IRC logs of #shogun for Friday, 2013-09-06

« Thursday, 2013-09-05 Index Saturday, 2013-09-07 »
--- Log opened Fri Sep 06 00:00:47 2013
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun00:02
shogun-notifier-shogun: lambday :develop * c54bd1d / src/shogun/mathematics/logdet/ (4 files): https://github.com/shogun-toolbox/shogun/commit/c54bd1d48331bebf818717ec5f83552af353f19700:02
shogun-notifier-shogun: precomputed eigenvalue set added in eigensolver (log-det)00:02
shogun-notifier-shogun: Heiko Strathmann :develop * 7434833 / src/shogun/mathematics/logdet/ (4 files): https://github.com/shogun-toolbox/shogun/commit/74348339decf631de45ba12287cb5a7d2f19044d00:02
shogun-notifier-shogun: Merge pull request #1536 from lambday/feature/log_determinant00:02
shogun-notifier-shogun:00:02
shogun-notifier-shogun: precomputed eigenvalue set added in eigensolver (log-det)00:02
shogun-buildbotbuild #1699 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1699  blamelist: lambday <heavensdevil6909@gmail.com>00:07
shogun-buildbotbuild #88 of osx1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/88  blamelist: lambday <heavensdevil6909@gmail.com>00:12
shogun-buildbotbuild #1616 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1616  blamelist: lambday <heavensdevil6909@gmail.com>00:14
shogun-buildbotbuild #1393 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1393  blamelist: lambday <heavensdevil6909@gmail.com>00:48
shogun-buildbotbuild #1700 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1700  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>00:51
shogun-buildbotbuild #89 of osx1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/89  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>00:53
shogun-buildbotbuild #1617 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1617  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>00:56
-!- travis-ci [~travis-ci@ec2-174-129-54-125.compute-1.amazonaws.com] has joined #shogun01:00
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/1103659901:00
-!- travis-ci [~travis-ci@ec2-174-129-54-125.compute-1.amazonaws.com] has left #shogun []01:00
@wikingyo01:30
@wiking:)01:30
shogun-buildbotbuild #1394 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1394  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>01:33
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 245 seconds]02:30
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]03:02
-!- sonne|osx_ [~sonne@f053045167.adsl.alicedsl.de] has joined #shogun03:31
-!- sonne|osx [~sonne@f053037017.adsl.alicedsl.de] has quit [Ping timeout: 260 seconds]03:33
-!- sonne|osx_ is now known as sonne|osx03:33
-!- foulwall [~zhengyang@114.255.40.12] has quit [Quit: leaving]03:44
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun03:44
shogun-buildbotbuild #515 of nightly_default is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/51504:06
-!- pickle27 [~Kevin@199.119.128.114] has quit [Quit: Leaving]05:48
-!- sonne|osx [~sonne@f053045167.adsl.alicedsl.de] has quit [Quit: sonne|osx]06:39
@sonney2kwiking, you are back :)06:42
-!- besser82|afk is now known as besser8207:48
-!- sonne|osx [~sonne@89.204.153.122] has joined #shogun08:08
sonne|osxwiking: you wanted me to create some subdomain of shogun-toolbox.org right?08:08
sonne|osxwhich one and where should it point to?08:08
sonne|osxwiking: we have now buildbot.s-t.org and fatbot.*08:09
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun08:26
shogun-notifier-shogun: Soeren Sonnenburg :develop * f8d2574 / README.cmake: https://github.com/shogun-toolbox/shogun/commit/f8d2574ad00848dee9357590ad3c8f3aeede11da08:26
shogun-notifier-shogun: mention how to specify swig executable08:26
shogun-notifier-shogun: Soeren Sonnenburg :develop * 03c0257 / examples/undocumented/python_modular/ (8 files): https://github.com/shogun-toolbox/shogun/commit/03c02571f146a1f83c7cd675a1731cfc34da28b408:26
shogun-notifier-shogun: convert more examples to CSVFile08:26
-!- sonne|osx [~sonne@89.204.153.122] has quit [Quit: sonne|osx]08:26
shogun-buildbotbuild #1701 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1701  blamelist: Soeren Sonnenburg <sonne@debian.org>08:48
shogun-buildbotbuild #1618 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1618  blamelist: Soeren Sonnenburg <sonne@debian.org>08:53
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:8cfc:dbdc:38a:9128] has joined #shogun09:06
-!- mode/#shogun [+o iglesiasg] by ChanServ09:06
@iglesiasgwiking, you are back!09:06
@iglesiasghi guys09:06
shogun-buildbotbuild #1702 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1702  blamelist: Soeren Sonnenburg <sonne@debian.org>09:17
-!- travis-ci [~travis-ci@ec2-23-20-125-26.compute-1.amazonaws.com] has joined #shogun09:18
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/1104744009:18
-!- travis-ci [~travis-ci@ec2-23-20-125-26.compute-1.amazonaws.com] has left #shogun []09:18
shogun-buildbotbuild #1619 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1619  blamelist: Soeren Sonnenburg <sonne@debian.org>09:21
shogun-buildbotbuild #1395 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1395  blamelist: Soeren Sonnenburg <sonne@debian.org>09:30
shogun-buildbotbuild #1396 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1396  blamelist: Soeren Sonnenburg <sonne@debian.org>10:12
@iglesiasghey hushell10:13
hushelliglesiasg: hey10:15
hushelliglesiasg: how is going?10:15
@iglesiasghushell, I am good, what about you?10:17
@iglesiasgwe should separate that pull request I think10:17
hushelliglesiasg: good, let's talk here :)10:17
hushelliglesiasg: I close this PR or just remove cmake files and make them a new PR?10:18
@iglesiasghushell, I will answer there your last question hehe10:18
@iglesiasghushell, the second option sounds cool10:18
@iglesiasgremove the CMake part from this pull request10:18
@iglesiasgand do another one from another branch with just the find cmake10:19
@iglesiasgthe find mosek, sorry10:19
hushelliglesiasg: I see how useful the branches :)10:19
hushelliglesiasg: okay, I do it right away10:19
@iglesiasgnice10:19
@iglesiasgthank you10:19
sonne|workiglesiasg: this morning it was around 200MB and sometimes 150MB10:20
sonne|workswitching between releases(!)10:20
sonne|worklets see what we get in the evening10:20
@iglesiasgmmm10:20
@iglesiasgsonne|work,  but is there such a big change between one commit and the next one?10:21
sonne|workiglesiasg: I will have to plot this10:21
@iglesiasggood idea10:22
sonne|workand I will add this to the buildbot10:22
@iglesiasgawesome10:22
sonne|workso that we always have some figure to look at10:22
sonne|workwe could include compile time too btw10:23
shogun-notifier-shogun: van51 :develop * f7bbfa1 / src/shogun/ (7 files): https://github.com/shogun-toolbox/shogun/commit/f7bbfa19c9ed338b6637f8716367f414a4c6fe6010:26
shogun-notifier-shogun: Fixed some documentation warnings10:26
shogun-notifier-shogun: Soeren Sonnenburg :develop * dcfcb7c / src/shogun/ (7 files): https://github.com/shogun-toolbox/shogun/commit/dcfcb7c8aa5748e1dbc1dfecb9ff14eef131d0ea10:26
shogun-notifier-shogun: Merge pull request #1537 from van51/feature/fix_warnings10:26
shogun-notifier-shogun:10:26
shogun-notifier-shogun: Fixed some documentation warnings10:26
shogun-buildbotbuild #1703 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1703  blamelist: van51 <vangelis_51@hotmail.com>10:36
besser82Hello to all! Is it known that currently ~ 300 tests are failing? See: http://ur1.ca/fecpl10:40
@wikingsonne|work: hey10:42
@wikingand morning all10:42
@wikingsonne|work: about the new osx buildbot, should it be called osx2?10:42
@iglesiasgwiking, yoyo! Hope you had a great time in the wedding!10:42
@iglesiasgbesser82, omg 300 is a lot :O10:43
@iglesiasgbesser82, is there an error trace?10:43
besser82unit-tests are segfaulting, btw...10:43
@wikingiglesiasg: hehehe yeah i had :)10:43
besser82iglesiasg: how to produce such trace?10:43
shogun-buildbotbuild #1620 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1620  blamelist: van51 <vangelis_51@hotmail.com>10:44
@iglesiasgbesser82, so you used ctest to run those tests, right?10:44
besser82iglesiasg: yes10:44
@iglesiasgbesser82, there is an option, let me check it10:44
@iglesiasgbesser82, ctest --output-on-failure I think10:45
@wikingiglesiasg: yes it's that one10:45
besser82iglesiasg, will re-run this on current commit && pastebin :)10:45
@iglesiasg wiking, I checked it in the mail you sent ^^10:46
@iglesiasgbesser82, cool, thanks10:46
besser82iglesiasg: may take up ~15 mins10:46
besser82iglesiasg: need some details about the lib-version i'm using?10:46
@iglesiasgbesser82, yes, those could be useful once we get to see the trace probably10:47
@iglesiasgbesser82, anyway, the Shogun version is?10:47
@iglesiasgcurrent develop?10:47
besser82iglesiasg: yes, current develop / libs are mostly take from F19 system10:47
sonne|workbesser82: errm doesn't look like anything fails https://travis-ci.org/shogun-toolbox/shogun/jobs/1105026110:48
sonne|workwiking: sec meeting10:48
@iglesiasgbesser82, F19?10:48
besser82iglesiasg: Fedora 19 ;)10:48
@iglesiasggot it, thanks!10:48
@wikingsonne|work: ok10:48
@wikingiglesiasg: heheh we need to install fedora on the new machine as i suspected ;P10:49
besser82sonne|work: those are only ~108, when I run testsuite there are ~570...10:49
@iglesiasghushell, I will have a look to the non-cmake PR ASAP, it is still sort of big so I will need to put apart fome minutes for it, probably this evening!10:49
sonne|workwiking: besser82 is packaging shogun for fedora10:51
sonne|workjust FYI10:52
@wikingbesser82: oh great, how do you want to do the packaging for fedora?10:53
besser82wiking: for the official fedora repo ;)10:54
hushelliglesiasg: no worries, I am going to sleep now, see you tomorrow10:54
@wikingbesser82: yes but what do you want to use?10:54
@iglesiasghushell, bye bye, good night10:54
hushellbye guys :)10:55
besser82wiking: kk, got it :)   Mostly the libs provided by system.  Already finished some pkgs for ColPack and nlopt.10:55
@wikingbesser82: no you don't get me10:55
@wikingbesser82: how do you want to create the packages of shogun for fedora?10:56
@wikingbesser82: some scripts or how?10:56
besser82wiking: rpm-spec10:56
shogun-notifier-shogun: Viktor Gal :develop * 85b5000 / cmake/external/ColPack.cmake: https://github.com/shogun-toolbox/shogun/commit/85b5000ed2ce4d0e04dbc89c4c4e2bd556e4567110:56
shogun-notifier-shogun: Little fix for external ColPack.cmake issue #152910:56
-!- iglesiasg_ [~iglesias@n139-p12.kthopen.kth.se] has joined #shogun10:56
@wikingbesser82: ok.... do you know cpack?10:57
besser82wiking: sure, but won't fit fedora's packaging guidelines ;)10:57
@wikingbesser82: why not?10:57
sonne|workwiking: rules10:57
@wikingi mean it just uses rpm-spec afaik10:57
@wikingsonne|work: yes i know10:57
sonne|workwiking: same with .deb's created by this10:58
@wikingbut you can do all the things afaik according to the riules10:58
@wikingrules10:58
@wikingi mean it's just a script language10:58
besser82wiking: yes, but it's not integrated in the build-system used by most distros ;)10:58
@wikingsonne|work: ok so what should be with the osx bot? as i'll setup there the static analyzer as it's far more updated than the one coming with ubuntu10:58
sonne|workwiking: what osx bot?10:59
sonne|workyou have one now?10:59
sonne|workfrom gunnar or VM or ...?10:59
@wikingbesser82: ok so basically you are saying that we'll need to have some extra files in our repo for supporting fedora packaging?10:59
@wikingsonne|work: VM... i'm just about to set it up10:59
besser82wiking: not really10:59
sonne|workwiking: nice10:59
@wikingsonne|work: i don't know anything about gunnar's11:00
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:8cfc:dbdc:38a:9128] has quit [Ping timeout: 245 seconds]11:00
@wikingsonne|work: so should it be osx2?11:00
sonne|workwiking: yea gunnar == /dev/null11:00
@wikingsonne|work: lol ok11:00
sonne|workwiking: yes osx211:00
@wikingsonne|work: do u control now shogun-toolbox.org?11:00
besser82wiking: but if you want I can create some rpm-template for use with cpack ;)11:00
@wikingbesser82: no i mean how one will be able to build the packages after?11:00
sonne|workwiking: not really but I can create a few subdomains11:00
sonne|workwiking: if we still want one ...11:00
-!- iglesiasg_ [~iglesias@n139-p12.kthopen.kth.se] has quit [Read error: Connection reset by peer]11:01
@wikingsonne|work: ok i'll think about it...11:01
@wikingbesser82: so basically what i'm saying is that we ourselves would like to build fedora packages nightly11:01
@wikingbesser82: how can we do that...?11:01
besser82wiking: You'll need some spec-template and src-tarball11:02
besser82wiking: build a srpm from that and pass it to some build-worker ;)11:03
-!- HeikoS [~heiko@nat-160-163.internal.eduroam.ucl.ac.uk] has joined #shogun11:03
-!- mode/#shogun [+o HeikoS] by ChanServ11:03
besser82wiking: I can provide you with access to fedora-build workers ;)11:03
@wikingbesser82: ok then we'll need to have those in our repo if we want to let other people as well to build their own rpm, right?11:03
@wikingbesser82: that sounds great11:03
shogun-buildbotbuild #1704 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1704  blamelist: Viktor Gal <viktor.gal@maeth.com>11:03
@HeikoSwiking: is back! :)11:03
@wikingHeikoS: indeed11:03
@wikingHeikoS: comment and close :)11:03
@HeikoShi, how are things?11:03
@HeikoShad a good time?11:04
@wikingmmm good back in shogun business :P aiming for 3.011:04
@wikingyeah it was really cool11:04
@HeikoSyeah :)11:04
@wikingspent the whole week in cote d'azur11:04
besser82wiking: there's a way to archive this by using cpack and some extra make-target ;)11:04
@HeikoSwiking: ah sweet!11:04
@wikingbesser82: that's alright11:04
besser82wiking: so people can build everything needed from sources11:04
@wikingbesser82: yes i got ya11:05
besser82wiking: like generating spec from template and such ;)11:05
@wikingbesser82: afaik we'll not add this directly into shogun's repo but rather a submodule...11:05
@wikingbut still i'd like to provide this11:05
@wikingfor everybody11:05
@wikinganyhow11:05
@wikingbesser82: i'll create now a fedora stable buildbot11:05
besser82wiking: would be just one more ascii-file in repo ;)11:05
besser82wiking: sounds good ;)11:06
@wikingbesser82: and we'll see what's happening when u try to build11:06
@wikingbesser82: what do you reccommend for a buildbot11:06
@wikingbesser82: i mean which fedora... ? is there like fedora-server?11:06
besser82wiking: Fedoa 19 would be good ;)11:06
besser82wiking: not really11:06
besser82wiking: you will just need minimal11:07
besser82wiking: btw. do you have some other redhat based stuff running?11:07
besser82wiking: ...brb...11:07
shogun-buildbotbuild #1621 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1621  blamelist: Viktor Gal <viktor.gal@maeth.com>11:08
@wikingbesser82: yeah this bot: http://buildbot.shogun-toolbox.org/builders/rpm1%20-%20libshogun11:08
@wikingbesser82: i have no idea what's this distro ....11:08
sonne|workwiking: old11:09
sonne|workI guess 6.311:09
sonne|workredhat11:09
sonne|workHeikoS: did you see the email from sergey?11:09
@HeikoSsonne|work: yes :(11:10
@HeikoSI answered11:10
sonne|workme too11:10
@wikingbesser82: i cannot find fedora minimal? where's that :)11:10
@HeikoSwiking: btw the PBS/SGE backend I wrote works pretty smooth, I will soon clean up the code a bit and put it online11:11
@wikingHeikoS: \o/11:12
@wikingthat's great!11:12
besser82wiking: take the fedora-install dvd ;)11:13
@HeikoSafk for 10 be back soon11:13
besser82wiking: and install the "minimal"-config ;)11:13
@wikingbesser82: ok11:14
@wikingbesser82: i'm on it11:14
besser82wiking: when finished, ping, plz. I can provide you with further infomation, then ;)11:14
@wikingbesser82: do you know how to install fedora in a chroot?11:15
besser82wiking: yes, sure11:15
besser82wiking: you'll need yum11:16
@wikingbesser82: no i have debian ;P11:16
besser82wiking: kk, shall I give you some F19 chroot tarball, then?11:16
shogun-buildbotbuild #1705 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1705  blamelist: Soeren Sonnenburg <sonne@debian.org>11:16
@wikingsonne|work: heheh this machine is great. downloading f19 iso with 11.1mb/sec11:16
@wikingbesser82: would be aaawwwwwwweeeesome11:17
sonne|workwiking: we only have 10MB/s11:17
sonne|workbut steady11:17
besser82wiking: kk, just minute ;)11:17
besser82wiking: need amd64 or i686 ?11:17
@wikingbesser82: amd6411:17
besser82wiking: allright on the run ;)11:17
shogun-buildbotbuild #1397 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1397  blamelist: van51 <vangelis_51@hotmail.com>11:19
shogun-buildbotbuild #1622 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1622  blamelist: Soeren Sonnenburg <sonne@debian.org>11:22
-!- travis-ci [~travis-ci@ec2-23-20-125-26.compute-1.amazonaws.com] has joined #shogun11:23
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/1105025911:23
-!- travis-ci [~travis-ci@ec2-23-20-125-26.compute-1.amazonaws.com] has left #shogun []11:23
@wikingsonne|work: kmod_module_new_from_loaded: could not open /proc/modules: No such file or directory11:26
@wikingError: could not get list of modules: No such file or directory11:26
@wikingsonne|work: is it intentional that we dont have /proc/modules?11:26
sonne|workwiking: you have to mount proc etc11:26
@wikingor it's just not set up yet?11:26
@wikingokok11:27
@wikingi'll do it then11:27
sonne|workwiking: you are in a chroot now right?11:27
@wikingnot yet11:27
sonne|workwiking: then mount proc and sys11:27
@wikingi'm doing11:27
@wikingvirtualbox now11:27
sonne|workwhy that?11:27
@wikingsonne|work: osx211:27
sonne|workahh ok11:27
@wikinglol wtf?11:28
@wikinghow is it possible that we dont have /proc/modules?11:28
@wikingi mean we have /proc11:29
@wikingproc on /proc type proc (rw,nosuid,nodev,noexec,relatime)11:29
@wikingsonne|work: isn't this some security stuff?11:29
sonne|workwiking: sure11:29
besser82wiking: http://besser82.fedorapeople.org/fedora-19-x86_64.tar.xz  is ~46 MBytes11:29
@wikingsonne|work: ideas for fix?11:30
@wikingbesser82: cool downloading11:30
@wikingbesser82: got it... instructions?11:30
besser82wiking: untar to some dir, mount /proc and such, go ;)11:31
@wikingbesser82: ok cool.... i'm on it11:31
@wikingsonne|work: aaah fucking grs11:31
besser82wiking: before you install anything with yum, you should `yum makecache` ;)11:32
besser82wiking: i'd recommend `yum group install 'C Development Tools and Libraries'` to get basic build-env ;)11:33
@wikingbesser82: oh great11:34
@HeikoSwiking: can I specify a custom g++ binary with cmake?11:34
@wikingCXX=whateveryouwant cmake ../11:34
besser82HeikoS: you can!11:34
@HeikoSsweet thanks!11:34
sonne|workHeikoS: please put that in README.cmake immediately!11:35
@HeikoSsonne|work: will do once it works11:35
@wikingsonne|work: these are simple cmake hacks... so rather just add a link to cmake FAQ11:36
@wikinghttp://www.cmake.org/Wiki/CMake_FAQ11:36
sonne|workHeikoS: and add a link to the cmake FAQ too in that file11:38
@HeikoSsorry to annoy you, but I get11:38
@HeikoS[ucabhst@splash build]$ /usr/bin/g++44 --version11:38
@HeikoSg++44 (GCC) 4.4.0 20090514 (Red Hat 4.4.0-6)11:38
@HeikoSCopyright (C) 2009 Free Software Foundation, Inc.11:38
@HeikoSThis is free software; see the source for copying conditions.  There is NO11:38
@HeikoSwarranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.11:38
@HeikoS[ucabhst@splash build]$ CC=/usr/bin/gcc44 CXX=/usr/bin/g++44 cmake ..11:38
@HeikoS-- Could NOT find CCache (missing:  CCACHE)11:38
@HeikoSCMake Error at CMakeLists.txt:83 (MESSAGE):11:38
@HeikoS  g++ version is too old11:38
sonne|worktrue11:39
sonne|worktoo old11:39
@wikingbesser82: ok i've got now the /proc /sys and /dev mounted11:39
@HeikoSsonne|work: but cmake say 4.3 at least11:39
@HeikoSin CMakeLists.txt11:39
besser82wiking: you'll need resolv conf and such als well ;)11:39
besser82wiking: so yum can "talk" to outside ;)11:40
sonne|workwiking: when you have a chroot like this please copy it over under chroot-template11:40
@wikingbesser82: hahahah11:40
besser82wiking: then yum makecache11:40
@wikingsonne|work: where?11:40
sonne|workwiking: so we can create some for ubuntu sid etc11:40
@wikingsonne|work: where's chroot-template? :)11:40
sonne|workwiking: /home/chroots11:40
sonne|workwiking: well you create one :)11:41
@wikingbesser82: sudo chroot /home/wiking/f19/fedora-19-x86_64 /usr/bin/bash11:41
@wikingbash-4.2# ps aux11:41
@wikingbash: ps: command not found11:41
@wiking:(11:41
sonne|workI mean you do currently so no need to redo this11:41
sonne|workwiking: PATH needs to be set11:41
@wikingok other things work11:41
@wikingP11:41
sonne|workbashrc11:41
besser82wiking: it's a very basic chroot ;) just the ABSOLUTELY minimum ;)11:41
sonne|work/etc/bash.bashrc11:41
sonne|workheh11:41
@wikingok11:42
besser82wiking: yum groups install 'C Development Tools and Libraries'11:42
@wikingyum makecache is working now11:42
@wikingafter this  yum groups install 'C Development Tools and Libraries' ?11:42
sonne|workwiking: I hope you did the template before that...11:42
@wikingsonne|work: what template? :)11:42
@wikingno i did nothing11:42
sonne|workexcellent :(11:42
@wikingi just untar-ed the tar.xz i've got from besser8211:42
@HeikoSguys I do not get it, CMakeList.txt want g++ >= 4.3 and I have 4.4 but it still says too old :(11:42
@wikingi mean what else would i need to template?11:43
sonne|workwiking: well we need to redo these steps (before the untar) for other chroots11:43
@wikingsonne|work: before?11:43
besser82wiking: depends on what you want to do?11:43
@wikingi didn't do anything before11:43
@wikingsonne|work: i'll just add 3 lines into fstab and that's all11:43
@HeikoSah it was cached11:43
@HeikoSthats not good, should be updated automatically11:43
@wikingother than that it's untar and then just runc chroot <path> <shell>11:43
besser82wiking: why don't you use schroot?11:43
@wikingbesser82: do i care? :)11:43
@wikingbesser82: i know i should11:44
besser82wiking: :D11:44
@wikingsonne|work: so there's really not much to template apart from the fstab lines11:44
@wikingthis image i've got besser82 is already done smoothly11:44
@wikingbesser82: ok i'm yum-ing the stuff11:44
besser82wiking: so after the yum parts there should be everything up for basic builds11:45
@wikingbesser82: afaik i should still install buildslave, cmake and ....?11:45
@wikinggit i guess11:45
besser82wiking: you might need to install some more stuff like cmake and additionally needed -devel pks ;)11:45
besser82wiking: I'll make a list just a moment...11:45
@wikingbesser82: well i just want to fire up a buildslave11:45
@wikingbesser82: and see all the errors :)11:46
@wikingok i guess i shouldnt run a buildslave with root :P11:47
@wikingi'll just run as nobody:nogroup11:48
sonne|workwiking: or create a user...11:48
besser82wiking: yum install -y arpack-devel atlas-devel blas-devel bzip2-devel ccache cmake doxygen-latex eigen3-devel fdupes glpk-devel graphviz hdf5-devel java-devel json-c-devel lapack-devel libxml-devel lpsolve-devel lua-devel lzo-devel mono-devel numpy octave-devel perl-devel pkgconfig python-docutils python-jinja2 python2-devel ruby-devel R-core-devel snappy-devel swig texlive-appendix texlive-lastpage texlive-todonotes xz-devel zlib-devel11:49
shogun-notifier-shogun: Heiko Strathmann :develop * 0d96b97 / README.cmake: https://github.com/shogun-toolbox/shogun/commit/0d96b9740be280760121ed558c00ae4bed54a77d11:49
shogun-notifier-shogun: added a comment on how to specify different compilers11:49
shogun-notifier-shogun: Heiko Strathmann :develop * 6e56a90 / README.cmake: https://github.com/shogun-toolbox/shogun/commit/6e56a90fce97c135974294b09f73537ea2b687cf11:49
shogun-notifier-shogun: Merge pull request #1541 from karlnapf/develop11:49
shogun-notifier-shogun:11:49
shogun-notifier-shogun: added a CMake Readme comment on how to specify different compilers11:49
besser82Where has iglesiasg gone????11:50
besser82Have some test-results for him....11:50
thoralfHmm.  Should it work to do "CBinaryLabels b = CBinaryLabels(10); something; b = CBinaryLabels(20);"?11:50
thoralfAnyone?11:50
sonne|workthoralf: what should this do?11:51
thoralfsonne|work: Just re-initializing an object with different values?11:52
@wikingbesser82: almost there... need to setup the buildmaster11:53
thoralfsonne|work: I'm using new[] to allocate a bunch of BinaryLabels, but they're all zero-length.  So I tried to reallocate like described and got 700 errors in valgrind.11:53
sonne|workthoralf: since these sgobjects use refcounting you shouldn't do stuff like this11:53
thoralfsonne|work: Okay, shouldn't do.  Thanks, then I'll use sth ilke SG_MALLOC.11:54
sonne|workthoralf: why not keep SGVectors?11:57
sonne|workand then set the labels?11:57
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has joined #shogun11:57
sonne|workvan51: good morning!11:57
van51sonne|work: hello :)11:57
besser82wiking: don't to forget to run a chronjob or sth. keep the f19-chroot updated ;)11:58
besser82wiking: There are updates around nearly everyday11:58
besser82wiking: yum makecache && yum -y distro-sync  ;)11:58
sonne|workwiking: and while you are at it please install a debian and ubuntu sid the same way :)11:58
besser82wiking: fedora-rawhide (equals debian sid) would be nice as well ;)11:59
besser82wiking: can provide a tarball for this as well ;)11:59
@wikingbesser82: waiting for it :)11:59
sonne|workbesser82: yes indeed - wiking just do it :D11:59
besser82wiking: just need to bootstrap, few minutes ;)12:00
sonne|workthere is no need for idle time on this machine12:00
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]12:00
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun12:00
shogun-buildbotbuild #1706 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1706  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>12:00
sonne|workwiking: we have to think about what we backup though... we don't want to spend days of setting up this thing when it gets down12:00
sonne|workvan51: what do you want to do next?12:01
@wikingsonne|work: do we have a backup space?12:01
sonne|workwiking: no12:01
sonne|workwe could create some .tgz12:02
@wikingsonne|work: provider does not provide one?12:02
sonne|worknope12:02
@wikinghttp://buildbot.shogun-toolbox.org/builders/rpm2%20-%20libshogun12:02
@wikingthere we go12:02
sonne|workwiking: look at the hd size12:02
shogun-buildbotbuild #1398 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1398  blamelist: Soeren Sonnenburg <sonne@debian.org>, Viktor Gal <viktor.gal@maeth.com>12:02
sonne|worktoo big12:02
@wikingsonne|work:  i mean a minimal size12:02
@wikingsonne|work: i've got 100gigs of backup from hetzner12:02
@wikingi'm backing up there the most important stuff12:02
@wikingok let's force a build on this new bot12:03
van51sonne|work: hmm good question.. I could either continue on random kitchen sinks or get started with the ipython notebooks and the demo12:03
@wikingsonne|work: force build --branch=develop 'rpm2 - libshogun'12:03
@wikingoh no u12:03
@wikingshogun-buildbot: force build --branch=develop 'rpm2 - libshogun'12:04
shogun-buildbotbuild #0 forced12:04
shogun-buildbotI'll give a shout when the build finishes12:04
shogun-buildbotbuild #0 of rpm2 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm2%20-%20libshogun/builds/012:05
@wikinghehehe that was faaast :)12:05
shogun-buildbotbuild #1707 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1707  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>12:05
shogun-buildbotbuild #1623 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1623  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>12:05
@wikingshogun-buildbot: force build --branch=develop 'rpm2 - libshogun'12:06
shogun-buildbotbuild #1 forced12:06
shogun-buildbotI'll give a shout when the build finishes12:06
shogun-buildbotbuild #1 of rpm2 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm2%20-%20libshogun/builds/112:06
@wikingshogun-buildbot: force build --branch=develop 'rpm2 - libshogun'12:07
shogun-buildbotbuild #2 forced12:07
shogun-buildbotI'll give a shout when the build finishes12:07
shogun-buildbotbuild #2 of rpm2 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm2%20-%20libshogun/builds/212:07
@wikingok let's start checking these errors :)12:07
@wikingbecause i think this is some missing HAVE_12:08
@wikingwtf12:08
@wikingwhy doesn't this have crontab?12:08
@wikingneither find12:09
@wiking:D12:09
@wikingwtf12:09
@wiking?!12:09
besser82wiking: ???12:09
@wikingbesser82: bash-4.2# find12:09
@wikingbash: find: command not found12:09
@wikingand i dont have it neither under /bin nor /usr/bin12:10
@wikingbesser82: same with crontab12:10
besser82wiking: yum groups install 'Minimal Install'12:11
shogun-buildbotbuild #1624 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1624  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>12:11
@wikingbesser82: ok trying12:12
@wikingError unpacking rpm package openssh-6.2p2-5.fc19.x86_6412:12
@wikingerror: unpacking of archive failed on file /usr/libexec/openssh/ssh-keysign: cpio: chmod12:12
@wikingError unpacking rpm package cronie-1.4.10-5.fc19.x86_6412:12
@wikingerror: unpacking of archive failed on file /usr/bin/crontab: cpio: chmod12:12
@wikingand some others...12:12
@wikingFailed: cronie.x86_64 0:1.4.10-5.fc19  initscripts.x86_64 0:9.47-1.fc19  openssh.x86_64 0:6.2p2-5.fc19    openssh-clients.x86_64 0:6.2p2-5.fc19  passwd.x86_64 0:0.79-1.fc19 polkit.x86_64 0:0.111-2.fc19   procmail.x86_64 0:3.22-32.fc19    sendmail.x86_64 0:8.14.7-1.fc19  sudo.x86_64 0:1.8.6p7-1.fc19           sysvinit-tools.x86_64 0:2.88-10.dsf.fc1912:13
@wikingbesser82: all for cpio chmod errors12:13
besser82wiking: did you run as root?12:13
@wikingbesser82: yes12:13
@wikingbash-4.2# whoami12:13
@wikingroot12:13
besser82wiking: ransom...12:13
@wikinga lot of other packages got installed w/o any problems12:14
@wikingbut these ones failed12:14
besser82wiking: then just retry another run ;) sometimes this might happen12:14
@wikingdoing it12:14
besser82wiking:tarball for rawhide: http://besser82.fedorapeople.org/fedora-rawhide-x86_64.tar.xz12:15
@wikingnning transaction Installing : cronie-1.4.10-5.fc19.x86_64                                                                                                                                      1/112:15
@wikingError unpacking rpm package cronie-1.4.10-5.fc19.x86_6412:15
@wikingerror: unpacking of archive failed on file /usr/bin/crontab: cpio: chmod12:15
@wikingbesser82: and it just does not want to fix by rerunning12:15
@wikingbesser82: is there like --force? :p12:16
@wikingbesser82: i think i know the prolbem12:17
besser82wiking: which?12:17
besser82ls12:17
@wikingno12:17
besser82yum reinstall fileystem -y12:17
besser82yum reinstall filesystem -y12:18
besser82typo12:18
-!- iglesiasg [~iglesias@2001:6b0:1:1041:e02c:79cd:fa8d:8b0b] has joined #shogun12:18
-!- mode/#shogun [+o iglesiasg] by ChanServ12:18
@wikingbesser82: same12:18
@wikingbesser82: oh fuuuck12:18
besser82iglesiasg: your test-suite logs: http://ur1.ca/fede012:18
besser82wiking: whot?12:18
@wikingbesser82: i should have untared the archive as root rigth?12:18
besser82wiking: yes ;)12:19
@wikingbesser82: coz now some of the files have my UID gid under /usr/bin12:19
@wikingok then redoing the whole thing12:19
besser82wiking: kk ;)12:19
besser82wiking: tar cpf12:19
@iglesiasgbesser82, you guys already talking about why did it go wrong?12:19
@wikingbesser82: i dont need to hald a chrooted os right?12:20
@iglesiasgbesser82, some of them are caused because the data files are missing, I guess you already realized of that12:20
besser82iglesiasg: yes12:20
@HeikoSwiking: I got a colpack problem, you probably know this error12:21
@HeikoSchecking whether make sets $(MAKE)... yes12:21
@HeikoSconfigure: error: source directory already configured; run "make distclean" there first12:21
@HeikoSmake[2]: *** [COLPACK/src/COLPACK-stamp/COLPACK-configure] Error 112:21
@HeikoSmake[1]: *** [CMakeFiles/COLPACK.dir/all] Error 212:21
besser82iglesiasg: forgot to fetch them ;)12:21
@wikingHeikoS: oh fuck12:21
@HeikoSfrom the bundled colpack12:21
@wikingHeikoS: revert my change...12:21
besser82wiking: there's no need to hald this...12:21
besser82iglesiasg: let me re-run with data-files present :)12:22
@iglesiasgbesser82, sure12:22
@HeikoSwiking: yet that did it12:23
shogun-notifier-shogun: Heiko Strathmann :develop * 409ab9d / cmake/external/ColPack.cmake: https://github.com/shogun-toolbox/shogun/commit/409ab9d795a68ab96ee9983ac5f07a109d3ed29212:24
shogun-notifier-shogun: Revert "Little fix for external ColPack.cmake issue #1529"12:24
shogun-notifier-shogun:12:24
shogun-notifier-shogun: This reverts commit 85b5000ed2ce4d0e04dbc89c4c4e2bd556e45671.12:24
shogun-notifier-shogun:12:24
shogun-notifier-shogun: This broke the build12:24
shogun-notifier-shogun: Heiko Strathmann :develop * 7352a35 / cmake/external/ColPack.cmake: https://github.com/shogun-toolbox/shogun/commit/7352a35753d851f209c4db7f8e63823d558b3c6c12:24
shogun-notifier-shogun: Merge pull request #1542 from karlnapf/develop12:24
shogun-notifier-shogun:12:24
shogun-notifier-shogun: fix colpack build12:24
@wikingHeikoS: sorry mate i thought we only need that for json12:24
@wikingseems colpack comes with the same error :P12:24
@HeikoSwiking: no worries, good that you know whats going on ;) for me, these error messages are just so cryptic :D12:24
@wikingbesser82: i think we have some weird shit going on12:25
@wikingbesser82: same error...12:25
besser82wiking: try chown -R 0:0 $chroot-dir12:25
besser82besser82: maybe my tarball has an error :(12:26
besser82wiking: or shall I create a new tarball?12:26
besser82wiking: meanwhile you can try with the rawhide one: http://besser82.fedorapeople.org/fedora-rawhide-x86_64.tar.xz12:27
@wikingwtf is this12:27
@wikingyum groups install 'Minimal Install'12:28
@wikingWarning: environment Minimal Install does not exist.12:28
@wikingNo packages in any requested group available to install or update12:28
@wikingi ran it before and it was ok12:28
@wikingok this does not work besser8212:28
@wiking:(12:28
shogun-buildbotbuild #1708 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1708  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>12:28
@wikingbesser82: what's rawhide?12:29
@wikingbesser82: ok i have rawhide running :P12:30
besser82wiking: rawhide in F is as Sid in debian ;)12:30
@wikingah ok12:31
@wikinglet's see if that works12:31
@wikingok12:31
@wikingthis fails as well12:31
besser82let me examine a bit, please...12:32
@wikingcan this be a feature of grs?12:32
shogun-buildbotbuild #1625 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1625  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>12:33
besser82wiking: possibly, can you please re-run yum command with -d 9 switch? and paste ouput somewhere?12:34
@wikingbesser82: ok just a sec12:34
besser82iglesiasg: here's oupt with data-files avail: http://ur1.ca/fedve12:35
@iglesiasgbesser82, import modshogun ImportError: No module named modshogun12:36
@iglesiasgbesser82, it seems to be a problem with python installation12:36
@wikingbesser82: i'll just cronio with -d 912:36
@wikingthat should trigger the problem12:36
besser82wiking: kk12:36
@wikingi mean cronie.x86_6412:37
besser82iglesiasg: testsuite is run without having (python-)modshogun installed on system...12:37
@wikingyum install -d 9 cronie.x86_64 ?12:37
@wikingor yum -d 9 install ?12:37
besser82wiking: does the same ;)12:37
besser82iglesiasg: so I need to set PYTHONPATH i guess?12:38
@wikingbesser82: http://pastebin.com/8SfKYuLb12:38
@iglesiasgbesser82, well then it is normal that Python examples break :)12:38
@iglesiasgbesser82, But that sounds weird, since the tests are not run for the interfaces that are not selected at install12:38
@wikingbesser82: owah man this i never seen before12:39
@iglesiasgbesser82, aaah ok. So you mean that Python modular is selected at installation but the installation is not visible to Python12:39
@wikingbesser82:12:39
@wikingbash-4.2# ls -la /usr/bin/crontab12:39
@wiking-rw-r--r-- 1 root root 53472 Sep  6 10:37 /usr/bin/crontab12:39
@wiking:)12:39
@iglesiasgbesser82, then yes, you need to set PYTHONPATH to look at the dist-packages directory and LD_LIBRARY_PATH to libshogun.so12:39
@iglesiasgbesser82, well or equivalent env for LD_LIBRARY_PATH in Fedora, no idea hehe12:40
@wikingbesser82: but i ran chmod 755 /usr/bin/crontab12:40
@wikingand it was all fine12:40
besser82wiking: now i got the reason: no groups file ;)12:41
besser82wiking: /etc/group-12:41
besser82and /etc/passwd-12:41
besser82;)12:42
@wikingok quickfix?12:42
besser82copy in the one from system's /etc12:42
besser82wiking: cp -a /etc/group* $chroot/etc12:43
@wikingbesser82: but /etc/passwd is there..12:43
@wikingas well /etc/group12:43
besser82and /etc/shadow?12:43
@wiking---------- 1 root root 405 Sep  6 09:23 /etc/shadow12:44
besser82wiking: just a second....12:44
@wikingand the *- are there12:44
shogun-buildbotbuild #1399 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1399  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>12:45
thoralfsonne|work: "thoralf: why not keep SGVectors?" <-- because my process returns BinaryLabels12:46
besser82wiking: are you running selinux on host?12:46
thoralfsonne|work: Anyway, it does not change my problem: Allocating memory for storing X objects and the setting each.12:47
thoralfsonne|work: Just didn't know that new[] is problematic.12:47
besser82wiking: or is chroot dir on nfs?12:47
thoralfsonne|work: And, btw., BinaryLabels::set_labels() copies the sgvector.12:48
besser82wiking: if chroot-dir in on NFS then you must mount it with "no_root_squash"12:49
@HeikoSsonne|work, travis is red, I know its since my merge, but its some ruby matrix problems, I have no idea how to fix this12:49
@wikingbesser82: i guess.. sonne|work would know it better12:50
besser82wiking: exec `echo "%_netsharedpath /" >> /etc/rpm/macros` inside the chroot and try yum again please12:50
@wikingbesser82: it's not nfs for sure12:50
@wikingbesser82: but i dont know what is the security shit12:51
besser82wiking: which security shit?12:51
@wikingbesser82: the stuff that makes /proc/modules disappear12:52
@wiking:)12:52
besser82which option you use to mount /proc12:52
besser82mount -o rbind ?12:52
besser82or mount -o bind ?12:52
sonne|workthoralf: well you can get the label vector from your binary labels (SGVector) and drop the label object12:55
sonne|workthoralf: hmmhh it should no longer be necessary to copy12:56
thoralfsonne|work: Thats right.  But won't I get the same problem with SGVector?  Allocating an array for storing it, and finally reassign the entries?12:56
sonne|workthoralf: well you could of course keep using binarylabels but then just store ptr's of the objects (CDynamicObjectArray ...)12:57
@HeikoSsonne|work, wiking do we want compile errors if lapacl is not present? currently this happens, or rather tell cmake to block if this is not there?12:59
@HeikoSiglesiasg: therre?13:00
@iglesiasgHeikoS, tell me13:00
@HeikoSiglesiasg: your stuff is using PCA13:00
@HeikoSwhich is not defined if LAPACK is not present13:00
@HeikoSso compile error13:00
@iglesiasgHeikoS, aaah I am missing the HAVE13:00
@iglesiasgsorry :(13:01
@HeikoSiglesiasg: no worries13:01
@HeikoScould you fix that now?13:01
@HeikoSotherwise I will do it13:01
@iglesiasgHeikoS, sure13:01
@iglesiasgHAVE_LAPACK I gues it is13:01
sonne|workHeikoS: no it should work w/o LAPACK13:01
-!- travis-ci [~travis-ci@ec2-174-129-54-125.compute-1.amazonaws.com] has joined #shogun13:01
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/1105114113:01
-!- travis-ci [~travis-ci@ec2-174-129-54-125.compute-1.amazonaws.com] has left #shogun []13:01
@HeikoSok cool, sonne|work do we have a buildbot without lapack to check this?13:02
@iglesiasgso why should it work without LAPACK as well?13:02
@iglesiasgI don't really get it13:02
@HeikoSiglesiasg: there should be at least no compile error13:03
@iglesiasgHeikoS, mmmm ok. I ignore how that is handled though :)13:03
@HeikoSrather a runtime error that your class is not available if LAPACK isnt13:03
@HeikoSor just jide it13:03
@HeikoSbut then you also have to do this in the examples/tests13:03
@iglesiasgaham13:03
@HeikoSiglesiasg: or13:03
@HeikoSreplace the lapack calls in PCA by eigen313:04
@HeikoSshould be less work actually13:04
sonne|workHeikoS: sure13:04
@HeikoS :)13:04
@HeikoSiglesiasg: and since you are using pca in your class which is tested, PCA is also implicitly tested so you got a guard whether these replacements change results13:04
@HeikoSiglesiasg: let me check CPCA and how much work that would be13:05
@iglesiasgok13:06
@HeikoS3 calls13:06
@HeikoSshould I change?13:06
@HeikoSno comments what the calls do, very nice13:07
@HeikoSbackward engineering :D13:07
@HeikoSiglesiasg: ah I dont know13:08
@wikingbesser82: -o bind13:08
@HeikoSiglesiasg: how much work is it to LAPACK guard your stuff?13:08
@HeikoSiglesiasg: could you do that asap? I currently cannot compile and am a bit in a hurry13:08
@iglesiasgHeikoS, I could do it if you tell me what is to be guarded13:09
@iglesiasgHeikoS, so sonne|work said the class is not necessary13:09
@HeikoSCPCA is not available if LAPACK is not13:09
@iglesiasgHeikoS, then, just the tests, or?13:09
@HeikoSso your class, which uses CPCA, also has to be unavailable if LAPACK isnt13:10
@HeikoSiglesiasg: same holds for all the examples/tests that include your CLMNNImpl class13:10
@HeikoStests not maybe13:11
@HeikoSbut its easy to add in unit tests13:11
@HeikoSpython is not that important I guess, just good practice13:11
@HeikoSiglesiasg: but if you could guard the CLMNNImpl and classes that include it, that would already allow me to compile libshogun13:11
@iglesiasgHeikoS, ok. Give me a couple of minutes then13:12
@iglesiasgI am on it13:12
@HeikoSiglesiasg: thanks a lot!13:12
shogun-buildbotbuild #1709 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1709  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>13:12
@HeikoSiglesiasg: maybe disable LAPACK and then you will see13:12
@wikingbesser82: i've tried rbind for /proc ... same13:13
@wikingsonne|work: what is the security stuff we on this machine?13:14
sonne|workwiking: ?13:14
@wikingsonne|work: grsec?13:14
sonne|worky13:14
@wikingsonne|work: we have problems with chroots13:14
sonne|workwiking: in which respect?13:14
@wikingsonne|work: well chmod is failing :)13:15
@wikingsonne|work: as well as /proc/modules is not present13:15
@wikingwe dont have /etc/grsec/policy? :)13:16
@wikinghow's this grsec setup...13:16
shogun-buildbotbuild #1626 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1626  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>13:17
@wikingsonne|work: u rolled yourself this kernel?13:17
@iglesiasgHeikoS, just sent a PR. But I didn't check it compiles without LAPACK13:17
sonne|workwiking: no13:18
sonne|workovh standard kernel13:18
sonne|workthey take care of security updates for it13:18
@wikingi see13:18
@HeikoSiglesiasg: thanks, let me try it :)13:19
@wikingi dont get this...13:19
shogun-notifier-shogun: Fernando Iglesias :develop * b1096b0 / / (6 files): https://github.com/shogun-toolbox/shogun/commit/b1096b0d778dda21efb4e0df68fd9c24519e290513:19
shogun-notifier-shogun: Guard LMNN with HAVE_LAPACK due to PCA use.13:19
shogun-notifier-shogun: Heiko Strathmann :develop * 3f9b71f / / (6 files): https://github.com/shogun-toolbox/shogun/commit/3f9b71f3bf5b16cc1f6a1804d089860eae2aefbf13:19
shogun-notifier-shogun: Merge pull request #1544 from iglesias/develop13:19
shogun-notifier-shogun:13:19
shogun-notifier-shogun: Guard LMNN with HAVE_LAPACK due to PCA use.13:19
@iglesiasgHeikoS, let's hope it works fine13:19
@HeikoScompiling ...13:20
@wikinghahahaha13:20
@wikingkernel.grsecurity.chroot_deny_chmod = 113:20
@wikingkernel.grsecurity.chroot_deny_chroot = 113:20
@wikingkernel.grsecurity.chroot_deny_fchdir = 113:20
@wikingkernel.grsecurity.chroot_deny_mknod = 113:20
@wikingkernel.grsecurity.chroot_deny_mount = 113:20
@wikingkernel.grsecurity.chroot_deny_pivot = 113:20
@wikingkernel.grsecurity.chroot_deny_shmat = 113:20
@wikingkernel.grsecurity.chroot_deny_sysctl = 113:20
@wikingkernel.grsecurity.chroot_deny_unix = 113:20
@wikingkernel.grsecurity.chroot_enforce_chdir = 113:20
@wikingkernel.grsecurity.chroot_findtask = 113:20
@wikingbesser82: i think i've found our problem :)))13:20
@wikingsonne|work: objections to kernel.grsecurity.chroot_deny_chmod=0 ?13:21
sonne|workheh13:21
sonne|workI never questioned one can break out of the chroot so do it13:22
@wikingsonne|work: ok i've created a grsec.conf under /etc/sysctl.d13:24
@wikingbesser82: now it works like a charm..13:24
shogun-buildbotbuild #1710 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1710  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>13:27
shogun-buildbotbuild #1400 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1400  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>13:27
shogun-buildbotbuild #1711 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1711  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>13:28
@HeikoSiglesiasg: seems to have worked :)13:28
@HeikoSthanks!13:28
@iglesiasgHeikoS, cool, thank you for pointing it out!13:28
shogun-buildbotbuild #1627 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1627  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>13:29
shogun-buildbotbuild #1628 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1628  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>13:34
@wikingok13:34
@wikingrpm2 is back13:34
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]13:37
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun13:37
@wikingwtf 14947 buildsl+  20   0 1729860 1.610g  11540 R  99.8  6.8   0:46.00 cc1plus13:39
shogun-buildbotbuild #3 of rpm2 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm2%20-%20libshogun/builds/3  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Fernando Iglesias <fernando.iglesiasg@gmail.com>13:40
@wikingiglesiasg: ok here's your problem ^13:40
@wikinghttp://buildbot.shogun-toolbox.org/builders/rpm2%20-%20libshogun/builds/3/steps/compile/logs/stdio13:40
@wikingiglesiasg: no lapack on this machine yet13:40
@wiking;)13:40
@wikingmmm13:41
@wikingbut this is rather the problem of the unit tests13:41
@iglesiasgno LMNN error there though13:42
@iglesiasgso I guess those are other classes not guarded with HAVE_LAPACK13:42
@wikinguyeah it was my bad sorry13:42
@iglesiasgah all right, no problem13:42
@wikingi just grepped for lapack:P13:42
thoralfsonne|work: I'm confused.  This "SGVector v=SGVector(10); v=SGVector(20);" works, although the same pattern does not work for CBinaryLabels.13:47
thoralfsonne|work: I know, it has a different base class.  But the issue was about reference counting, which applies to both.13:48
sonne|workthoralf: well there is some magic to support this o_O13:48
sonne|workfor SGVector, sparsevector13:48
sonne|workand matrix13:48
thoralfsonne|work: And may I rely on this magic?13:49
sonne|workcurrently yes13:49
@wikingshogun-buildbot: force build --branch=develop 'rpm2 - libshogun'13:49
shogun-buildbotbuild #4 forced13:49
shogun-buildbotI'll give a shout when the build finishes13:49
sonne|workbut surprises are never a good idea so I guess on the long run we have to get rid of this13:49
besser82wiking: nice to hear.  have been afk for lunch ;)13:50
@wikingsonne|work: ok just a question: if i would add the shogun-fedora1 bot under rpm1 - libshogun, would it always build, or it's like a round-robin then?13:50
thoralfsonne|work: But actually it's the only we to get certain things done.13:50
@wikingbesser82: http://buildbot.shogun-toolbox.org/builders/rpm2%20-%20libshogun/builds/413:50
sonne|workwiking: random13:51
sonne|workwiking: we have that for deb modular13:51
sonne|workand static etc13:51
sonne|workwiking: I no longer think it is a good idea. setups deviate slightly13:52
@wikingsonne|work: ok so if i add the sid version of fedora then it should be rpm3 - libshogun or something13:52
sonne|workso we have weird behaviour13:52
sonne|workwiking: I would rather use one rpm for RHEL6.313:52
sonne|workand one for FC19 etc13:52
@wikingsonne|work: i'll add rpmX - modular13:52
@wikingok so what should be the name13:52
@wikingas i'm about to create the other buildbot targets13:52
@wikinglike modular f1913:52
@wikingand modular fedora 'sid'13:53
@wikingbtw i've just add 2 threads for the buildslave as we'll have some amount of buildslaves running now on that machine ;)13:53
sonne|workwiking: or even distribution names lice RH63 / FC1913:54
@wikingsonne|work: ok13:54
@wikingso i'll start renaming it to FC1913:55
@wikingand shit like that13:55
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]13:56
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun13:56
@wikingshogun-buildbot: force build --branch=develop 'FC19 - libshogun'13:58
shogun-buildbotThe build has been queued, I'll give a shout when it starts13:59
thoralfsonne|work: Sorry, I was lying.  set_labels() does not copy.  Only the constructor does before initializing set_labels() and set_values().  FYI14:00
sonne|workthoralf: ahh ok14:01
-!- travis-ci [~travis-ci@ec2-23-20-125-26.compute-1.amazonaws.com] has joined #shogun14:05
travis-ci[travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/1105254714:05
-!- travis-ci [~travis-ci@ec2-23-20-125-26.compute-1.amazonaws.com] has left #shogun []14:05
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]14:05
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun14:05
shogun-buildbotbuild #1401 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1401  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Fernando Iglesias <fernando.iglesiasg@gmail.com>14:09
-!- vgorbati [c3ee5cb1@gateway/web/freenode/ip.195.238.92.177] has joined #shogun14:15
besser82wiking: is the fc rawhide thing ready, too?14:20
besser82wiking: i'd suggest naming it `FCrh`....14:20
@wikingbesser82: http://buildbot.shogun-toolbox.org/builders/FC19%20-%20libshogun/builds/0/steps/test/logs/stdio14:20
@wikingbesser82: so the libshogun is built successfully14:21
besser82wiking: +114:21
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has quit [Quit: Leaving.]14:23
@wikingshogun-buildbot: force build --branch=develop 'FC19 - modular_interfaces'14:27
shogun-buildbotbuild #0 forced14:27
shogun-buildbotI'll give a shout when the build finishes14:27
shogun-buildbotbuild #0 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/014:28
-!- travis-ci [~travis-ci@ec2-174-129-54-125.compute-1.amazonaws.com] has joined #shogun14:40
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/1105330414:40
-!- travis-ci [~travis-ci@ec2-174-129-54-125.compute-1.amazonaws.com] has left #shogun []14:40
shogun-buildbotbuild #1402 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1402  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>14:51
-!- michail [9bcf0401@gateway/web/freenode/ip.155.207.4.1] has joined #shogun14:55
-!- michail [9bcf0401@gateway/web/freenode/ip.155.207.4.1] has quit [Quit: Page closed]15:05
-!- sonne|work [~sonnenbu@91-64-72-127-dynip.superkabel.de] has left #shogun []15:31
-!- blaubaer_ [8d4c3eb2@gateway/web/freenode/ip.141.76.62.178] has joined #shogun15:36
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]15:36
@wikingshogun-buildbot: force build --branch=develop 'FC19 - libshogun'15:36
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun15:36
@wikingshogun-buildbot: force build --branch=develop 'FC19 - libshogun'15:36
shogun-buildbotbuild forced [ETA 14m23s]15:36
shogun-buildbotI'll give a shout when the build finishes15:36
blaubaer_Hello15:37
@wikinghey15:37
blaubaer_I have a question regarding the model selection feature of shogun15:37
besser82wiking: thanks to you for setting this up ;)15:37
besser82wiking: the FC19 builder, I mean ;)15:38
@wikingbesser82: well we still need some time15:38
blaubaer_I have my line: best_parameters=model_selection.select_model()15:38
@wikingbesser82: go ahead15:38
blaubaer_which works fine15:38
blaubaer_now i would like to access the found parameters as strings or whatever to use them elsewhere in my program15:38
blaubaer_so something like: best_parameters.get_parameter("name")15:38
blaubaer_as it seems my machine parameter configuration contains children15:39
blaubaer_root with C with values: vector=[1] kernel:"GaussianKernel" normalizer:"IdentityKernelNormalizer" width with values: vector=[1]15:39
blaubaer_best_parameters.print_tree()15:40
blaubaer_root with15:40
blaubaer_C with values: vector=[1]15:40
blaubaer_kernel:"GaussianKernel"15:40
blaubaer_normalizer:"IdentityKernelNormalizer"15:40
blaubaer_width with values: vector=[1]15:40
blaubaer_how can i access C, kernel name and normalizer?15:40
blaubaer_i am using the python bindings15:41
blaubaer_best_parameters.apply_to_machine(machine)15:42
blaubaer_print best_parameters.get_parameter('C',machine)15:43
blaubaer_None15:43
@wikingbesser82: should i have yum makecache && yum -y distro-sync for fedora 19 weekly?15:44
@wikingor just for the 'sid'?15:44
@wikingblaubaer_: mmm HeikoS would know the answer for this15:45
blaubaer_ok15:45
@wikingHeikoS: ping15:45
blaubaer_thanks for the reply. how or when could i ask him15:45
@wikingblaubaer_: he should be around15:46
@wikingnow :)15:46
blaubaer_:)15:46
shogun-buildbotHey! build FC19 - libshogun #1 is complete: Success [build successful]15:48
shogun-buildbotBuild details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20libshogun/builds/115:48
besser82wiking: daily ;)15:48
@wikingbesser82: both?15:48
besser82wiking: both ;)15:49
@wikingbesser82: crond is not running, and i cannot find /etc/init.d/crond or something to start it15:49
besser82wiking: just a minute ;)15:49
@iglesiasgblaubaer_, I also think that get_parameter as you were doing is the way to obtain the value of the parameter15:53
blaubaer_the line print best_parameters.get_parameter('C',machine)15:54
@iglesiasgblaubaer_, do you have a minimal code sample where this is not working fine?15:54
blaubaer_i tried to show you above15:54
blaubaer_print_tree() shows the correct values15:54
@iglesiasgblaubaer_, something I could download and execute in my machine so I can try around15:54
besser82wiking: yum install cronie && systemctl enable chrond && service chrond start15:54
blaubaer_however get_parameter() does not let me access it15:54
@iglesiasgblaubaer_, pastebin, gist, or something similar could help15:54
@iglesiasgblaubaer_, I see15:55
@wikingbesser82: root@ks312251 /]# systemctl enable chrond && service chrond start15:55
@wikingOperation failed: Invalid argument15:55
@wikingisn't it crond15:55
@wiking?:)15:55
@wikingnoup15:55
@iglesiasgblaubaer_,  I have not really used a lot the model selection framework, but if you have some minimal code I can try to see what's going on15:55
@wikingthat did not help15:55
blaubaer_it's now part of a larger codebase. i would take a while15:55
@iglesiasgblaubaer_, ok, then maybe wait until Heiko can answer you15:55
besser82wiking: sure, sry for the typo it's crond ;)15:55
@wikingbesser82: same error15:56
blaubaer_@iglesiasg: thanks for the try15:56
@wikingbesser82: systemctl gives this15:56
@wikingbesser82: systemctl15:57
@wikingFailed to get D-Bus connection: No connection to service manager.15:57
besser82wiking: ok, let me investigate...15:57
@wikingbesser82: thnx a lot15:57
besser82wiking: try `systemctl enable crond.service`15:58
besser82wiking: the dbus thing can be ignored15:58
besser82wiking: and don't forget to `service chrond start`15:59
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]16:00
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun16:00
@wikingbesser82: the first command worked16:00
@wikingthe second16:00
@wikingservice crond start16:00
@wikingRedirecting to /bin/systemctl start  crond.service16:00
@wikingFailed to get D-Bus connection: No connection to service manager.16:00
@wikingshogun-buildbot: force build --branch=develop 'FCRH - libshogun'16:01
shogun-buildbotbuild #0 forced16:01
shogun-buildbotI'll give a shout when the build finishes16:01
shogun-buildbotbuild #0 of FCRH - libshogun is complete: Failure [failed git]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/016:01
blaubaer_iglesiasg, i tried the first model selection example on the tutorial website: http://www.shogun-toolbox.org/doc/en/current/python_modular_examples.html#python_modular_modelselection_examples16:01
@wikingshogun-buildbot: force build --branch=develop 'FCRH - libshogun'16:01
shogun-buildbotbuild #1 forced16:01
shogun-buildbotI'll give a shout when the build finishes16:01
blaubaer_if i try to access  best_combination.get_parameter('C1',classifier) it gives me 'None'16:02
@iglesiasgblaubaer_, all right. Let me try that one16:02
shogun-buildbotbuild #1 of FCRH - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/116:03
@wikingshogun-buildbot: force build --branch=develop 'FCRH - libshogun'16:04
shogun-buildbotbuild #2 forced16:04
shogun-buildbotI'll give a shout when the build finishes16:04
shogun-buildbotbuild #2 of FCRH - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/216:04
@wikingbesser82: i'll need your help with this: http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/2/steps/configure/logs/stdio16:04
@wiking:D16:04
@wikingsorry to bother u so much16:04
besser82wiking: np. you're welcome ;)16:08
besser82wiking: `yum erase liblapack-devel liblapack` should fix16:10
besser82wiking: some things might have changed in rawhide recently ;)16:10
besser82wiking: and you'll need to `yum install libxml2-devel`16:11
@wikingNo Match for argument: liblapack-devel16:12
@wikingNo Match for argument: liblapack16:12
@wikingNo Packages marked for removal16:12
@wikingshogun-buildbot: force build --branch=develop 'FCRH - libshogun'16:13
shogun-buildbotbuild #3 forced16:13
shogun-buildbotI'll give a shout when the build finishes16:13
shogun-buildbotbuild #3 of FCRH - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/316:13
besser82wiking: `yum erase lapack-devel lapack` should be the ones16:13
@wikingshogun-buildbot: force build --branch=develop 'FCRH - libshogun'16:13
shogun-buildbotbuild #4 forced16:13
shogun-buildbotI'll give a shout when the build finishes16:13
shogun-buildbotbuild #4 of FCRH - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/416:14
@wikingok almost there16:14
@wikingnow there's still something with HDF516:14
@wikingbesser82: i think i'll have to fix the LAPACK detection imho16:14
@wikingas now it's -- A library with LAPACK API not found. Please specify library location.16:14
@wikingwhich is not the best..16:14
besser82wiking: should be ;)  btw. atlas provides lapack ;)16:15
besser82wiking: or better described as clapack ;)16:15
@wikingmmm i need to do something about lua detection as well16:16
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]16:19
@iglesiasgblaubaer_, I am not sure, but I think that the print you are trying to do in the example doesn't work because the parameters do not belong to the classifier, but to the kernel16:20
blaubaer_i am not sure if i made my point correctly. the print works. i want to access the VALUES of the parameter object to save them outside of shogun. i want to remember the result of the model selection.16:21
@iglesiasgblaubaer_, well the print was showing None, right?16:21
blaubaer_print_tree() correctly shows the whole tree16:22
blaubaer_i just do not know how to access the value of one parameter16:22
blaubaer_as a string or something16:22
@iglesiasgblaubaer_, print best_combination.get_parameter() I meant, sorry16:22
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has joined #shogun16:22
blaubaer_print is just an example16:23
@iglesiasgblaubaer_, I understand that you don't want to print them but get the value16:23
blaubaer_:) good16:24
blaubaer_i totally do not get, why i cannot access the values of C1 and C2. besides that the kernel values degree distance name_of_kernel are another problem. as i understand they are children of the class ParameterCombination and there seems to be no python wrapper function for get_child() or something similar16:27
-!- travis-ci [~travis-ci@ec2-54-234-214-31.compute-1.amazonaws.com] has joined #shogun16:30
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/1105447916:30
-!- travis-ci [~travis-ci@ec2-54-234-214-31.compute-1.amazonaws.com] has left #shogun []16:30
@iglesiasgblaubaer_, It seems that the parameters that you can get have to be associated with an object16:36
blaubaer_ok. can you tell me with what kind of an object?16:37
@iglesiasgblaubaer_, so check for instance param_gaussian_kernel inside the create_param_tree function16:37
@iglesiasgblaubaer_, the kernel parameter in there is associated with the Gaussian kernel16:38
@iglesiasgthen you can do get_parameter using gaussian_kernel16:38
besser82wiking: are the debug-versions of HDF5 really needed?16:44
blaubaer_@iglesiasg, please verify that i am doing the correct thing: i inlined the whole function create_param_tree for simplicity.  now i can call best_combination.get_parameter(param_name, mystic object). what is the mystic object for param_name="C1" to get 1.0?16:45
@wikingbesser82: mmm i dont even know why is it looking for it16:45
@iglesiasgblaubaer_, no idea how can you get C1 since when it is created using ModelSelectionParameters there is no other object given16:45
blaubaer_oh16:46
blaubaer_btw print best_combination.get_parameter('degree',gaussian_kernel) does not work ether16:46
blaubaer_either16:46
@iglesiasgwhat about 'kernel',gaussian_kernel?16:47
besser82wiking: let me have a look on the .cmake16:47
blaubaer_nope16:47
@iglesiasgblaubaer_, is degree of gaussian_kernel, btw?16:47
@iglesiasgI think it is of power kernel16:47
blaubaer_ups16:47
@iglesiasgwidth should be of Gaussian16:47
blaubaer_no luck again16:48
blaubaer_the output of best_params.print_tree() is the following:16:49
blaubaer_root "C1"=1.000000 "C2"=2.000000  "kernel":PowerKernel at 0x136e570  "degree"=1.000000  "distance":MinkowskiMetric at 0x125db80  "k"=1.00000016:49
blaubaer_so it finds the powerKernel as best kernel with degree 1.0 distance minkowski metric and k=116:49
blaubaer_so shouldn't i be able to access the values without knowing beforehand which kernel will be selected as best?16:50
@iglesiasgblaubaer_, what values?16:50
-!- sonne|osx [~sonne@89.204.138.34] has joined #shogun16:50
@iglesiasgblaubaer_, in best_params only the best should appear, right?16:51
blaubaer_values for degree, distance etc16:51
blaubaer_yup16:51
@iglesiasgblaubaer_, so get_parameter("kernel",PowerKernel)16:51
@iglesiasgshould work I guess16:51
@iglesiasgor power_kernel, I don't remember the name of the variable16:52
blaubaer_but if i have to give a kernel object, i must know which one (power?, gaussian?) and that is what i want to find out16:52
blaubaer_i initialize the parameter tree with different kernels. one gets selected. i want to find out which one?16:52
sonne|osxwiking: what subdomain did you want me to point where?16:54
shogun-buildbotbuild #90 of osx1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/90  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Soeren Sonnenburg <sonne@debian.org>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, van51 <vangelis_51@hotmail.com>, Viktor Gal16:54
shogun-buildbot<viktor.gal@maeth.com>16:54
@wikingsonne|osx: do we have a www server on the new machine?16:56
@iglesiasgblaubaer_, I see what you mean, but I am not sure how to do it correctly, sorry16:56
@wikingor can we have one?16:56
sonne|osxwiking: I would prefer not to have one - what do you need it for?16:56
@wikingsonne|osx: cdash but ok16:56
sonne|osxwiking: I mean we have one dedicated webserver16:56
@wikingsonne|osx: http://maeth.com/cdash/16:56
@wikingthis16:56
@wikingbut if not u can direct it under my ip16:57
@wikingand i can host it it's ok16:57
blaubaer_@iglesiasg, thank you very much for your time! i think i will write a question to the mailing list.16:57
@iglesiasgblaubaer_, good idea. HeikoS probably knows how to do this stuff as he wrote the model selection framework16:57
@iglesiasgI found it weird though that in any of the examples it doesn't appear how to do this16:58
-!- thoralfk [~thoralf@enki.zib.de] has joined #shogun16:58
blaubaer_i do too ;-)16:58
@iglesiasgblaubaer_, it might be also useful to check in the mailing some of the mails that mention something about model selection. Someone might have already tried to do that16:58
-!- thoralf [~thoralf@enki.zib.de] has quit [Write error: Broken pipe]16:58
@iglesiasgit seems something quite reasonable indeed16:59
* sonne|osx off17:00
-!- sonne|osx [~sonne@89.204.138.34] has quit [Quit: sonne|osx]17:00
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun17:15
shogun-notifier-shogun: Viktor Gal :develop * d7c27d1 / CMakeLists.txt,cmake/ShogunUtils.cmake: https://github.com/shogun-toolbox/shogun/commit/d7c27d194e0cb0eb8bf013c97e1f2ef4f523b8fe17:15
shogun-notifier-shogun: Fix Lua detection17:15
shogun-notifier-shogun: Fix JBlas detection17:15
shogun-notifier-shogun: Simplify the interface status printing by using macro17:15
shogun-notifier-shogun: Fix JavaModular status detection17:15
shogun-notifier-shogun: Add LuaModular status printing17:15
@wikingok17:32
@wikingwhy was it necessary to break RubyModular?17:32
-!- vgorbati [c3ee5cb1@gateway/web/freenode/ip.195.238.92.177] has quit []17:36
shogun-buildbotbuild #1712 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1712  blamelist: Viktor Gal <viktor.gal@maeth.com>18:06
-!- travis-ci [~travis-ci@ec2-54-224-93-41.compute-1.amazonaws.com] has joined #shogun18:08
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/1106283218:08
-!- travis-ci [~travis-ci@ec2-54-224-93-41.compute-1.amazonaws.com] has left #shogun []18:08
shogun-buildbotbuild #5 of FCRH - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/5  blamelist: Viktor Gal <viktor.gal@maeth.com>18:08
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has quit [Quit: Leaving.]18:08
shogun-buildbotbuild #1629 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1629  blamelist: Viktor Gal <viktor.gal@maeth.com>18:12
shogun-buildbotbuild #1 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/1  blamelist: Viktor Gal <viktor.gal@maeth.com>18:18
-!- pickle27 [~Kevin@199.119.128.114] has joined #shogun18:27
@wikingpickle27: dude that nmatrix is fucking not working18:34
@wikingpickle27: so please tell me how u compiled it18:34
pickle27wiking, hmmm I had a bit of a rough go as well18:35
pickle27I thought they had posted some new instructions on how to build it18:35
@wikingi dont care18:35
@wikinghow did u do it18:35
@wikingit's completely unstable18:36
@wikingand untested18:36
@wikingit was a rather bad idea to do this before release18:36
@wikinganyhow18:36
@wikingelaborate18:36
pickle27to be honest I wasn't expecting it be merged for a while18:36
@wikinghow did u compile18:36
@wikingbecause now i have to fucking redo this shit18:36
@wikingon travis18:36
@wikingbuildbots18:36
pickle27I seem to recall that I had to have libgsl installed and used their version of blas18:37
@wikingno fucking way18:38
@wikingok18:38
@wikingthis stuff is going to be reverted18:38
@wiking...18:38
pickle27I built on my desktop about a month ago, I need to build on my laptop in the next few days so I can figure it out and get back to you18:38
@wikingno time for that sorry18:38
pickle27honestly thats probably what I would do until they make a deb package for it18:39
pickle27even building numpy is a huge mess18:39
@iglesiasgall right guys, see you later!18:41
pickle27wiking, I was planning on just having it as a branch of my own until they released a stable nmatrix18:41
@wikingok fucking hell thi sis18:42
@wikingbecause this is basically 2 fucking commit18:42
@wikings18:42
pickle27wiking, reverting it you mean?18:43
-!- iglesiasg [~iglesias@2001:6b0:1:1041:e02c:79cd:fa8d:8b0b] has quit [Quit: Ex-Chat]18:44
shogun-notifier-shogun: Viktor Gal :develop * 6b1d4c6 / / (29 files): https://github.com/shogun-toolbox/shogun/commit/6b1d4c646bae3ddf4f8bbcfccc9343aadecb420718:45
shogun-notifier-shogun: Revert "updated ruby modular to use nmatrix, initial commit"18:45
shogun-notifier-shogun:18:45
shogun-notifier-shogun: This reverts commit f713eec5465a000e1d452d1dea5dfed2a940e78b.18:45
shogun-notifier-shogun:18:45
shogun-notifier-shogun: nmatrix is faaaar from being stable...18:45
@wikingthere we go18:45
pickle27wiking, okay right on - sorry about that I didn't expect it to be merged, also I thought they had made the build smoother since I last did it18:46
@wikingpickle27: not your fault18:46
@wikingpickle27: can u give me that commit18:46
@wikingso i can start a feature branch18:46
pickle27sure - how do I do that?18:46
shogun-buildbotbuild #1403 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1403  blamelist: Viktor Gal <viktor.gal@maeth.com>18:50
@HeikoSwiking, sonney2k, I got more error, any ideas?18:51
@HeikoS/home/ucabhst/shogun/build/src/interfaces/python_modular/modshogunPYTHON_wrap.cxx: In function 'PyTypeObject* SwigPyStaticVar_Type()':18:51
@HeikoS/home/ucabhst/shogun/build/src/interfaces/python_modular/modshogunPYTHON_wrap.cxx:3378: warning: deprecated conversion from string constant to 'char*'18:51
@HeikoSIn file included from /share/apps/python/lib/python2.7/site-packages/numpy/core/include/numpy/ndarrayobject.h:26,18:51
@HeikoS                 from /share/apps/python/lib/python2.7/site-packages/numpy/core/include/numpy/arrayobject.h:14,18:51
@HeikoS                 from /home/ucabhst/shogun/build/src/interfaces/python_modular/modshogunPYTHON_wrap.cxx:5973:18:51
@HeikoS/share/apps/python/lib/python2.7/site-packages/numpy/core/include/numpy/__multiarray_api.h: In function 'int _import_array()':18:51
@HeikoS/share/apps/python/lib/python2.7/site-packages/numpy/core/include/numpy/__multiarray_api.h:1535: warning: deprecated conversion from string constant to 'char*'18:51
@HeikoS/share/apps/python/lib/python2.7/site-packages/numpy/core/include/numpy/__multiarray_api.h:1542: warning: deprecated conversion from string constant to 'char*'18:51
@wikingnoup18:51
@HeikoSg++4.418:51
@HeikoSargh18:51
@HeikoSannoying!18:51
@HeikoSdoes swig need any includes?18:51
@wikinga lot18:52
@wiking:018:52
@HeikoSI gave cmake my swig bin path18:52
@wikinggot for it18:53
@wiking:)18:53
pickle27wiking, how do I give you the commit for a feature branch?18:53
@wikingi mean good18:53
@HeikoSmaybe an include path missing?18:53
shogun-buildbotbuild #6 of FCRH - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/6  blamelist: Viktor Gal <viktor.gal@maeth.com>18:53
@wikingpickle27: i think i'll get it from the PR18:54
pickle27wiking, okay let me know if you need anything18:54
@HeikoS/home/ucabhst/shogun/build/src/interfaces/python_modular/modshogunPYTHON_wrap.cxx:8941: error: 'Py_buffer' was not declared in this scope18:54
pickle27wiking, after I'l put an email out on nmatrix maillist basically saying hey we have this feature branch please help with more examples and make your build easier18:55
@HeikoSgoogle doesnt help18:55
shogun-buildbotbuild #1630 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1630  blamelist: Viktor Gal <viktor.gal@maeth.com>18:58
shogun-notifier-shogun: Kevin :feature/nmatrix * 3a20e1d / / (29 files): https://github.com/shogun-toolbox/shogun/commit/3a20e1d291624c103185ea2097d29851a74f3e0c19:00
shogun-notifier-shogun: updated ruby modular to use nmatrix, initial commit19:00
@wikingthere you go19:00
pickle27cool, I'll mail the sciruby guys19:02
shogun-buildbotbuild #2 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/2  blamelist: Viktor Gal <viktor.gal@maeth.com>19:07
@HeikoSwiking: could you have another look here?19:10
@HeikoShttps://gist.github.com/karlnapf/646675319:10
@HeikoSsonney2k: ^19:10
@HeikoSany ideas?19:10
@wikingHeikoS: not until you try to make with19:11
@wikingmake VERBOSE=119:11
@wikinglike this i cannot tell sorry19:11
@HeikoSok let me update19:11
@wikingshogun-buildbot: force build --branch='develop' 'FC19 - modular_interfaces'19:11
shogun-buildbotbuild #3 forced19:11
shogun-buildbotI'll give a shout when the build finishes19:11
shogun-buildbotbuild #3 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/319:12
@HeikoSwiking: https://gist.github.com/karlnapf/646675319:14
@wikingdude19:15
@wiking/usr/include/python219:15
@wiking-I/usr/include/python2.419:15
@wikingwtf is this?19:15
@HeikoSah here we go19:15
@wikingpython --version19:16
@HeikoSso the thing is19:16
@wikingwhat's the version?19:16
@HeikoSpython is installed in/some/dir19:16
@HeikoSand I added the bin to PATH19:16
@HeikoSwhat else do I need to do there?19:16
@wikingHeikoS: read cmake faq19:16
@wikinghow to set properly paths for python19:16
@wikingHeikoS: a good start: /usr/share/cmake/Modules/FindPythonLibs.cmake19:17
@HeikoSwiking: cool thanks for the hint19:18
@wiking# If you'd like to specify the installation of Python to use, you should modify19:18
@wiking# the following cache variables:19:18
@wiking#  PYTHON_LIBRARY             - path to the python library19:18
@wiking#  PYTHON_INCLUDE_DIR         - path to where Python.h is found19:18
@HeikoSwiking: is this also done with this -DPYTHON_LIBRARY?19:19
@wikingHeikoS: try it19:19
@HeikoSand how can I see whether this worked apart form compiling?19:21
@HeikoSI will add these things to the readme afterwards19:21
@wikingHeikoS: cmake output?19:21
@wikingor dig somewhere19:21
@wikingin the udner the build dir19:22
@HeikoScmake detects python2.7 for me with running it plan19:22
@HeikoSthats what it sais19:22
@wikingwell19:22
@wikingwhat did it say before?19:22
@wikingnot for the interpreter but for the library19:22
@wikingbesser82: here?19:23
-!- thoralfk [~thoralf@enki.zib.de] has quit [Quit: Konversation terminated!]19:25
-!- vgorbati [~vgorbati@91.216.173.29] has joined #shogun19:31
besser82wiking: sry, been afk. no back ;)19:32
besser82wiking: s/no/&w/19:33
@wikingbesser82: do you know if there's narray package for ruby in fedora?19:33
@wikingi couldn't find it but maybe you know better19:33
besser82wiking: currently there isn't, but I'm on it ;)19:34
@wikingbesser82: ah ok... now the next question is then how to find the right path for the narray.h and narray.so when installing it via gem :P19:34
@wikingbecause on other distros19:34
@wikinggem which narray.h19:34
@wikinggives the right path19:34
@wikingbut not under fedora :S19:35
besser82wiking: hmm. tough one. let me have a look. about 5 mins. ;)19:35
@wikingokok that's alright19:35
@wikingim just currently asking things that i have no clue about :)19:35
shogun-buildbotbuild #1404 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1404  blamelist: Viktor Gal <viktor.gal@maeth.com>19:36
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 245 seconds]19:37
shogun-buildbotbuild #1713 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/1713  blamelist: Viktor Gal <viktor.gal@maeth.com>19:39
@HeikoSwiking: yeah, I finally got it installed :D19:41
@HeikoSafter 100 compile errors and other things :)19:41
@wikingyey19:43
-!- vgorbati [~vgorbati@91.216.173.29] has quit [Ping timeout: 264 seconds]19:44
-!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has quit [Ping timeout: 245 seconds]19:45
-!- vgorbati [~vgorbati@91.216.173.29] has joined #shogun19:47
besser82wiking: would it be enough if I prepare rubygem-narray.rpm over the we?19:50
@wikingi'm fine with it19:54
besser82wiking: message was for me?19:54
@wikingbesser82: yes yes19:54
besser82wiking: kk ;) thx19:55
besser82wiking: will provide you with rpm on monday, then ;)19:55
@wikingok i might fix it till then19:56
besser82wiking: no need to waste your time ;)19:57
besser82So guys, Friday night 8 o'clock local-tz ;)  I wish you all a nice eve and some refreshing we ;)20:04
@wikingthnx for all the help20:10
@wiking!20:10
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has joined #shogun20:25
@sonney2kHeikoS, wiking I am currently building each possible revision of shogun checking sizes20:41
@sonney2kof libshogun*.so.*20:41
@sonney2kback in 2012 libshogun was just 105MB20:42
@sonney2k'just' ...20:42
@sonney2know we are at 250MB20:43
@wikingsonney2k: why do u do hat? :)20:43
@sonney2klet me try to do some plot20:43
@sonney2kwiking, well don't you find it a bit suspicious that we had ~100MB with shogun 2.120:43
@sonney2kand now have 250?20:43
@wikinghehehe20:44
@sonney2kbesides the loading time of the library of course20:44
@wikingsonney2k: striping?20:44
@wikingsonney2k: and removing -g20:44
@wikingwould help a lot20:44
@sonney2kwiking, sure but still >20MB20:44
@sonney2kwhich is still a lot20:44
@sonney2kunder /usr/lib/ only libwebkit is about the same size20:46
@sonney2kso we are producing the biggest lib ever o_O20:46
@sonney2kstupidly I forgot to measure compile time for each release20:47
@wiking:>20:48
@sonney2kwiking, and we have like 20% of the commits not compiling btw20:51
@wiking?20:54
@sonney2kwiking, it just doesn't compile that's all20:59
-!- travis-ci [~travis-ci@ec2-54-224-93-41.compute-1.amazonaws.com] has joined #shogun21:06
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/1106674021:06
-!- travis-ci [~travis-ci@ec2-54-224-93-41.compute-1.amazonaws.com] has left #shogun []21:06
@HeikoSsonney2k, cool21:16
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has joined #shogun21:19
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]21:20
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun21:20
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]21:26
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun21:26
@sonney2kHeikoS, wiking : https://dl.dropboxusercontent.com/u/31480471/libshogun_size.png21:46
@HeikoSerror 40421:46
@sonney2kHeikoS, try again21:47
@HeikoSsonney2k: do you know how to start a custom python binary in such way that it finds all the libraries? Do I only need to setup the LD_LIBRARY_PATH? I get loads of undefined symbols in importing things21:48
@HeikoSshogun works but numpy, scikits.sparse dont21:48
@sonney2kHeikoS, LD_LIBRARY_PATH and PYTHON_PATH are sufficient21:48
shogun-notifier-shogun: Viktor Gal :develop * 019ac5c / CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/019ac5c0fd7f6e631ebf1b756ee419211c05b3b021:48
shogun-notifier-shogun: Force to use optimized HDF5 libraries21:48
@HeikoSok thanks21:48
@sonney2kHeikoS, I find it rahter shocking that we went up from 100MB -> 240MB21:49
@HeikoSI dont really get the plot21:49
@HeikoSno descriptions ;)21:49
@sonney2kHeikoS, y axis is compiled size of libshogun in MB21:49
@wikingsonney2k: i guess it's backwards right?21:49
@sonney2kx is now till november 201221:50
@sonney2kall revisions21:50
@wikingsonney2k: have u used any bundle stuff?21:51
@sonney2kwiking, ?21:51
@wikingsonney2k: when u created the librar21:51
@wikingwith cmake21:51
@wiking-DBUNDLE_*=ON?21:51
@wikinghave u specified any of those?21:51
@sonney2kwiking, where in this script?21:51
@sonney2kno that is all still configure based21:51
@wikingah ok21:51
@sonney2kwe didn't have cmake back then21:52
@wikingyeah yeah21:52
@sonney2kbut it certainly explains the increase in compile time21:52
@wikingnow = cmake21:52
@sonney2kwiking, now is just before cmake21:52
@wikingah ok21:52
@sonney2know is maybe 10MB higher even21:52
@HeikoSwow I see21:54
@HeikoSany ideas where this symbol is found?21:54
@HeikoSImportError: /home/ucabhst/.python-eggs/scikits.sparse-0.1-py2.7-linux-x86_64.egg-tmp/scikits/sparse/cholmod.so: undefined symbol: amd_printf21:54
@HeikoSamd_printf?21:54
@sonney2kHeikoS, use grep amd_printf * --include='*.so.*'21:55
@sonney2k-R21:55
@sonney2kto find it21:55
shogun-buildbotbuild #7 of FCRH - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/7  blamelist: Viktor Gal <viktor.gal@maeth.com>21:55
@HeikoSsonney2k: I dont know these commands to well, what would be the full command?21:55
@HeikoSappreciating your help21:56
@sonney2kHeikoS, grep amd_printf * --include='*.so.*' -R21:56
@sonney2kwell the first * would be the dir name(s)21:56
@HeikoSok so best to look in shared/lib/dir21:56
@sonney2kgrep amd_printf /possible/dir --include='*.so.*' -Rl21:57
@HeikoSsonney2k: thanks, waiting for output ....21:58
@sonney2kHeikoS, wiking  - regarding the figure: when we released shogun2.1.0 we were at 121MB21:58
@HeikoSmany undefined symbols when trying to import things ...21:58
@HeikoSsonney2k: maybe we should start thinking about compiling in modules21:59
@sonney2kso we doubled the size for some reason (not saying that 121MB is not a lot already...)21:59
@HeikoSin the way that one only specifys a collection of classes one wants21:59
@HeikoSand then only the dependencies are compiled21:59
@sonney2kHeikoS, it won't solve the size problem but sure it makes sense (but it is tough!)22:00
shogun-buildbotbuild #1631 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1631  blamelist: Viktor Gal <viktor.gal@maeth.com>22:01
@HeikoSsonney2k: isnt the file smaller when not everything is compiled?22:02
@HeikoSand one hardly needs everything22:02
@HeikoSor is it possble to creat multiple library files?22:02
@sonney2kHeikoS, all possible and yes22:02
@sonney2kbut if we compile e.g. shogun for a modular interface we need everythong again22:02
@sonney2kthing22:03
@HeikoSah22:03
@HeikoSdamn22:03
@HeikoSsonney2k: I got another library question22:05
@HeikoSif I get this here: https://gist.github.com/karlnapf/646891022:05
@HeikoSis there a way to fix it without re-linking the library? Which means for me asking the sysadmin to re-install/re-compile22:05
@HeikoSor can I solve that with LD_LIBRARY_PATH?22:06
shogun-buildbotbuild #4 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/4  blamelist: Viktor Gal <viktor.gal@maeth.com>22:08
@sonney2kHeikoS, yes it is just missing the proper LD_LIBRARY_PATH22:08
@HeikoSok so all I need to find is this file that includes all the symbols?22:08
@HeikoSsonney2k: is there some way to find the lib files of all missing symbols?22:09
@HeikoSI dont really know what to google for in that case22:09
@sonney2kHeikoS, well all of them are name *.so.*22:09
@sonney2kso just do find /dir -name '*.so.*'22:09
-!- travis-ci [~travis-ci@ec2-54-234-214-31.compute-1.amazonaws.com] has joined #shogun22:09
travis-ci[travis-ci] it's Kevin'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/1106737222:09
-!- travis-ci [~travis-ci@ec2-54-234-214-31.compute-1.amazonaws.com] has left #shogun []22:09
@sonney2kand then add all these paths to LD_LIBRARY_PATH22:10
@HeikoSehm, isnt that too much?22:10
@sonney2kHeikoS, sure but it seems like the only thing you can do ATM :)22:11
@sonney2kthe other option was grep but you seem not to be happy about it22:11
@HeikoSsonney2k: ok cool, so I put all those into a text files, now how to add all those paths to LD_LIBRARY_PATH?22:11
@HeikoSsonney2k: it did not terminate after 15 mins22:11
@sonney2kHeikoS, sure takes ages22:11
@HeikoSsonney2k: the problem is that there are many many .so. files which are conflicting, so maybe the grep is better22:12
@sonney2kHeikoS, how did you manage to create such a mess(tm)22:12
@HeikoSsonney2k: cluster, which has no software installed globally, but everything is in /share/apps22:12
@HeikoSI managed to compile shogun though22:13
@HeikoSnow I just need this scikits.sparse.cholmod module22:13
@HeikoSand then I am fine22:13
@HeikoSsonney2k: I found all these files with the right names but .a   what are those doing?22:16
@HeikoScan I use them?22:17
@HeikoSstatically linked stuff22:17
@sonney2kHeikoS, no you cannot use them22:19
@sonney2kthese are useful only at compile time22:19
@sonney2kbut you want to start things and use dynamic linking22:19
@HeikoSargh22:19
@sonney2kso you need *.so.*'s22:19
@HeikoSok thanks22:20
-!- HeikoS [~heiko@nat-160-163.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]22:29
@wikingsonney2k: here?22:31
@wikingsonney2k: do u remember what does LPX_22:32
@wiking LPX_22:32
@wikingLPX_K_PRESOL do?22:32
@wikingok i've found it22:34
shogun-notifier-shogun: Viktor Gal :develop * a5a115c / src/shogun/classifier/mkl/MKL.cpp,src/shogun/classifier/mkl/MKL.h: https://github.com/shogun-toolbox/shogun/commit/a5a115cdaa30ca4f17435de1382ab8c7a6b3916422:36
shogun-notifier-shogun: Use glp_ functions in MKL instead of deprecated lpx_ ones22:36
shogun-buildbotbuild #1405 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1405  blamelist: Viktor Gal <viktor.gal@maeth.com>22:37
-!- vgorbati [~vgorbati@91.216.173.29] has quit [Quit: vgorbati]22:38
shogun-buildbotbuild #1090 of rpm1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm1%20-%20libshogun/builds/1090  blamelist: Viktor Gal <viktor.gal@maeth.com>22:42
shogun-buildbotbuild #1632 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1632  blamelist: Viktor Gal <viktor.gal@maeth.com>22:54
shogun-notifier-shogun: hushell :develop * f337508 / src/shogun/base/class_list.cpp.py: https://github.com/shogun-toolbox/shogun/commit/f3375086c92f21478e00d0e399d30653b4df9f4022:54
shogun-notifier-shogun: fix a tiny issue when checking virtual functions in class_list.cpp.py22:54
shogun-notifier-shogun: Soeren Sonnenburg :develop * 65b0f53 / src/shogun/base/class_list.cpp.py: https://github.com/shogun-toolbox/shogun/commit/65b0f531268b1ac9e04ebd9dce1526e5d916ecd222:54
shogun-notifier-shogun: Merge pull request #1540 from hushell/class_list22:54
shogun-notifier-shogun:22:54
shogun-notifier-shogun: fix a tiny issue when checking virtual functions in class_list.cpp.py22:54
shogun-buildbotbuild #5 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/5  blamelist: Viktor Gal <viktor.gal@maeth.com>23:08
shogun-buildbotbuild #8 of FCRH - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/823:09
shogun-buildbotbuild #1633 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1633  blamelist: Soeren Sonnenburg <sonne@debian.org>, hushell <hushell@hushell-U510.(none)>23:13
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun23:26
-!- mode/#shogun [+o iglesiasg] by ChanServ23:26
shogun-buildbotbuild #6 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/6  blamelist: Soeren Sonnenburg <sonne@debian.org>, hushell <hushell@hushell-U510.(none)>23:26
@iglesiasgHi all!23:27
@iglesiasghushell, out of curiosity, why in your PR did tid changed from a SGString<char> to an int32_t in a few places?23:28
shogun-buildbotbuild #1406 of cyg1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1406  blamelist: Viktor Gal <viktor.gal@maeth.com>23:31
shogun-buildbotbuild #1091 of rpm1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm1%20-%20libshogun/builds/1091  blamelist: Viktor Gal <viktor.gal@maeth.com>23:34
shogun-buildbotbuild #1715 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/1715  blamelist: Viktor Gal <viktor.gal@maeth.com>23:53
hushelliglesiasg: hey, I am back from a football game. Just in time :)23:53
@iglesiasghushell, nice! Was it nice?23:54
hushelliglesiasg: yeah, pretty good, but in US only elder people play football, :(23:54
@iglesiasghehe23:55
hushellregarding the type id, because I try to maintain a w_map in the factorgraph model23:55
hushellso the value in the w_map is the id23:55
hushellwhich is more convenient than a string23:55
hushellin that way, I have to use hash map23:56
hushellnow m_w_map is a vector, has the same length as model weights23:56
@iglesiasgok23:57
hushellbecause model parameters are actually stored in FactorType, so I have to define a way to combine them to the global weight vector23:57
hushelliglesiasg: I'll update the PR ASAP :) Thanks for the comments!23:58
@iglesiasgno problem, thanks to you23:58
--- Log closed Sat Sep 07 00:00:48 2013
« Thursday, 2013-09-05 Index Saturday, 2013-09-07 »

Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!