thoralfHmm.  All these OSX builds are stuck in travis.18:06
thoralfQueued, but not started since 6 hours.18:07
shogun-buildbotbuild #1481 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1481  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:08
@wikingthoralf: osx bots are down i guess18:10
@wikingHeikoS: have u tested CIndependentComputationEngine?18:16
@HeikoSwiking: nope, but lamday should have18:16
@HeikoShe is using it18:16
@HeikoSwhy?18:16
@HeikoSonly serial is impleented yet though18:17
@HeikoSfor clusters we will use clone, but not sure for multicore18:17
@HeikoSsince shared memory would be cool for that18:17
shogun-buildbotbuild #17 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/17  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:17
@wikingHeikoS: clone?18:17
@HeikoSwiking: yes serialisation clones18:17
@wikingHeikoS: ah ok but what is going to be the backend? :)18:18
@wikingi mean it needs to be interfaced somehow18:18
@HeikoSyes, but thats all yet in the stars18:18
@wikingyeye i'm just thinking about it now loudly18:18
@wikingthat we could of course serialise the cloned object into fs18:19
@wikingbut we could as well use/support somekind of a cluster backend18:19
@wikingthat actually does this via sockets18:19
@wikingi mean the passing around of serialised objects18:19
@HeikoSwiking: sorry again, any way to tell shogun which python to use?18:20
@wikingHeikoS: -DCMAKE_PREFIX_PATH is actually a good bet18:20
@wikingif u have your python somewhere funky18:20
@HeikoSPYTHON_PACKAGES_PATH18:21
@wikingsay i have one python in /usr/18:21
@wikingbut then i have another python under /opt/local18:21
@wikingi do cmake -DCMAKE_PREFIX_PATH=/opt/local18:21
@wikingso that it first searches there the stuff18:21
@HeikoSwiking: I see, thanks18:21
@wikingbecause we could actually start supporting Apache AVRO18:22
@wikingvote 7 down vote accepted18:23
@wiking18:23
@wikingSo when Hadoop was being written by Doug Cutting he decided that the standard Java method of serializing Java object using Java Object Serialization (Java Serialization) didn't meet his requirements for Hadoop. Namely, these requirements were:18:23
@wiking    Serialize the data into a compact binary format.18:23
@wiking    Be fast, both in performance and how quickly it allowed data to be transfered.18:23
@wiking    Interoperable so that other languages plug into Hadoop more easily.18:23
@wikingAs he described Java Serialization:18:23
@wiking    It looked big and hairy and I though we needed something lean and mean18:23
@HeikoSwiking: maybe18:24
@wikingand it's pretty cool because18:24
@wikingit has like all the interfaces18:24
@wikingthat we support18:24
@HeikoSbut we should be very careful about not starting too much18:24
@wikingHeikoS: no i mean it's just an api18:25
@wikingHeikoS: i mean interfacing18:25
@HeikoSwiking: this python trick did not work :(18:25
@wikingHeikoS: and avro is cool in sense that the serialised objects are selfcontained. i.e. the schema of the object is in the serialised file18:25
@wikingand it's defined by JSON18:26
@wikingthat we already support18:26
@wiking;P18:26
@HeikoSthat sounds sweet18:26
@HeikoSwhat does python need?18:26
@HeikoSI just have a pip install of python and some libs18:27
@HeikoSbut no libpython.so18:27
@HeikoS-- Found PythonLibs: /usr/lib/libpython2.7.so (found version "2.7.2+")18:27
@HeikoSthis one is not around in my installation18:27
@wikingHeikoS: it's system lib18:28
@wikingu need that18:28
@HeikoSI see18:28
@wikingw/o that u cannot have pythonmodular18:28
@wikingHeikoS: i mean if u want to have that installed locally18:28
@HeikoSok but does it have to match my interpreter version?18:28
@wikingHeikoS: then you have to install the the whole python from src yourself18:28
@HeikoSargh really?18:28
@wikingHeikoS: yes otherwise it will get messy18:28
@HeikoSso can I combine this stuff?18:28
@HeikoSI just need my local ipython in fact18:29
@wikingHeikoS: ipython is just on top of python18:29
@HeikoSso then I will rather use system python18:29
@HeikoSok18:29
@HeikoSocol18:29
@wikingok brb18:29
@wikingi'll check on this avro shit laterz18:29
@HeikoSwiking: the mpi thing still blocks my compilation, what to do?18:44
@HeikoSI can install packages for now though18:45
@HeikoSbut libopenmpi-dev is already there18:45
shogun-buildbotbuild #1273 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1273  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:47
pickle27iglesiasg:18:49
pickle27i fixed clone in jedidiag18:49
@iglesiasgpickle27, hey, wassup18:49
@iglesiasgpickle27, cool! what was it?18:49
pickle27obj name was wrong lol18:49
@iglesiasgoh! subtle bug :)18:49
@iglesiasgpickle27, all right, so we let travis take care of it and then I will merge18:50
pickle27iglesiasg: travis hasnt started on it, its been 2 hours18:51
@iglesiasgpickle27, mmm annoying18:51
@HeikoSjust run tests locally :)18:51
pickle27all green :)18:52
@HeikoSso the clone was resolved? :)18:52
pickle27yup18:52
@HeikoScool!18:52
@iglesiasglet's do it then!18:52
@HeikoSwiking: we should really have these tests ^18:52
@HeikoSget_name18:52
shogun-notifier-shogun: Kevin :develop * 113fe62 / / (11 files): https://github.com/shogun-toolbox/shogun/commit/113fe622525aecc6618731932ce5bdbfaf5bd49718:52
shogun-notifier-shogun: added JediSep ICA algorithm18:52
shogun-notifier-shogun: Kevin :develop * f334832 / src/shogun/ (6 files): https://github.com/shogun-toolbox/shogun/commit/f334832298fffb352e203c655607c9be7ac3195218:52
shogun-notifier-shogun: JediSep cleanup18:52
shogun-notifier-shogun: Kevin :develop * b5ade41 / src/shogun/converter/ica/JediSep.h: https://github.com/shogun-toolbox/shogun/commit/b5ade41b3628f1e84b1ac56d8b772a30d12750a818:52
shogun-notifier-shogun: fixed JediSep object name18:52
shogun-notifier-shogun: Kevin :develop * 25f81d5 / src/shogun/mathematics/ajd/JediDiag.h: https://github.com/shogun-toolbox/shogun/commit/25f81d5123b53a1b37ebdfbdc760f97a1a9c0e0818:52
shogun-notifier-shogun: fixed JediDiag obj name18:52
shogun-notifier-shogun: Fernando Iglesias :develop * 1e7f32a / / (14 files): https://github.com/shogun-toolbox/shogun/commit/1e7f32af83790413a6902170d451c7629b70a96f18:52
shogun-notifier-shogun: Merge pull request #1433 from pickle27/develop18:52
shogun-notifier-shogun:18:52
shogun-notifier-shogun: added JediSep ICA algorithm18:52
pickle27hopefully travis starts on this pr18:57
shogun-buildbotbuild #1451 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1451  blamelist: Kevin <kevinhughes27@gmail.com>18:59
shogun-buildbotbuild #1452 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1452  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com>18:59
shogun-buildbotbuild #1482 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1482  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com>19:03
shogun-buildbotbuild #18 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/18  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com>19:11
-!- lambday [67157c37@gateway/web/freenode/ip.103.21.124.55] has joined #shogun19:17
-!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has quit [Ping timeout: 256 seconds]19:23
-!- lisitsyn1 [~lisitsyn@fb2-lo1.global63.net] has joined #shogun19:27
lisitsyn1wazzaaa19:27
shogun-buildbotbuild #1844 of deb1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1844  blamelist: Kevin <kevinhughes27@gmail.com>19:40
shogun-buildbotbuild #1274 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1274  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com>19:41
-!- HeikoS [~heiko@nat-167-65.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]19:52
@wikinglisitsyn1: superlu?19:53
@wikingheokok i'm on it19:53
@wikinglisitsyn1: do u still need it?19:53
-!- iglesiasg [~iglesias@89-77-4-38.dynamic.chello.pl] has quit [Quit: Ex-Chat]20:08
-!- travis-ci [~travis-ci@ec2-54-224-242-169.compute-1.amazonaws.com] has joined #shogun20:13
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/1040630620:13
-!- travis-ci [~travis-ci@ec2-54-224-242-169.compute-1.amazonaws.com] has left #shogun []20:13
@wikingoh yeah20:14
@wikingtravis is starting to wake up20:14
lisitsyn1wiking: no20:26
@wikinglisitsyn1: can i throw then out the whole SuperLU detection?20:26
lisitsyn1wiking: yes20:26
@wikinggreat thnx20:26
shogun-notifier-shogun: Viktor Gal :develop * e97ed82 / / (4 files): https://github.com/shogun-toolbox/shogun/commit/e97ed82d515bec9adb6b6fa3e327b7f924c6d49420:28
shogun-notifier-shogun: Remove SuperLU detection as it not used anymore20:28
shogun-buildbotbuild #1453 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1453  blamelist: Viktor Gal <viktor.gal@maeth.com>20:33
-!- van511 [~van51@ppp-94-66-27-251.home.otenet.gr] has joined #shogun20:37
shogun-buildbotbuild #1483 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1483  blamelist: Viktor Gal <viktor.gal@maeth.com>20:37
-!- van51 [~van51@ppp-2-87-0-126.home.otenet.gr] has quit [Ping timeout: 264 seconds]20:38
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has joined #shogun20:38
shogun-buildbotbuild #19 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/19  blamelist: Viktor Gal <viktor.gal@maeth.com>20:45
-!- van511 [~van51@ppp-94-66-27-251.home.otenet.gr] has quit [Quit: Leaving.]20:51
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 276 seconds]21:00
shogun-buildbotbuild #1275 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1275  blamelist: Viktor Gal <viktor.gal@maeth.com>21:15
-!- gsomix [~gsomix@95.67.149.216] has joined #shogun21:37
lambdaywiking: there?21:49
lambdaywiking: I need to use an external library for graph coloring.. need shogun to configure with it21:50
@wikinglambday: yes21:50
@wikinglambday: ok give me details21:50
@wikingwhat's the lib u need21:51
-!- lambday_ [67157d37@gateway/web/freenode/ip.103.21.125.55] has joined #shogun21:51
lambday_err got ddc21:51
@wikingnow worries just gimme the lib name21:52
lambday_wiking: its ColPack graph coloring library...21:52
@wikinglol21:53
lambday_:-/21:53
@wikinglthis is the one:21:53
@wikinghttp://www.cscapes.org/coloringpage/21:53
@wiking?21:53
lambday_wiking: yep21:53
@wikingmmm ok21:54
shogun-buildbotbuild #1569 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/156921:54
@wikingthis is going to be rather intersting21:54
lambday_http://cscapes.cs.purdue.edu/download/ColPack/ColPack-1.0.9.tar.gz21:54
lambday_wiking: why so?21:54
-!- lambday [67157c37@gateway/web/freenode/ip.103.21.124.55] has quit [Ping timeout: 250 seconds]21:55
@wikingwell this seems to be a library that's really not wide spread21:55
lambday_wiking: ummm yeah21:55
@wikinglambday_: do u know what version u would need at least?21:55
lambday_umm not sure... I am using 1.0.921:56
@wikinghehe ok21:56
lambday_wiking: if this is not there in debian repo, should we use something else?21:56
@wikinglambday_: no not necessary21:56
@wikingi mean it's ust that this feature is going to be used by very few peopl21:57
@wikingu know that21:57
@wiking ;)21:57
@wikingas it's highly improbable that someone is going to have this library installed21:57
lambday_wiking: yeah.. only those people with a huge sparse matrix to compute log-det of :(21:57
@wikingunless he really wants the feature u r creating with it21:57
lambday_wiking: yeah... that's true21:57
@wikinglambday_: but no worries i can add the detection for it21:57
lambday_wiking: thanks man! :)21:58
@wikinglambday_: what does it install actually?21:58
lambday_headers and lib21:58
lambday_libColPack.so21:58
lambday_in /usr/local/lib/21:58
lambday_and /usr/local/include/ColPack/ [default]21:59
@wikingwhat's the header? ColPack.h?21:59
lambday_wiking: ColPackHeaders.h21:59
@wikingokok i see22:00
lambday_I just gotta include this one and rest all are included22:00
lambday_err I will get dc22:00
-!- lambday_ [67157d37@gateway/web/freenode/ip.103.21.125.55] has quit [Ping timeout: 250 seconds]22:04
shogun-notifier-shogun: Viktor Gal :develop * 137d664 / CMakeLists.txt,cmake/FindColPack.cmake: https://github.com/shogun-toolbox/shogun/commit/137d664fb40032ffd0f916afed886a08fbebe29022:16
shogun-notifier-shogun: Add FindColPack script for graph coloring22:16
shogun-notifier-shogun: dedicated to lambday22:16
@wikingsorry for the double commit :S22:17
shogun-notifier-shogun: Viktor Gal :develop * 5640485 / src/shogun/lib/config.h.in: https://github.com/shogun-toolbox/shogun/commit/5640485109184fb2dc81b570d7c73af480f6814422:17
shogun-notifier-shogun: add HAVE_COLPACK to the config.h22:17
shogun-buildbotbuild #1454 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1454  blamelist: Viktor Gal <viktor.gal@maeth.com>22:24
shogun-buildbotbuild #1484 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1484  blamelist: Viktor Gal <viktor.gal@maeth.com>22:24
shogun-buildbotbuild #20 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/20  blamelist: Viktor Gal <viktor.gal@maeth.com>22:35
-!- zxtx [~zv@rrcs-74-62-200-195.west.biz.rr.com] has joined #shogun22:42
-!- travis-ci [~travis-ci@ec2-23-20-152-199.compute-1.amazonaws.com] has joined #shogun23:05
travis-ci[travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/1041306023:05
-!- travis-ci [~travis-ci@ec2-23-20-152-199.compute-1.amazonaws.com] has left #shogun []23:05
shogun-buildbotbuild #1276 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1276  blamelist: Viktor Gal <viktor.gal@maeth.com>23:06
shogun-buildbotbuild #1455 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1455  blamelist: Viktor Gal <viktor.gal@maeth.com>23:06
shogun-buildbotbuild #1485 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1485  blamelist: Viktor Gal <viktor.gal@maeth.com>23:06
shogun-buildbotbuild #21 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/21  blamelist: Viktor Gal <viktor.gal@maeth.com>23:11
@wikingshogun-buildbot: force rebuild --branch develop 'bsd1 - libshogun'23:12
shogun-buildbottry 'force build [--branch=BRANCH] [--revision=REVISION] [--props=PROP1=VAL1,PROP2=VAL2...]  <WHICH> <REASON>'23:12
@wikingshogun-buildbot: force rebuild --branch=develop 'bsd1 - libshogun'23:12
shogun-buildbottry 'force build [--branch=BRANCH] [--revision=REVISION] [--props=PROP1=VAL1,PROP2=VAL2...]  <WHICH> <REASON>'23:12
@wikingshogun-buildbot: force build --branch=develop 'bsd1 - libshogun'23:12
shogun-buildbotbuild forced [ETA 8m25s]23:12
shogun-buildbotI'll give a shout when the build finishes23:12
shogun-buildbotbuild #1486 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/148623:17
shogun-notifier-shogun: Viktor Gal :develop * 6c21825 / tests/unit/base/clone_unittest.cc.py: https://github.com/shogun-toolbox/shogun/commit/6c218252a4f0e03b5243cb6b5692a0c9327ed16123:18
shogun-notifier-shogun: Remove BaggingMachine from ignores in clone_unittes.cc.py23:18
-!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has quit [Quit: Leaving.]23:19
shogun-buildbotbuild #953 of rpm1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/953  blamelist: Viktor Gal <viktor.gal@maeth.com>23:20
shogun-buildbotbuild #1456 of deb2 - static_interfaces is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1456  blamelist: Viktor Gal <viktor.gal@maeth.com>23:24
shogun-buildbotbuild #1487 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1487  blamelist: Viktor Gal <viktor.gal@maeth.com>23:27
shogun-buildbotbuild #22 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/22  blamelist: Viktor Gal <viktor.gal@maeth.com>23:28
shogun-buildbotbuild #1277 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1277  blamelist: Viktor Gal <viktor.gal@maeth.com>23:49
--- Log closed Wed Aug 21 00:00:24 2013