IRC logs of #shogun for Tuesday, 2013-09-10

« Monday, 2013-09-09 Index Wednesday, 2013-09-11 »
--- Log opened Tue Sep 10 00:00:52 2013
shogun-buildbotbuild #1762 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/1762  blamelist: Thoralf Klein <thoralf.klein@zib.de>00:11
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has joined #shogun00:17
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/1116616600:17
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has left #shogun []00:17
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Ex-Chat]00:33
-!- zhengyangl [~zhengyang@114.255.40.22] has joined #shogun00:35
-!- zhengyangl is now known as foulwall00:38
shogun-buildbotbuild #1763 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/1763  blamelist: Soeren Sonnenburg <sonne@debian.org>, Thoralf Klein <thoralf.klein@zib.de>00:39
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun00:54
-!- foulwall [~zhengyang@114.255.40.22] has quit [Quit: leaving]01:07
shogun-buildbotbuild #1761 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1761  blamelist: lambday <heavensdevil6909@gmail.com>, Soeren Sonnenburg <sonne@debian.org>01:43
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has joined #shogun02:18
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/1116882702:18
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has left #shogun []02:18
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]02:25
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has joined #shogun03:00
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/1116888103:00
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has left #shogun []03:00
-!- sonne|osx_ [~sonne@f053039166.adsl.alicedsl.de] has joined #shogun03:23
-!- sonne|osx [~sonne@f053041001.adsl.alicedsl.de] has quit [Ping timeout: 264 seconds]03:26
-!- sonne|osx_ is now known as sonne|osx03:26
-!- pickle27 [~Kevin@199.119.128.114] has quit [Quit: Leaving]03:34
-!- travis-ci [~travis-ci@ec2-54-242-103-128.compute-1.amazonaws.com] has joined #shogun03:35
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/1116895503:35
-!- travis-ci [~travis-ci@ec2-54-242-103-128.compute-1.amazonaws.com] has left #shogun []03:35
-!- travis-ci [~travis-ci@ec2-54-242-103-128.compute-1.amazonaws.com] has joined #shogun04:19
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/1116903304:19
-!- travis-ci [~travis-ci@ec2-54-242-103-128.compute-1.amazonaws.com] has left #shogun []04:19
-!- foulwall [~zhengyang@114.255.40.22] has joined #shogun06:09
-!- sonne|osx [~sonne@f053039166.adsl.alicedsl.de] has quit [Quit: sonne|osx]07:30
-!- besser82 [~besser82@fedora/besser82] has joined #shogun07:48
-!- sonne|osx [~sonne@82.113.121.129] has joined #shogun08:23
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun08:39
shogun-notifier-shogun: Viktor Gal :feature/SerialUTests * ac0c181 / tests/unit/io/SerializationJSON_unittest.cc.jinja2: https://github.com/shogun-toolbox/shogun/commit/ac0c1810c948e0c21907b6959345f37996eaa66108:39
shogun-notifier-shogun: fix typo in SerializationJSON_unittest.cc.jinja208:39
shogun-notifier-shogun: Thoralf Klein :develop * 7cc148a / src/shogun/mathematics/Math.h: https://github.com/shogun-toolbox/shogun/commit/7cc148ac68888a64b043e7321c14bc3d22f8118f08:42
shogun-notifier-shogun: Fixing failed template inference: specialize function template instead of overloading function.08:42
shogun-notifier-shogun: (To avoid complex64_t warnings when using min/max/clamp with non-complex64_t types).08:42
shogun-notifier-shogun: Soeren Sonnenburg :develop * 671248a / src/shogun/mathematics/Math.h: https://github.com/shogun-toolbox/shogun/commit/671248ad601ea0904391068b12f30791902fe29a08:42
shogun-notifier-shogun: Merge pull request #1560 from tklein23/cmath_max_complex64_t08:42
shogun-notifier-shogun:08:42
shogun-notifier-shogun: Fixing failed template inference: specialize function template08:42
shogun-buildbotbuild #50 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/50  blamelist: Soeren Sonnenburg <sonne@debian.org>08:42
-!- sonne|osx [~sonne@82.113.121.129] has quit [Quit: sonne|osx]08:45
shogun-buildbotbuild #2077 of deb1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2077  blamelist: Soeren Sonnenburg <sonne@debian.org>08:53
shogun-buildbotbuild #1687 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1687  blamelist: Thoralf Klein <thoralf.klein@zib.de>09:04
shogun-notifier-shogun: Viktor Gal :feature/SerialUTests * e0f7b92 / src/shogun/latent/LatentModel.cpp: https://github.com/shogun-toolbox/shogun/commit/e0f7b92eef87858c6ecc32ecce80e464795ee58b09:06
shogun-notifier-shogun: Fix parameter name in LatentModel09:06
shogun-notifier-shogun: space are not allowed09:06
shogun-buildbotbuild #51 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/51  blamelist: Thoralf Klein <thoralf.klein@zib.de>09:13
@wikingsonney2k:09:17
@wikingsonney2k: sparse vector writing with HDF5 is unsupported?09:19
sonne|workwiking: no idea might be09:24
shogun-buildbotbuild #1765 of deb3 - modular_interfaces is complete: Failure [failed install test libshogun test python_modular test java_modular test csharp_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1765  blamelist: Thoralf Klein <thoralf.klein@zib.de>09:37
-!- besser82 [~besser82@fedora/besser82] has quit [Remote host closed the connection]09:38
sonne|workwiking: what I know for sure is that hdf5 itself has no representation for sparse09:41
shogun-buildbotbuild #1139 of rpm1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm1%20-%20libshogun/builds/1139  blamelist: Soeren Sonnenburg <sonne@debian.org>09:50
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has joined #shogun09:51
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has quit [Changing host]09:51
-!- besser82 [~besser82@fedora/besser82] has joined #shogun09:51
-!- besser82 [~besser82@fedora/besser82] has quit [Client Quit]09:51
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has joined #shogun09:51
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has quit [Changing host]09:51
-!- besser82 [~besser82@fedora/besser82] has joined #shogun09:51
@wikingarg09:55
@wikingi fucking hate grsec09:55
@wikingsonne|work: do we really need this?09:55
sonne|workwiking: haha :)09:56
@wikingi cannot run any sane stuff on that machine09:56
sonne|workit was the default09:56
@wikingnow it's basically any sanitizer tool by google is not running at all09:56
sonne|workdo whatever you want. for me this is a throwaway scratch machine09:57
@wikingit's for sure some memory protection + address randomization shit09:57
sonne|workwe should back it up though once we are done setting it up09:57
@wikingbut i couldn't find the way to disable it so that the tools starts to work09:57
@wikingsonne|work: ok so we can install the native kernel of debian?09:57
sonne|workwiking: have you seen? buildbot kaboom action with everything failing? Could it be that we are running things twice on the new buildbot and it does some kind of massacre?09:57
sonne|workwiking: no idea how though09:58
sonne|workwiking: whatever you install will not be used IIRC09:58
@wikingbecause?09:58
@wikingwe cannot have our own kernel?09:58
@wikingthat would be really rediciolous09:59
@wikingwell there's grub10:00
@wikingand grub.cfg10:00
-!- besser82 [~besser82@fedora/besser82] has quit [Quit: Verlassend]10:04
sonne|workwiking: no we can but it is netbooted10:06
sonne|workcurrently I mean10:06
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has joined #shogun10:07
-!- besser82 [~besser82@77-22-24-208-dynip.superkabel.de] has quit [Changing host]10:07
-!- besser82 [~besser82@fedora/besser82] has joined #shogun10:07
-!- travis-ci [~travis-ci@ec2-107-20-27-75.compute-1.amazonaws.com] has joined #shogun10:09
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/1118311210:09
-!- travis-ci [~travis-ci@ec2-107-20-27-75.compute-1.amazonaws.com] has left #shogun []10:09
@wikingok10:15
@wikingcan i try a reboot?10:15
@wikinglinux-image-3.2.0-4-amd6410:16
sonne|workwiking: sure I cannot help though10:16
@wikingthat should be fine for us10:16
@wikingok installing10:16
@wikingsonne|work: what happens if it doesn't boot up10:16
@wikingis there like a remote access thing for it?10:16
sonne|workwiking: then I can do sth about it in the evening or tomorrow10:16
@wikingsonne|work: there's no remote access for that machine?10:17
sonne|workno idea10:17
@wikingbtw it's installed and grub.cfg setup10:17
@wikingi highly doubt that this kernel would not boot10:18
sonne|worknever needed it10:18
@wikingsonne|work: ovh.com?10:18
sonne|workyes10:20
@wikingthere is this10:21
@wikinghttp://www.ovh.de/dedicated_server/reboot.xml10:21
@wikingok10:23
@wikingi'll reboot10:23
@wikingcross fingers10:23
@wikingok10:28
@wikingit's back10:28
@wikingLinux ks312251.kimsufi.com 3.10.9-xxxx-grs-ipv6-64 #1 SMP Wed Aug 21 11:51:59 CEST 2013 x86_64 GNU/Linux10:28
@wikinglol10:28
@wikingah ok10:30
@wikingit all makes sense now10:31
@wiking/etc/grub.d/06_OVHkernel10:31
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:691e:24cb:6948:fcc6] has joined #shogun10:33
-!- mode/#shogun [+o iglesiasg] by ChanServ10:33
@iglesiasggood morning10:35
shogun-notifier-shogun: Kevin :develop * 749899e / / (6 files): https://github.com/shogun-toolbox/shogun/commit/749899e0d274c16d399fb1b8b8c90cab6e2c4ed210:35
shogun-notifier-shogun: added fastica algorithm10:35
shogun-notifier-shogun: Kevin :develop * 02cd990 / src/shogun/converter/ica/FastICA.cpp,src/shogun/converter/ica/FastICA.h: https://github.com/shogun-toolbox/shogun/commit/02cd990424927ea6d80e93696efd6db100d36b7410:35
shogun-notifier-shogun: added some more documentation to fastica10:35
shogun-notifier-shogun: Fernando Iglesias :develop * 9347a37 / / (6 files): https://github.com/shogun-toolbox/shogun/commit/9347a37835fc2e227fe02d0d13a69dae4d4b346610:35
shogun-notifier-shogun: Merge pull request #1563 from pickle27/fastica10:35
shogun-notifier-shogun:10:35
shogun-notifier-shogun: added fastica algorithm10:35
shogun-notifier-shogun: Evgeniy Andreev :develop * 073ae61 / / (14 files): https://github.com/shogun-toolbox/shogun/commit/073ae61615b9bb758ec738677a66793a326b74ac10:35
shogun-notifier-shogun: fix the bug related to indexing in a svmlight file format; clean up CSVFile; convert features and some examples for use LibSVMFile10:35
shogun-notifier-shogun: Soeren Sonnenburg :develop * cf3397c / / (14 files): https://github.com/shogun-toolbox/shogun/commit/cf3397cb3a675327bc8146415e8e0c74c557f36d10:35
shogun-notifier-shogun: Merge pull request #1559 from gsomix/feature/libsvm10:35
shogun-notifier-shogun:10:35
shogun-notifier-shogun: Pack of changes related to CSVFile and LibSVMFile10:35
@wikingsonne|work: do you have some admin webpage for this server10:36
@wikingas there we can change the boot type10:36
@wikinghttp://help.ovh.com/KernelNetboot10:37
@wikingsomething like manager v310:37
sonne|workwiking: yes sure but I cannot access it from here10:37
@wikinghttps://www.ovh.de/managerv3/10:37
sonne|workyes10:38
shogun-buildbotbuild #1140 of rpm1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm1%20-%20libshogun/builds/1140  blamelist: Thoralf Klein <thoralf.klein@zib.de>10:38
sonne|workbut don't know my pwd10:38
@wiking:D10:38
@wikingreset it10:38
@wiking:>10:38
@wikingHaben Sie Ihr Passwort verloren?10:39
@wiking:>10:39
@wikingok10:46
@wikingi'll do another reboot10:46
@wikinghere we go10:48
@wikingdo i get an extra star for sysadmining this machine? :)10:50
shogun-buildbotbuild #2078 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/207810:51
@wikingooooh yeaaah10:51
@wiking3.10.10-xxxx-std-ipv6-6410:51
@wiking:D10:51
@wikingno fucking grsec10:51
@wikingaaaaaaand it works!10:52
@wiking5013== WARNING: MemorySanitizer: use-of-uninitialized-value #0 0x7fd2c3100de1 (/root/a.out+0x31de1) #1 0x7fd2c22e2994 (/lib/x86_64-linux-gnu/libc.so.6+0x21994) #2 0x7fd2c3100c3c (/root/a.out+0x31c3c)10:52
@wikingExiting10:52
@wiking==================10:52
@wikingWARNING: ThreadSanitizer: data race (pid=5099)10:52
@wiking  Write of size 4 at 0x7f19b5bee5ac by thread T2:10:52
@wiking    #0 Thread2(void*) /root/simple_race.cc:12 (exe+0x000000060f26)10:52
@wiking  Previous write of size 4 at 0x7f19b5bee5ac by thread T1:10:52
@wiking    #0 Thread1(void*) /root/simple_race.cc:7 (exe+0x000000060eb6)10:52
@wikingyeeeeey10:52
@wikingok briging back the bots10:52
shogun-buildbotbuild #1764 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1764  blamelist: Soeren Sonnenburg <sonne@debian.org>10:55
shogun-buildbotbuild #62 of FCRH - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/62  blamelist: Kevin <kevinhughes27@gmail.com>10:57
shogun-buildbotbuild #1688 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1688  blamelist: Kevin <kevinhughes27@gmail.com>11:00
-!- lisitsyn1 [~lisitsin@mxs.kg.ru] has joined #shogun11:00
sonne|workwiking is playing with his little botnet11:03
sonne|workbuildbotnet that is :P11:03
lisitsyn1hah sonne|work I bet we now have one of the most powerful building infrastructure11:07
sonne|worklisitsyn1: I don't know any better for sure11:07
sonne|workairplane approaching...11:08
sonne|workman that is kind of scary11:08
sonne|workyou sit in a tower11:08
sonne|workand an airplane is coming straigth towards you11:08
-!- Netsplit *.net <-> *.split quits: besser82, foulwall, @sonney2k, shogun-buildbot, hushell11:11
lisitsyn1sonne|work: what airplane? :)11:14
sonne|workthere is some seaplane moving around tourists11:15
sonne|workit is starting in the river spree11:15
sonne|workjust passing the tower I am sitting in when starting and landing11:16
lisitsyn1sonne|work: haha11:18
sonne|workthoralf: hmmhh your patch seems to have broken some test11:19
sonne|workhmmhh all good again ?!?!11:19
-!- Netsplit over, joins: foulwall, @sonney2k, shogun-buildbot, hushell11:20
shogun-buildbotbuild #52 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/52  blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com>11:23
shogun-buildbotbuild #63 of FCRH - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/6311:24
@wikingok tsen analyzer is coming up11:27
-!- travis-ci [~travis-ci@ec2-107-20-27-75.compute-1.amazonaws.com] has joined #shogun11:38
travis-ci[travis-ci] it's Fernando Iglesias's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/1118529111:38
-!- travis-ci [~travis-ci@ec2-107-20-27-75.compute-1.amazonaws.com] has left #shogun []11:38
@wikingah maan this is pain:11:41
@wiking18932 buildsl+  20   0 3094284 2.914g   7148 R  99.9 12.4   1:50.89 cc1plus11:41
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has joined #shogun11:42
van51hello11:42
shogun-notifier-shogun: Viktor Gal :develop * a7153a3 / CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/a7153a3b620ed2bd2d36f033de314f98bdda633c11:46
shogun-notifier-shogun: Add ENABLE_TSEN cmake option for thread-sanitizing11:46
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]11:50
-!- shogun-buildbot_ [~shogun-bu@7nn.de] has joined #shogun11:50
@wikingmmm the buildbot grid got a bit big11:52
shogun-buildbot_build #53 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/53  blamelist: Viktor Gal <viktor.gal@maeth.com>11:53
@wikingshogun-buildbot_: force build --branch=develop 'clang34 - thread analysis'11:54
shogun-buildbot_build #0 forced11:54
shogun-buildbot_I'll give a shout when the build finishes11:54
shogun-buildbot_Hey! build clang34 - thread analysis #0 is complete: Success [build successful]12:00
shogun-buildbot_Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/012:00
sonne|workvan51: good morning!12:01
van51sonne|work: hey :)12:02
van51sonne|work: sorry for being absent12:02
-!- shogun-buildbot_ [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]12:02
sonne|workvan51: is everything ok?12:02
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun12:02
van51sonne|work: yeah, we just had gone to get my grandmother from the village12:02
van51sonne|work: and we kind of liked it there and we got back yesterday :)12:03
van51sonne|work: I did read on random kitchen sinks though and I am trying to see where they fit better in shogun12:03
shogun-buildbotbuild #1767 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/1767  blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin12:04
shogun-buildbot<kevinhughes27@gmail.com>12:04
@wikingoh man12:04
sonne|workvan51: so we have 2 things on the todo right now? the demo and the kitchen sinks right?12:05
@wiking-- A library with LAPACK API not found. Please specify library location.12:05
@wikingon debian sid12:05
@wikingfucking lapack+blas shit12:05
van51sonne|work: yeap12:05
sonne|workwiking: well install them!12:05
@wikingit's installed12:05
@wikingof course... :D12:05
-!- HeikoS [~heiko@nat-168-139.internal.eduroam.ucl.ac.uk] has joined #shogun12:06
-!- mode/#shogun [+o HeikoS] by ChanServ12:06
sonne|workvan51: I was thinking that maybe you should talk to foulwall and get the demo going so you have the parts (that are not easy for you) prepared12:06
@wikingbut this is eeeven better12:06
sonne|workwiking: did locations change?12:06
@wikingmmm12:07
@wikingnoooow it suddenly found it12:07
@wikingwhen i ran it by hand12:07
@wikingWTF?!12:07
van51sonne|work: ok sure! I can start with that12:07
van51foulwall: ping12:07
sonne|workvan51: I mean the part where you get an interface for typing sth and you pass it on12:08
@wikingok there's something wrong defeinitely with sid12:08
sonne|workwiking: sounds like the cyg1 / osx bot behaviour12:08
@wikingsonne|work: :D12:08
@wikingsonne|work: great12:08
@wikingsonne|work: if i use gcc then it cannot detect basically anything...12:09
@wikingsonne|work: neither threads nor lapack...12:09
@wikingsonne|work: but if i use clang then it's all good12:09
@wikingmaaadness12:10
van51sonne|work: yeah.. I'll check the readme again now and some demos to get the basic idea12:10
-!- travis-ci [~travis-ci@ec2-107-20-27-75.compute-1.amazonaws.com] has joined #shogun12:13
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/1118534212:13
-!- travis-ci [~travis-ci@ec2-107-20-27-75.compute-1.amazonaws.com] has left #shogun []12:13
@wikingmmm we have some new unit tests failing12:15
@wiking:)12:15
@wikingok12:16
@wikingthoralf: u broke utests12:16
@wikingsonne|work: should have waited travis...12:21
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has quit [Ping timeout: 248 seconds]12:25
thoralfsonne|work: Hmm.  Let me see.12:26
thoralfwiking: Yes.12:26
thoralf[22:30] <wiking> thoralf: what did u use <-- I didn't understand.12:26
@wikingthoralf: nada that was another question12:26
@wikingthoralf: i have now this problem12:27
@wikingjust a sec regenerating the error12:27
@wikingsuddenly some utests started to fail12:28
@wikinglike BaggingMachine12:29
@wikingunknown file: Failure12:29
@wikingUnknown C++ exception thrown in the test body.12:29
@wiking[  FAILED  ] MajorityVote.binary_combine_vector (0 ms)12:29
@wikingand no idea why12:29
@wikingit's after your commmit12:29
thoralfwiking: Which one?  I had three PRs.12:30
thoralfBut probably this SG_REF-thing?12:30
@wikingthoralf: after all your prs merged12:30
thoralfOkay.12:30
@wikingthoralf: basically MajorityVote and WeightedMajorityVote fails in all utests12:30
@wikingthoralf: ./tests/unit/shogun-unit-test --gtest_filter=*Major*12:31
@wikingall the 8 different tests fails12:31
-!- besser82 [~besser82@fedora/besser82] has joined #shogun12:32
-!- besser82 [~besser82@fedora/besser82] has quit [Client Quit]12:32
shogun-buildbotbuild #1141 of rpm1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm1%20-%20libshogun/builds/114112:35
thoralfwiking: I'm just checking locally.  My system is quite busy, so give me 10 minutes.  But I'm on it.12:35
thoralfwiking: On upstream/develop everything is fine here.12:40
thoralfwiking: Can you give more details?12:40
@wikinghehe12:40
@wikingw812:40
thoralf"Works for me" syndrome ;)12:40
thoralfsonne|work: Which tests were been?  *Major*-Tests as well?12:48
thoralfsonne|work: Btw., I observed something strange which might be a bug.  But I'm not sure what happens there... it's related to base/class_list.cpp12:50
thoralfsonne|work: Sometimes the PT_COMPLEX64 are all "return NULL;", sometimes they're not.12:51
thoralf"case PT_COMPLEX64: return NULL;" vs. "case PT_COMPLEX64: return new CSparseFeatures<complex64_t>();"12:51
@wikingthoralf: haha yeah it's funny because it only fails for me if i enable thread-sanitizing12:59
thoralfwiking: How can I enable this?13:00
@wikingthoralf: u need gcc 4.8 or clang 3.413:00
@wikingif u have that then: cmake -DENABLE_TSEN=ON -DENABLE_TESTING=ON -DENABLE_CCACHE=OFF -DCMAKE_BUILD_TYPE=Debug ..13:00
shogun-buildbotbuild #1768 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/1768  blamelist: Viktor Gal <viktor.gal@maeth.com>13:03
thoralfwiking: Only gcc 4.7 (ubuntu 13.04)13:03
thoralfand clang 3.213:03
@wiking:<13:03
@wikingthoralf: i'll try to find out13:04
@wikingmaybe it's some threadsanitizer bug or something13:04
thoralfwiking: I think it's related to the SG_REF stuff.  Maybe you can step back until it works?13:04
thoralfwiking: Btw., which distro are you using?13:05
@wikingthoralf: sid13:07
thoralfic13:07
* thoralf installing gcc 4.8.113:08
@wikingthoralf: oh clang 3.2 should have it as well13:11
thoralfwiking: no worries - having gcc 4.8.1 now13:12
shogun-buildbotbuild #1766 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1766  blamelist: Kevin <kevinhughes27@gmail.com>13:12
@wikingthoralf: when u ran cmake can u plz copy-paste the output to pastebin or something13:12
shogun-notifier-shogun: Viktor Gal :develop * 3da70f4 / examples/undocumented/python_modular/classifier_multiclass_ecoc_random.py: https://github.com/shogun-toolbox/shogun/commit/3da70f40f02742ba9b20c61bb680d5f14bba2e0313:21
shogun-notifier-shogun: Fix indent in classifier_multiclass_ecoc_random.py13:21
thoralfwith gcc-4.8 linking of discover_gtest_tests fails13:23
thoralflots of unresolved symbols: __tsan*13:24
@wikingdunno really13:24
@wikingstill trying to figure out13:25
@wikingunknown file: Failure13:25
@wikingUnknown C++ exception thrown in the test body.13:25
@wiking[  FAILED  ] SGObject.clone_equals_WeightedMajorityVote (1 ms)13:25
@wikingi'm getting errors like this13:25
@wikingand gdb does not really help13:25
shogun-buildbotbuild #62 of FC19 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20libshogun/builds/62  blamelist: Viktor Gal <viktor.gal@maeth.com>13:31
shogun-buildbotbuild #54 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/54  blamelist: Viktor Gal <viktor.gal@maeth.com>13:31
shogun-buildbotbuild #1691 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1691  blamelist: Viktor Gal <viktor.gal@maeth.com>13:31
shogun-buildbotbuild #1142 of rpm1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm1%20-%20libshogun/builds/1142  blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com>, Viktor Gal <viktor.gal@maeth.com>13:34
@wikingoh interesting :)13:36
thoralfDiscovered magic hidden 9th core: Computing with 8 cores is 8.37 as fast as 1 core?13:41
thoralfOkay, wiking.  Major*-things failing with clang 3.2 as well.13:43
thoralfFATAL: ThreadSanitizer can not mmap the shadow memory (something is mapped at 0x555555554000 < 0x7cf000000000)13:44
thoralfFATAL: Make sure to compile with -fPIE and to link with -pie.13:44
thoralfIs what gdb tells13:44
@wiking:@13:45
@wikingthoralf: echo 2 >/proc/sys/kernel/randomize_va_space13:45
@wikingooor13:46
@wikingif u gdb then13:46
@wikinggdb -ex 'set disable-randomization off' --args ./a.out13:46
thoralfProgram received signal SIGSEGV, Segmentation fault.13:48
thoralf0x00007fd38741d080 in __tsan_write4 ()13:48
thoralf--gtest_filter=MajorityVote.combine_matrix --gtest_break_on_failure13:48
@wiking:>>>13:48
@wikingcool13:48
thoralfAnd a stack trace.13:49
thoralfBut it's all gtest related.13:49
@wiking:>>>mmmm13:49
@wikingi think i know what's the problem13:49
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has joined #shogun13:49
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/1118740713:49
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has left #shogun []13:49
@wiking:S13:49
@wikingthoralf: i've got gmock and gtest instrumented :(13:50
thoralfShouldn't be a problem, though.13:50
@wikingwell who know13:50
@wikings13:50
@wikingbut anyhow13:50
@wikingwe dont want a thread analysis of gmock/gtest13:50
@wikingso i'll add it to the blacklist13:51
@wikingjust a sec13:51
shogun-buildbotbuild #1769 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/1769  blamelist: Viktor Gal <viktor.gal@maeth.com>13:52
shogun-buildbotbuild #1143 of rpm1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/rpm1%20-%20libshogun/builds/114314:10
sonne|workfoulwall: ping?14:12
@wikingsonne|work: only 2 integration tests fails!14:16
sonne|workerrm why does any fail?14:17
sonne|workwiking: they all worked last time I checked14:17
sonne|workwhat fails?14:17
@wikingno they dont14:17
@wikinghttps://travis-ci.org/shogun-toolbox/shogun/jobs/11190100#L438414:18
@wikingwe should just allow an error margin there14:20
sonne|workwiking: -w SGVector<float64> 2 ({-0.2386181320370405}{-0.7375129670701867})14:20
sonne|work+w SGVector<float64> 2 ({-0.2386181320370403}{-0.7375129670701869})14:20
sonne|workso hmmhh14:20
sonne|workhttp://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1768/steps/test%20python_modular/logs/stdio14:21
sonne|workhere they work14:21
thoralfsonne|work: Different compiler options?  -O2, -O3, -ffast-math?14:26
thoralfwiking: a7153a3 is the first bad commit14:29
thoralfwiking: Add ENABLE_TSEN cmake option for thread-sanitizing ;)14:29
sonne|workno idea14:35
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has quit [Quit: Leaving.]14:44
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has joined #shogun14:46
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/1119009614:46
-!- travis-ci [~travis-ci@ec2-204-236-247-16.compute-1.amazonaws.com] has left #shogun []14:46
thoralfwiking: Ping?14:55
thoralfwiking: I rebased with all my commits removed and get the same errors.14:55
@wikingmmm yeah it's some tsen+gmock variation error14:57
-!- tuxa [808337b4@gateway/web/freenode/ip.128.131.55.180] has joined #shogun15:09
-!- tuxa [808337b4@gateway/web/freenode/ip.128.131.55.180] has left #shogun []15:09
thoralfsonne|work: Did you read my observation about class_list.cpp?15:19
@iglesiasglisitsyn1, hi there15:53
@iglesiasglisitsyn1, I need of your Eigen wisdom :)15:53
-!- thoralf [~thoralf@enki.zib.de] has quit [Ping timeout: 256 seconds]16:05
-!- thoralf [~thoralf@enki.zib.de] has joined #shogun16:06
lisitsyn1iglesiasg: yes?16:10
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has joined #shogun16:11
@iglesiasglisitsyn1, I am thinking how LMNN could be speeded up when the feature dimension is large16:17
@iglesiasglisitsyn1, this is the bottleneck right now: https://github.com/shogun-toolbox/shogun/blob/develop/src/shogun/metric/LMNNImpl.cpp#L21916:17
@iglesiasgthose for loops in there16:17
@iglesiasglisitsyn1, they have to do a couple outer products with large vectors (the dx*dx.tranpose() part) each loop per iteration16:18
@iglesiasglisitsyn1, and it seems that it gets costly16:18
@iglesiasglisitsyn1, the Matlab code from LMNN's author does this computations in batch, instead of vector per vector. Have a look http://pastebin.com/S2m1BY4P16:19
@iglesiasglisitsyn1, do you know if this kind of strategy would speed up computations made with Eigen?16:20
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]16:21
thoralfiglesiasg: Looks funny to me, since dx*dxT would be a matrix.  Hope that dx is a row vector ;)16:23
@iglesiasgthoralf, dx*dxT is a matrix, yes16:23
thoralfiglesiasg: But X.col() looks likt column vector?16:23
thoralfiglesiasg: Oh.16:23
@iglesiasgthoralf, outer product16:23
thoralfiglesiasg: That's bad. :D16:23
@iglesiasgthoralf, hehe well it is the gradient of a matrix16:24
lisitsyn1iglesiasg: in a minute16:24
foulwallping sonne|work16:24
@iglesiasglisitsyn1, sure, no rush16:25
thoralfiglesiasg: No way to tune O(n^2) to be less than O(n^2), since you at least need to read/write O(n^2) memory. ;)16:25
foulwallping van5116:25
lisitsyn1iglesiasg: this is called rank update16:25
@iglesiasgthoralf, but you can make it faster of course16:25
lisitsyn1iglesiasg: G must be symmetric, right?16:26
van51hey foulwall16:26
@iglesiasglisitsyn1, let me think of it16:26
foulwallHey van51 :)16:26
van51foulwall: I'll probably need your help developing a language detection demo :)16:26
lisitsyn1iglesiasg: G.selfadjointView<Upper>().rankUpdate(dx1, dx1);16:27
lisitsyn1try that16:27
foulwallOk van51 , are there any existing examples on that?16:27
@iglesiasglisitsyn1, that should make it faster thatn making the product by hand?16:28
lisitsyn1iglesiasg: yes I believe so16:28
van51foulwall: in python specifically, right?16:28
@iglesiasglisitsyn1, I will try asap16:28
@iglesiasglisitsyn1, thanks!16:28
foulwallvan51: sure16:28
van51foulwall: if so, then not yet16:28
lisitsyn1iglesiasg: but do not forget to not use raw G afterwards16:28
lisitsyn1iglesiasg: use selfadjointView16:28
van51foulwall: but still the model training takes some time so I think I will just be de-serializing it for the demo16:29
lisitsyn1if it is not symmetric that won't work16:29
@iglesiasglisitsyn1, aham I see16:29
@iglesiasglisitsyn1, I have to make sure it is symmetric first then16:29
van51foulwall: let me get a bit more into the framework and I'll bug you later with solid questions :)16:30
@iglesiasglisitsyn1, I think you must be right and the matrix must be symmetric16:32
foulwallvan51: ok, just leave the js/css/html work to me.16:32
@iglesiasglisitsyn1, since it is just composed of sums of outer products16:32
lisitsyn1iglesiasg: if it is only rank updated16:32
lisitsyn1then it is16:33
@iglesiasglisitsyn1, I was checking its initialization to make sure of it. But yeah, outer products to initialize it as well16:34
van51foulwall: ah if that's the case then I can tell you what I had in mind for the interface16:36
van51foulwall: but if you have other stuff to do I 'll manage I think :)16:37
@iglesiasglisitsyn1, but what about what you told me of not using it raw16:37
lisitsyn1iglesiasg: it will be corrupted after that16:37
lisitsyn1iglesiasg: only Upper or Lower (according to template parameter) part is stored16:37
foulwallok van51, I planed to do the demos for gsocers this week, so just leave it to me:)16:40
van51foulwall: cool :D16:42
van51foulwall: so all I have to do is the python handler in demos/ right?16:42
foulwallvan51: right16:43
foulwallvan51: :D16:43
van51foulwall: do you want me to tell you anything about the input I'll be expecting or will wediscuss it at another time?16:44
foulwallvan51: language detection demo, is it a textbox and input some text and some arguments and the server do the detect and response with the result?16:46
foulwallvan51: like this http://nn.7nn.de:8000/ocr/entrance ? but with text input?16:47
van51foulwall:  yes exactly :)16:47
foulwallvan51: That's easy to do, just tell me how to do the detect (or give me the ipynb) and that's all I need16:50
@wikingaaaaaaasdf16:53
@wikinglisitsyn1: soon u should fire a doodle for the next dev meeting16:54
@wikingplz16:54
van51foulwall: ok, I 'll get back to you with that16:54
van51foulwall: thanks :)16:54
lisitsyn1wiking: ohh thanks for reminder16:55
lisitsyn1wiking: I am a bit overwhelmed, never going to take additional job while having a full-time job :D16:57
@iglesiasglisitsyn1, oh, are you with two jobs atm?16:59
lisitsyn1iglesiasg: second one is just kind of freelance and I am approaching all deadlines17:01
@iglesiasglisitsyn1, aham I see. The video detection thing?17:01
lisitsyn1iglesiasg: but it is CV, I believe need experience in CV17:01
lisitsyn1iglesiasg: yes17:01
@iglesiasglisitsyn1, nice, I wish you good luck with it!17:01
lisitsyn1iglesiasg: thanks17:02
lisitsyn1wiking: have you already tried that -Os?17:04
lisitsyn1or how it is called17:04
lisitsyn1that super fast mode of GCC17:04
@wikinglisitsyn1: oh no ... i think Os is size no?17:07
lisitsyn1wiking: let me google17:08
@wikingyes17:08
@wikinglisitsyn1: according to gcc man17:08
lisitsyn1wiking: -Og!17:08
@wiking-Os Optimize for size, but not at the expense of speed.17:08
@wikingmmm my gcc man doesn't have that flag :P17:08
lisitsyn1wiking: since 4.817:08
lisitsyn1wiking: ;)17:08
@wikingah i see17:09
lisitsyn1wiking: way too new I guess17:09
@wikinghttps://travis-ci.org/shogun-toolbox/shogun.png17:10
@wikingheh17:10
@wiking:)17:10
lisitsyn1wiking: yes there is a pic17:10
lisitsyn1for any project17:10
lisitsyn1wiking: you know crazy thoughts come into my mind sometimes17:11
@wikinglike git rm * && git commit . && git push17:11
@wiking?: D17:11
lisitsyn1wiking: how crazy would you say I am17:11
@wiking:>17:11
lisitsyn1if I said I'd like to reduce LoC to 50k17:11
lisitsyn1:D17:11
lisitsyn1well not that release17:12
lisitsyn1just some random thought17:12
@wikingheheh17:12
@wikinghave fun with it17:12
@wiking:)17:12
@wikingi would rather start getting rid of some custom created stuff17:12
@wikinglike refCount17:13
lisitsyn1wiking: yes17:13
lisitsyn1wiking: that's a part of the great plan17:13
lisitsyn1;)17:13
lisitsyn1wiking: you can't get to 50k with custom stuff17:13
@wikingindeed17:13
@wikingbut as of first17:13
@wikingwe need to get all the static analyzer + valgrind + buildbot errors down to 017:13
lisitsyn1wiking: yes that's for this release17:14
lisitsyn1I am talking about next one probably17:14
@wikingyeah otherwise17:14
@wikingyeah clear up shit17:14
lisitsyn1I wouldn't suggest to do that now17:14
@wikinggit flow feature start ThrowShitOut17:14
@wiking;)17:14
lisitsyn1wiking: I had a chance to try my keywords lib in action17:14
lisitsyn1wiking: I changed some things but it is pretty ok17:15
@wikingthis release should be called bB[C[C[C[C[C[CBunyloated17:15
@wiking:D17:15
@wikingi mean BloatedBunny17:15
lisitsyn1wiking: :D17:15
lisitsyn1wiking: ah we need a name17:16
lisitsyn1okay putting that to agenda ;)17:16
@wiking:>17:17
lisitsyn1wiking: I've been thinking about one more thing too17:19
-!- sonne|osx [~sonne@89.204.138.138] has joined #shogun17:19
lisitsyn1wiking: people afraid of gpl you know17:19
lisitsyn1wiking: and it is no way to use 130mb library when you have 1mb application17:20
sonne|osxlisitsyn1: 250!17:20
lisitsyn1sonne|osx: oh yeah17:20
sonne|osxlisitsyn1: we are bigger than firefox will ever be17:20
lisitsyn1sonne|osx: wiking: so my idea (that appeared from my needs)17:20
lisitsyn1is to separate 'apply' module17:20
lisitsyn1with no training17:20
lisitsyn1and BSD/MIT it or whatever17:21
thoralf250MB?  Do you have any traffic limitations? ;)17:21
sonne|osxI didn't follow the discussion so no idea what is going on here17:21
sonne|osxthoralf: no of course not :D17:21
lisitsyn1sonne|osx: it started exactly when you joined the room17:21
thoralfsonne|osx: Phew ;)17:21
lisitsyn1sonne|osx: I mean the idea17:21
sonne|osxthoralf: is the failure fixed or did it reappear?17:22
lisitsyn1sonne|osx: see: now I have 1mb application17:22
lisitsyn1I need machine learning there17:22
sonne|osxlisitsyn1: use shogun for learning for applying write your own17:22
sonne|osxthat's what I do17:22
thoralfsonne|osx: The only failure I could reproduced is about thread sanitizing.17:23
thoralfsonne|osx: Which one did you see?17:23
sonne|osxapply is usually 100LoC but the rest is waaah17:23
lisitsyn1sonne|osx: but can't we have compact fast and non-gpl 'apply' module?17:23
thoralfsonne|osx: MajorVote?17:23
sonne|osxthoralf: I meant this one http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2077/steps/test/logs/stdio17:24
sonne|osxthoralf: build 2077 was failing but 2076 was still fine http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/207717:24
sonne|osxthoralf: but it didn't re-appear...17:24
thoralfsonne|osx: Hmm.  It's a numerical issue - shouldn't be related.17:25
sonne|osxthoralf: it is the only change isn't it?17:25
sonne|osxhmmhh17:25
thoralfsonne|osx: Talking about /home/buildslave/deb1_-_libshogun/build/tests/unit/features/StreamingHashedDocDotFeatures_unittest.cc:143: Failure, right?17:25
sonne|osxthoralf: yes17:25
-!- HeikoS [~heiko@nat-168-139.internal.eduroam.ucl.ac.uk] has quit [Remote host closed the connection]17:25
sonne|osxvan51, thoralf or is this test running with multiple threads?17:25
sonne|osxso we have a Heisenbug?17:26
thoralf  Actual: 4.86341e+1517:26
thoralfWhich is: 4.86341e+1517:26
thoralfLooks like numerical stuff.17:26
thoralfHeisenbugs are more funny. ;)17:26
lisitsyn1thoralf: wat is the difference17:26
lisitsyn1am I blind :D17:26
sonne|osxlisitsyn1: rounding17:27
thoralflisitsyn1: Something < E+10?17:27
lisitsyn1I see17:27
lisitsyn1but that's beyond significant digits already17:27
sonne|osxrounding in gtest sucks17:27
@wikingthoralf: expect_double_eq?17:27
sonne|osxeither results are same or not17:27
lisitsyn1sonne|osx: exceptions do as well17:27
lisitsyn1sonne|osx: I spend a few years trying to check what wrong exception is thrown17:28
sonne|osxwiking: could you try to remove the external json dir for the bsd1 buildbot prior to building?17:28
lisitsyn1sonne|osx: ahh and other crazy idea I had17:28
lisitsyn1sonne|osx: 50K LoC or whatever17:28
sonne|osxlisitsyn1: ??17:29
lisitsyn1sonne|osx: we reduce to 100k or 50k next release17:29
sonne|osxlisitsyn1: doing what?17:29
lisitsyn1with same functionality17:29
thoralfsonne|work: Btw., I observed something strange which might be a bug.  But I'm not sure what happens there... it's related to base/class_list.cpp17:29
thoralfsonne|work: Sometimes the PT_COMPLEX64 are all "return NULL;", sometimes they're not.17:29
lisitsyn1sonne|osx: reduce LoC with proper code17:29
thoralf"case PT_COMPLEX64: return NULL;" vs. "case PT_COMPLEX64: return new CSparseFeatures<complex64_t>();"17:29
thoralfsonne|osx: ^ I wrote it a few hours ago.  Any idea?17:30
sonne|osxthoralf: looks like sth with complex64 is not supported17:30
thoralfIt's toggling on the same system.17:30
thoralfNo updates.17:30
thoralfDoesn't yield to errors, but looks suspicious.17:31
thoralfwiking: Btw., class_list.cpp will be generated in the source tree, not in the build directory.  You're aware of that?17:32
@wikingintentional17:32
sonne|osxthoralf, woking: ohh ok in that test there is indeed a expect_eq on floats and there is a mixture of float32 and 64 so that might explain it17:35
-!- lisitsyn1 [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.]17:35
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun17:36
shogun-notifier-shogun: Soeren Sonnenburg :develop * ddbe5f1 / tests/unit/features/StreamingHashedDocDotFeatures_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/ddbe5f1d08341ce2369d937170cc4e2bdb80dcee17:36
shogun-notifier-shogun: use expect_near instead of eq for double/float17:36
shogun-notifier-shogun: Soeren Sonnenburg :develop * 1f0a927 / src/shogun/lib/external/gpm.cpp: https://github.com/shogun-toolbox/shogun/commit/1f0a927837e8097549e22a255c527b9d6336bd3e17:37
shogun-notifier-shogun: fix clang warning17:37
shogun-notifier-shogun: Soeren Sonnenburg :develop * 3f14cbf / examples/undocumented/python_modular/ (3 files): https://github.com/shogun-toolbox/shogun/commit/3f14cbf56b2c7356ed8d1fb37d15255a16c5905817:37
shogun-notifier-shogun: convert more examples to use CSVFile17:37
sonne|osxwiking: is slave1 fatbot? http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/176917:38
sonne|osxwiking: it keeps failing weirdly on make install17:38
@wikingyeah i think we need to set some chmods17:38
sonne|osxwiking: should you please chown buildslave /usr/lib/x86_64-linux-gnu/octave/site/oct/api-v48+/x86_64-pc-linux-gnu -R17:41
sonne|osxwiking: that should fix it17:41
sonne|osxalright17:41
* sonne|osx off17:41
-!- sonne|osx [~sonne@89.204.138.138] has quit [Quit: sonne|osx]17:41
shogun-buildbotbuild #1692 of bsd1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/169217:42
shogun-buildbotbuild #63 of FC19 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20libshogun/builds/6317:45
shogun-buildbotbuild #55 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/55  blamelist: Soeren Sonnenburg <sonne@debian.org>17:45
-!- pickle27 [~Kevin@199.119.128.114] has joined #shogun17:56
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has quit [Quit: Leaving.]18:00
@wikingooh maaan this sanitizer shit needs more hacking18:08
@wikingbut let's see if it was worth it... :P18:10
shogun-buildbotbuild #1770 of deb3 - modular_interfaces is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/177018:13
shogun-buildbotbuild #1771 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/1771  blamelist: Soeren Sonnenburg <sonne@debian.org>18:13
@wiking/usr/bin/ld: /home/buildslave/clang34_-_thread_analysis/build/third_party/libs/gmock/libgmock.a(gtest-all.cc.o): relocation R_X86_64_32S against `.rodata.str1.1' can not be used when making a shared object; recompile with -fPIC18:14
@wikingyeeey18:14
-!- sonne|osx [~sonne@f053039166.adsl.alicedsl.de] has joined #shogun18:17
-!- sonne|osx [~sonne@f053039166.adsl.alicedsl.de] has quit [Client Quit]18:18
thoralfwiking: Already had this ;)18:20
thoralfwiking: Recompile gmock :)18:20
@wikingthoralf: imo it should really be compiled with fPIC :P18:21
shogun-buildbotbuild #56 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/56  blamelist: Soeren Sonnenburg <sonne@debian.org>18:31
-!- vgorbati [~vgorbati@91.216.173.29] has joined #shogun18:32
@wikingwtf is this /usr/bin/ld: /usr/lib/debug/usr/lib/x86_64-linux-gnu/Scrt1.o(.debug_info): relocation 9 has invalid symbol index 1318:37
-!- sonne|osx [~sonne@f053039166.adsl.alicedsl.de] has joined #shogun18:38
@wikingsonne|osx: do u know wtf is this18:39
@wiking/usr/lib/debug/usr/lib/x86_64-linux-gnu/Scrt1.o(.debug_info): relocation 9 has invalid symbol index18:39
thoralfiglesiasg: StructuredLabels::set_labels() does something different than the documentation tells18:41
@iglesiasgthoralf, let me check18:41
thoralfiglesiasg: It inserts new labels18:41
thoralfiglesiasg: Which means num_labels+=118:41
@iglesiasgthoralf, StructuredLabels::set_label() you mean?18:42
@iglesiasgsingular18:42
thoralfiglesiasg: Yes, sorry.18:42
@iglesiasgthoralf, just to be sure of it :)18:42
@iglesiasgthoralf, mmm are you sure?18:43
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has joined #shogun18:43
thoralfiglesiasg: Yes.  It inserts labels instead of substituting.18:43
@iglesiasgthoralf, mm ok, let me check insert_element in dynamic object array18:43
thoralfiglesiasg: set_element() would be correct.18:44
@wikinglisitsyn: did u try: http://clang-omp.github.io/18:44
@wiking?18:44
lisitsynwiking: no18:44
lisitsynwiking: should be in trunk soon, right?18:44
@wikinglisitsyn: apparently yes18:44
@wikingoh man wtf is this shiat18:44
@wikingnow it's fPIC missing18:45
@wikingbefore it was that weird Scrt error18:45
@iglesiasgthoralf, yes, I think you are entirely right18:46
@iglesiasgthoralf, do you want to fix it?18:47
thoralfiglesiasg: Not yet.  set_element() doesn't do what I thought - how to replace an element?18:47
thoralfI don't want to shift ;)18:47
@iglesiasgthoralf, you can use set_element and check the index given to set_label18:48
@iglesiasgthoralf, if it is out of the current number of labels, then raise error, otherwise set18:49
thoralfiglesiasg: Sorry, off-by-one.  It's alright.18:51
thoralfiglesiasg: Will manage to replace an array element now. ;)18:51
@iglesiasgthoralf, off-by-one?18:51
thoralfiglesiasg: Yeah, set_element(2) is off-by-one than set_element(1).  Explains why it didn't work :D18:52
@iglesiasgthoralf, aaah ok :)18:52
@iglesiasgthoralf, thank you for pointing the bug out. You are an expert on finding these things! Impressive18:53
thoralfiglesiasg: I'd like to avoid. ;)18:53
@iglesiasgthoralf, avoid bugs or avoid finding them? :P18:53
thoralfiglesiasg: either :)18:54
@iglesiasghaha18:54
-!- travis-ci [~travis-ci@ec2-54-227-133-66.compute-1.amazonaws.com] has joined #shogun18:54
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/1120036818:54
-!- travis-ci [~travis-ci@ec2-54-227-133-66.compute-1.amazonaws.com] has left #shogun []18:54
thoralfiglesiasg: But, pleeeaaasse, write fucking-please  more tests. :D18:54
@iglesiasgthoralf, yeeeah... all my structure thing is without tests18:55
@iglesiasgthoralf, have focused on LMNN since we started with tests18:55
shogun-buildbotbuild #1772 of deb3 - modular_interfaces is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/177219:00
@wikingwoah modular interface actually fucking success19:15
* wiking wonders why static analysis is not started...19:16
@wikingdoh this wont work :(19:20
@wikingoh lOOOOL19:23
@wikinghttp://gcc.gnu.org/bugzilla/show_bug.cgi?id=4305219:23
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:691e:24cb:6948:fcc6] has quit [Quit: Ex-Chat]19:31
@sonney2kwiking, have you seen almost all buildslaves green19:31
@sonney2konly the FC* stuff and your static analysis not yet19:31
@wikingsonney2k: i've put a dependency for modular on static analysis19:37
@wikingbut i dont understand why is it not fired19:37
@wikingah ok i know19:38
@wikingsince FC is failing19:38
@wikingbut yeah static analysis will fail19:38
@wikingas there's like 100 errors that needs to be fixed :)19:38
@wikinghttp://buildbot.shogun-toolbox.org/static_analysis/2013-09-09-1/19:39
@wikingshogun-buildbot: force build --branch=develop 'clang34 - static analysis'19:39
shogun-buildbotbuild #9 forced19:39
shogun-buildbotI'll give a shout when the build finishes19:39
thoralfwiking: Yeah, cool stats.19:39
-!- travis-ci [~travis-ci@ec2-54-227-133-66.compute-1.amazonaws.com] has joined #shogun19:40
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/1120040119:40
-!- travis-ci [~travis-ci@ec2-54-227-133-66.compute-1.amazonaws.com] has left #shogun []19:40
@sonney2kwiking, but these are the only shogun ones right? The FC* bots are just buggy19:41
@wikingsonney2k: yesh fc19 modular has to be fixed19:41
@wikingbut that means that i have to slice up the master.cfg19:41
@wikingas its already too hectic19:41
@wikingsonney2k: ah and we have tons of valgrind problesm ;)19:41
@wikingand i haven't even merged this: SerialUTest branch19:42
@wikingwhich will bring some other unit test down19:42
@wikingbut i've tried to put together a thread analyzer bot19:42
@wikingbut seems like an impossible mission with thread-sanitizer19:42
thoralfwiking: Yeah, but most of the valgrind things are wrong tests - mixing up array indices, setting up too short stuff, not initializing.19:42
@wikingthoralf: still .... needs to be fixed :)19:43
thoralfwiking: Indeed.19:43
@sonney2kwiking, it is not earth shattering19:43
@wikingyeah i know19:43
@sonney2kthis is quality wise much much better than previously19:43
@sonney2kand with that buildbotnet in the background things will continue to improve19:44
@wikingthoralf: so what did u use last time for thread analyzing? drd?19:44
thoralfwiking: drd and helgrind, I think helgrind was more useful.19:46
thoralfwiking: Detecting possible races.19:46
shogun-buildbotbuild #9 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/920:02
@wikinglisitsyn: here?20:08
lisitsynwiking: yes20:08
@wikinglisitsyn: http://pastebin.com/03U111vJ20:08
@wikingdoes this make any sense to u/20:09
@wiking?20:09
lisitsynwiking: not really20:09
@wikingso it's a bug u say?20:10
lisitsynwiking: well that's just insertion to map20:11
lisitsynwiking: so it is just std::map20:12
@wikingyeah i get that20:12
@wikingwell anyhow it was just a test... :)20:14
thoralfAargh.  Objections against removing CLabels::get_label_type() and labels/LabelTypes.h?20:17
thoralfIt's only used in LabelsFactory and almost useless when knowing dynamic_cast20:18
-!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has quit [Ping timeout: 245 seconds]20:21
-!- vgorbati [~vgorbati@91.216.173.29] has quit [Quit: vgorbati]20:21
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 246 seconds]20:22
@wikingshogun-buildbot: force build --branch=developer 'clang34 - thread analysis'20:22
shogun-buildbotbuild forced [ETA 6m14s]20:22
shogun-buildbotI'll give a shout when the build finishes20:22
shogun-buildbotbuild #1 of clang34 - thread analysis is complete: Failure [failed git]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/120:22
@wikingshogun-buildbot: force build --branch=develop 'clang34 - thread analysis'20:23
shogun-buildbotbuild forced [ETA 6m14s]20:23
shogun-buildbotI'll give a shout when the build finishes20:23
shogun-buildbotbuild #2 of clang34 - thread analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/220:25
-!- vgorbati [~vgorbati@91.216.173.29] has joined #shogun20:31
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]20:37
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has joined #shogun21:00
-!- vgorbati [~vgorbati@91.216.173.29] has quit [Quit: vgorbati]21:23
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun21:30
-!- mode/#shogun [+o iglesiasg] by ChanServ21:31
@wikingthoralf: here?21:46
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun21:53
shogun-notifier-shogun: Thoralf Klein :develop * 53440c6 / / (2 files): https://github.com/shogun-toolbox/shogun/commit/53440c62e07ebf2949575c7e53ce9491d7b6791321:53
shogun-notifier-shogun: Fixed bug in StructuredLabels - using set_element() instead of insert_element().  Added unit test.21:53
shogun-notifier-shogun: Thoralf Klein :develop * 506f5cf / tests/unit/labels/StructuredLabels_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/506f5cfece757b20dd675aee3d004c4ce86c839421:53
shogun-notifier-shogun: Freeing memory in StructuredLabels unit test.21:53
shogun-notifier-shogun: Fernando Iglesias :develop * 9b4bac2 / / (2 files): https://github.com/shogun-toolbox/shogun/commit/9b4bac28d7dc472040020118f42c888f2f53ed4c21:53
shogun-notifier-shogun: Merge pull request #1568 from tklein23/struct_labels_bug21:53
shogun-notifier-shogun:21:53
shogun-notifier-shogun: Fixed bug in StructuredLabels21:53
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has joined #shogun21:57
@iglesiasgwiking, ping21:57
@wikingiglesiasg: pong21:58
@iglesiasgwiking, what do you think about getting back static interfaces?21:58
@iglesiasgbefore release21:58
@wikingiglesiasg: they are back21:58
@iglesiasgfuck, I am always missing important stuff :'(21:58
@wikingnw21:59
@wikingok i remember there was a discussion about properly init-ing class parameters... and registering21:59
@wikingwhat was the right order?21:59
@iglesiasgSG_ADD and assigning default values?21:59
@wikingi mean i have thi sclass and i know that there's somewthing wrong with init+register param22:00
@wikingbut i dont know what22:00
shogun-buildbotbuild #1695 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1695  blamelist: Thoralf Klein <thoralf.klein@zib.de>22:00
@iglesiasgmm I think I normally SG_ADD and then initialize22:00
@iglesiasgbut I thought it didn't make a difference22:00
@wikinghttps://github.com/shogun-toolbox/shogun/blob/develop/src/shogun/ensemble/WeightedMajorityVote.cpp#L2122:01
@iglesiasgwiking, yep. try the other way round maybe22:01
@wikingiglesiasg: same...22:01
@iglesiasgmm22:01
@wikingi mean i dont really know wtf is the problem... but22:02
@iglesiasgwhat's the error?22:02
@wikingsince threadsanitizer is failing on this22:02
@wikingw/o real error22:02
shogun-buildbotbuild #57 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/57  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Thoralf Klein <thoralf.klein@zib.de>22:02
@iglesiasgwiking, what about doing m_weights=SGVector<>() instead of the resize??22:02
@iglesiasgI guess it shouldn't make a difference but...22:03
@wikingthe init fucks it up for sure22:03
shogun-buildbotbuild #1696 of bsd1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/169622:04
@wikingwoah man22:04
@wikingthread sanitizer actually FUCKING WORKS22:04
@wiking:DDDD22:04
@wikingthoralf: where r u? :)))22:04
@iglesiasgwiking, after the change to the constructor?22:05
@wikingyeah22:05
@iglesiasgcool22:05
@wikingiglesiasg: the problem is that errors are not being written at all22:05
@iglesiasgaham22:06
@wikingso it dies with an error22:06
@wikingbut no mention about wtf is the error22:06
@iglesiasgwiking, aham22:06
@iglesiasgwiking, actually I think that SGVector and SGMatrix do not need to be initialized at all22:06
@iglesiasgdefault constructor is called automatically as they are objects and no primitive data type22:06
@wikingmmm the only thing i could think about22:07
@wikingthat actually the errors are supressed22:07
@wikingby gtest22:07
@wikingi'd be very much interested why this error only comes up when threadsanitizer is running22:19
@wikingthoralf: if u r around would be great22:20
shogun-buildbotbuild #1774 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/1774  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>22:27
shogun-buildbotbuild #71 of FCRH - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/FCRH%20-%20libshogun/builds/71  blamelist: Thoralf Klein <thoralf.klein@zib.de>22:49
shogun-buildbotbuild #58 of FC19 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/58  blamelist: Thoralf Klein <thoralf.klein@zib.de>22:49
shogun-buildbotbuild #1775 of deb3 - modular_interfaces is complete: Failure [failed test libshogun test ruby_modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1775  blamelist: Thoralf Klein <thoralf.klein@zib.de>23:38
--- Log closed Wed Sep 11 00:00:54 2013
« Monday, 2013-09-09 Index Wednesday, 2013-09-11 »

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