Open in new window / Try shogun cloud
--- Log opened Mon Sep 09 00:00:51 2013
-!- Netsplit *.net <-> *.split quits: @sonney2k, shogun-notifier-, pickle27, @iglesiasg, shogun-buildbot, sonne|osx, hushell00:07
-!- Netsplit over, joins: @iglesiasg00:18
-!- Netsplit over, joins: shogun-buildbot, shogun-notifier-, pickle27, sonne|osx, @sonney2k00:19
-!- Netsplit over, joins: hushell00:19
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]00:55
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun00:55
@wikingshogun-buildbot: force build --branch=develop 'debian wheezy - memcheck'00:55
shogun-buildbotbuild #4 forced00:55
shogun-buildbotI'll give a shout when the build finishes00:55
hushelliglesiasg: hey01:02
@iglesiasghushell, hi there ,whats' up?01:02
hushelliglesiasg: good, I am working on the python stuffs01:02
@iglesiasghushell, examples?01:03
@iglesiasgsonney2k, so what's the stuff that needs to be done for the webpage?01:04
hushelliglesiasg: yeah, and I try to make an initial version of ipython notebook before implementing other solvers01:08
@iglesiasghushell, it sounds cool!01:08
hushellonly 2 weeks left :(01:08
@iglesiasghehe yep01:09
hushelliglesiasg: Have you finished everything on the schedule?01:09
@iglesiasghushell, pretty much yes01:09
@iglesiasghushell, I have to test the method more though01:10
@iglesiasgfor instance I am facing now some late convergence issue in my code01:10
@iglesiasgfor a particular data set01:10
@iglesiasgthe more data sets I use, the more flaws I find :D01:10
hushelliglesiasg: that may be a problem for the LMNN, not your implementation01:12
@iglesiasghushell, fortunately not, I checked in the author's implementation ;)01:12
hushellWhen we should do the peer checks?01:12
@iglesiasgI think that they should be after this pencil downs date01:13
@iglesiasgso one week more for coding, and after that01:13
hushelliglesiasg: strange issues always happen, this is why we are needed :)01:14
@iglesiasghushell, indeed01:14
hushellcool, then I have 7 days for preparing myself01:15
@iglesiasghushell, I am not entirely sure though01:15
@iglesiasglet's check in the meeting logs01:15
shogun-buildbotbuild #4 of debian wheezy - memcheck is complete: Failure [failed memory check]  Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/401:15
@iglesiasgI think it is like that01:16
hushellI checked the timeline on gsoc, I think that's what supposed to do in the last week01:19
@iglesiasggood then01:19
hushellis the PR good to merge?01:19
@iglesiasghushell, I think the code is good yeah01:20
@iglesiasghushell, but something is wrong with Travis01:20
hushelliglesiasg: but Travis failed for some reason01:20
@iglesiasghushell, maybe it is unrelated to your PR01:20
@iglesiasgwe should be sure of that01:21
hushelliglesiasg: yeah, let me check for a while01:21
@iglesiasghushell, all right, thanks!01:21
@iglesiasgcompare with the build history in travis, if other builds fail with the same error01:21
@iglesiasgthen I think it is a great indicative that it is not something in your PR01:21
hushelliglesiasg: I rebased again, how to check the travis history?01:24
hushelliglesiasg: I saw it, build history01:25
@iglesiasghushell, there is a tab called build history01:25
@iglesiasggood01:25
hushelliglesiasg: thanks01:25
@iglesiasghushell, you are welcome01:25
hushelliglesiasg: seems the issues start from a day ago, recent commits all failed the same problem: python-integration-test and a lot of failures in java modular01:32
@iglesiasghushell, ok! Then we are ready to go01:33
hushelliglesiasg: let's wait the new commit finish. It's very late in your time01:34
@iglesiasghushell, a little bit hehe01:34
hushelliglesiasg: no hurry, we can do it tomorrow, I guess you are sleepy01:34
@iglesiasghushell, yep, I was planning to shut down soon01:35
@iglesiasghushell, ok cool01:35
@iglesiasgI might write some comments again though01:35
@iglesiasgI was not very thorough checking the pull request honestly01:35
hushelliglesiasg: nice job!01:35
@iglesiasghushell, what about Patric? Is he also having a look that everything goes in the right direction?01:36
hushellI know checking code is even painful01:36
hushelliglesiasg: haven't hear from him for a while, maybe I should send him an email before merging01:36
@iglesiasghushell, that would be great, if he can have a look too01:37
hushelliglesiasg: He just moved to Berlin01:37
@iglesiasghushell, yep, I read something about his new job01:37
hushellokay, let's postpone the merge01:37
@iglesiasghushell, is it fine for you?01:37
@iglesiasghushell, I don't know if you plan to submit something else soon that depends on this PR01:38
hushelliglesiasg: it's fine for me. I have started a new branch to do python code01:38
@iglesiasghushell, ok, but I gues you won't be able to merge until this one gets in, right?01:38
@iglesiasgI mean, your current branch starts from the tip of your develop in the PR I guess01:39
hushelliglesiasg: yeah, but the new PR will be a small one01:39
@iglesiasgwell the thing is that those commits that are not in shogun develop yet but yes in your develop will pop up in the other PR as well01:39
hushellyes, it depends on this one, but I guess rechecking is necessary, and you may not check the new PR until tomorrow01:39
hushelliglesiasg: so you mean, I'll have to wait this one merged01:41
@iglesiasghushell, yes01:41
@iglesiasgI think so, at least01:41
hushellbut since the python PR will be small, if there are conflicts, I can do vimdiff01:42
@iglesiasggood01:42
@iglesiasgall right, time to sleep then01:42
@iglesiasggood night!01:42
hushellI was worrying about this, but will be ok01:42
hushellgood nite!01:43
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Ex-Chat]01:43
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]01:51
-!- sonne|osx_ [~sonne@f053041001.adsl.alicedsl.de] has joined #shogun03:25
-!- sonne|osx [~sonne@f053041058.adsl.alicedsl.de] has quit [Ping timeout: 268 seconds]03:27
-!- sonne|osx_ is now known as sonne|osx03:27
-!- sonne|osx [~sonne@f053041001.adsl.alicedsl.de] has quit [Quit: sonne|osx]07:14
-!- pickle27 [~Kevin@199.119.128.114] has quit [Quit: Leaving]07:37
-!- sonne|osx [~sonne@89.204.130.32] has joined #shogun08:00
-!- sonne|osx [~sonne@89.204.130.32] has quit [Client Quit]08:05
-!- sonne|osx [~sonne@89.204.130.32] has joined #shogun08:10
-!- sonne|osx [~sonne@89.204.130.32] has quit [Client Quit]08:12
-!- sonne|work [~sonnenbu@91-64-72-127-dynip.superkabel.de] has joined #shogun08:48
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]08:52
@wikingshogun-buildbot: force build --branch=develop 'debian wheezy - memcheck'08:52
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun08:52
@wikingshogun-buildbot: force build --branch=develop 'debian wheezy - memcheck'08:52
shogun-buildbotbuild #5 forced08:52
shogun-buildbotI'll give a shout when the build finishes08:52
sonne|workwiking: torturing fatbot already?08:58
lisitsynsonne|work: what is fatbot?09:02
sonne|worklisitsyn: fatbot.shogun-toolbox.org09:02
sonne|worklisitsyn: http://futurama.wikia.com/wiki/Fatbot09:02
sonne|workthis guy09:02
lisitsynsonne|work: hmm this address doesn't work for me09:03
sonne|worklisitsyn: what do you mean? it is not a web server09:03
lisitsynoh09:03
lisitsynokay :)09:03
shogun-buildbotbuild #5 of debian wheezy - memcheck is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/509:16
-!- gsomix [~gsomix@85.26.232.253] has joined #shogun09:17
gsomixhi09:17
sonne|workgsomix: hey there09:17
sonne|workgsomix: how is it going?09:17
gsomixsonne|work, fine. but I'm sick a little.09:18
gsomixsonne|work, can you help me? I want to run integration tests.09:19
gsomixfor unit-tests I just execute 'ctest' command09:20
sonne|workgsomix: unit or integration tests?09:20
@wikingsonne|work: yes09:21
gsomix>> integration tests | this09:21
sonne|workgsomix: ctest -R python_modular ?09:22
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]09:23
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun09:23
@wikingshogun-buildbot: force build --branch=develop 'clang34 - static analysis'09:23
shogun-buildbotbuild #7 forced09:23
shogun-buildbotI'll give a shout when the build finishes09:23
gsomixsonne|work,09:25
gsomixctest -R python_modular09:25
gsomixTest project /home/gsomix/Workspace/shogun/build09:25
gsomixNo tests were found!!!09:25
sonne|workgsomix: but you did enable python modular did you?09:26
gsomixlet me check09:26
gsomixsonne|work, cool, thanks. it was an confusion. =___=09:33
gsomixsonne|work, but now fails tests that's different from Travis' output. ok, need check.09:38
sonne|workgsomix: how is your protobuf stuff progressing?09:39
sonne|workwiking: how difficult is it to add a protobuf check to CMake?09:39
sonne|workall we need is test if protoc executable is available and can do protoc --cpp_out09:40
@wikingprotobuf?09:40
@wikingah you mean like a find_package?09:41
@wikingi would google first for09:41
sonne|workwiking: https://code.google.com/p/protobuf/09:41
@wikingFindProtobuf.cmake09:41
sonne|workyes09:41
@wikingi'm 100% sure09:41
sonne|worklike this09:41
@wikingthat somebody already has written it09:41
sonne|workhttps://github.com/trendmicro/tme/blob/master/src/cmake/FindProtobuf.cmake09:41
sonne|workis it already in cmake?09:42
@wikingcheck it out09:42
@wiking/usr/share/cmake-2.8/Modules09:42
@wikingindeed it is09:43
@wiking(at least my cmake has it09:43
sonne|workcool09:43
shogun-buildbotbuild #7 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/709:43
gsomixsonne|work, not many. I was trying make some examples outside shogun. if we have protobuf support in shogun's cmake I can start porduce code for shogun.09:44
sonne|workgsomix: no you can do it w/o that already09:44
gsomixbut now I want finish LibSVMFile support in examples.09:45
gsomixI sent PR, but integration tests failde in Travis09:45
gsomix*failed09:45
gsomixhttps://travis-ci.org/shogun-toolbox/shogun/jobs/1113945409:47
gsomixthat's very different from that I see on my local machine09:47
sonne|workwiking: is there any chance to get a summary of failed tests with ctest? I have a hard time parsing the output :/09:49
gsomixsonne|work, wiking intergrations tests generates from undocumented examples, right?09:54
@wikingsonne|work: ?09:55
@wikingsonne|work: end of the ctest is a summary09:56
sonne|workwiking: ahh but then please run tester.py with single python script as argument otherwise it is impossible to see09:56
@wikingsonne|work: change it as u wish and commit09:57
sonne|workwiking: now all integration tests are run in one go with tester.py09:57
sonne|workthat is why it is a mess(tm)09:57
sonne|workwiking: cannot from work09:57
@wikingok09:57
@wikingthen when u can09:57
@wikinghttp://buildbot.shogun-toolbox.org/static_analysis/2013-09-09-1/report-b33cb4.html#EndPath09:58
@wikingoooh yeah09:58
@wikingmemory leak09:58
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun10:05
shogun-notifier-shogun: Viktor Gal :develop * 821b511 / src/shogun/lib/external/libocas.cpp: https://github.com/shogun-toolbox/shogun/commit/821b511d70ca932bf9e2df2ce1f03494abb334ef10:05
shogun-notifier-shogun: Fix dead assignment errors in libocas10:05
shogun-notifier-shogun: Viktor Gal :develop * b89a541 / src/shogun/kernel/Kernel.cpp: https://github.com/shogun-toolbox/shogun/commit/b89a541ca7a1122a707e3fde4e67e8e60fd4b16410:05
shogun-notifier-shogun: Fix possible division by zero in Kernel.cpp by assertation10:05
shogun-notifier-shogun: Viktor Gal :develop * 4586d39 / src/shogun/structure/libppbm.cpp: https://github.com/shogun-toolbox/shogun/commit/4586d39c8f3419f03f7e8510ca780eab4d93fd6610:05
shogun-notifier-shogun: Remove dead assigments in libppbm10:05
shogun-notifier-shogun: Viktor Gal :develop * 75d2e78 / src/shogun/labels/LatentLabels.cpp: https://github.com/shogun-toolbox/shogun/commit/75d2e789d41f50c457031b734ff2acf6fed89ced10:05
shogun-notifier-shogun: Fix possible null pointer dereference in LatentLabels10:05
hushellHi guys, with swig python interfaces, if a class A needs a pointer (e.g. CFactorType*) as argument, can I pass an object B = FactorType(...)?10:07
hushellpointer, reference in c++ are the same in python, right?10:07
hushellI got an error saying: TypeError: in method 'new_TableFactor', argument 1 of type 'CTableFactorType *', but I have no idea how to deal with it10:09
shogun-buildbotbuild #3 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/3  blamelist: Viktor Gal <viktor.gal@maeth.com>10:09
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has joined #shogun10:10
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has quit [Changing host]10:10
-!- besser82 [~besser82@fedora/besser82] has joined #shogun10:10
-!- besser82 [~besser82@fedora/besser82] has quit [Client Quit]10:10
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has joined #shogun10:11
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has quit [Changing host]10:11
-!- besser82 [~besser82@fedora/besser82] has joined #shogun10:11
shogun-buildbotbuild #4 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/4  blamelist: Viktor Gal <viktor.gal@maeth.com>10:11
shogun-buildbotbuild #1668 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1668  blamelist: Viktor Gal <viktor.gal@maeth.com>10:13
-!- gsomix [~gsomix@85.26.232.253] has quit [Ping timeout: 276 seconds]10:14
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has left #shogun []10:15
shogun-buildbotbuild #1669 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1669  blamelist: Viktor Gal <viktor.gal@maeth.com>10:16
shogun-buildbotbuild #36 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/36  blamelist: Viktor Gal <viktor.gal@maeth.com>10:16
shogun-buildbotbuild #1749 of deb3 - modular_interfaces is complete: Failure [failed install test libshogun test python_modular test java_modular test csharp_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1749  blamelist: Viktor Gal <viktor.gal@maeth.com>10:38
-!- gsomix [~gsomix@85.26.232.253] has joined #shogun10:41
sonne|workhushell: I am not sure you can pass around references with swig10:46
shogun-buildbotbuild #1430 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1430  blamelist: Viktor Gal <viktor.gal@maeth.com>10:54
shogun-buildbotbuild #5 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/5  blamelist: Viktor Gal <viktor.gal@maeth.com>10:55
sonne|workwiking: you are killing it http://buildbot.shogun-toolbox.org/waterfall10:56
shogun-buildbotbuild #37 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/37  blamelist: Viktor Gal <viktor.gal@maeth.com>10:56
sonne|workman kind of a x-mas tree!10:56
shogun-buildbotbuild #1670 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1670  blamelist: Viktor Gal <viktor.gal@maeth.com>10:58
@wikingsonne|work: sonne|work10:58
@wikingquestion about MulticlassOCAS10:58
@wikingyou might remember it better10:58
sonne|workask10:58
@wikingdo you think it'd be a problem10:58
@wikinguser_data.W = SG_MALLOC(float64_t, (int64_t)num_features*num_classes); user_data.oldW = SG_MALLOC(float64_t, (int64_t)num_features*num_classes); user_data.new_a = SG_MALLOC(float64_t, (int64_t)num_features*num_classes); user_data.full_A = SG_MALLOC(float64_t, (int64_t)num_features*num_classes*m_buf_size); user_data.output_values = SG_MALLOC(float64_t, num_vectors);10:59
@wikingi mean at least user_data.W10:59
@wikingwith calloc instead of mallco?10:59
@wikingthis way i can supress a lot of conditional jump on uninitialised data errors with valgrind10:59
@wikingand possibly fix the hectic (sometimes fail) bug with the unit test11:00
sonne|workwiking: if you do that you should check that this was a bug11:00
@wikingwell we have this11:00
sonne|workwiking: is it single core the test I mean?11:00
shogun-buildbotbuild #1748 of deb3 - modular_interfaces is complete: Failure [failed install test libshogun test python_modular test java_modular test csharp_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1748  blamelist: Viktor Gal <viktor.gal@maeth.com>11:00
sonne|workbesser82: btw you can now run cmake when the .git dir is not available - please try!11:00
@wikinghttps://github.com/shogun-toolbox/shogun/issues/141111:00
besser82wiking, will do and report ;)11:01
besser82wiking, thx!11:01
@wikingbesser82: you mean sonne|work11:01
@wikingsonne|work: there's no multithread with ocas :)11:01
besser82wiking, yes, just overlooked that, because you posted some link to gh issues ;)11:02
sonne|workerrm is it disabled?11:02
@wikingsonne|work: and yeah we had a lot of Conditional jump or move depends on uninitialised value(s)11:02
besser82sonne|work, thanks will do! ;)11:02
@wikingwith only malloc11:02
@wikingif i change user_data.W = SG_CALLOC(float64_t, (int64_t)num_features*num_classes);11:02
sonne|workwiking: well ocas is parallelized11:02
@wikingthen it's all gone11:02
@wikingwhere?11:03
sonne|workso if we don't fix it to use a single thread we get different result11:03
@wikingbut multiclassOcas.cpp has 0 lines of pthread11:03
gsomixcool, evernote is very usable for storing parts of shogun's irc logs11:04
sonne|workwiking: sure it is under the hod11:04
sonne|workwiking: all encapsulated in dotfeatures11:04
@wikingsonne|work: i dont get it11:04
sonne|workwiking: ?11:04
@wikingwhat does this have to do with having W initialised with random data11:04
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:299e:473d:d0ad:c8d5] has joined #shogun11:04
-!- mode/#shogun [+o iglesiasg] by ChanServ11:04
@wikingor full 011:04
sonne|workwiking, has nothing to do with w11:05
sonne|workwiking: but with a test failing from time to time11:05
@wikingsonne|work: ah that's another thing11:05
@wikingsonne|work: first i want to fix this Conditional jump or move depends on uninitialised value(s)11:05
@wikingso11:05
sonne|workwell you gave me the link11:05
sonne|workso I was trying to find a reason11:05
sonne|worknot so much about your W=011:05
@wikingso would it be a problem if W=0 and not W=(random)11:06
-!- besser82 [~besser82@fedora/besser82] has quit [Quit: Verlassend]11:06
sonne|workwiking: not a problem for sure but IDK if this was a bug!11:06
@wikingwell valgrind was bitching about it11:07
@wikingi'm going through the valgrind errors11:07
@wikingand this was the first on my list11:07
-!- travis-ci [~travis-ci@ec2-54-226-222-38.compute-1.amazonaws.com] has joined #shogun11:07
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/1114274911:07
-!- travis-ci [~travis-ci@ec2-54-226-222-38.compute-1.amazonaws.com] has left #shogun []11:07
sonne|workwiking: well set it to use CALLOC but we should check if it gives the same result like the orig mc ocas11:09
@wikingsonne|work: imho if i do SGVector::random(W) would surpress the error as well11:10
sonne|workwiking: sure.11:11
sonne|workmy guess is that W=0 was intended11:11
sonne|workbut I am blind here - neither looking at code or anything11:11
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has joined #shogun11:16
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has quit [Changing host]11:16
-!- besser82 [~besser82@fedora/besser82] has joined #shogun11:16
@wikingok we need to meeerge feature/SerialUTests11:17
sonne|workwiking: btw could we integrate our unit tests w/ cmake the same way then? I mean that just a single unit test is run instead of all - so we see the proper cmake summary of test failures?11:19
@wikingsonne|work: i've done something different11:21
@wikingsonne|work: you'll see when i merge11:21
sonne|workwhich is?11:21
@wikingbasically i supress OK unittest prints11:21
@wikingand just print the error11:21
@wikingif there's11:21
sonne|workcurrently the ctest output is not useful11:21
@wikingsonne|work: how so?11:21
sonne|workI simply cannot figure out what fails11:21
@wikingthere's the summary11:21
@wikingon the end of it11:22
sonne|workwell unit tests failed11:22
sonne|workor integration tests failed is a bit too coarse11:22
@wikingThe following tests FAILED: 1 - unit-test (SEGFAULT)11:22
@wikingand then u go there11:22
@wikingand check what really failed11:22
@wikingi mean what's so hard to find out11:23
sonne|workI would rather want to see the exact name of the unit test11:23
@wikingsonne|work: not possible11:23
@wikingas unit test is just one ctest11:23
sonne|workwhy not? we could run it with proper gtest filers11:23
@wikingeither create as many unit test binaries as many tests there are11:24
@wikingfiltering i dont see working11:24
@wikingor maybe11:24
@wikingbut no11:24
sonne|workif I look at this I don't see what fails http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1670/steps/test/logs/stdio11:24
@wikingso you cannot do the filtering11:24
@wiking 1 - unit-test (SEGFAULT)11:24
@wikingno?11:24
-!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun11:25
sonne|workyes some unit test fails and then?11:25
@wikingfiltering is very cumbersom because you would need to know somehow the filter names prior having the binary (unit test) compiled...11:25
@wikingsonne|work: go to that test and see what happened there11:25
sonne|workwell which test?11:26
@wikingunit-test11:26
@wikingthe first one11:26
@wikingyou wont be able to know better if it's a segfault11:26
sonne|workwiking: yes but which of the unit tests11:26
sonne|workwiking: yeah with the current setup not11:26
@wikingsonne|work: go to the unit-test ctest11:26
@wikingand check it out11:27
sonne|workif we would filter we would know exactly11:27
@wikingas i said before11:27
@wikingfiltering is not possible11:27
@wikingyou'll have to roll a unit test binary per11:27
sonne|workand all we would need to do is filter the .cpp files for the tests11:27
@wikingsonne|work: no11:27
@wikingsonne|work: as it's not really the case...11:27
@wikingthe name of the cpp is not the name of the test11:27
sonne|workwhy not?11:27
@wikingwell check it11:27
@wikingand you'll see11:27
@wikingand how do you do the template generated naming...11:28
sonne|workwell I checked - for the tests I've written ti would work11:28
@wikingsonne|work: yeah but there's like kazillion other tets11:28
@wikingthat you have not written11:28
@wikingand see the templating problem11:28
sonne|workyou autogenerate this right so you can get the names of the tests11:29
@wikingok11:29
@wikingdo as u feel11:29
@wikingi'm just saying11:29
@wikingthat this is going to be very cumbersom11:29
@wikinge11:29
@wikingrather go with compiling a unit test binary per cpp11:29
@wikingbtw: Serialization is collidig11:30
sonne|workI am also fine with that - anything that helps to pinpoint what failed11:30
sonne|workwiking: you mean cannot be run in parallel?11:30
sonne|workwhy not?11:31
-!- gsomix [~gsomix@85.26.232.253] has quit [Ping timeout: 245 seconds]11:31
-!- gsomix [~gsomix@85.26.232.253] has joined #shogun11:31
@wikingyou have io/Serialisation and io/Serialization_unittest.cc and base/Serialization_unittest.cc11:31
@wikinghence11:32
@wikingfiltering is not really doing what u want either11:32
@wikingyeah u can run it in parallel ctest -j811:32
sonne|workwhat is the issue?11:35
@wikingi'm saying that the name is the same11:35
@wikinghence collision11:35
@wikingin naming11:35
@wikinganyhow i'm sure somebody already did this11:36
@wikingso it's wroth trying a google for it11:36
shogun-buildbotbuild #1431 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1431  blamelist: Viktor Gal <viktor.gal@maeth.com>11:37
@wikingmmm we cannot write complex with HDF5?11:38
hushellHi iglesiasg, have you had the TypeError with swig python interfaces?11:38
@iglesiasghushell, Hi. No, at least I don't remember having it.11:39
@iglesiasghushell, where is it appearing?11:39
hushelliglesiasg: there is an annoying error, when I want to create a CFactor(FactorType, ...) in python, it says the type of FactorType is wrong11:40
@iglesiasghushell, mm I see.11:41
@iglesiasglet me check how FactorType looks like11:41
@iglesiasghushell, does the error say something else?11:41
hushellTypeError: in method 'new_TableFactor', argument 1 of type 'CTableFactorType *'11:41
hushellsonne|work: I just need pass a pointer11:42
@iglesiasghushell, what?11:42
hushelliglesiasg: I asked the question before you joining :) sonne|work gave me some advice11:42
@iglesiasghushell, so CFactor needs in the constructor a CTableFactorType*11:43
sonne|workwiking: complex is very new in shogun...11:43
hushelliglesiasg: you want to check the script I am testing?11:43
sonne|workwiking: but better file a bug so we don't forget11:43
@iglesiasghushell, are you giving it a CTableFactorType or a FactorType?11:43
@iglesiasghushell, I am not sure if these two are related, you tell me :)11:44
sonne|workhushell: that works then11:44
hushellI am giving a TableFactorType, wait a second, I push this to a branch11:44
sonne|workhushell: but you need to make TableFactoryType known in swig prior to TableFactor11:44
sonne|workhushell: note that order in swigs .i files matters!11:44
hushellthe example is here: https://github.com/hushell/shogun/blob/py_fg/examples/undocumented/python_modular/structure_factor_graph_model.py11:46
hushelllook at line 3211:46
sonne|workhushell: shure that works with swig11:47
sonne|workhushell: I guess your include order in the .i files is wrong11:47
hushellsonne|work: yes, I included the Factor.h before FactorType.h11:48
sonne|worksee kaboom :)11:48
hushelllet me try the reverse, I didn't expect this will cause errors11:48
-!- gsomix [~gsomix@85.26.232.253] has quit [Ping timeout: 240 seconds]11:51
hushellsonne|work, iglesiasg: that error fixed! I spent a night to search for it...11:59
hushellSo much thanks!12:00
@iglesiasgsonne|work is the master12:00
sonne|workhushell: yeah that include order matters in swig is a HUGE trap12:00
hushellseems many tricks hiding there12:00
hushelldamn, more problems ...12:02
-!- gsomix [~gsomix@r206-10.smr.ru] has joined #shogun12:05
@iglesiasgwiking, sup12:08
@iglesiasgwiking, any idea what's wrong with Python and Java modular in Travis?12:09
@wikingiglesiasg: python is broken because of integration tests12:17
@wikingiglesiasg: java is broken because of issue https://github.com/shogun-toolbox/shogun/issues/155712:17
@iglesiasgwiking, do we need this +#ifdef EIGEN_MAPPED_SPARSEMATRIX_H?12:19
shogun-buildbotbuild #1432 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1432  blamelist: Viktor Gal <viktor.gal@maeth.com>12:20
@wikingiglesiasg: dunno it's heiko's stuff12:21
@iglesiasglet's wait then12:21
-!- HeikoS [~heiko@nat-163-180.internal.eduroam.ucl.ac.uk] has joined #shogun12:26
-!- mode/#shogun [+o HeikoS] by ChanServ12:26
-!- HeikoS [~heiko@nat-163-180.internal.eduroam.ucl.ac.uk] has left #shogun []12:26
-!- blaubaer_ [8d4c3eb2@gateway/web/freenode/ip.141.76.62.178] has quit [Ping timeout: 250 seconds]12:41
hushellhaha, the swig python example is running good, I feel much more efficient when you guys online12:45
hushellwell, time to go to bed, have fun!12:47
-!- gsomix [~gsomix@r206-10.smr.ru] has quit [Remote host closed the connection]13:04
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]13:05
-!- besser82 [~besser82@fedora/besser82] has quit [Quit: Verlassend]13:05
shogun-buildbotbuild #1750 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1750  blamelist: Viktor Gal <viktor.gal@maeth.com>13:09
-!- thoralf [~thoralf@enki.zib.de] has joined #shogun13:23
thoralfHey.13:23
@iglesiasgHello thoralf!13:23
thoralfHey iglesias13:23
@wikingsonne|work: around?14:06
sonne|workwiking: the build error for java is this EIGEN_MAPPED_SPARSEMATRIX_H14:06
@wikingsonne|work: yes14:06
sonne|workit is nowhere defined or?14:06
@wikingsonne|work: dunno what's it about... ask heiko he did that14:07
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun14:09
shogun-notifier-shogun: Soeren Sonnenburg :develop * 048c316 / src/interfaces/modular/Transfer.i: https://github.com/shogun-toolbox/shogun/commit/048c316e0757e437ae4f5869f2c6849a0d52cf3c14:09
shogun-notifier-shogun: drop ifdef EIGEN_MAPPED_SPARSEMATRIX_H causing java modular to fail14:09
shogun-notifier-shogun: Soeren Sonnenburg :develop * 6ed1224 / src/interfaces/modular/Transfer.i: https://github.com/shogun-toolbox/shogun/commit/6ed12240dc554df726f3dcb32e074940f902afdf14:11
shogun-notifier-shogun: use ifdef HAVE_LAPACK instead14:11
sonne|workwiking: ok that should fix it14:11
sonne|workwiking: what did you want to ask?14:14
sonne|workwiking: ohh and I have a Q14:14
sonne|workwiking: where is TEST_SRC defined?14:14
shogun-buildbotbuild #6 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/6  blamelist: Soeren Sonnenburg <sonne@debian.org>14:14
shogun-buildbotbuild #38 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/38  blamelist: Soeren Sonnenburg <sonne@debian.org>14:16
sonne|workwiking: scratch this Q14:17
shogun-buildbotbuild #1671 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1671  blamelist: Soeren Sonnenburg <sonne@debian.org>14:18
sonne|workthoralf: what does -DENABLE_TESTING=ON do?14:23
@wikingenables tests14:24
@wikingunit test14:24
@wikingsonne|work: btw here's your solution https://code.launchpad.net/~compiz-team/compiz/compiz.compiz_discover_tests/+merge/11332414:24
sonne|workwiking: and how do I enable integration tests?14:25
@wikingsonne|work: it's autoenabled ...14:26
@wikingexcept if u explicitly stat that u dont want14:26
@wiking*state14:26
sonne|workwiking: ahh nice14:28
@wikingsonne|work: ok i think i've managed it14:39
@wikingsonne|work: currently with that script14:39
@wikingand binary14:39
sonne|workwiking: ?14:39
@wikingall little tests are becoming one ctest14:39
@wikingADD_TEST (CSVFileTest.read_matrix "./tests/unit/shogun-unit-test""--gtest_filter=CSVFileTest.read_matrix")14:39
@wikingADD_TEST (CSVFileTest.write_matrix_real "./tests/unit/shogun-unit-test""--gtest_filter=CSVFileTest.write_matrix_real")14:39
@wikingADD_TEST (CSVFileTest.write_matrix_int "./tests/unit/shogun-unit-test""--gtest_filter=CSVFileTest.write_matrix_int")14:39
@wikingADD_TEST (CSVFileTest.write_vector_int "./tests/unit/shogun-unit-test""--gtest_filter=CSVFileTest.write_vector_int")14:39
@wikingADD_TEST (CSVFileTest.read_write_string_list "./tests/unit/shogun-unit-test""--gtest_filter=CSVFileTest.read_write_string_list")14:39
sonne|workwiking: I am doing that for python integration test scurrently14:39
@wikinglike this14:39
sonne|workwiking: did you do this already now or shall I continue?14:40
sonne|workwiking: nice ! lets hope it works14:40
@wikingsonne|work: well this is specifically for gtest14:40
@wikingsonne|work: yeah but i thought you want like 1 ctest per 'module' e.g. CSVFileTest.* would be 1 ctest14:41
@wikingor?14:41
sonne|workwiking: I am happy with anything pointing to at least a file to look at14:41
@wikingok i'm testing now..14:42
@wikingheheh there are like 3000+ tests14:42
@wikingwe'll have a problem :(14:42
@wiking(travis namely)14:42
shogun-buildbotbuild #1433 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1433  blamelist: Soeren Sonnenburg <sonne@debian.org>14:57
@wikingthe log is limited to 10000 lines and currently we already have like 5000 lines14:57
@wikingso if we have like 3000+ ctest and each ctest outputs (if it succeed) 2 lines => 2*3000 the best case..14:58
@wikingif it fails... well we are fucked :)14:58
@wikingbut already with the non-failing case we exceed 10000 lines14:58
@wikingand then travis automatically shuts down the job and marks it as failed14:59
shogun-buildbotbuild #7 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/7  blamelist: Soeren Sonnenburg <sonne@debian.org>14:59
shogun-buildbotbuild #39 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/39  blamelist: Soeren Sonnenburg <sonne@debian.org>15:00
shogun-buildbotbuild #1672 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1672  blamelist: Soeren Sonnenburg <sonne@debian.org>15:02
shogun-notifier-shogun: Soeren Sonnenburg :develop * 2dd0285 / tests/ (3 files): https://github.com/shogun-toolbox/shogun/commit/2dd02858eee3d87bb036aeda9f2658ae27fc6ee615:03
shogun-notifier-shogun: fix sgsparsematrix unittest when EIGEN is not available and make python modular test separate15:03
-!- travis-ci [~travis-ci@ec2-107-21-152-89.compute-1.amazonaws.com] has joined #shogun15:11
travis-ci[travis-ci] it's Soeren Sonnenburg'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/1114853115:11
-!- travis-ci [~travis-ci@ec2-107-21-152-89.compute-1.amazonaws.com] has left #shogun []15:11
shogun-notifier-shogun: Soeren Sonnenburg :develop * 74756ec / tests/integration/python_modular/CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/74756ecb0b69f8ffc53c53a0cdd58805be3e275c15:13
shogun-notifier-shogun: add forgotton .py suffix15:13
shogun-buildbotbuild #1751 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular test ruby_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1751  blamelist: Soeren Sonnenburg <sonne@debian.org>15:16
shogun-buildbotbuild #8 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/8  blamelist: Soeren Sonnenburg <sonne@debian.org>15:17
sonne|workwiking: hmmhh so what then?15:18
sonne|workwiking: show only failing tests?15:18
sonne|workwiking: we have 636 python integration tests15:18
shogun-buildbotbuild #40 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/40  blamelist: Soeren Sonnenburg <sonne@debian.org>15:19
@wikingsonne|work: yeah but unit tests are exploding... and as said 2 lines per ctest, if it's succeeds...15:20
@wikingif it fails it's even more...15:20
@wikingi'm trying to modify that filter in a way that per 'module'15:21
shogun-buildbotbuild #1673 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1673  blamelist: Soeren Sonnenburg <sonne@debian.org>15:21
@wikingi.e. we would have tests like unit-SVMOcasTest15:21
@wikingunit-EuclideanDistance15:21
sonne|workwiking: well anything that helps pointing at the failing file does tremendously help15:21
@wikingetc..15:21
@wikingok i did this and now only 126 tests like this...15:22
@wikingsonne|work: summary like this is ok:15:22
@wikingThe following tests FAILED: 150 - unit-RationalApproximationCGMJob. (Failed) 217 - unit-BeliefPropagation. (Failed) 218 - unit-FactorGraph. (Failed) 223 - unit-SerializationHDF5. (Failed)15:22
@wikingThe following tests FAILED: 150 - unit-RationalApproximationCGMJob. (Failed) 217 - unit-BeliefPropagation. (Failed) 218 - unit-FactorGraph. (Failed) 223 - unit-SerializationHDF5. (Failed)15:23
@wikingErrors while running CTest15:23
@wiking?15:23
@wikingalthough this will be shit as well15:23
@wikingbecause unit-SerializationHDF5 is fucking huge15:23
@wikingi mean the output of that15:23
sonne|workwiking: well good enough15:29
sonne|workwiking: I am concerned about why the osx bot is failing to compile ... http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/95/steps/compile/logs/stdio15:29
sonne|workseems like it is compiling gtest and failing15:29
@wikingsonne|work: i think glpk is old15:30
@wikingas it runs just fine on my machine15:30
@wikingand it's osx as well15:30
@wikingsonne|work: dude15:31
sonne|workwiking: ahh ok - so lets pimp the required glpk version then15:31
@wikingintegration-python_modular-tester-classifier_averaged_perceptron_modular15:31
@wikingwhy does cmake prints this?15:31
@wikingforgot something?15:31
sonne|workwiking: oops15:31
sonne|workdebug print msg still in15:31
sonne|workremove it at will or I do15:32
@wikingi'll do a commit now15:32
@wikingwhere so i'll do it15:32
@wikingwhere do i look?15:32
sonne|workwiking: in integration/python_modular15:32
sonne|worktests/integration/python_modular/CMake*.txt15:32
sonne|workwiking: could you pimp the glpk version aswell please?15:32
@wikingok found it15:32
@wikingsonne|work:dunno what is the good first version :)15:33
sonne|workme neither15:33
sonne|worklet me check what I have on osx15:33
@wiking  glpk @4.48_0 (active)15:33
@wikingand it runs fine with it15:33
sonne|work4.4815:33
sonne|workhah15:33
@wikinglol15:33
@wikingok then something else but we dont know yet what15:34
sonne|workwiking: I had this weird issue on osx that I compiled for python15:34
sonne|workand then had a clash with the system python15:34
-!- travis-ci [~travis-ci@ec2-54-226-222-38.compute-1.amazonaws.com] has joined #shogun15:36
travis-ci[travis-ci] it's Soeren Sonnenburg'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/1115001515:36
-!- travis-ci [~travis-ci@ec2-54-226-222-38.compute-1.amazonaws.com] has left #shogun []15:36
sonne|workwiking: ahh the new failures are perfect :)15:37
shogun-buildbotbuild #1434 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1434  blamelist: Soeren Sonnenburg <sonne@debian.org>15:42
shogun-buildbotbuild #1753 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1753  blamelist: Soeren Sonnenburg <sonne@debian.org>15:44
shogun-notifier-shogun: Viktor Gal :develop * dd3eef5 / / (4 files): https://github.com/shogun-toolbox/shogun/commit/dd3eef5332e7ca657095fe2e3e7d131cd3a4a6c515:45
shogun-notifier-shogun: Run each unit test module as a separate ctest15:45
@wikingthere15:45
shogun-notifier-shogun: remove debugging output from python modular intergration's cmake file15:45
sonne|workwiking: I think we should do ctest -j X on the bots now too...15:46
shogun-buildbotbuild #9 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/9  blamelist: Viktor Gal <viktor.gal@maeth.com>15:49
shogun-notifier-shogun: Soeren Sonnenburg :develop * cc4d411 / tests/integration/python_modular/ (10 files): https://github.com/shogun-toolbox/shogun/commit/cc4d411eda78186b540b356083eb47f0c7f7d59915:49
shogun-notifier-shogun: use modshogun in old integration tests15:49
sonne|workit is wow fast with ctest -j 8 :)15:49
shogun-buildbotbuild #41 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/41  blamelist: Viktor Gal <viktor.gal@maeth.com>15:51
shogun-buildbotbuild #1674 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1674  blamelist: Viktor Gal <viktor.gal@maeth.com>15:52
@wikinghttp://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1674/steps/test/logs/stdio15:52
@wikingsonne|work: there15:52
@wikingnow u know15:52
sonne|workwiking: yeah perfect15:53
sonne|workwiking: now please fix it :)15:54
@wikingicant15:54
@wikingi dont know what really fails15:55
@wikingthe only option is to disable mocking tests on freebsd15:55
shogun-notifier-shogun: Soeren Sonnenburg :develop * 8587982 / examples/undocumented/python_modular/ (2 files): https://github.com/shogun-toolbox/shogun/commit/8587982a846c022dec054fb51dacce4c19074a6015:59
shogun-notifier-shogun: dont' fail when lapack is not available15:59
sonne|workwiking: ok then disable mocking tests on fbsd15:59
shogun-buildbotbuild #1752 of deb3 - modular_interfaces is complete: Failure [failed test python_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1752  blamelist: Soeren Sonnenburg <sonne@debian.org>16:00
sonne|worka failing test that cannot be fixed has negative value16:00
shogun-buildbotbuild #10 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/10  blamelist: Soeren Sonnenburg <sonne@debian.org>16:03
shogun-buildbotbuild #42 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/42  blamelist: Soeren Sonnenburg <sonne@debian.org>16:04
shogun-buildbotbuild #11 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/11  blamelist: Soeren Sonnenburg <sonne@debian.org>16:06
shogun-buildbotbuild #1675 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1675  blamelist: Soeren Sonnenburg <sonne@debian.org>16:06
shogun-buildbotbuild #43 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/43  blamelist: Soeren Sonnenburg <sonne@debian.org>16:07
@wikingheh16:08
@wikinga lot of integration tests fail16:08
shogun-buildbotbuild #1676 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1676  blamelist: Soeren Sonnenburg <sonne@debian.org>16:10
shogun-buildbotbuild #1754 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1754  blamelist: Viktor Gal <viktor.gal@maeth.com>16:10
sonne|workwiking: errm 5 ?16:19
shogun-buildbotbuild #12 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/12  blamelist: Soeren Sonnenburg <sonne@debian.org>16:21
thoralflisitsyn: A few weeks ago you told a "magic" format string to use for integers of different types - no matter if signed/unsigned short/long/longlonglong... didn't find it on the IRC logs. Could you tell again?16:22
shogun-buildbotbuild #44 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/44  blamelist: Soeren Sonnenburg <sonne@debian.org>16:23
shogun-buildbotbuild #1435 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1435  blamelist: Soeren Sonnenburg <sonne@debian.org>16:24
shogun-buildbotbuild #1677 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1677  blamelist: Soeren Sonnenburg <sonne@debian.org>16:24
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:299e:473d:d0ad:c8d5] has quit [Quit: Ex-Chat]16:25
@wikingsonne|work: https://github.com/shogun-toolbox/shogun/issues/134816:27
@wikingsonne|work: i dont think it's a good idea to put the whole function into the header... you?16:27
-!- travis-ci [~travis-ci@ec2-107-21-152-89.compute-1.amazonaws.com] has joined #shogun16:27
travis-ci[travis-ci] it's Soeren Sonnenburg'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/1115038516:27
-!- travis-ci [~travis-ci@ec2-107-21-152-89.compute-1.amazonaws.com] has left #shogun []16:27
sonne|workwiking: we should try to have basically only definitions in headers nothing else16:28
@wikingwell16:28
@wikingthis will be shit anyhow16:28
@wikingbecause this .h has to include all the other shogun headers...16:28
sonne|workwiking: no16:30
sonne|workwiking: it can use forward declarations16:31
sonne|workbut the .cpp will be wtf16:31
shogun-buildbotbuild #1755 of deb3 - modular_interfaces is complete: Failure [failed test python_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1755  blamelist: Soeren Sonnenburg <sonne@debian.org>16:40
sonne|workwiking: can you disable the mocking based tests on bsd1?16:43
@wiking#ifndef FREEBSD :)16:46
sonne|workwiking: ok :D16:51
shogun-notifier-shogun: Viktor Gal :develop * af0165f / tests/unit/ (2 files): https://github.com/shogun-toolbox/shogun/commit/af0165f6e2b509afccae9cdeeb5230ed9e1765de16:51
shogun-notifier-shogun: Disable mocking tests on freebsd16:51
* wiking wonders what do he gets for the 3.0 release ....16:51
@wikingsonne|work: do u know anything what's happening on that cygwin machine?16:52
@wikingsonne|work: and it woudl be great to install ccache on cygwin...16:52
sonne|workwiking: no idea it works when I run the examples outside of cmake...16:54
shogun-buildbotbuild #13 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/13  blamelist: Viktor Gal <viktor.gal@maeth.com>16:54
@wikingsonne|work: no segfault at all?16:54
shogun-buildbotbuild #52 of FCRH - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/52  blamelist: Viktor Gal <viktor.gal@maeth.com>16:56
shogun-buildbotbuild #45 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/45  blamelist: Viktor Gal <viktor.gal@maeth.com>16:56
shogun-buildbotbuild #1678 of bsd1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/167816:57
@wikingok16:59
@wikingthen let's just drop tests on cygwin...16:59
shogun-buildbotbuild #1436 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1436  blamelist: Soeren Sonnenburg <sonne@debian.org>, Viktor Gal <viktor.gal@maeth.com>17:07
-!- sonne|osx [~sonne@89.204.153.86] has joined #shogun17:09
shogun-buildbotbuild #96 of osx1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/96  blamelist: Soeren Sonnenburg <sonne@debian.org>, Viktor Gal <viktor.gal@maeth.com>17:09
sonne|osxwiking: I just compiled under osx17:19
sonne|osxwiking: it compiles fine17:19
sonne|osxbut not on the builds lave though it is the same machine (wtf!?)17:19
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]17:19
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun17:19
@wikingshogun-buildbot: force build --branch=developer 'precise - libshogun'17:20
shogun-buildbotbuild #14 forced17:20
shogun-buildbotI'll give a shout when the build finishes17:20
shogun-buildbotbuild #14 of precise - libshogun is complete: Failure [failed git]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/1417:20
@wikingwtf17:20
@wikingshogun-buildbot: force build --branch=developer 'precise - libshogun'17:20
shogun-buildbotbuild #15 forced17:20
shogun-buildbotI'll give a shout when the build finishes17:20
shogun-buildbotbuild #15 of precise - libshogun is complete: Failure [failed git]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/1517:20
@wikingwhat the fuck17:20
@wikingwhat the fuck17:20
@wikingshogun-buildbot: force build --branch=developer 'precise - libshogun'17:21
shogun-buildbotbuild #16 forced17:21
shogun-buildbotI'll give a shout when the build finishes17:21
shogun-buildbotbuild #16 of precise - libshogun is complete: Failure [failed git]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/1617:21
@wikingshogun-buildbot: force build --branch=develop 'precise - libshogun'17:21
shogun-buildbotbuild #17 forced17:21
shogun-buildbotI'll give a shout when the build finishes17:21
shogun-buildbotbuild #17 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/1717:21
@wikingshogun-buildbot: force build --branch=develop 'precise - libshogun'17:22
shogun-buildbotbuild #18 forced17:22
shogun-buildbotI'll give a shout when the build finishes17:22
@wikingwtf... cc1plus: warning: unrecognized command line option "-Wno-c++11-narrowing" [enabled by default]17:23
sonne|osxwiking: I've logged out and in on osx maybe that helps...17:25
@wikingsonne|osx: ahhahaha17:25
sonne|osxshogun-buildbot: force build --branch=develop 'osx1 - libshogun'17:25
shogun-buildbotbuild forced [ETA 10m00s]17:25
shogun-buildbotI'll give a shout when the build finishes17:25
@wikingyeah maybe i should do the same...17:25
shogun-buildbotbuild #1757 of deb3 - modular_interfaces is complete: Failure [failed test python_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1757  blamelist: Viktor Gal <viktor.gal@maeth.com>17:25
sonne|osxwiking: no seriously maybe some paths or so were not set17:25
sonne|osxbut lets see17:25
@wikingsonne|osx: but it has found it...17:25
@wikingi mean glpk17:25
sonne|osxyeah :/17:26
sonne|osxkaboom17:26
sonne|osxdied17:26
@wikingsame here17:26
sonne|osxok let me try as the exact same user that starts the buildslave17:26
shogun-buildbotbuild #97 of osx1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/9717:26
sonne|osxwiking: argh it compiles17:28
* sonne|osx clueless17:28
sonne|osxgtg17:28
-!- sonne|osx [~sonne@89.204.153.86] has quit [Quit: sonne|osx]17:28
shogun-buildbotbuild #18 of precise - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/1817:29
-!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.]17:37
-!- sonne|osx [~sonne@f053041001.adsl.alicedsl.de] has joined #shogun17:45
@wikingsonne|work: kernel_distance_modular.py setting 1/2                       EXCEPTION 'ascii' codec can't decode byte 0xf1 in position 1: ordinal not in range(128)17:47
@wikingideas? :)17:47
sonne|osxwiking: yes17:50
sonne|osxwiking: errm no :)17:51
shogun-buildbotbuild #1437 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1437  blamelist: Viktor Gal <viktor.gal@maeth.com>17:51
thoralfwiking: Great error ;)17:52
@wikingyeah python3 is fun17:53
@sonney2kwiking, ohh strings are unicode now17:54
@sonney2kwiking, it means that the current alphabet (I guess DNA cannot contain a string 0xf1)17:57
@sonney2kbut only A C G T chars17:57
@wikingmm17:57
@wikingwell i guess there's a magic conversion function17:57
@wikingsomewhere :)17:57
thoralfFound a funny bug pattern, which could affect other classes as well: https://github.com/shogun-toolbox/shogun/pull/156217:58
thoralfJust in case you see it somewhere else... anyway, is not fatal as long as you're passed objects have refcount > 1.17:59
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has joined #shogun18:00
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/1115158518:00
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has left #shogun []18:00
thoralfpassing objects with refcount > 118:00
-!- vgorbati [~vgorbati@91.216.173.29] has joined #shogun18:25
shogun-buildbotbuild #1756 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1756  blamelist: Soeren Sonnenburg <sonne@debian.org>18:36
-!- pickle27 [~Kevin@199.119.128.114] has joined #shogun18:39
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:299e:473d:d0ad:c8d5] has joined #shogun18:46
-!- mode/#shogun [+o iglesiasg] by ChanServ18:46
@iglesiasgHi all!18:47
@iglesiasghushell, hey! Did you already have a rest?18:47
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has joined #shogun18:59
travis-ci[travis-ci] it's Soeren Sonnenburg'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/1115171218:59
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has left #shogun []18:59
@sonney2kwiking, argh! I accidentally killed ssh on the buildslave19:06
thoralfsonney2k: Ugh.19:08
thoralfsonney2k: No open ssh session left?19:08
@sonney2kthoralf, cygwin19:09
@sonney2k...19:09
thoralfsonney2k: Beside this, I think I found a good solution to the CMath template issue we talked about on friday: https://github.com/shogun-toolbox/shogun/pull/156019:12
thoralfsonney2k: Could've gone further, but wanted to avoid breaks somewhere else.19:13
@wikingsonney2k: which ?19:14
@wikingah ok... well it was failing anyways :D19:14
@wikingshogun-buildbot: force build --branch=develop 'FCRH - libshogun'19:14
shogun-buildbotbuild forced [ETA 1m45s]19:14
shogun-buildbotI'll give a shout when the build finishes19:14
@wikingshogun-buildbot: force build --branch=develop 'clang34 - static analysis'19:15
shogun-buildbotbuild #8 forced19:15
shogun-buildbotI'll give a shout when the build finishes19:15
shogun-buildbotbuild #53 of FCRH - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/5319:15
-!- lambday [67157c36@gateway/web/freenode/ip.103.21.124.54] has joined #shogun19:18
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:299e:473d:d0ad:c8d5] has quit [Quit: Ex-Chat]19:20
-!- pickle27 [~Kevin@199.119.128.114] has quit [Remote host closed the connection]19:31
shogun-buildbotbuild #8 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/819:36
vgorbatisonney2k: hi. About the PR: are there any more issues besides ones commented by you?19:44
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]19:51
@sonney2kvgorbati, we need some expert to look at this which I am not20:10
@sonney2kvgorbati, and I am knee deep in tons of other issues :/20:11
vgorbatisonney2k: I see:) And who can be the expert?)20:12
@sonney2kwiking, great you managed to get precise to green20:12
@sonney2kvgorbati, some serious NN guy20:12
vgorbatisonney2k: do you know such a guy?)20:13
@sonney2kvgorbati, the best would be if you could ask around in the community20:13
@sonney2kvgorbati, pretty clear that the big big shots won't listen but maybe they have some suggestion so asking won't hurt20:13
vgorbatisonney2k: Ok, maybe I'll try:) But implementing NNs in general and embedding them in shogun is a bit different things. I mean, you could have noticed some wrong use of existing shogun classes or violation of some policy etc. - was there anything (besides commented issues)?20:20
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has joined #shogun20:21
@sonney2kvgorbati, well rule is to not use std:: stuff in headers of classes that you swig wrap20:22
@sonney2kvgorbati, that's about it20:22
vgorbatisonney2k: ok) Then I'll close the PR. Maybe I will open a new one if I eventually reach a point when there is something working (like some basic neural network topology, that can be created/trained/applied from interfaces)?20:27
@sonney2kvgorbati, to be really useful you should somehow connect to a NN expert who can guide you through the tricks of trade20:28
@sonney2kvgorbati, every ML algo and NNs in particular requires some tricks to work well20:29
vgorbatisonney2k: well, I did take a look at existing python implementations (including projects, developed by some experts) + there are online courses on NNs, taught by Geoffrey Hinton, where he gives tips and trick on them20:33
vgorbatisonney2k: as well as describing the algorithms pretty much in detail (course link - https://class.coursera.org/neuralnets-2012-001/class/index)20:35
@sonney2kvgorbati, well it certainly is a good start20:36
-!- travis-ci [~travis-ci@ec2-54-242-103-128.compute-1.amazonaws.com] has joined #shogun20:36
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/1115406920:36
-!- travis-ci [~travis-ci@ec2-54-242-103-128.compute-1.amazonaws.com] has left #shogun []20:36
@wikingsonney2k: fuckshiiiiit20:36
@sonney2kvgorbati, I can try to ask around if sb. is interested in pushing it further20:36
@wikingsonney2k: i'm getting relly tired :(20:36
@wikingsonney2k: but i know i have to push this fucking release :<20:37
@sonney2kwiking, well you have been messing this all day20:37
@wikingsonney2k: ok so question20:37
@wikingsonney2k: we have the SerialUtest branch20:37
@sonney2kand we've come a loong way today I would say20:37
@wikingwhere we added automatic tests20:37
@wikingfor all the serialization outputs20:37
@wikingsome of them fail obviously20:37
@wikinglike HDF5 trying to write complex_t20:38
@wikingwhat do you way what should we do if that utest fail20:38
@wikingmake it really fail20:38
@wikingor just return from the function silently?20:38
@sonney2kwiking, we need a blacklist20:38
vgorbatisonney2k: I am sorry, did not get it - what do you mean by 'pushing it further'?20:38
@wikingsonney2k: and disable the test?20:38
@wikingwhich fails?20:38
@wikingah ok20:38
@sonney2kvgorbati, to get it to be comparable to the state of the art20:38
@wikingthat sounds reasonable20:38
shogun-buildbotbuild #98 of osx1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/9820:39
@wikingsonney2k: ok i'll add DISABLED20:39
@sonney2kwiking, we need a way to disable tests20:39
@wikingand then gtest will complain that there are disabled tests20:39
@wikingsonney2k: TEST(whatever, testinfunction_DISABLE)20:39
@wikingautomatically disables a unit test20:39
@sonney2kwiking, and have this list of disabled somewhere so we can always continue fixing things20:39
@wikingand gtest complains about that20:39
@sonney2kwiking, like we did before gsoc20:39
@sonney2kotherwise we would not have gotten anything merged20:39
@wikingok20:39
@sonney2kshogun-buildbot, f... you!20:39
shogun-buildbotWhat you say!20:39
@wikingi'll do this now then20:40
@wikingand then merge that branch20:40
@sonney2khow can this osx machine fail20:40
@wikingbecause there's quite some fix there by heiko20:40
@wikingthat solves some of the valgrind errors20:40
vgorbati:D20:40
@sonney2kI did compare environment variables20:40
@sonney2ksame20:40
@sonney2k^%$@%^#@^%#%^!!!20:40
@wikingsonney2k: run that bot with MAmake VERBOSE=120:41
@sonney2kwiking, what was the command again to get a verbose outptut from cmake20:41
@sonney2kexactly20:41
@wiking mean make VERBOSE=120:41
vgorbatisonney2k: what do you mean by 'it'?20:42
@sonney2kvgorbati, it == the/your implementation20:43
vgorbatisonney2k: but I want to push it further:) Or do you mean that I am not skilled enough to make it?20:44
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 240 seconds]20:52
@sonney2kvgorbati, well you are not hinton so you will reach a point where someone who is doing this all his live will be of good use20:53
@sonney2kwiking, I found the reason (but not the cause): on the buildbot we do -isysroot /Applications/...blabla../MacOSX10.8.sdk21:02
@sonney2kwiking, if I add this then the compile failes21:02
vgorbatisonney2k: I definitely do not pretend to do it as good as Hinton does:) But if I at least implement what I can (using existing implementations as reference) and if it works 'relatively well' - won't it be suitable to include in shogun?21:02
@sonney2kvgorbati, sure it would be sufficient21:03
@wikingsonney2k: lol...21:06
vgorbatisonney2k: this is what I was talking about - if I implement some working topology like a basis to move forward (I won't bother you until I have some results that can be evaluated - like python example of usage on real data etc.) - can I then issue a PR with this implementation?21:08
@sonney2kwiking, what do we need that for?21:09
@sonney2kvgorbati, sure21:09
@sonney2kwiking, can't we just get rid of this or will package building fail then?21:09
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun21:11
-!- mode/#shogun [+o iglesiasg] by ChanServ21:11
@sonney2kwiking, ?21:12
shogun-buildbotbuild #99 of osx1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/9921:31
shogun-buildbotbuild #100 of osx1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/10021:33
shogun-buildbotbuild #101 of osx1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/10121:37
@sonney2kwiking, finally found it: for some reason   MACOSX_DEPLOYMENT_TARGET=10.8 is set under the buildbot21:45
@sonney2kthis causes kabooms21:45
shogun-buildbotbuild #102 of osx1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/10221:47
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]21:52
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun21:52
hushelliglesiasg: hi, I am here :)21:54
@iglesiasghushell, hello hello :)21:54
hushelliglesiasg: the python and java issues have been solved?21:55
@iglesiasghushell, I think the java issue was solved, let me check about Python21:56
@iglesiasghushell, Python is still not good21:56
hushelldo I have to rebase the PR before merging?21:57
hushelliglesiasg: I guess Patrick doesn't have time at this point, maybe we should move on21:58
@iglesiasghushell, sure!21:59
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun21:59
shogun-notifier-shogun: hushell :develop * 4244455 / / (22 files): https://github.com/shogun-toolbox/shogun/commit/4244455e4e03563eeb823d9e6b989bb97def31ee21:59
shogun-notifier-shogun: learning FactorGraphModel21:59
shogun-notifier-shogun: Fernando Iglesias :develop * ec619b3 / / (22 files): https://github.com/shogun-toolbox/shogun/commit/ec619b31f5f430f88a1856508e1dda8849d3104321:59
shogun-notifier-shogun: Merge pull request #1538 from hushell/develop21:59
shogun-notifier-shogun:21:59
shogun-notifier-shogun: learning FactorGraphModel21:59
hushelliglesiasg: thanks!21:59
@iglesiasghushell, thanks to you!22:00
hushelliglesiasg: after I finished the 1st ipython notebook, I'll send another PR22:00
@iglesiasghushell, good thanks.22:00
@iglesiasghushell, paste it also in a gist or sth and put it here so I am can see it please :) http://nbviewer.ipython.org/22:01
shogun-notifier-shogun: Thoralf Klein :develop * ae6c7cf / src/shogun/ (2 files): https://github.com/shogun-toolbox/shogun/commit/ae6c7cf8c866ac4ac9002cb54af3897edb3e916322:01
shogun-notifier-shogun: SG_REF/SG_UNREF in wrong order can cause segfaults when setting the same mode again.22:01
shogun-notifier-shogun: Soeren Sonnenburg :develop * b25c744 / src/shogun/ (2 files): https://github.com/shogun-toolbox/shogun/commit/b25c7446fb17164274294e713b9b4ed90fb0cac222:01
shogun-notifier-shogun: Merge pull request #1562 from tklein23/struct_output_model_segfaults22:01
shogun-notifier-shogun:22:01
shogun-notifier-shogun: SG_REF/SG_UNREF in wrong order can cause segfaults22:01
hushelliglesiasg: You used the web browser GUI to edit it, right?22:02
hushellor you have a QT console22:03
hushelliglesiasg: Sure. I'll publish it as html22:03
@iglesiasghushell, I use the browser22:04
shogun-buildbotbuild #103 of osx1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/10322:05
hushelliglesiasg: nice, okay I'll jump to finish the paper draft for my supervisor22:07
@iglesiasghushell, cool, what is about?22:07
@iglesiasgIn case it is not secret hehe22:07
hushelliglesiasg: a workshop paper for domain adaptation, about transferring part based model, it's not a secret of course, too trivial haha22:08
@sonney2kparty22:08
@sonney2kosx1 went green22:09
hushellcool!22:09
@sonney2kour buildbot armada got crazily big http://buildbot.shogun-toolbox.org/waterfall22:10
@sonney2k16 buildbots...22:10
@iglesiasghehe22:14
@iglesiasglet's aim at getting all green!22:14
shogun-buildbotbuild #1679 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1679  blamelist: hushell <hushell@hushell-U510.(none)>22:15
@wikingsonney2k: pong22:16
shogun-buildbotbuild #1680 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1680  blamelist: Soeren Sonnenburg <sonne@debian.org>, Fernando Iglesias <fernando.iglesiasg@gmail.com>22:16
@wikingsonney2k: i dont know why do we need that.... and why aAGIN the fucking freebsd bot fails22:16
@wikingoh niice22:17
@wikingcompile error22:17
@wiking:D22:17
shogun-buildbotbuild #105 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/105  blamelist: Soeren Sonnenburg <sonne@debian.org>, Fernando Iglesias <fernando.iglesiasg@gmail.com>22:21
@sonney2kwiking, I hate this RationalApproximationCGMJob  - it fails every now and then... lambday you looked into this right?22:23
@sonney2klambday, I upgraded to eigen 3.2 but same thing22:23
@sonney2kwiking, do we have a valgrind analysis of this thing?22:24
@wikingsonney2k: in base64+gzip format yet :)22:26
@wikingyes22:26
@wikinghttp://buildbot.shogun-toolbox.org/memcheck/20130909-0653/DynamicAnalysis.xml22:26
@wiking;)22:26
@wikingsonney2k: should we run some thread checker as well?22:26
@sonney2kwiking, freebsd sounds like a make -j 128 issue22:26
lambdaysonney2k: sorry was away22:26
lambdaysonney2k: what fails?22:26
@sonney2kwiking, there seems to be a race condition22:26
@sonney2klambday, still the RationalApproximationCGMJob22:27
lambdayunit-test?22:27
@wikingsonney2k: another buildbot job in order...? :)22:27
@sonney2klambday, only osx22:27
@sonney2klambday, http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/10522:27
shogun-buildbotbuild #46 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/46  blamelist: Soeren Sonnenburg <sonne@debian.org>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, hushell <hushell@hushell-U510.(none)>22:27
lambdaysonney2k: checking22:27
@sonney2klambday, is there anything random in the test / did you valgrind it?22:27
lambdaysonney2k: few tests are random, ye22:28
lambdayyes22:28
@sonney2kwiking, [  3%] Performing install step for 'JSON'22:28
@sonney2kMaking install in .22:28
@sonney2ktest -z "/usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/third_party/libs/json" || /usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/build/JSON/src/JSON/install-sh -c -d "/usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/third_party/libs/json"22:28
@sonney2ktest -z "/usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/third_party/libs/json/pkgconfig" || /usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/build/JSON/src/JSON/install-sh -c -d "/usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/third_party/libs/json/pkgconfig"22:28
@sonney2kmkdir: /usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/third_party/libs: File exists22:28
@sonney2kmkdir: /usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/third_party/libs/json: File exists22:28
@sonney2kwiking, so what is going on here?22:28
@sonney2klambday, did you set a seed?22:28
@wikingsonney2k: i dont really know :)22:29
@wikingsonney2k: try force rebuild22:29
lambdaysonney2k: checking if this one actually has random anything - this fails because of eigen322:29
@sonney2kwiking, well not really a fix isn't it?22:29
@wikingsonney2k: try it22:29
@wiking:)22:29
@sonney2klambday, I already upgraded to eigen3 3.2.0 and it didn't help22:29
lambdaysonney2k: I remember you telling me this... I should give it a fixed value22:29
lambdaysonney2k: fixing22:30
@sonney2klambday, and from time to time it compiled fine22:30
@wikingsonney2k: ok so what is a good thread analyzer22:30
@wikingthoralf: what did u use22:31
@wikinghelgrind?22:31
@wikingor drd22:31
@sonney2kwiking, how about installing json-c on that bsd bot permanently?22:31
@wikingsonney2k: there's a good reason22:31
@wikingwhy i dont do that22:31
@wiking:)22:31
@sonney2kwiking, I mean make install of JSON failing because the dir is already there22:31
@sonney2kwiking, could you then please remove the install dir before installing?22:32
@wikingsonney2k: why it does not fail on other bots?22:32
@wiking:)22:32
@wikingsonney2k: i mean until now it was working fine22:32
@sonney2kwiking, because bsd is magic22:33
@sonney2kwiking, just rm -rf your way22:33
lambdaysonney2k: its really weird, since it doesn't have anything random22:33
@wikingshogun-buildbot: force build --branch=develop 'bsd1 - libshogun'22:33
shogun-buildbotbuild forced [ETA 5m55s]22:33
shogun-buildbotI'll give a shout when the build finishes22:33
@sonney2klambday, argh22:33
@wikinglet's see if it repeats itself :)22:33
@wikingok time to see the templating magic i did in action22:33
lambdaysonney2k: I am fixing it to a fixed value22:33
shogun-buildbotbuild #1681 of bsd1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/168122:36
@wikingsonney2k: see22:42
@wikingsonney2k: random glitch22:42
@sonney2klambday, thanks22:42
@sonney2kwiking, which will re-occur soon22:42
@sonney2koh great statistics_quadratic_time_mmd.py is giving different results all the time22:42
lambdaysonney2k: sent the PR22:43
@sonney2keasy to fix - another random.seed was missing22:44
@wikingsonney2k: so thread analyzer bot... shall we?22:45
lambdaysonney2k: np :) still clueless what might be causing this random behavior, specially the eigen3 direct solver :-/22:45
lambdayerr a mistake22:45
-!- pickle27 [~Kevin@199.119.128.114] has joined #shogun22:46
@sonney2klambday, could you valgrind it please?22:48
lambdaysonney2k: yeah already did22:48
@sonney2klambday, maybe there is some uninited memory involved22:48
lambdaysonney2k: no errors22:48
@sonney2klambday, and?22:48
@sonney2khmmhh22:48
lambdaysonney2k: anyway, this one will do as of now22:48
lambdayah, found the mistake22:50
lambdayanyway, this test will do fine22:51
lambdaysetting direct value won't harm22:51
lambdaysonney2k: could you please merge if travis passes?22:52
shogun-buildbotbuild #56 of FCRH - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/56  blamelist: Thoralf Klein <thoralf.klein@zib.de>22:54
shogun-buildbotbuild #47 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/47  blamelist: Thoralf Klein <thoralf.klein@zib.de>22:54
@sonney2klambday, what was it?22:54
lambdaysonney2k: my mistake! I mapped a SGVector to eigen3 first and then assigned that to a different vector22:55
@sonney2kwiking, examples/undocumented/python_modular/kernel_anova_modular.py  and classifier_multiclass_ecoc_random seem really broken22:55
lambdayand later using that map22:55
lambdayI wonder why it passed at all!22:55
lambdayfor few cases that is :|22:55
@sonney2klambday, lucky!22:55
@wikingsonney2k: yes indeed22:56
lambdaysonney2k: its badluck to be this lucky! tragically osx had to catch it not linux :|22:56
@wikingsonney2k: anova has various issues22:56
@wikingsonney2k: like serialization22:56
lambdaysonney2k: I am going off... good night :)22:57
lambdaygood night wiking :)22:57
shogun-buildbotbuild #106 of osx1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/10622:57
@wikingnightynight22:57
@wikingok22:57
@wikingwe are going to use this22:57
@wikinghttp://code.google.com/p/thread-sanitizer/22:57
-!- lambday [67157c36@gateway/web/freenode/ip.103.21.124.54] has quit []22:58
@wikingas soon as i get fixed this error22:58
-!- vgorbati [~vgorbati@91.216.173.29] has quit [Quit: vgorbati]22:58
@wikingFATAL: ThreadSanitizer CHECK failed: ../../../../src/libsanitizer/sanitizer_common/sanitizer_allocator.h:310 "((kSpaceBeg)) == ((reinterpret_cast<uptr>(Mprotect(kSpaceBeg, kSpaceSize))))" (0x7d0000000000, 0xffffffffffffffff)22:58
@wikingFATAL: ThreadSanitizer: failed to intercept pthread_mutex_lock22:58
@wiking:D22:58
shogun-buildbotbuild #1758 of deb3 - modular_interfaces is complete: Failure [failed install test libshogun test python_modular test java_modular test csharp_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1758  blamelist: hushell <hushell@hushell-U510.(none)>22:59
@sonney2kohh anova is not thread safe22:59
@sonney2kkaboom kaboom22:59
-!- sonney2k [~shogun@7nn.de] has left #shogun ["Ex-Chat"]23:00
-!- sonney2k [~shogun@7nn.de] has joined #shogun23:00
-!- mode/#shogun [+o sonney2k] by ChanServ23:00
@sonney2kand whatnot arrays23:00
@wikingSerializationAscii.GUIClassifier23:01
@wikingthis is bound to fail23:01
@wikingright?23:01
shogun-buildbotbuild #1136 of rpm1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm1%20-%20libshogun/builds/1136  blamelist: Thoralf Klein <thoralf.klein@zib.de>23:01
@sonney2kwiking, so serialization issues will be gone in a sec23:02
@sonney2kwiking, ohh man ignore the GUI* crap23:02
@wikingokey23:02
@wikingtrying to do that23:02
@wikingand get tsan work as well23:02
@sonney2kwiking, I mean it23:02
@wikingmmm i guess this thing is another grsec feature :S23:05
@wikingSerializationHDF5.ANOVAKernel23:07
@wikingfails23:07
@wikingIn file /Users/wiking/shogun/src/shogun/kernel/ANOVAKernel.cpp line 113: assertion lhs && rhs failed in void shogun::CANOVAKernel::allocate_arrays() file /Users/wiking/shogun/src/shogun/kernel/ANOVAKernel.cpp line 11323:07
@wiking[WARN] In file /Users/wiking/shogun/src/shogun/base/SGObject.cpp line 497: ANOVAKernel::load_serializable_post(): ShogunException: [ERROR] In file /Users/wiking/shogun/src/shogun/kernel/ANOVAKernel.cpp line 113: assertion lhs && rhs failed in void shogun::CANOVAKernel::allocate_arrays() file /Users/wiking/shogun/src/shogun/kernel/ANOVAKernel.cpp line 11323:07
shogun-buildbotbuild #1759 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1759  blamelist: Soeren Sonnenburg <sonne@debian.org>, Fernando Iglesias <fernando.iglesiasg@gmail.com>23:08
@sonney2kwiking, no this is just totally el stupida anova kernel crap23:09
shogun-notifier-shogun-data: Soeren Sonnenburg :master * 08c63b3 / testsuite/tests/ (8 files): https://github.com/shogun-toolbox/shogun-data/commit/08c63b369d6435774f8aa9ab7167bac7d1b201a023:18
shogun-notifier-shogun-data: convert linear svms to use dense instead of sparse features.23:18
shogun-notifier-shogun-data:23:18
shogun-notifier-shogun-data: adjust reference output data accordingly (trained svms didn't change!)23:18
shogun-notifier-shogun-data:23:18
shogun-notifier-shogun-data: cpu time should be < 0.5 so True is the correct output23:18
shogun-notifier-shogun: Soeren Sonnenburg :develop * 5f14bc2 / / (4 files): https://github.com/shogun-toolbox/shogun/commit/5f14bc2269333799765c5cbeb5b276de8084a92923:20
shogun-notifier-shogun: - fix statistics*mmd test to be reproducible (seed numpy rng)23:20
shogun-notifier-shogun: - fix anova kernel (was not thread safe)23:20
shogun-notifier-shogun: lambday :develop * 48f2b11 / tests/unit/lib/computation/RationalApproximationCGMJob_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/48f2b115967f7642a244181a9849ba66dafedde223:21
shogun-notifier-shogun: added direct value in unit-test to avoid fail on osx23:21
shogun-notifier-shogun: Soeren Sonnenburg :develop * 49891ec / tests/unit/lib/computation/RationalApproximationCGMJob_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/49891ec45d63b65afc536f8e122a6aa6f320da1f23:22
shogun-notifier-shogun: Merge pull request #1565 from lambday/feature/log_determinant23:22
shogun-notifier-shogun:23:22
shogun-notifier-shogun: added direct value in unit-test to avoid fail on osx23:22
shogun-notifier-shogun: Thoralf Klein :develop * 872c168 / src/shogun/ (9 files): https://github.com/shogun-toolbox/shogun/commit/872c168d50abf2b224d763d2073ba87fb0d3db4823:23
shogun-notifier-shogun: Fixed setters, where SG_UNREF was called before SG_REF.23:23
shogun-notifier-shogun: Thoralf Klein :develop * 2a2ddbd / src/shogun/ (12 files): https://github.com/shogun-toolbox/shogun/commit/2a2ddbd3d84b89815d0f3521346d7c6c8379e80923:23
shogun-notifier-shogun: Found even more critical places.23:23
shogun-notifier-shogun: Thoralf Klein :develop * 2fd1598 / src/shogun/ (6 files): https://github.com/shogun-toolbox/shogun/commit/2fd15985362068e1b86918b83dbcf8776fd11f3923:23
shogun-notifier-shogun: Even more classes with displaced SG_REF/SG_UNREF23:23
shogun-notifier-shogun: Thoralf Klein :develop * 2e18f46 / src/shogun/ (6 files): https://github.com/shogun-toolbox/shogun/commit/2e18f461dadee308c9fe51eee4fafc9c46449ea023:24
shogun-notifier-shogun: Swapping SG_REF / SG_UNREF in critical places.23:24
shogun-notifier-shogun: Soeren Sonnenburg :develop * 35e82ba / src/shogun/ (33 files): https://github.com/shogun-toolbox/shogun/commit/35e82bad8b379ec0534704b0de6b77fe7924096523:24
shogun-notifier-shogun: Merge pull request #1564 from tklein23/sgref_before_sgunref23:24
shogun-notifier-shogun:23:24
shogun-notifier-shogun: SG_REF before SG_UNREF23:24
shogun-notifier-shogun: Viktor Gal :feature/SerialUTests * 9ebe629 / tests/unit/CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/9ebe629adbeefccae0acc856bb744345f9fcbf1423:25
shogun-notifier-shogun: Make templated unit tests regenerate if the template or the generator23:25
shogun-notifier-shogun: script is changed23:25
shogun-notifier-shogun: Viktor Gal :feature/SerialUTests * 2b3c37f / tests/unit/ (7 files): https://github.com/shogun-toolbox/shogun/commit/2b3c37f467dc8e69bcd5536b55bc55b7768a36f923:25
shogun-notifier-shogun: Move ignore list of templated unit test to the template itself23:25
@wikingok let's see what travis does with this23:25
@sonney2kwiking, I fixed all of the integration tests but one23:25
@sonney2kclassifier_multiclass_ecoc_random is still flaky23:26
shogun-buildbotbuild #57 of FCRH - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/5723:26
@sonney2kand a the time test is so so... I mean it is rather dangerous to assume that it takes <0.5s CPU ...23:27
shogun-buildbotbuild #48 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/48  blamelist: lambday <heavensdevil6909@gmail.com>, Soeren Sonnenburg <sonne@debian.org>23:28
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has joined #shogun23:30
travis-ci[travis-ci] it's Fernando Iglesias'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/1116605423:30
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has left #shogun []23:30
shogun-buildbotbuild #1760 of deb3 - modular_interfaces is complete: Failure [failed test python_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1760  blamelist: Thoralf Klein <thoralf.klein@zib.de>23:30
shogun-buildbotbuild #49 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/49  blamelist: Soeren Sonnenburg <sonne@debian.org>, Thoralf Klein <thoralf.klein@zib.de>23:33
shogun-buildbotbuild #1137 of rpm1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm1%20-%20libshogun/builds/113723:35
shogun-buildbotbuild #108 of osx1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/108  blamelist: Thoralf Klein <thoralf.klein@zib.de>23:36
shogun-buildbotbuild #109 of osx1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/10923:43
--- Log closed Tue Sep 10 00:00:52 2013