Open in new window / Try shogun cloud
--- Log opened Tue Aug 13 00:00:13 2013
shogun-buildbotbuild #1747 of deb1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1747  blamelist: Kevin <kevinhughes27@gmail.com>00:04
shogun-buildbotbuild #1748 of deb1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1748  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>00:05
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has joined #shogun00:15
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/1012933100:15
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has left #shogun []00:15
@iglesiasggood night people00:35
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Ex-Chat]00:36
-!- travis-ci [~travis-ci@ec2-23-22-20-189.compute-1.amazonaws.com] has joined #shogun00:47
travis-ci[travis-ci] it's Sergey Lisitsyn'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/1013166100:47
-!- travis-ci [~travis-ci@ec2-23-22-20-189.compute-1.amazonaws.com] has left #shogun []00:47
shogun-notifier-shogun: Roman Votyakov :develop * 2ea8e6b / / (30 files): https://github.com/shogun-toolbox/shogun/commit/2ea8e6b4b89685729a8a7a962ac67cd267e3b8f501:05
shogun-notifier-shogun: refactor GP framework01:05
shogun-notifier-shogun: Heiko Strathmann :develop * 8ab70aa / / (30 files): https://github.com/shogun-toolbox/shogun/commit/8ab70aaf1908c5dabd0dfffb14a50ac112bffe0a01:05
shogun-notifier-shogun: Merge pull request #1398 from votjakovr/feature/gp_refactoring01:05
shogun-notifier-shogun:01:05
shogun-notifier-shogun: Refactor GP framework01:05
shogun-buildbotbuild #1750 of deb1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1750  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>01:09
shogun-buildbotbuild #1749 of deb1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1749  blamelist: Roman Votyakov <votjakovr@gmail.com>01:11
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has joined #shogun01:17
travis-ci[travis-ci] it's Sergey Lisitsyn'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/1013214401:17
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has left #shogun []01:17
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has quit [Ping timeout: 260 seconds]02:12
-!- travis-ci [~travis-ci@ec2-23-22-20-189.compute-1.amazonaws.com] has joined #shogun02:17
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/1013263902:17
-!- travis-ci [~travis-ci@ec2-23-22-20-189.compute-1.amazonaws.com] has left #shogun []02:17
-!- hushell [~hushell@c-24-21-169-136.hsd1.or.comcast.net] has joined #shogun02:40
-!- van51 [~van51@athedsl-224969.home.otenet.gr] has joined #shogun02:46
shogun-buildbotbuild #432 of nightly_none is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/43203:00
shogun-buildbotbuild #422 of nightly_all is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/42203:00
shogun-buildbotbuild #487 of nightly_default is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/48703:01
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has joined #shogun03:09
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/1013380203:09
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has left #shogun []03:09
van51sonney2k: sent a PR about the refcounting03:46
van51sonney2k: have a look when you can03:46
van51cu in the morning03:46
-!- van51 [~van51@athedsl-224969.home.otenet.gr] has left #shogun ["QUIT :Leaving."]03:47
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]04:05
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has quit [Ping timeout: 256 seconds]04:24
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun05:49
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 264 seconds]05:57
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun05:58
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 264 seconds]06:11
-!- gsomix_ is now known as gsomix06:30
gsomixgood morning06:30
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun06:31
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun06:33
shogun-notifier-shogun: Viktor Gal :develop * 27a1cee / / (6 files): https://github.com/shogun-toolbox/shogun/commit/27a1cee134cd68f50231de427c6da6395a7384f806:33
shogun-notifier-shogun: Add SYSTEM_INCLUDES list in cmake for supporting -isystem (like eigen)06:33
shogun-notifier-shogun: Fixing FindCBLAS.cmake for atlas based cblas libs06:33
shogun-notifier-shogun: readded -Wshadow for more details see issue #138506:33
@wikingshogun-buildbot: force build --branch=develop 'deb1 - libshogun'06:40
shogun-buildbotbuild forced [ETA 16m01s]06:40
shogun-buildbotI'll give a shout when the build finishes06:40
@wikingshogun-buildbot: force build --branch=develop 'bsd1 - libshogun'06:41
shogun-buildbotbuild forced [ETA 8m25s]06:41
shogun-buildbotI'll give a shout when the build finishes06:41
shogun-buildbotbuild #1752 of deb1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/175206:41
shogun-buildbotbuild #1751 of deb1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1751  blamelist: Viktor Gal <viktor.gal@maeth.com>06:46
shogun-buildbotbuild #1405 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/140506:48
gsomixwiking, around?06:49
@wikingyes06:50
gsomixwiking, how to work with unit-tests now?06:50
@wikingwhy? :)06:50
@wikingwhat u want06:50
@wikinggdb?06:50
@wikingbeecause then06:50
@wikinggdb ./build/tests/unit/shogun-unit-tests06:51
gsomixwiking, nope. just compile and run.06:51
gsomixwith new cmake system06:51
@wikinggsomix: do the cmake06:51
@wikinggsomix: make sure you've given -DENABLE_TESTING=ON06:51
@wikingfor cmake06:51
@wikingi.e in build: cmake -DENABLE_TESTING=ON ..06:51
@wikingand then just make06:51
@wikingand then06:51
@wikingctest --output-on-failure06:52
gsomixctest --output-on-failure06:53
gsomixTest project /home/gsomix/Workspace/shogun/build06:53
gsomix        Start   1: unit-test06:53
gsomix  1/108 Test   #1: unit-test ..........................................................   Passed    6.12 sec06:53
@wikinglol06:54
gsomixwiking, hm. it seems there is no problem on my machine06:54
@wikingyes06:54
@wikingbecuae i think it's disabled now :)06:54
gsomixoh06:54
@wikingno it's not06:55
@wikingbut heh it's fun06:55
@wikingit just segfaulted on my machine06:55
gsomixwiking, ok. I'll check it twice.06:55
@wikinggsomix: and it really does segfault on most machine06:56
@wiking:*06:56
@wiking:(06:56
@wikingok06:57
@wikingon another machine it's not segfault06:57
@wikingit's like this06:57
@wikinghttp://pastebin.com/H9N9VT6E06:58
@wikingbut now it segfaulted as well06:58
@wikinggsomix: try to run it several times06:58
gsomixnice06:59
@wikinggsomix: if u want to run only unit tests: ctest --output-on-failure -R unit07:01
shogun-notifier-shogun: Viktor Gal :develop * 1274a74 / .travis.yml,tests/unit/io/CSVFile_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/1274a74ee5eb463165004cc9f80fbc5cb15e6aed07:12
shogun-notifier-shogun: Move OctaveModular travis job out of allowed to fail since aa628f436007:12
shogun-notifier-shogun: fixed the segfault07:12
shogun-notifier-shogun:07:12
shogun-notifier-shogun: temporarily disable CSVFileTest.read_write_string_list unit test (see07:12
shogun-notifier-shogun: issue #1395)07:13
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has joined #shogun07:14
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/1014188807:14
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has left #shogun []07:14
@wikingshogun-buildbot: force build --branch=develop 'bsd1 - libshogun'07:14
shogun-buildbotbuild forced [ETA 8m25s]07:14
shogun-buildbotI'll give a shout when the build finishes07:14
shogun-buildbotbuild #1753 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/175307:16
shogun-buildbotbuild #1492 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/1492  blamelist: Viktor Gal <viktor.gal@maeth.com>07:16
shogun-buildbotbuild #1378 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/1378  blamelist: Viktor Gal <viktor.gal@maeth.com>07:17
shogun-buildbotbuild #872 of rpm1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/872  blamelist: Viktor Gal <viktor.gal@maeth.com>07:18
shogun-buildbotbuild #1194 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1194  blamelist: Viktor Gal <viktor.gal@maeth.com>07:18
@wikinggsomix: if u update now your git repo, you wont have errors anymore as i've disabled the failing test... so make sure you enable it locally....07:19
-!- gsomix [~gsomix@95.67.162.140] has quit [Ping timeout: 248 seconds]07:19
shogun-buildbotbuild #1406 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/140607:22
shogun-buildbotbuild #1407 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1407  blamelist: Viktor Gal <viktor.gal@maeth.com>07:29
-!- gsomix [~gsomix@85.26.234.140] has joined #shogun07:32
-!- lambday [67157f36@gateway/web/freenode/ip.103.21.127.54] has joined #shogun08:00
lambdaywiking: there?08:00
lambdaywiking: I have libcblas.so.3 in /usr/lib64/atlas/08:01
lambdaycmake can't detect it, can I manually specify this?08:01
lambdaywiking: checked config.h I get `/* #undef HAVE_LAPACK */`08:02
lambdayoh and in one of the undocumented examples (examples/undocumented/classifier_lda.cpp) I think it should have checked for HAVE_LAPACK08:02
lambdayit fails when it tries to use CDataGenerator::generate_gaussians(..), may be pickle27 and iglesias can take a look08:03
-!- travis-ci [~travis-ci@ec2-23-22-20-189.compute-1.amazonaws.com] has joined #shogun08:03
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/1014261308:03
-!- travis-ci [~travis-ci@ec2-23-22-20-189.compute-1.amazonaws.com] has left #shogun []08:03
gsomixwiking, I'm freaking retard. There is fix for #1395, let's see what Travis will say.08:06
@wikinggsomix: have u removed DISABLE_?08:10
@wiking:)08:10
gsomixwiking, yep.08:10
@wikingu did okey08:10
@wikinglet's see08:10
@wikinglambday: ping?08:14
lambdaywiking: pong08:14
lambdaywiking: no symlinks in /usr/lib or /usr/lib6408:14
lambday:(08:14
@wikingmmm okey08:14
lambdayI guess default installation doesn't create this one08:14
@wikinglambday: what's your distro?08:15
lambdayfedora16.x86+6408:15
@wikingaha08:15
@wikingokey08:15
@wikingi'll add atlas detection08:15
@wikingjust a sec08:15
lambdaywiking: thanks man08:15
@wikingor at least a preliminary one08:15
@wikingbut you'lll have to test it because i dont have a machine to do so08:15
lambdaywiking: oh and we completely got rid of the consfigure script?08:15
lambdaywiking: no problem I will test it08:15
lambdayconfigure*08:16
@wikinglambday: indeed08:16
@wikinglambday: sonney2k was in a hurry it seems08:16
lambdaywiking: I don't know how this cmake works.. can I manually specify an interface to build, say, cmdline_static, or java_modular?08:17
lambdaywiking: or, say I have an older version of eigen3 somewhere and I want to make using that one instead of what I have in /usr/include/eigen308:18
@wikinglambday: the interface you can manually specify08:18
lambdaywiking: how?08:19
@wikinglambday: well currently only libshogun is enabled by default08:19
@wikinglambday: the rest you read my email08:19
@wiking(see shogun mailing list08:20
@wiking)08:20
lambdaywiking: alright... I'll check it.. thanks08:20
@wikinglambday: [shogun] shogun with cmake08:21
@wikingthis was the subject of the email08:21
lambdaywiking: thanks man :)08:22
-!- gsomix [~gsomix@85.26.234.140] has quit [Ping timeout: 264 seconds]08:26
-!- gsomix [~gsomix@85.26.234.140] has joined #shogun08:27
lambdaythis is amazing!08:28
@wikingwhat?08:29
lambdaywhat cmake can do...08:29
lambdayspecially the pack thing..08:29
lambdayso fedora people will get shogun.rpm soon08:29
lambday:D08:29
lambdaywiking: oh but regarding specifying a different eigen3 include dir, currently it can't be done, right?08:30
@wikingyes it can bee08:31
@wiking-DCMAKE_PREFIX_PATH08:31
lambdaywiking: that's not for target installation dir prefix?08:32
@wikingnoup08:32
lambdayalright I am checking08:32
shogun-notifier-shogun: Viktor Gal :develop * 4c903e4 / data: https://github.com/shogun-toolbox/shogun/commit/4c903e4c669c675d7b7c46931a3c2854b2d0e32808:33
shogun-notifier-shogun: refernce the right version of data for ica notebook08:33
lambdayoh it says unsuitable version 3.0.6, made my life easier that I don't have to worry about older eigen3 anymore08:35
@wikingyeah forget everything that is older than 3.1.208:36
lambdayah what a relief!08:36
shogun-buildbotbuild #1379 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/1379  blamelist: Viktor Gal <viktor.gal@maeth.com>08:36
lambdayno more YES_I_KNOW_SPARSE_MODULE_NOT_STABLE_YET :D08:36
shogun-buildbotbuild #1493 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/1493  blamelist: Viktor Gal <viktor.gal@maeth.com>08:37
-!- gsomix [~gsomix@85.26.234.140] has quit [Ping timeout: 264 seconds]08:37
shogun-buildbotbuild #1195 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1195  blamelist: Viktor Gal <viktor.gal@maeth.com>08:38
shogun-buildbotbuild #873 of rpm1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/873  blamelist: Viktor Gal <viktor.gal@maeth.com>08:39
sonne|worklambday: is it stable now?08:40
lambdaysonne|work: I think so08:41
lambdaysonne|work: moin moin by the way :)08:41
sonne|worklambday: moin moin too08:41
sonne|worklambday: did wiking fix the cblas issue then?08:42
sonne|worklisitsyn: why did you remove conjugateindex?08:42
lisitsynsonne|work: it is not a classifier and it is untested08:42
lambdaysonne|work: well, I had libcblas.so.3 in /usr/lib64/atlas... so currently it can't detect that08:42
lambdaysonne|work: wiking is adding atlas support08:42
lambdaythere is an undocumented example in /libshogun/classifier_lda.cpp, which should have a HAVE_LAPACK check08:43
shogun-buildbotbuild #1408 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1408  blamelist: Viktor Gal <viktor.gal@maeth.com>08:44
lambdayCDataGenerator::generate_gaussian(..) is defined when HAVE_LAPACK08:44
sonne|worklisitsyn: so? add a test08:44
sonne|worklambday: well fix and PR!08:44
lambdaysonne|work: alright I'll add it08:44
lisitsynsonne|work: I don't want to spend any time on something that won't work anyway :)08:46
-!- gsomix [~gsomix@85.26.234.140] has joined #shogun08:46
sonne|workmoin gsomix!08:47
gsomixsonne|work, hello08:47
gsomixsonne|work, about vlen+1. don't need bacause our strings are not zero-terminated.08:48
gsomixbecause08:48
@wikinglambday: working on it08:48
gsomixbut I think we should store '\0' at end of strings. it will be very useful for functions from <string.h>08:49
sonne|workgsomix: please document that - it might very well be that some functions depends on a zero terminated string08:49
sonne|workdepend08:50
shogun-notifier-shogun: Evgeniy Andreev :develop * e6d63be / src/shogun/io/CSVFile.cpp,tests/unit/io/CSVFile_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/e6d63be2e156fd0b8764b8aaebdf929b01078b1f08:54
shogun-notifier-shogun: fix for issue #139508:54
shogun-notifier-shogun: Soeren Sonnenburg :develop * ae55acc / src/shogun/io/CSVFile.cpp,tests/unit/io/CSVFile_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/ae55acc16341d4b24424a0e96e24ca50f8d5fe6508:54
shogun-notifier-shogun: Merge pull request #1400 from gsomix/develop08:54
shogun-notifier-shogun:08:54
shogun-notifier-shogun: Fix for issue #139508:54
gsomixsonne|work, ok. btw we should change CFile interface to use SGVector insted SGString, right?08:55
sonne|workgsomix: not yet. we would have to convert SGString -> SGVector first and this is a massive undertaking08:56
shogun-buildbotbuild #1756 of deb1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1756  blamelist: Soeren Sonnenburg <sonne@debian.org>08:59
sonne|workwiking: is it possible with cmake to split up the compile step and the test step?09:00
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has quit [Quit: Leaving.]09:00
sonne|workso that one can run make tests separately?09:00
@wikingsonne|work: yes but then we have a problem09:01
@wikingsonne|work: say you do make && ctest --output-on-failure09:01
shogun-buildbotbuild #1380 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/1380  blamelist: Evgeniy Andreev <gsomix@gmail.com>09:01
@wikingsonne|work: then ctest will fail because make did not created the tests...09:01
shogun-buildbotbuild #1494 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/1494  blamelist: Evgeniy Andreev <gsomix@gmail.com>09:02
@wikingsonne|work: if we accept this. then we can create a custom target say make test09:02
@wikingsonne|work: i mean make tests or something like that as make test is already reserved for ctest09:02
sonne|workwiking: why doesn't make test compile the test and run them?09:02
@wikingsonne|work: because this is how it is09:03
@wikingsonne|work: there's tons of discussions about htis..09:03
@wikingbut it is like this09:03
shogun-buildbotbuild #1196 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1196  blamelist: Evgeniy Andreev <gsomix@gmail.com>09:03
sonne|workwiking: you mean cmake related discussions?09:03
@wikingsonne|work: yes09:03
sonne|workyeah sure that was expected09:03
shogun-buildbotbuild #874 of rpm1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/874  blamelist: Evgeniy Andreev <gsomix@gmail.com>09:04
sonne|workit takes a while until everythign is back to normal09:04
sonne|workit is a very intrusive change09:04
shogun-buildbotbuild #1409 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1409  blamelist: Evgeniy Andreev <gsomix@gmail.com>09:08
@wikinglambday: here?09:09
@wikinglambday: i have a preliminary patch for fixing your lapack/atlas problem but before commiting i'd like u to test it. what's your email where i can send it to?09:10
lambdaywiking: yes sure... heavensdevil6909@gmail.com09:11
-!- gsomix [~gsomix@85.26.234.140] has quit [Ping timeout: 264 seconds]09:12
lambdaywiking: got it09:12
@wikinglambday: sent it... just git apply <patchname> and then rerun cmake09:12
@wikinglambday: what's the output of your cmake?09:15
lambdaywiking: Could NOT find ATLAS (missing:  ATLAS_LIBRARIES)  -- Could NOT find CBLAS (missing:  CBLAS_LIBRARY)09:15
lambday:(09:15
@wikingah ok09:15
@wikinglambday: i'm sending now you a new one09:16
@wikingsent09:18
@wikingbut in order to apply it09:18
lambdaywiking: checking09:18
@wikinggit reset --hard && rm cmake/FindAtlas.cmake09:18
@wikingand then just after that do the git apply09:18
lambdaywiking: I just did git apply -R <previous-patch>,09:19
lambdaywiking: will it affect?09:19
@wikingdunno09:19
lambdayokay.. trying09:19
@wikingthe only thing i've changed is that in FindAtlas at some lines i've added /usr/lib6409:19
@wikingcheck your cmake/FindAtlas.cmake09:20
@wikingfind_file(ATLAS_LIB libatlas.so.3 PATHS /usr/lib /usr/lib64 $ENV{ATLASDIR})09:20
@wikingif it's like this09:20
lambdayit is missing /usr/lib64...09:22
lambday(I didn't apply the new patch yet)09:23
lambdaywiking:09:23
@wikinglambday: but you have anyways /usr/lib64/atlas/.... right/09:23
@wikinglambday: but you have anyways /usr/lib64/atlas/.... right?09:23
lambdayyep09:23
-!- gsomix [~gsomix@88.200.245.98] has joined #shogun09:25
@wikinglambday: ok i think i have a fix then09:27
@wikinglambday: just sent a latest fix09:28
lambdaywiking: alright checking09:28
@wikinglambday: as said: git reset --hard && rm cmake/FindAtlas.cmake09:28
@wikingand then run the cmake09:28
@wikingand show me the output09:28
lambdayokie09:28
@wikinglambday: do you have a liblapack_atlas.so.3 under /usr/lib64/atlas ?09:29
lambdaywiking: its actually in two different places, both /usr/lib64/ and /usr/lib64/atlas09:29
lambdayyes /usr/lib64/atlas/ has liblapack.so.309:30
@wikinglambday: you have no liblapack_atlas.so.3 ?09:30
@wikingonly liblapack.so.309:30
@wiking?09:30
lambdaywiking: yep09:30
lambdaywiking: but it found09:31
lambdayA library with LAPACK API found. -- Found ATLAS: /usr/lib64/atlas/liblapack.so.3;/usr/lib64/atlas/libcblas.so.3;/usr/lib64/atlas/libf77blas.so.3;/usr/lib64/atlas/libatlas.so.3;/usr/lib64/liblapack.so.309:31
@wikinglambday: ok09:31
lambdaymaking09:31
@wikinglambday: can u plz now make the whole thing09:31
@wikingand try to run the tests09:31
@wikinglet's hope this fixed it09:31
@wikingand then i can push the commit09:31
lambdayalright09:31
lambdaymaking09:31
@wikingok let me know of the output09:31
@wikinggsomix: can we mark #1395 fixed then?09:32
gsomixwiking, hmhm, there are no segfaults on your machine?09:33
@wikinggsomix: noup :)09:33
@wikinggsomix: i was just wondering if the discussion of the PR is settled...?09:33
lambdaywiking: yooooooooo09:34
@wikinglambday: what?09:34
lambdaywiking: make successful09:34
lambdaywiking: but now running the unit-tests09:35
lambdayumm09:35
@wikinglambday: ctest --output-on-failure09:35
gsomixwiking, nope. you can close #1395.09:35
lambdayalright09:35
@wikinglambday: in the ./build directory where you did cmake and make09:35
lambdayokie09:35
lambdaywiking: oh and I need to cmake with -DENABLE_TESTING=ON09:36
@wikinglambday: indeed :)09:36
lambdayalright09:36
@wikingif u did not then u wont have tests09:36
sonne|workgsomix: so what are your next steps?09:36
lambdaywiking: making again...09:37
lambdaywiking: by the way, make valgrind and valgrind-per-module things are still same, right?09:37
lambdayfor unit tests09:37
@wikinglambday: no09:38
@wikinglambday: they are all gone09:38
lambdayno valgrind :(09:38
gsomixsonne|work, libsvm, refactoring old code for use csv reader.09:39
@wikinglambday: u can have valgrind if u add -DBUILD_DASHBOARD_REPORTS=ON to cmake arguments09:39
@wikinglambday: then after cmake and make u can do: ctest -D ExperimentalMemCheck09:39
@wikingand that'll do u a valgrind on all the tests09:40
@wikingif u want only on the unit test then: ctest -D ExperimentalMemCheck -R unit09:40
lambdaywiking: ah alright09:40
gsomixsonne|work, do we need sparse access methods for csv reader?09:40
gsomixI mean this methods will be in libsvm reader.09:40
lambdaywiking: but I got this while making... svn: Server sent unexpected return value (307 Temporary Redirect) in response to OPTIONS request for 'http://googlemock.googlecode.com/svn/trunk'09:41
lambdaywiking: where shall I put gtest/gmock so that it doesn't try to download each time09:41
lambdayits in /usr/src09:41
sonne|workgsomix: I could imagine a case where a sparse matrix is stored as a .csv but I guess it is sufficient to just convert it from a dense matrix09:41
-!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun09:42
@wikinglambday: nowhere... it will download it once09:42
@wikinglambday: try to run make again09:42
lambdaywiking: okay I will have to set svn proxy first09:43
-!- gsomix_ [~gsomix@85.26.234.140] has joined #shogun09:45
-!- gsomix [~gsomix@88.200.245.98] has quit [Ping timeout: 256 seconds]09:46
gsomix_sonne|work, are there some functions to convert dense to sparse matrices?09:47
sonne|workgsomix_: yes09:47
gsomix_sonne|work, cool.09:56
@wikinggsomix_: bugs for ya09:57
gsomix_wiking, c'mon09:59
gsomix_I have very bad internet connection now. :(09:59
@wikinggsomix_:  :(10:00
@wikingi'm just saying that i've found some misterious bugs10:00
@wikinglambday: any progress?10:00
lambdaywiking: yes.. ran unit-tests, now cmake-ing with dashboard reports on10:01
gsomix_wiking, do you create issue on github? or so?10:01
lambdayso, check I do ctest -R unit, it just checks one unit-test10:02
@wikinggsomix_: yep i've even assigned you10:02
@wikinglambday: ctest -R unit runs the whole unit test10:03
@wikingi.e. all the unit tests that are currently in the repo10:03
@wikinglambday: can u plz run ctest --output-on-failure -R libshogun10:03
lambdaywiking: oki... but it doesn't report each one individually.. just said 100% tests passed10:03
@wikinglambday: just to see if everything is linked in a correct way?10:03
lambdaywiking: okay I will... waiting for make to complete10:04
@wikinglambday: ok.... --output-on-failure10:04
gsomix_wiking, thanks!10:04
@wikinglambday: it reports if there's a failure10:04
@wikinglambday: otherwise it doesn't show the stdout/stdcout10:04
lambdaywiking: I tried removing that, but still same.. may be I did something wrong10:04
lambdaywiking: checking again10:04
@wikinglambday: if u really want it give -V or -VV flags for ctest10:04
lambdayalright10:05
@wikinglambday: if all the tests passed10:05
@wikingthen there's not unit test output because there was nothing wrong with it10:05
@wikingi.e. --output-on-failure only shows the output if there's something wrong (as the cmd args says)10:05
lambdayoki10:05
@wikinglambday: ok so cblas/atlas is fixed actually with that patch10:05
@wikingright?10:05
lambdaywiking: yep :D10:05
@wikinglambday: cool10:05
shogun-notifier-shogun: Viktor Gal :develop * d4baee9 / CMakeLists.txt,cmake/FindAtlas.cmake: https://github.com/shogun-toolbox/shogun/commit/d4baee974e20fd615e6cf58e99b46a5cd28155ec10:06
shogun-notifier-shogun: Add Atlas detection to cmake10:06
shogun-notifier-shogun: this fixes issue #139410:06
lambdaythis is amazing... no manual download needed.. cmake does it all for us..10:06
lambdayjust need a few days to get used to it10:07
-!- gsomix_ [~gsomix@85.26.234.140] has quit [Ping timeout: 264 seconds]10:09
shogun-buildbotbuild #1757 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/175710:12
shogun-buildbotbuild #1381 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/1381  blamelist: Viktor Gal <viktor.gal@maeth.com>10:12
shogun-buildbotbuild #1495 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/1495  blamelist: Viktor Gal <viktor.gal@maeth.com>10:12
shogun-buildbotbuild #1197 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1197  blamelist: Viktor Gal <viktor.gal@maeth.com>10:14
shogun-buildbotbuild #875 of rpm1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/875  blamelist: Viktor Gal <viktor.gal@maeth.com>10:14
sonne|workwiking: will this work when atlas is not available but blas/lapack?10:18
shogun-buildbotbuild #1410 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1410  blamelist: Viktor Gal <viktor.gal@maeth.com>10:19
@wikingsonne|work: yes... but still cblas has to be present...10:21
@wikingsonne|work: if there's no libcblas then LAPACK is not enabled...10:21
sonne|workwiking: that is checked though right?10:21
@wikingsonne|work: yes10:21
sonne|workok perfect!10:21
sonne|workwiking: now for the bots to work I still need an option to manually specify gmock/gtest paths10:22
sonne|workwiking: for the other libs I guess there is some standard if they are not in standard include etc locations right?10:22
@wikingsonne|work: why we cannot have svn?10:22
sonne|workwiking: you cannot assume machines with internet connectivity to build shogun10:23
@wikingsonne|work: no i dont assume that10:23
@wikingsonne|work: but then just dont do the tests10:23
@wikingand that's all10:23
sonne|workwiking: why is bsd still failing http://shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun ?10:23
@wikingsonne|work: there's some weird gmock segfault10:24
sonne|workwiking: no I want to run the tests on cygwin10:24
@wikingand it doesn't have internet?10:24
sonne|workwiking: it still says ../../src/shogun/libshogun.so.14.0: undefined reference to `clapack_dgetrf'10:24
sonne|workwiking: that is not the point - consider someone on an island with shogun sources10:25
sonne|worktrying to get it to work10:25
sonne|workthis 'island' test is sth that has to pass for a software to be in debian10:26
@wikingmmm10:28
@wikingthat'll be interesting10:28
@wikingbecause i dont know how i can add there a REVISION check10:29
sonne|workwiking: what do you mean with REVISION check?10:31
@wikingsonne|work: well i have to assure that a given svn rev is provided for the gmock/gtest10:32
@wikingas src10:32
@wikingbecause for example the 1.6.0.zip is not sufficient...10:33
sonne|workwiking: why not 1.6?10:34
sonne|worknot going with official releases will cause pain10:35
@wikingsonne|work: well that one is not sufficient10:35
@wikingthat's why10:35
sonne|workwhat is missing?10:35
@wikingi mean there's a good reason i was doing this this way10:35
@wikinga lot...10:35
sonne|workwiking: I don't question that but what are the resons?10:37
sonne|workwiking: will it present itself as 1.7 sth?10:38
@wikingdunno10:38
@wikingi'm not a gtest/gmock developer10:38
@wikingsonne|work: as far as i can see there's a constantly upgraded CHANAGES log where 1.7.0 is mentioned10:46
@wikingbut that's not yet releaed10:46
@wikingreleased10:46
sonne|workwiking: I have no idea why we need 1.7 but hey...10:47
@wikingsonne|work: c++11 and others10:47
sonne|workahh ok that is a good reason10:47
@wiking1.6 is like from 2001110:47
@wiking201110:47
lisitsynwiking: it is like they are doing something rock-stable10:53
@wikingi guess so :P10:53
@wikingi wonder10:56
@wiking../../src/shogun/libshogun.so.14.0 -lgmock -lgtest /usr/local/lib/liblapack.so /usr/local/lib/libcblas.so /usr/local/lib/libf77blas.so /usr/local/lib/libatlas.so /usr/local/lib/liblapack.so /usr/local/lib/libcblas.so /usr/local/lib/libf77blas.so /usr/local/lib/libatlas.so /usr/local/lib/libglpk.so /usr/local/lib/libsuperlu_3.0.so ../../../third_party/libs/json/libjson-c.a /usr/local/lib/libxml2.so /usr/local/lib/libhdf5.so -lz -lm -lz -lm -lbz2 -llzma /10:56
@wiking../../src/shogun/libshogun.so.14.0: undefined reference to `clapack_dgetrf'10:56
@wiking../../src/shogun/libshogun.so.14.0: undefined reference to `clapack_dgetri'10:56
@wiking../../src/shogun/libshogun.so.14.0: undefined reference to `clapack_dpotrf'10:56
@wiking../../src/shogun/libshogun.so.14.0: undefined reference to `clapack_dposv'10:56
lisitsynwiking: so all wrappers?10:57
lisitsynthat we have10:57
sonne|workwiking, lisitsyn - it makes sense to do some rock stable for a test framework right?11:02
lisitsynsonne|work: yes if it wasn't I'd not use it11:03
@wikingsonne|work: i think i know what's the problem on cygwin11:09
@wikingsonne|work: i'll try to fix it...11:09
-!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has joined #shogun11:12
-!- mode/#shogun [+o iglesiasg] by ChanServ11:12
@wikingiglesiasg: you should have now lapack...11:15
-!- gsomix [~gsomix@95.67.166.251] has joined #shogun11:15
@iglesiasgwiking: good, thank you11:16
@iglesiasgwiking: I am working now in another machine to the one I tried yesterday. I will tell you about it in the afternoon/evening11:16
sonne|workgsomix: what reminds me - is the libsvm reader finished? if so please also allow readers for dense vectors/matrices just by using a converter in the end11:16
gsomixsonne|work, very close to finish.11:19
-!- gsomix [~gsomix@95.67.166.251] has quit [Read error: Connection reset by peer]11:21
sonne|workgsomix you might want to check http://sourceforge.net/projects/waffles/files/waffles/2013-04-06/ waffles - it can also read arff files11:23
-!- van51 [~van51@athedsl-224969.home.otenet.gr] has joined #shogun11:24
-!- gsomix [~gsomix@95.67.165.3] has joined #shogun11:33
-!- gsomix [~gsomix@95.67.165.3] has quit [Ping timeout: 240 seconds]11:37
-!- gsomix [~gsomix@80.234.59.144] has joined #shogun11:39
gsomixsonne|work, thanks for link.11:42
@wikingsonne|work: btw for havig a build for debian u dont need unit testing enabled.... we would do that anyways based on the master/stable branch12:05
sonne|workwiking: I would prefer to run the unit tests. You know debian has 10 architectures and I have seen the *weirdest* things12:07
@wikingthen i'm sorry12:07
@wikingask gmock ppl to release12:07
lisitsynwiking: you shoulda sign a petition like in postal 212:10
shogun-notifier-shogun: Viktor Gal :develop * 5fc1b59 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/5fc1b595b0312b59e99dc2e1fe07671685bf84b112:20
shogun-notifier-shogun: Fix Atlas detection on FreeBSD12:20
thoralfsonne|work: I'd like to change the sort_features() method in SGSparseVector a bit.  The usage of realloc introduces problems with sparse readers, since it frees the passed reference and creates a new one - but the sparse reader will never notice.12:20
@wikinglambday: let me know if this latest patch broke our atlas detection :P12:23
thoralfsonne|work: So first of all, when using "vec = stream_features->get_vector();" - are we allowed to use this vector like our own or do we need to copy it before touching?12:24
shogun-buildbotbuild #1758 of deb1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1758  blamelist: Viktor Gal <viktor.gal@maeth.com>12:24
lambdaywiking: alright... what changed?12:25
@wikinglambday: something about lapack_atlas detection12:25
lambdaywiking: okay.. I'll let you know12:25
-!- Netsplit *.net <-> *.split quits: naywhayare12:40
-!- Netsplit over, joins: naywhayare12:46
-!- van51 [~van51@athedsl-224969.home.otenet.gr] has quit [Quit: Leaving.]13:05
sonne|workwiking: seems like debian people have 1.7 already installed13:31
sonne|workwiking: but still if you have an idea how to specify a custom gmock/gtest path - that would be sufficient13:32
sonne|workthoralf: I don't get where a realloc is needed?13:36
@wikingsonne|work: no that wouldn't be .... and there's no 1.7 release so it's interesting how can debian have 1.713:37
@wikingsonne|work: i don't know if maaybe gmock has a versionheader.h13:37
@wikingor soemthing like that13:37
sonne|workwiking: http://packages.debian.org/sid/google-mock13:37
@wikingbut basically13:37
@wikingi fucking dont care13:37
@wikingit's working13:38
@wikingi'm sorry13:38
thoralfsonne|work: Acutally it's two issues: The first one is that we're not writing the sorted data into the original reference, so that the original is getting invalid.13:38
@wikingit's just a bit too much now13:38
sonne|workwiking: all I need is an option to specify the location where google mock is13:38
@wikingsonne|work: i canno tdo it13:38
sonne|workhow does one do it then?13:38
thoralfsonne|work: The second is that (introduced by me) we're compressing the sparse vector and then doing a realloc to shrink.13:38
@wikingsonne|work: as i said i wouldn't know how to do it, with the constraints i've told you13:39
sonne|workwiking: why not just do a compile check with some test and see if google mock/test works?13:40
@wikingsonne|work: why not reinvent the wheel?13:40
sonne|workthoralf: the second makes a lot of sense though13:40
@wikingsonne|work: i'm sorry i'm too tired, and i see this work pointless13:40
sonne|workwiking: well we are in all broken state right http://shogun-toolbox.org/buildbot/waterfall ?13:40
@wikingsonne|work: and that's mostly not because of gmock13:41
@wikingsonne|work: apt-get/rpm install subversion13:41
sonne|workwiking: sure also other stuff13:41
@wikingsonne|work: moreover on the rpm machine just build cmake from src13:41
sonne|workon bsd it is cblas not working13:41
sonne|workon cygwin and debian google mock/test13:42
@wikingsonne|work: fixt it13:42
@wikingsonne|work: yes because there's no svn13:42
@wikingsonne|work: but there's like svn for anything nowdays13:42
@wikingshogun-buildbot: force build --branch=develop 'bsd1 - libshogun'13:43
shogun-buildbotbuild forced [ETA 8m25s]13:43
shogun-buildbotI'll give a shout when the build finishes13:43
@wikingsonne|work: i dont think it's unreasonable to have svn on a buildbot13:43
sonne|workthoralf: the realloc when compressing makes sense. but sorting could be done in place13:45
thoralfsonne|work: I should just copy the vector returned by stream_features.13:47
thoralfSince there is no proper way of signaling size changes to the caller.13:47
@iglesiasgsee you later guys, heading to a PhD interview right now, wish me luck!13:48
thoralfEven if we keep the reference - the caller will have a wrong num_entries count.13:48
sonne|workiglesiasg: good luck :)13:48
thoralfiglesiasg: :)13:48
@iglesiasgthanks!13:48
@wikingiglesiasg: gluck13:49
-!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has quit [Quit: Leaving]13:49
@wikingsonne|work: for some reason the bsd bot is not being triggered for rebuild13:49
@wikingi mean after a commit13:49
sonne|workwiking: it is only triggered if the linux libshogun one passes13:50
shogun-buildbotbuild #1411 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/141113:50
@wikingsonne|work: ah i see... ok now 'only' test failed13:51
sonne|workwiking: I don't see which test13:51
@wikingsonne|work: unit... segfaults13:52
@wiking  1/108 Test   #1: unit-test ..........................................................***Exception: SegFault  1.55 sec13:52
@wiking[ RUN      ] LatentModel.argmax_h13:52
@wikingi've tried to see wtf is with that13:52
@wikingbut that seems to be some gmock+os+arch+compiler combo13:53
@wikingerror13:53
sonne|workwiking: how do you find that even? I only see that something failed but not what13:59
@wikingsonne|work: http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1411/steps/test/logs/stdio14:00
@wikingand search for: Start   2: libshogun-basic_minimal14:00
sonne|workwiking: 1/108 Test   #1: unit-test ..........................................................***Exception: SegFault  1.55 sec14:00
@wikingand then you'll see that the unit-test stdout ended there14:00
@wiking[ RUN      ] LatentModel.argmax_h14:00
@wiking-> the segfault happened there14:00
sonne|workI see14:01
sonne|workhard to see14:01
sonne|workI guess you can only do valgrind/gdb to figure out what goes on14:01
@wikingsonne|work: indeed14:01
@wikingsonne|work: if u cmake with -DBUILD_DASHBOARD_REPORTS=ON14:02
@wikingthen u can do14:02
@wikingctest -D ExperimentalMemCheck14:02
@wikingbut maybe we should just by default add this ExperimentalMemCheck option for ctest14:02
@wiking(this will run valgrind for all the tests)14:02
sonne|workwiking: for python modular before we did list a gdb backtrace when things fail14:03
@wikingsonne|work: that's preserved14:03
sonne|workwiking: could we conditionally do this only when a test fails?14:04
@wikingsonne|work: dunno14:04
@wikingsonne|work: but i dont think so14:04
@wikingsonne|work: this needs fair amount of ctest/cmack hacking14:04
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]14:24
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun14:24
@wikingoooh14:32
@wikinginteresting14:32
@wikinglisitsyn: r u using gcc for shogun?14:32
sonne|workwiking: I am no cmake expert but with ordinal make you could just touch some file '_failed_tests_exists_' which you remove when all is good and otherwise in the same target run the tests again with valgrind14:33
lisitsynwiking: yes usually14:33
@wikinglisitsyn: mmm i think valgrind + openmp + unit test does some funky shit :)))14:35
@wikingin MultidimensionaScalingTest.distance_preserving14:35
@wikingthe process goes to 'usem' STATE14:35
@wikingand doesn't move until i kill it14:35
lisitsynwiking: hmm weird14:39
lisitsynwiking: well openmp leaks in valgrind btw14:39
lisitsynbut no it should not go to some hanging state14:39
@wikingit is 4 some reason14:41
lisitsynwiking: what's the version?14:44
lisitsynof gcc14:44
@wikinggcc 4714:46
@wikingseems like a race condition14:46
lisitsynwiking: can you paste it anywhere?15:19
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]15:20
-!- van51 [~van51@athedsl-224969.home.otenet.gr] has joined #shogun16:14
-!- iglesiasg [~Fernando@2001:6b0:1:1041:517b:fd15:3056:85e8] has joined #shogun16:15
-!- mode/#shogun [+o iglesiasg] by ChanServ16:15
@wikinglisitsyn: where is the code?16:18
@wikinglisitsyn: i mean, do you need the precompiled code?16:18
lisitsynwiking: ?? I mean the output of helgrind or whatever you use16:24
@wikinglisitsyn: it doesn't say anything16:25
@wikingit's gets stuck there16:25
lisitsynoh okay16:25
@wikingi'm retesting it now16:26
@wikinglisitsyn: if ran it w/o valgrind all is good :S16:30
-!- iglesiasg [~Fernando@2001:6b0:1:1041:517b:fd15:3056:85e8] has quit [Quit: Leaving]16:38
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]16:38
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun16:38
@wikingso i dont get it why valgrind kills it16:54
lisitsynwiking: it might be openmp issue16:55
lisitsynopenmp is not very friendly with valgrind for some reason16:56
-!- foulwall [~user@2001:da8:215:c252:48a7:bec4:9387:251a] has joined #shogun16:59
@sonney2kvan51, the only thing to make the code smaller would be macros right?17:20
van51sonney2k: I guess, what do you not like?17:21
@sonney2kvan51, the other thing I don't understand is the CSGObject stuff is not yet converted right?17:22
van51sonney2k: can you elaborate a bit?17:23
@sonney2kvan51, only the if (m_refcount==NULL)17:23
@sonney2k+    return -1;17:23
@sonney2k+17:23
@sonney2k+  return m_refcount->ref();17:23
@sonney2kit could be 1 line doing the check with a macro17:23
@sonney2kvan51, well CSGObject also has a refcount internally17:24
@sonney2kvan51, and it is doing the exact same thing ref/unref/ref_count17:24
@sonney2kvan51, so it would be natural to just have a single implementation17:24
van51sonney2k: ah I see.17:25
@sonney2kvan51, same with unref() etc17:25
van51sonney2k: should I try to have it use the RefCount class?17:25
@sonney2kvan51, yes you can just use it17:25
@sonney2kvan51, and I think you could for ref have a static function in Refcount17:26
@sonney2kthat gets as argument a RefCount*17:26
@sonney2kwhich then does ref()17:26
@sonney2kso int ref_static(RefCount*)17:26
@sonney2kvan51, and you could in ref() { return ref_static(m_refcoutn))17:27
@sonney2kvan51, OK?17:27
van51sonney2k: that's the SGObject ref?17:27
@sonney2kvan51, no17:27
@sonney2kvan51, well these are 2 things17:27
@sonney2k1) reduce code in ref() / ...17:28
@sonney2k2) use RefCount in CSGObject too17:28
van51sonney2k: ok17:29
@sonney2kfor 1) you can use the static function17:30
@sonney2k2) would just be the same you did already in SGIO17:30
@sonney2kvan51, ok then. this really simplifies the code a lot...17:32
van51sonney2k: ok :)  I have one more question, that I commented on the code17:33
@sonney2kvan51, well what error?17:37
@sonney2kvan51, you need to include lib/common.h whenever you use int32_t17:37
@sonney2kvan51, you can do VERBOSE=1 make to see the real command17:37
@sonney2kvan51, have to leave the train now...17:38
van51sonney2k: ok17:38
van51sonney2k: i'll keep working on it17:38
-!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun17:49
@wikinglisitsyn: yeah most probably it's something like that18:15
lambdaywiking: the new patch breaks lapack detection on my machine :(18:20
lambdayA library with LAPACK API found. -- Could NOT find ATLAS (missing:  ATLAS_LIBRARIES)  -- Could NOT find CBLAS (missing:  CBLAS_LIBRARY)18:20
lambdaywiking: I didn't test that earlier because I was working on something that requires lapack :D... just did a rebase and got this error18:21
-!- foulwall [~user@2001:da8:215:c252:48a7:bec4:9387:251a] has quit [Remote host closed the connection]18:23
@wikinglambday: ARGS :(18:26
lambday:(18:27
@wikinglambday: is there something about clapack_dpotrf maybe?18:27
@wikingin the cmake output?18:27
lambdayI didn't make after that18:27
lambdaychecking18:27
@wikingnono you dont need make18:28
@wikingjust cmake18:28
lambdayjust cmake didn't give anything particular18:30
@wiking:((18:30
@wikingcan u plz just try this18:30
@wikingin shogun's root: mkdir build_atlas && cd build_atlas18:31
@wikingand there rerun your full cmake command18:31
lambdaywiking: okay trying18:31
lambdaywiking: same18:32
lambday:(18:32
@wikingthere's no line at all18:32
@wikingabout clapack_dpotrf18:32
@wiking?18:32
lambdaywiking: nope18:33
lambday A library with LAPACK API found. -- Could NOT find ATLAS (missing:  ATLAS_LIBRARIES)  -- Could NOT find CBLAS (missing:  CBLAS_LIBRARY)18:33
lambdaythis is all that is18:33
lambday:(18:33
@wikinglambday: if so could you please send me this file: build_atlas/CMakeFiles/CMakeError.log18:33
lambdaywiking: mailing you18:34
@wikingthnx18:34
lambdaywiking: sent18:36
lambdayI am leaving for dinner... will be back in an hour or so18:36
@wikinglambday: hahah18:37
@wikinglambday: cc1plus: error: unrecognized command line option '-std=c++11'18:37
lambdaywiking: -std=c++0x should work then?18:38
@wikingi'm afraid not18:38
lambdaywiking:18:39
@wikinglambday: r u sure that this was generated after the first run of your cmake?18:39
lambday Performing Test _HAS_CXX11_FLAG -- Performing Test _HAS_CXX11_FLAG - Failed -- Performing Test _HAS_CXX0X_FLAG -- Performing Test _HAS_CXX0X_FLAG - Success18:39
lambdayyes18:39
@wikingas from this there's no way i can find out anything18:40
lambdaywiking: I just created this build dir before doing cmake :(18:40
lambdaywhen you said18:40
lambdaywait I am removing and creating another one18:40
@wikingCMakeFiles/CMakeOutput.log18:42
@wikingsend me this plz18:42
lambdaywiking: oh okay sending18:43
@wikingdo a grep -r clapack_dpotrf in your build dir18:43
@wikingas i really cannot understand how come it's not written anywhere18:43
lambdaywiking: nope.. okay finally trying git clean and then cmake18:48
@wikingthere's something weird18:48
@wikingthis has to be there18:48
@wikingyeah try a git reset --hard18:48
@wikinggit pull18:48
@wikingand then send again those 2 files18:48
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun18:49
shogun-notifier-shogun: Viktor Gal :develop * 48f5a85 / / (5 files): https://github.com/shogun-toolbox/shogun/commit/48f5a85e4a9ab74430e0362fa4b0a5c44e5c7e0e18:49
shogun-notifier-shogun: Add options for malloc replacement18:49
@wikinglambday: there should be actually 2 lines18:50
@wikingwhen u run cmake18:50
@wiking-- A library with LAPACK API found.18:51
@wiking-- Looking for clapack_dpotrf in /usr/lib/liblapack_atlas.so18:51
@wiking-- Looking for clapack_dpotrf in /usr/lib/liblapack_atlas.so - found18:51
@wiking-- Found ATLAS: /usr/lib/liblapack_atlas.so;/usr/lib/libcblas.so;/usr/lib/atlas-base/libf77blas.so;/usr/lib/libatlas.so;/usr/lib/liblapack.so18:51
@wikingsomething like this18:51
@wikingi dont see any tests on your machine for clapack_dpotrf18:51
lambday-- A library with LAPACK API found. -- Could NOT find ATLAS (missing:  ATLAS_LIBRARIES)  -- Could NOT find CBLAS (missing:  CBLAS_LIBRARY)  -- Found GLPK: /usr/lib64/libglpk.so18:51
@wikingwhat's your HEAD commit version18:51
lambdaywiking: even after git clean and git reset and git pull18:51
@wikingwhat's the last commit version in your git18:52
@wikinggit log18:52
lambday624a683259c8c74c5c54b753dce25cabad3b880f18:52
@wikinghe one before that18:52
@wiking?18:52
@wikingbecause i guess that's your PR commit18:52
lambdaythat's my commit but git log shows your commits18:52
@wikingyes before that what's the last commit hash18:53
lambdaycommit 48f5a85e4a9ab74430e0362fa4b0a5c44e5c7e0e Author: Viktor Gal <viktor.gal@maeth.com> Date:   Tue Aug 13 18:48:51 2013 +0200      Add options for malloc replacement  commit 5fc1b595b0312b59e99dc2e1fe07671685bf84b1 Author: Viktor Gal <viktor.gal@maeth.com> Date:   Tue Aug 13 11:22:19 2013 +0200      Fix Atlas detection on FreeBSD  commit d4baee974e20fd615e6cf58e99b46a5cd28155ec Author: Viktor Gal <viktor.gal@maeth.com> Date:   Tue18:53
@wikingmmm i have nooo idea how's at all this possible18:53
shogun-buildbotbuild #1759 of deb1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1759  blamelist: Viktor Gal <viktor.gal@maeth.com>18:53
@wikingdo you have liblapack_atlas.so.318:53
@wiking?18:53
@wikinglike /usr/lib64/atlas/liblapack_atlas.so.318:53
@wiking?18:53
lambdaywiking: no its all liblapack.so.3 only... in /usr/lib64/ and in /usr/lib64/atlas/18:54
@wikinglambday: ok show me the output18:54
@wikinglambday: of ls -la /usr/lib64/liblapack.so.318:55
@wikingand18:55
@wikingls -la /usr/lib64/atlas/liblapack.so.318:55
lambdaylrwxrwxrwx 1 root root 18 Aug 11 16:12 /usr/lib64/liblapack.so.3 -> liblapack.so.3.3.118:55
lambdaylrwxrwxrwx. 1 root root 18 Mar 28  2012 /usr/lib64/atlas/liblapack.so.3 -> ./liblapack.so.3.018:55
@wikingls -la /usr/lib64/atlas/liblapack.so.3.018:56
@wikingand as well18:56
@wikingls -la /usr/lib64/liblapack.so.3.3.118:56
lambday-rwxrwxr-x. 1 root root 5566152 Jun 16  2011 /usr/lib64/atlas/liblapack.so.3.018:56
@wikingand btw: all motherfucking packagers of all distros should fucking rot in hell18:56
lambday-rwxrwxr-x 1 root root 5502720 Jun  3  2011 /usr/lib64/liblapack.so.3.3.118:56
@wikingooookeeey18:56
lambdayhehe18:57
@wikinglet me do u the fix18:57
lambdayI should really switch to ubuntu it seems :-/18:58
@wikingnono18:59
@wikingit's great that u use fedora18:59
@wikingat least somebody is testing it for that18:59
lambdaywiking: yes I was just going to say that.. I complain about everything :D18:59
lambdaywiking: I will be back in an hour or so..19:00
@wikingok19:01
@wikingcan u test this before19:01
@wikingplz19:01
@wikinglambday: ping19:01
lambdaywiking: okay19:01
lambdaywiking: yes still here19:01
@wikingso19:01
lambdaywiking: sure..19:01
@wikinghttp://pastebin.com/Y3Fqmr4B19:01
@wikingjust paste this into your cmake/FindAtlas.cmake19:01
@wikingbetween line 21 and 2219:02
lambdaywiking: alright19:02
@wikingline 21 should be now:   unset(CMAKE_REQUIRED_LIBRARIES CACHE)19:02
@wikingand line 22 is endif()19:02
@wikingright?19:02
lambdaywiking: yes19:03
lambdaycurrently19:03
lambdaypasting19:03
@wikingokok19:03
@wikingand then rerun cmake19:03
@wikingthere has to be something now about clapack_dpotrf19:03
lambdaywiking:19:05
lambdayA library with LAPACK API found. -- Looking for clapack_dpotrf in /usr/lib64/atlas/liblapack.so.3 -- Looking for clapack_dpotrf in /usr/lib64/atlas/liblapack.so.3 - found -- Found ATLAS: /usr/lib64/atlas/liblapack.so.3;/usr/lib64/atlas/libcblas.so.3;/usr/lib64/atlas/libf77blas.so.3;/usr/lib64/atlas/libatlas.so.3;/usr/lib64/liblapack.so.319:05
lambdayyohoooo19:05
@wikingyes19:05
@wikingokey19:05
@wiking:D19:05
lambday:D19:05
@wikinglet's hope that now everything compiles! :)19:06
lambdayyeah19:06
shogun-notifier-shogun: Viktor Gal :develop * 2466c06 / cmake/FindAtlas.cmake: https://github.com/shogun-toolbox/shogun/commit/2466c06829f044533999f1486a3f05761f618c4519:06
shogun-notifier-shogun: Fix atlas lapack detection on Fedora19:06
@wikinghehe19:06
lambdaylol19:06
lambdaywill check once again before I leave19:07
shogun-notifier-shogun: Soeren Sonnenburg :develop * 6f81260 / examples/undocumented/python_modular/ (3 files): https://github.com/shogun-toolbox/shogun/commit/6f812602e3a8efaa23d4da5c14c5308b3ce25fae19:07
shogun-notifier-shogun: further CSVFile / modshogun conversion19:07
@sonney2kvan51, what did VERBOSE=1 make say?19:08
shogun-buildbotbuild #1760 of deb1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1760  blamelist: Viktor Gal <viktor.gal@maeth.com>19:08
@wikingmmm we need a fedora buildbot19:08
lambdaythere is a rpm one right?19:09
van51sonney2k: something with ccache.. I removed some stuff from my .bashrc and seems to be ok now19:09
van51sonney2k: I'm now recompiling atm19:09
van51sonney2k: btw I looked again at the code19:09
lambdaywiking: pull-ed and cmake worked.. now making19:09
@wikingvan51: -DENABLE_CCACHE=OFF19:09
@wikingin worst case19:09
van51sonney2k: I don't think that static method is necessary19:09
@sonney2kvan51, how else could you reduce the code?19:10
van51wiking: thanks :) I keep de-activating from ccmak19:10
@sonney2kI have never seen ccache to fail though19:10
shogun-buildbotbuild #1761 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/176119:10
van51sonney2k: well, in SGIO, Version and Parallel there is no way that the ref count will be null19:10
shogun-buildbotbuild #1496 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/1496  blamelist: Soeren Sonnenburg <sonne@debian.org>19:11
van51sonney2k: the check was required in SGReferencedData because it could be initialized with ref_counting=false19:11
@sonney2kvan51, true and in SGObject also not19:11
van51sonney2k: so I'll just remove it19:11
shogun-buildbotbuild #1382 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/1382  blamelist: Soeren Sonnenburg <sonne@debian.org>19:11
shogun-buildbotbuild #1198 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1198  blamelist: Soeren Sonnenburg <sonne@debian.org>19:12
shogun-buildbotbuild #876 of rpm1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/876  blamelist: Soeren Sonnenburg <sonne@debian.org>19:13
@sonney2kvan51, hmmhh I am not sure now19:13
* wiking wonders if malloc/realloc/calloc is replaced, does it mean that he malloc for new operator is replaced as well...?19:13
@sonney2kvan51, IIRC the check was necessary in SGReferencedData too19:14
@sonney2kwiking, we tried overloading new /delete but it is not always possible19:14
van51sonney2k: yea. what I'm saying is that, from what I understand, that's the only place where it's necessary19:14
@wikingsonney2k: but isn't new eventually just calling system's malloc?19:14
@wikingor c++ has it's on memory allocator itself...19:15
@wiking?19:15
@sonney2kvan51, exactly in unref() the refcount could be set to NULL so it is possible19:16
@wikingbtw shouldnt we start adding support for c++11's shared_ptr?19:16
@sonney2kwiking, well we don't 'overload' malloc etc19:16
@sonney2kwiking, well everything is broken right now so lets rather get it stable so we have a release post-gsoc19:17
@wikingsonney2k: ok because jemalloc and hoard actually overloads malloc19:17
@wikingsonney2k: travis is not broken :)19:17
@sonney2kwiking, travis tests just a tiny bit of stuff19:17
@sonney2kthe easy part19:17
@sonney2kno different platforms not all interfaces...19:18
@sonney2kalways same configurations19:18
shogun-buildbotbuild #1412 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1412  blamelist: Soeren Sonnenburg <sonne@debian.org>19:19
@wikingyea i'm trying to find a fedora server19:19
@sonney2kvan51, the big difference though is the unref() function - it calls delete this; and in the SGVector case it just frees data but never the vector itself19:19
@sonney2kwiking, is the redhat one not enough?19:19
@wikingsonney2k: well it's not the same but that redhat is like ancient19:19
@wikingsonney2k: it's like from 199519:19
@wikingwith cmake 2.6.19:19
@wikingx19:19
@sonney2kwiking, yeah well redhat19:20
van51sonney2k: how was that tackled?19:20
@sonney2kvan51, don't quite get it... SGVector is just keeping references to data19:21
van51sonney2k: yea and you said that with the delete it would free the data19:21
@wikingthe last time i used RH was when it was still version 4.219:21
@sonney2kvan51, all assignments etc would just copy references and increase refcounts19:21
@sonney2kvan51, no it will not free the SGVector object but just the data (some virtual free_data function...)19:22
@sonney2kvan51, you keep sgvector objects on stack solely19:23
-!- van51 [~van51@athedsl-224969.home.otenet.gr] has left #shogun ["JOIN #shogun"]19:23
@wikingBiltmore.... with Linux 2.0.30-219:23
@wiking:DDD19:23
-!- van51 [~van51@athedsl-224969.home.otenet.gr] has joined #shogun19:23
van51oops19:23
van51sonney2k: should it be doing something else as well?19:23
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 248 seconds]19:24
@sonney2kvan51, no19:25
@sonney2kvan51, but difference w/ SGObject is that it calls delete this19:25
@sonney2kvan51, with SGVector the object will remain but the refcount is gone19:26
van51sonney2k: aah ok I get that19:27
@sonney2kvan51, so you cannot fully replace the unref function everywhere but you have to call delete this etc when the refcount is <=019:28
@sonney2khmmhh19:28
-!- hushell [~hushell@c-24-21-169-136.hsd1.or.comcast.net] has quit [Ping timeout: 240 seconds]19:28
@sonney2kmaybe there is another race condition19:28
@sonney2ksay 2 objects reach refcount=0 at the same time19:29
@sonney2kthen one gets -1 and the other might get -2 and both will attempt to free data19:30
@sonney2kso I think it should only free when refcount==019:32
van51sonney2k: hmm makes sense19:34
@sonney2ksigh I think we need a stress test for SGVector just doing that - accessing a vector in parallel19:35
van51sonney2k: I could do that19:35
van51sonney2k: however, in SGReferenced data, the starting ref_count is 0 and so if unref is called, the value will be decreased by one and then checked against ==019:36
@sonney2kvan51, I am now wondering if this can be fixed at all with an atomic int - I mean the problem is that m_refcount is freed!19:37
@sonney2kvan51, unref cannot be manually called...19:37
@sonney2kvan51, it is a private function used only in copy constructors19:37
@sonney2kvan51, so only when you do a=b;19:37
@sonney2kyou will get refcount = 119:38
@sonney2kbut true19:38
@sonney2kit should be 1 from the beginning19:38
@sonney2kso when you have SGVector a19:38
@sonney2kyou have refcount =119:38
@sonney2kthen when you do19:38
@sonney2kb=a19:38
@sonney2kyou have 219:38
@sonney2kwhen b is destroyed you get 1 again19:38
@sonney2kand when a is destroyed you get the data freed19:39
van51yea19:39
van51sonney2k: when would m_refcount be freed?19:39
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun19:39
-!- mode/#shogun [+o iglesiasg] by ChanServ19:39
@sonney2kvan51, in the destructor19:40
@iglesiasgwiking, I like the dancing woman19:41
@iglesiasgtech me how to do it sensei19:41
van51sonney2k: well yeah, but it's a class member, not static. If refcounting works as intended it should be destroyed then19:42
van51sonney2k: no?19:42
@wikingiglesiasg: :dancer:19:42
@sonney2kvan51, well unref is only called in destructor19:42
@sonney2kvan51, actaully also in copy constructor when the object is overwritten19:43
@sonney2kas in a=b19:43
@sonney2ka would be unref()'d19:43
@wikingwoah man19:44
@wikingjemalloc is the maaaan19:44
* wiking is now doing some malloc comparisions.... stay tunes ;)19:44
@iglesiasgwiking, what are you doing with it? Let me check logs19:44
@wikingiglesiasg: mmm no i haven't written here anything19:45
@wikingiglesiasg: i've just had an oooold plan to add malloc replacements for shogun19:45
@wikingiglesiasg: but then i went around with this cmake hell19:45
@wikingbut now finally i've managed to get back to it19:45
@iglesiasgit sounds nice19:45
@iglesiasgwiking, in what sense do you think it can improve stuff?19:45
@iglesiasgreading jemalloc page atm19:45
@wikingiglesiasg: what's your running time for libshogun-modelselection_grid_search_kernel19:46
@wikingiglesiasg: ctest -R libshogun-modelselection_grid_search_kernel19:46
@iglesiasgwiking, 3.28 seconds19:46
@sonney2kwiking, well we are still in hell no? We need the buildbot back to work, so static interfaces etc..19:46
@wikingiglesiasg: Debug mode right?19:46
@iglesiasgwiking, yes19:47
@wikingiglesiasg:  43/107 Test #141: libshogun-modelselection_grid_search_kernel ........................   Passed    1.85 sec19:47
@iglesiasgwiking, is that release or debug?19:47
@wikingiglesiasg: i was around 4.x before jemalloc19:47
@wikingiglesiasg: debug19:47
@iglesiasgnice!19:47
@wikingiglesiasg: i'll do a comparison now19:47
@wikingiglesiasg: good thing that ctest has xml output19:47
@wikingso i can fast process the running times ;)19:47
@iglesiasg:)19:47
@wikingiglesiasg: but i'll use Release19:48
@iglesiasgwiking, does it entail big dependency to use jemalloc?19:48
@sonney2kvan51, only problem is that SGObject for sure needs a 0 as initial refcount since on return it will be increased by swig19:48
van51sonney2k: hmm19:52
@wikingiglesiasg: well i just did sudo port install jemalloc19:52
@wikingiglesiasg: afaik there's apt-get for it19:52
van51sonney2k: let me wrap some stuff up first in what I've already started19:53
@wikingiglesiasg: http://packages.ubuntu.com/raring/libjemalloc119:53
@sonney2kvan51, I think the same race condition exists in CSGObject unref() just between m_ref_lock.unlock(); and delete this;19:56
@sonney2kvan51, so I think both have a -- though very unlikely -- race condition19:57
van51sonney2k: maybe I should make the conversion and then we can use a tool like helgrind to check for possible race conditions by using a test like you said?20:01
van51sonney2k: I updated with some changes, have another look when you can20:11
van51sonney2k: I'll start now working on SGObject20:11
lambdayback20:12
lambdaywhy modular interfaces fails on travis?20:12
@wikinglambday: do they?20:12
@wikinglambday: which build you refer to/20:12
-!- lambday [67157f36@gateway/web/freenode/ip.103.21.127.54] has quit [Quit: Page closed]20:12
-!- lambday [67157c36@gateway/web/freenode/ip.103.21.124.54] has joined #shogun20:13
lambdaywiking: https://travis-ci.org/shogun-toolbox/shogun/builds/1016261420:14
lambdayoh no not modular interfaces but allowed features20:14
lambday(what are these by the way?)20:14
lambdayah, pink floyd :D20:15
@wikinglambday: ?20:16
@wikinglambday: you mean *Allowed Failures*20:16
@wikingFAILURES20:16
@wikingright? :)20:16
lambdaywiking: argh! yes yes!20:17
lambdaysorry :(20:17
@wikingthen u understand what are those right?20:17
van51going to grab a bite20:21
@wikingsonney2k: afaik i think i'm doing the examples builds wrong... am i right that the libshogun examples were always built with -O0 and w/o any optimization flags?20:22
@wikingsonney2k: about the buildbots... well most of them need love by svn and cmake20:25
@wikingonce those are fixed i can get on with the various cmake flags...20:25
@wikinghttp://worrydream.com/dbx/20:27
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has joined #shogun20:32
@sonney2kvan51, yeah  let me rather merge yours first and then you can do SGObject20:33
shogun-notifier-shogun: van51 :develop * 6d417dc / src/shogun/ (10 files): https://github.com/shogun-toolbox/shogun/commit/6d417dcb7b5e054a8a937349f02aae6009d56d8120:37
shogun-notifier-shogun: Moved RefCount to own file and used that class for ref counting instead of a plain int in the classes SGIO,Version and Parallel20:37
shogun-notifier-shogun: Soeren Sonnenburg :develop * e323d35 / src/shogun/ (10 files): https://github.com/shogun-toolbox/shogun/commit/e323d35bd2245806bf579becdf8056b6209619f820:37
shogun-notifier-shogun: Merge pull request #1399 from van51/feature/refcount20:37
shogun-notifier-shogun:20:37
shogun-notifier-shogun: Made thread-safe refcounting in SGIO, Version and Parallel20:37
shogun-buildbotbuild #1762 of deb1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1762  blamelist: van51 <vangelis_51@hotmail.com>20:38
-!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has quit [Ping timeout: 246 seconds]20:39
shogun-buildbotbuild #1199 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/119920:44
shogun-buildbotbuild #1763 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/176320:50
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has joined #shogun20:51
shogun-buildbotbuild #1200 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1200  blamelist: Soeren Sonnenburg <sonne@debian.org>20:51
shogun-buildbotbuild #1383 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/1383  blamelist: Soeren Sonnenburg <sonne@debian.org>20:51
shogun-buildbotbuild #1497 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/1497  blamelist: Soeren Sonnenburg <sonne@debian.org>20:52
shogun-buildbotbuild #877 of rpm1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/877  blamelist: Soeren Sonnenburg <sonne@debian.org>20:52
shogun-buildbotbuild #1413 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1413  blamelist: Soeren Sonnenburg <sonne@debian.org>20:55
van51sonney2k: I started a PR for SGObject20:58
shogun-buildbotbuild #1764 of deb1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/176421:00
van51I got to go now21:04
van51I'll continue tonight21:04
van51cya21:04
-!- van51 [~van51@athedsl-224969.home.otenet.gr] has quit [Quit: Leaving.]21:04
@wikingiglesiasg: hoard seems to be terrible for me21:11
@iglesiasgwiking, hoard?21:19
@wikinganother malloc replacement21:28
@wikinganyhow i have the data ... i'll do the processing+plotting21:30
@wikingand you'll see21:30
-!- travis-ci [~travis-ci@ec2-54-225-61-153.compute-1.amazonaws.com] has joined #shogun21:45
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/1016414321:45
-!- travis-ci [~travis-ci@ec2-54-225-61-153.compute-1.amazonaws.com] has left #shogun []21:45
-!- lambday [67157c36@gateway/web/freenode/ip.103.21.124.54] has quit [Ping timeout: 250 seconds]22:03
@wikingand that broke it22:03
* wiking loves py-pandas \o/22:30
-!- lisitsyn1 [~lisitsyn@fb2-lo1.global63.net] has joined #shogun22:43
shogun-notifier-shogun: Roman Votyakov :develop * 2853311 / src/shogun/mathematics/Integration.cpp: https://github.com/shogun-toolbox/shogun/commit/285331141f83daa8e275e6a265331583a546b63d22:44
shogun-notifier-shogun: improve memory usage in quadgk integration method22:44
shogun-notifier-shogun: Heiko Strathmann :develop * f938fc0 / src/shogun/mathematics/Integration.cpp: https://github.com/shogun-toolbox/shogun/commit/f938fc0963f28443ffce5bed6bc5a38e93fe8f7b22:44
shogun-notifier-shogun: Merge pull request #1403 from votjakovr/feature/numerical_integration22:44
shogun-notifier-shogun:22:44
shogun-notifier-shogun: Improve memory usage in quadgk integration method22:44
lisitsyn1wat is going on22:47
lisitsyn1wiking: if we have such a malloc overhead we should go for pooling22:47
-!- travis-ci [~travis-ci@ec2-54-225-61-153.compute-1.amazonaws.com] has joined #shogun22:48
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/1016709722:48
-!- travis-ci [~travis-ci@ec2-54-225-61-153.compute-1.amazonaws.com] has left #shogun []22:48
shogun-notifier-shogun: lambday :develop * a0fb40b / / (48 files): https://github.com/shogun-toolbox/shogun/commit/a0fb40b038a5e4e2a111fb15271837af67280b5b22:50
shogun-notifier-shogun: Linear operators redesigned (log-det)22:50
shogun-notifier-shogun: Heiko Strathmann :develop * 5c3f0b8 / / (48 files): https://github.com/shogun-toolbox/shogun/commit/5c3f0b8e18de6bf0a4c7f30e0edb425bf6257ff822:50
shogun-notifier-shogun: Merge pull request #1404 from lambday/feature/log_determinant22:50
shogun-notifier-shogun:22:50
shogun-notifier-shogun: Linear operators redesigned (log-det)22:50
@wikinglisitsyn1: just a sec..22:50
@wikinglisitsyn1: i'm trying to heat up d3.js22:51
@wikinglisitsyn1: i'll go for boxplots.. but if u have a better idea of representation/comparison let me know22:51
@wikingsonney2k: ^22:51
lisitsyn1no I don't have any clue what is you are doing ;)22:51
shogun-buildbotbuild #1384 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/1384  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>22:52
shogun-buildbotbuild #1498 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/1498  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>22:52
shogun-buildbotbuild #1201 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1201  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>22:52
shogun-buildbotbuild #878 of rpm1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/878  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>22:52
shogun-buildbotbuild #1414 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1414  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>22:53
@wikinglisitsyn1: comparing running time of test with various malloc libraries22:54
lisitsyn1wiking: in logs you said with debug enabled22:54
lisitsyn1right?22:54
lisitsyn1well it doesn't really matter though22:54
@wikinglisitsyn1: no it's now Release22:55
@wikingi've just tested first with debug22:55
@sonney2kwiking, have you seen http://shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1201/steps/configure/logs/stdio22:56
@sonney2kCMake Error at src/shogun/CMakeLists.txt:59 (INSTALL):22:56
@sonney2k  install Library TARGETS given no DESTINATION!22:56
@wikingsonney2k: no not yet22:58
@wikingooooh yeah i rememberz22:58
@wikingw822:58
shogun-buildbotbuild #1202 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1202  blamelist: lambday <heavensdevil6909@gmail.com>22:58
@sonney2kwiking, and could you please add or tell me the flags to compile all modular interfaces?22:59
shogun-buildbotbuild #879 of rpm1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/879  blamelist: lambday <heavensdevil6909@gmail.com>22:59
@wikingsonney2k: see .travis.yml22:59
@wikingsonney2k: -DPythonModular=ON -DRubyModular=ON -DLuaModular=ON -DRModular=ON -DOctaveModular=ON -DCSharpModular=ON -DJavaModular=ON23:00
@wikinghave i missed any?23:00
shogun-buildbotbuild #1385 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/1385  blamelist: lambday <heavensdevil6909@gmail.com>23:02
shogun-buildbotbuild #1203 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1203  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>23:02
shogun-buildbotbuild #1499 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/1499  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, lambday <heavensdevil6909@gmail.com>23:02
shogun-buildbotbuild #1386 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/1386  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>23:03
shogun-buildbotbuild #880 of rpm1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/880  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>23:03
shogun-buildbotbuild #1415 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1415  blamelist: lambday <heavensdevil6909@gmail.com>23:04
shogun-buildbotbuild #1416 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1416  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>23:08
@sonney2kwiking, can I just compile for one interface at a time?23:13
@sonney2kotherwise I don't see in the buildbot which interfaces fails but only the end result23:13
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]23:17
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun23:17
@wikingsonney2k: yes sure23:18
@wikingsonney2k: just give it the one -D that u wanna compile23:18
shogun-buildbotbuild #1500 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/150023:18
@wikingas in case of travis23:18
shogun-notifier-shogun: Soeren Sonnenburg :develop * 75f8cf1 / src/shogun/lib/ (3 files): https://github.com/shogun-toolbox/shogun/commit/75f8cf1ab342ee8fad9df862f9fa35f42d92c4f723:23
shogun-notifier-shogun: use forward declaration for RefCount and fix name clash ref_count() vs ref_count23:23
@sonney2kwiking, but then I have to configure for each setting again right? I cannot do them all in one go but just have a different target for each interface I can make <target>?23:25
@wikingdoh there's no out of box boxplot by d3js23:25
@wikingsonney2k: you can compile the all the interfaces at once23:25
@wikingbtw "buildbot which interfaces fails but only the end result" why?23:26
@wikingyou mean which test fails? or which compile fails?23:26
@sonney2kwiking, look at the master.cfg23:26
@sonney2kwiking, I had one Compile step per interface23:26
@sonney2kbut only one configure step23:26
@wikingmmmm23:26
@wikingyes the compile step will not be separate23:27
@sonney2kleaving that aside I have trouble with this now http://shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1500/steps/configure/logs/stdio23:27
@wikingthe test step can be made different23:27
@sonney2kCould NOT find JNI (missing: JAVA_AWT_LIBRARY JAVA_JVM_LIBRARY23:27
@sonney2kwiking, not the target for the compile?23:27
@wikingmmm why is there no jni :(23:27
@sonney2kI mean have sth like targets: all my interfaces?23:27
@wikingsonney2k: well23:28
@sonney2kit used to work with configure but maybe it is this JAVA_AWT_INCLUDE_PATH?23:28
@sonney2kwhat is this?23:28
@wikingsonney2k: actually there is a target23:28
@wikingsonney2k: make shogun23:28
@wikingthat is the libshogun23:28
@wikingmake python_modular23:28
@wikingmake <interface>_modular23:28
@wikingshould work23:28
@sonney2kwiking, exactly what I want23:28
@wikingas they are custom targets23:28
@wikingso it should work i think23:28
@wikingof course if you just type23:28
@wikingmake23:28
@wikingthen i'll do make all23:29
@wikingand that'll build u everything23:29
@wikingmmm where's your java?23:29
@wikingstandard openjdk by debian?23:30
@wikingor some funky shit?23:30
@sonney2kwiking, yes standard openjdk23:30
@wikingmmm neverseen this before23:31
@wikingheheh23:31
@wikinghttp://bugs.debian.org/cgi-bin/bugreport.cgi?bug=71829023:32
@wikingmaybe u have the same problem?23:32
shogun-buildbotbuild #1765 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/176523:33
shogun-buildbotbuild #1387 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/1387  blamelist: Roman Votyakov <votjakovr@gmail.com>23:33
@sonney2kwiking it is make _python_modular23:33
@sonney2kwiking, but it worked with configure?23:33
@wikingsonney2k: yes for pyton23:33
@wikingsonney2k: for python23:33
@wikingother interfaces are <interface>_modular23:33
@wikingsonney2k: swig+cmake has some funky thing going on with python23:34
@sonney2kwiking, no _<interface>_modular23:34
@wikingsonney2k: export JAVA_HOME=/23:34
shogun-buildbotbuild #1501 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/1501  blamelist: Roman Votyakov <votjakovr@gmail.com>23:34
@sonney2kahh23:34
@wikingsonney2k:23:34
@wikingexport JAVA_HOME=23:34
@wikingtry that23:34
shogun-buildbotbuild #1204 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1204  blamelist: Roman Votyakov <votjakovr@gmail.com>23:34
@wikingi mean the right path23:34
shogun-buildbotbuild #881 of rpm1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/881  blamelist: Roman Votyakov <votjakovr@gmail.com>23:35
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]23:37
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun23:37
shogun-buildbotbuild #1502 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/150223:37
@sonney2kwiking, you mean /usr/lib/jvm/java-6-openjdk-amd64 ?23:38
shogun-buildbotbuild #1769 of deb1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1769  blamelist: Soeren Sonnenburg <sonne@debian.org>23:38
@sonney2kwiking, / worked but now ruby fails with narray23:39
@sonney2kCMake Error at /usr/share/cmake-2.8/Modules/FindPackageHandleStandardArgs.cmake:97 (MESSAGE):23:39
@sonney2k  Could NOT find RubyNArray (missing: RUBY_NARRAY_INCLUDE_DIR23:39
@sonney2k  RUBY_NARRAY_LIBRARY)23:39
@sonney2kCall Stack (most recent call first):23:39
@sonney2k  /usr/share/cmake-2.8/Modules/FindPackageHandleStandardArgs.cmake:288 (_FPHSA_FAILURE_MESSAGE)23:39
@sonney2k  cmake/FindRubyNArray.cmake:42 (FIND_PACKAGE_HANDLE_STANDARD_ARGS)23:39
shogun-buildbotbuild #1417 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1417  blamelist: Roman Votyakov <votjakovr@gmail.com>23:39
@sonney2k  CMakeLists.txt:586 (FIND_PACKAGE)23:39
@wikingsonney2k: where's your narray23:41
@wikingsonney2k: i've been dealing with that quite some23:41
@sonney2kwiking, please ssh into the buildbot and check23:42
@wikingok i see23:43
@wikingwhere's the cmake?23:43
@wikingit's not under/var/lib/buildbot/slaves23:44
@wikingor it's not that machine actually... i guess23:44
@wikingsonney2k: can u send me the build/CMakeFiles/CMakeError.log plz23:45
shogun-notifier-shogun: Soeren Sonnenburg :develop * 32535dc / src/shogun/ (6 files): https://github.com/shogun-toolbox/shogun/commit/32535dc7bebd38726dd16378ae590021ab66c3de23:47
shogun-notifier-shogun: use forward declarations for RefCount also for other base objects23:47
-!- travis-ci [~travis-ci@ec2-54-225-61-153.compute-1.amazonaws.com] has joined #shogun23:49
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/1017182223:49
-!- travis-ci [~travis-ci@ec2-54-225-61-153.compute-1.amazonaws.com] has left #shogun []23:49
@sonney2kwiking, I have another big issue: when one has shogun installed in /usr/local/ with includes - cmake will take the wrong includes (the one from the system). could we somehow get an -I../src into the flags?23:55
--- Log closed Wed Aug 14 00:00:14 2013