Open in new window / Try shogun cloud
--- Log opened Thu Jan 02 00:00:33 2014
shogun-buildbotbuild #157 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at  blamelist: Soeren Sonnenburg <>00:05
shogun-buildbotbuild #172 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at  blamelist: Soeren Sonnenburg <>00:48
-!- new_lido [~walid@] has quit [Ping timeout: 272 seconds]01:13
-!- new_lido [~walid@] has joined #shogun01:18
-!- new_lido [~walid@] has quit [Ping timeout: 245 seconds]01:24
-!- new_lido [~walid@] has joined #shogun01:25
-!- new_lido [~walid@] has quit [Ping timeout: 240 seconds]01:46
-!- new_lido [~walid@] has joined #shogun01:46
-!- new_lido [~walid@] has quit [Ping timeout: 246 seconds]01:51
-!- new_lido [~walid@] has joined #shogun02:03
-!- new_lido [~walid@] has quit [Ping timeout: 260 seconds]02:25
-!- new_lido [~walid@] has joined #shogun02:36
-!- new_lido [~walid@] has quit [Ping timeout: 245 seconds]02:41
-!- sonne|osx_ [] has joined #shogun03:04
-!- sonne|osx [] has quit [Ping timeout: 245 seconds]03:05
-!- sonne|osx_ is now known as sonne|osx03:05
shogun-buildbotbuild #668 of nightly_default is complete: Failure [failed notebooks]  Build details are at
-!- new_lido [~walid@] has joined #shogun04:50
-!- new_lido [~walid@] has quit [Ping timeout: 272 seconds]04:54
-!- new_lido [~walid@] has joined #shogun04:57
-!- new_lido [~walid@] has quit [Ping timeout: 265 seconds]05:12
-!- new_lido [~walid@] has joined #shogun05:26
-!- new_lido [~walid@] has quit [Ping timeout: 252 seconds]05:31
-!- new_lido [~walid@] has joined #shogun05:48
-!- new_lido [~walid@] has quit [Ping timeout: 260 seconds]05:54
-!- lambday [67157c37@gateway/web/freenode/ip.] has joined #shogun06:20
lambdaysonne|osx: moin moin!06:21
lambdayI was using the same machine that I used for gsoc.. x86_64 kernel with gcc 4.6 :(06:22
lambdaycompiled it in x86_64 fedora 19... gcc 4.8 and it worked fine06:22
-!- lambday [67157c37@gateway/web/freenode/ip.] has quit [Ping timeout: 272 seconds]06:40
-!- sonne|osx [] has quit [Quit: sonne|osx]06:51
-!- lambday [67157f37@gateway/web/freenode/ip.] has joined #shogun07:15
-!- new_lido [~walid@] has joined #shogun07:32
-!- new_lido [~walid@] has quit [Quit: Leaving]08:17
-!- sonne|work [] has joined #shogun09:01
sonne|worklisitsyn: now you on vacations and me busy!09:02
lisitsynsonne|work: that's the balance09:38
sonne|worklisitsyn: you mean ying and yang, tom and jerry etc?10:35
lisitsynsonne|work: yes10:37
lisitsynsonne|work: okay %deity bless me I started10:37
sonne|workgo go :D10:56
-!- zxtx_ [] has quit [Ping timeout: 246 seconds]13:29
-!- zxtx_ [] has joined #shogun14:51
sonne|workbesser82: is fedora using /bin/bash as $SHELL?15:31
sonne|workor dash like the others?15:32
besser82sonne|work:  /bin/sh --->  /bin/bash  like ALL rpm-based distros  ;)15:32
sonne|workyeah... rpmbuild is full of bashisms and of course fails creating rpms when /bin/sh is dash15:33
besser82sonne|work:  yes, rpm* heavily depends on having bash  as default sh15:33
besser82sonne|work:  looks like a bug in debian then....15:34
sonne|workwhy debian? shouldn't rpmbuild call /bin/bash ?15:34
besser82sonne|work:  that depends on the options it was configured during build....15:35
besser82sonne|work:  on fedora it is configured using /bin/bash as shell15:35
besser82sonne|work:  on debian it uses /bin/sh  ;)15:35
sonne|workahh ok15:39
@wikingbesser82: /usr/bin/env bash15:43
@wikingbesser82: as that's really only for linux systems15:43
@wiking /bin/bash and alike15:44
@wiking /usr/bin/env bash should be consistent regardless of a distro15:45
besser82wiking:  ???15:47
@wikingbesser82: /bin/bash -> /usr/bin/env bash15:47
besser82wiking:  using /usr/bin/env might be somewhat problematic, I'm afraid15:48
@wikingon what distro?15:48
besser82wiking:  at least by means of rpmbuild15:48
besser82wiking:  basically all15:48
@wikingso fedora just doesn't support /usr/bin/env bash or what?15:48
besser82wiking:  on fedora there are guidelines for NOT using /usr/bin/env in hashbang and all15:49
@wikingmmm interesting15:49
@wikingso in fedora it's a guideline to not to be portable?15:49
besser82wiking:  No, but since /usr/bin/env spawns a new shell  with copying the env instead of using profile.d defaults  ;)15:50
@wikingwell actually yes15:50
besser82wiking:  so the resulting env might not be prestine and can cause problems  ;)15:50
@wikingif it's a guideline not to go with /usr/bin/env15:50
@wikingthen it actually implies that that shell script will not be portable15:51
besser82wiking:  portable by means of?15:51
@wikingbesser82: to be used on any distro where /bin/bash does not exist15:52
besser82wiking:  basically any distro I know has the shells in /bin15:52
@wikingbesser82: sh and csh15:52
@wikingbesser82: but not bash15:52
@wikingmaybe tcsh as well15:52
besser82wiking:  who ships bash in /usr/bin?15:53
@wikingbesser82: you would be amazed15:53
@wikingbesser82: even /usr/local/bin15:53
@wikingbesser82: bash is by far not a widespread shell15:53
@wikingonly linux distros has it15:53
@wikingbut not even all of it15:53
@wikingas default15:53
besser82wiking:  yes, but why should be a rpm for fedora be portable to let's say netBSD?!?15:54
@wikingso this guideline is purely about rpm generation?15:55
besser82wiking:  there is /usr/bin/env on fedora, too;  But when packagaging software in rpm one is required to change the hashbang from /usr/bin/env to the interpreter bin directly  ;)15:55
besser82wiking:  yes15:55
@wikingmea culpa15:56
besser82wiking:  np  ;)15:56
besser82wiking:  still look for a job?16:06
besser82wiking:  you're on SF Bay, right?16:07
@wikingheheh no16:09
@wikingbut thnx16:09
besser82wiking:  :D16:09
besser82wiking:  no, job or not SF bay?16:09
@wikingnot looking for a job atm16:09
@wikingor let's say i'm not looking for it that intensively16:09
@wikingmaybe if i really like something then16:10
@wikingbut thnx anyways16:10
besser82wiking:  Allrighty :D16:10
besser82lisitsyn:  ^^^^16:10
besser82lambday:  ^^^^16:11
@wikingbtw boston is really not a nice place ...16:11
besser82wiking:  the look for Mountain View CA, too  ;)16:11
@wikingor at least personally i did not like it16:11
besser82sonne|work:  ^^^^  do you know someone for that job? (big-data stuff)16:13
-!- lambday [67157f37@gateway/web/freenode/ip.] has quit [Ping timeout: 272 seconds]16:14
-!- lambday [67157c4d@gateway/web/freenode/ip.] has joined #shogun16:14
besser82lambday:  ^^^^16:15
lambdaybesser82: hey!... sorry my connection went off16:15
besser82lambday:  Hey!  np16:15
besser82lambday:  looking for a job?!?16:15
lambdaybesser82: hehe in the US?16:16
besser82lambday:  yes16:16
besser82lambday:  if interessted goto #fedora-bigdata and ping `mattf`.  Tell besser82 has send you  ;)16:16
lambdaybesser82: cool!!16:17
lambdaybesser82: wow!16:17
besser82lambday:  :D16:17
@wikingbesser82: he get's us a green card as well? :)16:17
besser82lambday:  ask him  ;)16:18
besser82lambday:  actually rh hires remote workers, too ;)16:18
besser82lambday:  they hire good plp where they are ;)16:18
besser82lambday:  and pay well16:18
lambdaybesser82: I know nothing about big data btw :(16:19
besser82lambday:  thats not required  ;)16:19
besser82lambday:  you need skills in 2 prog-lang and NoSQL  ;)16:20
besser82lambday:  and expirience in contributiong to FOSS-Projects  ;)16:20
lambdaybesser82: then I am falling short - no nosql :-/16:20
besser82lambday:  go over and ask  ;)16:21
lambdaybesser82: yeah thanks to shogun!16:21
besser82lambday:  noone bites you16:21
besser82lambday:  :D16:21
besser82lambday:  or write hin an emain16:22
besser82lambday:  email16:22
besser82lambday:  mattf@redhat.com16:22
lambdaybesser82: haha :D thanks.. I will... if I get to go US for a year or two, I'll dump oracle cause its not much money!16:22
lambdaybesser82: excellent16:22
lambdaybesser82: but I won't be able to join for another 6 months!16:22
besser82lambday:  still occupied?!?16:23
lambdaybesser82: well, last 1 semester left...16:23
lambdaywill graduate in June-July16:23
besser82lambday:  that won't be a problem  ;)16:23
besser82lambday:  rh hires students as well ;)16:24
besser82lambday:  you wouldn't be the only one  ;)16:24
lambdayhehe! :D16:24
lambdayalright... will mail matt :)16:24
lambdaybesser82: btw are you from US too?16:25
besser82lambday:  nope, Germany, like sonney2k16:25
lambdaythe big boss! :)16:25
besser82lambday:  hehe16:25
besser82lambday:  If you want to can ping by irc as well  ;)16:25
besser82lambday:  ^^ mattf16:26
besser82lambday:  he's currently in #fedora-bigdata16:26
besser82lambday:  waiting for YOU!  :-P16:26
besser82lambday:  (or in fact other people who want a cool job)16:26
lambdaybesser82: lol.. alright! he'll love me I am a dedicated fedora user for last 5 years :P16:27
besser82lambday:  That's actually a nice thing too  :D16:27
besser82lambday:  What'S your FAS btw?16:27
besser82lambday:  Fedora-Account-System ;)16:29
besser82lambday:  Every serious fedora-user has one  ;)16:29
lambdaybesser82: ...then I guess I am not that serious.. I thought I was :(16:30
besser82lambday:  was more some fun  about that serious16:30
besser82lambday:  in fact the contributors sign up there ;)16:30
lambdaybesser82: yeah checking that16:31
besser82lambday:  If want to help me maintaining shogun on Fedora, lemme know  ;)16:31
lambdaybesser82: yeah! but I know little about fedora specific things though - you'll have to teach me!16:33
besser82lambday:  that wouldn't be the problem  ;)16:33
lambdaycurrently I am not that occupied... so if there's something I can help with.. let me know16:34
besser82lambday:  currently not much...16:34
besser82lambday:  except if you want to bring some new stuff into fedora16:34
lambdaybesser82: I will be around.. hope I can be of some help!16:34
lambdaybesser82: new stuffs as in?16:35
lambdayshogun specific?16:35
besser82lambday:  like some nifty libs || apps16:35
besser82lambday:  if you want  ;)16:35
besser82lambday:  shogun specific as well, like fixing build of java-modular  for fedora16:35
lambdaybesser82: it doesn't work??16:35
lambdaybesser82: I used java-modular on fedora16:36
besser82lambday:  unfortunately no  :(16:36
lambdaywell, it was a year back!16:36
besser82lambday:  hehe16:36
besser82lambday:  since with cmake it doesn't....16:36
lambdaybesser82: I am using f19 and I don't like it btw!16:38
besser82lambday:  ugrade to F20  ;)16:39
besser82lambday:  It is much better16:39
besser82lambday:  I actually used F19 for just 4 weeks after released and switched to f20 then16:40
lambdaybesser82: ahan!16:40
besser82lambday:  F20 in pre-alpha was even better than F19  ;)16:40
lambdaybesser82: so yum upgrade should work or I should download the iso and?16:40
besser82lambday:  use `fedup`  ;)16:41
besser82lambday:  yum install fedup16:41
besser82lambday:  fedup --network 2016:41
lambdaywow!! :D loving the name....16:41
besser82lambday:  then just reboot and wait for fedup to finish  ;)16:42
lambdaybesser82: alrighty... running..16:42
besser82lambday:  +1  :D16:43
lambdaybesser82: I am behind proxy... should I configure fedup first or it can use env vars $http_proxy etc?16:46
lambdaynevermind... its running it seems :D16:47
besser82lambday:  It doesn't need network after reboot  ;)16:47
besser82lambday:  just for preping the sys16:47
besser82lambday:  and AFAIK it uses the proxies from NetworkManager ;)16:48
lambdayready to go then!16:48
besser82lambday:  kk16:48
besser82lambday:  after reboot the upgrade might take ~20 mins16:50
lambdaybesser82: alright.. can't wait to try it out!16:51
besser82lambday:  then go go goooooo !  :D16:51
lambdaydownloading 300~ out of 2036!16:59
lisitsynbesser82: hah #shogun-job17:20
-!- lisitsyn [] has quit [Quit: Leaving.]17:23
-!- lisitsyn [] has joined #shogun17:36
besser82lambday:  all went fine with the upgrade?!?18:34
lisitsynbesser82: are your compilation speedup commits in already?18:58
-!- iglesiasg [] has joined #shogun19:11
-!- mode/#shogun [+o iglesiasg] by ChanServ19:11
besser82lisitsyn:  not yet :(19:12
lambdaybesser82: hey.. sorry, was away!19:20
besser82lambday:  did upgrade run smooth?19:20
lambdaybesser82: something went wrong! after downloading all pkgs it was importing keys and failed there19:20
lambdayDownloading failed: Didn't install any keys19:21
besser82which fedup-version?19:21
besser82yum list fedup19:21
lambdaybesser82: thanks! its  0.8.0-3.fc1919:22
@iglesiasgsup guys!19:22
@iglesiasghappy new year to all :)19:22
lambdayiglesiasg: happy new year :)19:23
lambdaybesser82: you're the man! checking... :)19:24
lambdaylisitsyn: hi :)19:24
lisitsynlambday: hey19:30
lambdaybesser82: fedora-release is the latest one... it must be the third party repos19:30
lambdaylisitsyn: how are you?19:30
lisitsynlambday: pretty fine what about you?19:30
besser82lambday:  which do you have enabled?!?19:30
gsomix_iglesiasg, happy ny!19:30
lambdaylisitsyn: great! :) got a job placement.. project stage 1 is over..19:31
lambdaybesser82: checking... chrome may be19:31
lisitsynlambday: yeah I have seen you are an oracle guy now :)19:31
lambdaylisitsyn: not quite yet... don't wanna be :(19:31
besser82lambday:  chrome from Spot's repo?!?19:32
lisitsynlambday: why?19:32
lambdaybesser82: have these in my /etc/yum.repos.d/... adobe-linux, google-chrome, and rpmfusions apart from fedora repos19:33
besser82lambday:  fedup --network 20 --disable-repo=adobe* --disable-repo=google*19:34
lambdaylisitsyn: job profile has nothing to do with ML... but probably will learn other things..19:35
lambdaybesser82: cool! trying.. rpmfusions are okay then?19:35
besser82lambday:  yes, rpmfusion should be ok19:35
besser82lambday:  otherwise just append another disable-repo19:36
besser82lambday:  and then you need to `yum distro-sync`  after upgrade has finished !!!!  IMPORTANT!!!19:37
lambdaydisable-repo not recognized with 0.819:37
lambdaylisitsyn: hey I was just checking out the multitask learning tutorial you wrote... do you know any good materials for the backgrounds?19:40
lisitsynlambday: not really but I am trying to write a notebook now :)19:40
lambdayI was checking out a video tutorial which was going above my head19:40
lambdaylisitsyn: for the same?19:41
lambdaygreat! :)19:41
lambdaybesser82: disable-repo not recognized with 0.8.. should I manually disable them and rerun fedup?19:42
besser82lambday:  remove da dash `-` between disable-repo19:43
besser82lambday:  that works  ;)19:43
lambdayyeah! it does :D19:43
-!- lambday_ [67157d36@gateway/web/freenode/ip.] has joined #shogun19:55
-!- lambday [67157c4d@gateway/web/freenode/ip.] has quit []19:55
-!- lambday_ is now known as lambday19:57
-!- iglesiasg [] has quit [Quit: Leaving]19:58
-!- Netsplit *.net <-> *.split quits: anshul20:00
lambdaybesser82: rebooted and fedup running...20:01
besser82lambday:  +1  ;)20:02
besser82lambday:  lemme know if there any issue20:02
lambdaybesser82: absolutely ;)20:04
-!- Netsplit over, joins: anshul20:05
lambdaybesser82: worked like a charm! loving it :)20:59
lambdaywill try java modular on this20:59
lambdaygoing to sleep.. good night guys21:02
besser82lambday:  good night, sleep tight  ;)21:05
besser82lambday:  nice to hear all went fine :D21:05
-!- gsomix_ [~gsomix@] has quit [Quit: Leaving]21:15
-!- sonne|osx [~sonne@] has joined #shogun21:48
sonne|osxhey folks!21:48
sonne|osxlisitsyn: how did you notebook progress?21:54
-!- zxtx_ [] has quit [Ping timeout: 252 seconds]22:00
sonne|osxlisitsyn: serious question - do converters when you call apply() return a new object or not?22:09
sonne|osxI was under the empression - yes they do...22:09
* sonne|osx off train22:10
-!- sonne|osx [~sonne@] has quit [Quit: sonne|osx]22:10
-!- lambday [67157d36@gateway/web/freenode/ip.] has quit [Ping timeout: 272 seconds]22:33
-!- zxtx_ [] has joined #shogun23:09
-!- tricksy [uid12458@gateway/web/] has quit [Ping timeout: 246 seconds]23:12
-!- tricksy_ [uid12458@gateway/web/] has joined #shogun23:15
@sonney2kman I wish iglesiasg was here...23:20
-!- tricksy_ [uid12458@gateway/web/] has quit [Ping timeout: 264 seconds]23:22
-!- xanthus_ [uid15721@gateway/web/] has quit [Ping timeout: 246 seconds]23:25
-!- tricksy_ [uid12458@gateway/web/] has joined #shogun23:25
-!- xanthus_ [uid15721@gateway/web/] has joined #shogun23:27
-!- xanthus_ [uid15721@gateway/web/] has quit [Ping timeout: 245 seconds]23:33
-!- tricksy_ [uid12458@gateway/web/] has quit [Ping timeout: 246 seconds]23:34
-!- iglesiasg [] has joined #shogun23:35
-!- mode/#shogun [+o iglesiasg] by ChanServ23:36
@iglesiasgsonney2k, going to try with apply_structured right away23:36
-!- tricksy_ [uid12458@gateway/web/] has joined #shogun23:36
@iglesiasgsonney2k, mmm something really funny going on23:41
@iglesiasgsonney2k, with apply_structured in the one step setup it does not seem to finish23:41
@iglesiasgsonney2k, with apply in one step it crashes right away23:41
@iglesiasgsonney2k, with two steps using apply it works fast and without any problem23:42
@sonney2kiglesiasg, found it more or less23:43
@iglesiasgsonney2k, SWIG code?23:44
@sonney2kI understand the problem but no idea what the solution is23:44
@sonney2kI cooked everything down to very few lines23:44
@sonney2kwhen doing apply23:44
@sonney2kor apply_structured23:44
@sonney2k(doesn't matter)23:44
@sonney2kiglesiasg, the returned refcount is 123:45
@sonney2kwhich is what we would expect23:45
@sonney2kiglesiasg, btw you can call obj.ref_count() to see its current refcount23:45
@sonney2kiglesiasg, now when you do to_structured()23:45
@sonney2kthe refcount remains 123:45
@sonney2kwhich is what we would expect from C++23:45
@sonney2kbut not from swig23:46
@iglesiasgiglesiasg, why not?23:46
@iglesiasgsonney2k, ^23:46
@iglesiasgtalking to me again23:46
@sonney2kbecause swig will put the object into to_structured23:46
@sonney2krealize upon return that the bmrm.apply() is no longer needed23:46
@sonney2kand unref it23:46
@sonney2kand so we loose23:47
@sonney2kiglesiasg, so when you split up the call into 223:47
@sonney2kiglesiasg, you still have one reference hanging around23:47
@sonney2kand python cannot free the object23:47
@sonney2khowever if you would do del lab23:47
@sonney2kwhere lab = bmrm.apply()23:47
@sonney2kafter you do to_structured(lab)23:48
@sonney2kand then get the refcount23:48
@sonney2kyou will see the same issue23:48
@iglesiasgI get it I think23:48
@sonney2kI don't see a solution currently :/23:49
@sonney2kiglesiasg, if we would return the label object within c++ with refcount 1 already23:49
@sonney2kswig would inc it to refcount 223:49
@sonney2kand so we would have a leak23:49
@sonney2kso we somehow need swig to inc the refcount in the to_structured ...23:50
@sonney2kiglesiasg, but I don't think it makes sense to ref() inside to_structured or?23:50
@iglesiasgsonney2k, can we do that in .i?23:50
@iglesiasgsonney2k, no, I don't think it is right to ref inside to_structured23:51
@sonney2kiglesiasg, I am just reading
@sonney2kmaybe %newobject can help...23:52
@sonney2kiglesiasg, yes that seems to be good23:54
@sonney2kiglesiasg, The %newobject feature is designed to indicate to the target language that it should take ownership of the returned object. When used in conjunction with a type that has the "ref" feature associated with it, it additionally emits the code in the "ref" feature into the C++ wrapper. Consider wrapping the following factory function in addition to the above:23:54
-!- anshul [uid12132@gateway/web/] has quit [Ping timeout: 246 seconds]23:55
@sonney2kiglesiasg, so %newobject to_structured() should fix it23:56
@sonney2kit is confusing though since this is not a new object but just a cast'd one23:56
-!- anshul_ [uid12132@gateway/web/] has joined #shogun23:57
@sonney2kiglesiasg, however to_structured is not even needed because apply() already returns CStructuredLabels23:58
--- Log closed Fri Jan 03 00:00:35 2014