Open in new window / Try shogun cloud
--- Log opened Fri Jan 03 00:00:35 2014
@iglesiasgsonney2k, yep, in case you use apply_structured that is true00:00
@sonney2kiglesiasg, even with apply() !00:01
@sonney2kiglesiasg, try it00:01
@iglesiasgoh really!? let mesee00:01
@iglesiasgsonney2k, true, indeed00:02
@iglesiasgsonney2k, well, the fix here then is not to use to_structured I guess00:02
@iglesiasgsonney2k, where is that handled though?00:02
@iglesiasgsonney2k, that the return value is casted to structured labels00:03
@iglesiasgsonney2k, SWIG code?00:03
@sonney2kiglesiasg, in Machine.i  - yes swig code00:03
@sonney2kpretty cool uh?00:04
@iglesiasgsonney2k, yep00:04
@iglesiasghaha yeah00:04
@sonney2kiglesiasg, so please commit the fix that doesn't use this conversion00:04
@sonney2kI will nevertheless00:04
@sonney2kadd %newobject to all methods in CLabelsFactory00:04
@iglesiasgwhy so?00:04
@iglesiasgsonney2k, don't we get a leak then?00:05
@sonney2kno00:07
@sonney2kit will return the object but ref'd00:07
@sonney2kso if the refcount was 0 before it will be 1 now00:08
@iglesiasgsonney2k, but that is right when doing to_structured(apply())00:09
@iglesiasgwhat if doing?00:09
@iglesiasgvar =  apply()00:09
@sonney2kalso then00:09
@iglesiasgto_structured(var)00:09
@sonney2kvar.ref_count() should be 1 right?00:09
@sonney2kand then00:09
@sonney2kvar2= to_structured(var)00:09
@sonney2kvar2's refcount should be 200:09
@sonney2kbecause you have 2 references00:09
@iglesiasgyep, right00:10
@iglesiasgout of curiosity00:10
@iglesiasgvar = to_structured(var)00:10
@iglesiasgwhat would it happen there? mmm00:10
@sonney2kvar would just be converted to type CStructuredLabels00:11
@sonney2kand refcount would be 100:11
@iglesiasgabout var2's refcount being 2, and var's refcount 1 in the setting before00:12
@sonney2kiglesiasg, but now you know why we always do SG_REF when returning some object (except for new ones...)00:12
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun00:15
shogun-notifier-shogun: Fernando Iglesias :develop * 991e3cd / doc/ipython-notebooks/structure/FGM.ipynb: https://github.com/shogun-toolbox/shogun/commit/991e3cdd904a99776518ac7e711a5849bf26da4900:15
shogun-notifier-shogun: Fix FGM notebook00:15
shogun-notifier-shogun: Do not use LabelsFactory in Python for apply's output, the cast is already00:15
shogun-notifier-shogun: handled by SWIG code in Machine.i00:15
shogun-notifier-shogun: Fernando Iglesias :develop * 2451111 / doc/ipython-notebooks/structure/FGM.ipynb: https://github.com/shogun-toolbox/shogun/commit/2451111d21c830053d23f91b86a3064820d4625d00:15
shogun-notifier-shogun: Merge pull request #1806 from iglesias/develop00:15
shogun-notifier-shogun:00:15
shogun-notifier-shogun: Fix FGM notebook00:15
shogun-notifier-shogun: Soeren Sonnenburg :develop * 39842ca / src/interfaces/modular/Features.i: https://github.com/shogun-toolbox/shogun/commit/39842ca8fb6aae4e9619c9dd10a809f62fccd69c00:20
shogun-notifier-shogun: use %newobject on label factory functions to fix premature memory freeing00:20
@sonney2kiglesiasg, alright nice teamwork nailing this down :D00:20
@iglesiasgsonney2k, well it feels I didn't actually help much :D00:21
@sonney2kiglesiasg, if you have time - scan the code for similar scary patterns...00:21
@sonney2kiglesiasg, finding the exact line is half of the work!00:21
@iglesiasgsonney2k, I just see one to_* in python_modular examples00:22
@iglesiasgin one factor graph example from Shell too00:23
@iglesiasgthis is done in two steps so the crash did not pop up00:23
@iglesiasgbut I guess it is nicer without the cast so I will take it out00:23
@sonney2kiglesiasg, yes00:23
@sonney2kiglesiasg, but there might be more in the C++ code ... functions that take some object as input and return it without the ref increased or functions returning object w/o increasing the refcount00:24
@sonney2kthat is more subtle to find... so in case you have time have a look00:25
@iglesiasgsonney2k, I am not sure why the first case should increase the ref count though :S00:25
@sonney2kiglesiasg, and please update the NEWS about this bug00:26
@sonney2kso we are now in release shape00:26
@iglesiasgsonney2k, I mean, a function takes an object in the input, why should it do SG_REF of this?00:26
@sonney2kiglesiasg, it should not only if it returns it and then only in the modular interfaces00:26
@iglesiasgokok00:27
@iglesiasgso00:27
@sonney2kI am overdue with going to bed so be quick00:27
@iglesiasgp* f(p*)00:27
@iglesiasgshould use this %newobject thing00:27
@iglesiasgthat's it00:27
shogun-notifier-shogun: Fernando Iglesias :develop * 3038519 / examples/undocumented/python_modular/structure_factor_graph_model.py: https://github.com/shogun-toolbox/shogun/commit/30385190862550f63beffe69bc79e0b2b2da761400:31
shogun-notifier-shogun: Remove LabelsFactory.to_structured in python modular example00:31
shogun-notifier-shogun: NO need to use this method to do the cast since it is already handled00:31
shogun-notifier-shogun: in Machine.i00:31
shogun-notifier-shogun: Fernando Iglesias :develop * a22dd82 / examples/undocumented/python_modular/structure_factor_graph_model.py: https://github.com/shogun-toolbox/shogun/commit/a22dd82bf954a7acb3dc7c244694956ff7ef33a200:31
shogun-notifier-shogun: Merge pull request #1807 from iglesias/develop00:31
shogun-notifier-shogun:00:31
shogun-notifier-shogun: Remove LabelsFactory.to_structured in python modular example00:31
shogun-buildbotbuild #424 of FC19 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20libshogun/builds/424  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>00:32
@sonney2kiglesiasg, yes and also things doing00:32
@sonney2kp* f()00:32
@sonney2kreturning a new object00:32
@iglesiasgunit-SGVectorTest failed?!00:33
@sonney2kand the others that return a reference should do SG_REF()00:33
@iglesiasgwe lost Fedora :D00:33
@sonney2kiglesiasg, some numerical thing only though00:33
shogun-buildbotbuild #425 of FC19 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20libshogun/builds/42500:36
@sonney2kiglesiasg, btw did you get my email about this? I am just wondering...00:36
@iglesiasgsonney2k, about the buildbot fail? No, I didn't00:37
@sonney2kiglesiasg, no the reply to your mail today00:38
@iglesiasgsonney2k, got it today about three hours ago00:40
@iglesiasg2 hours actually, 21:55, just checked it00:40
@iglesiasgsonney2k, why do you wonder about it? Did you send it before that today?00:40
@iglesiasgsonney2k, just added a line in NEWS https://github.com/shogun-toolbox/shogun/pull/1808/files00:43
@iglesiasgjust merge it please if that's how you wanted it00:43
@sonney2kiglesiasg, I had some issue recently so I was just wondering00:44
* iglesiasg afk00:49
* iglesiasg back01:04
shogun-buildbotbuild #162 of clang34 - undefined behaviour analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/162  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>01:20
-!- iglesiasg [~iglesiasg@211.Red-83-40-129.dynamicIP.rima-tde.net] has quit [Quit: Leaving]01:32
shogun-buildbotbuild #158 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/158  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>01:36
-!- travis-ci [~travis-ci@ec2-54-211-35-213.compute-1.amazonaws.com] has joined #shogun01:38
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/1628795801:38
-!- travis-ci [~travis-ci@ec2-54-211-35-213.compute-1.amazonaws.com] has left #shogun []01:38
shogun-buildbotbuild #173 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/173  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>02:21
shogun-buildbotbuild #163 of clang34 - undefined behaviour analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/163  blamelist: Soeren Sonnenburg <sonne@debian.org>02:34
shogun-buildbotbuild #159 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/159  blamelist: Soeren Sonnenburg <sonne@debian.org>02:51
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]03:31
shogun-buildbotbuild #174 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/174  blamelist: Soeren Sonnenburg <sonne@debian.org>03:34
-!- travis-ci [~travis-ci@ec2-54-227-186-41.compute-1.amazonaws.com] has joined #shogun03:40
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/1628889303:40
-!- travis-ci [~travis-ci@ec2-54-227-186-41.compute-1.amazonaws.com] has left #shogun []03:40
shogun-buildbotbuild #164 of clang34 - undefined behaviour analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/164  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>03:45
shogun-buildbotbuild #160 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/160  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>04:02
shogun-buildbotbuild #669 of nightly_default is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/66904:41
shogun-buildbotbuild #175 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/175  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>04:44
-!- lambday [67157f36@gateway/web/freenode/ip.103.21.127.54] has joined #shogun06:27
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 245 seconds]06:31
-!- wiking [~wiking@info2k1.hu] has joined #shogun06:32
-!- lambday [67157f36@gateway/web/freenode/ip.103.21.127.54] has quit [Ping timeout: 272 seconds]07:53
-!- wiking [~wiking@info2k1.hu] has quit [Changing host]08:06
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun08:06
-!- mode/#shogun [+o wiking] by ChanServ08:07
sonne|workheya all good morning!08:53
sonne|workwiking, besser82 ^08:54
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun10:07
shogun-notifier-shogun: Fernando Iglesias :develop * 7dccde8 / NEWS: https://github.com/shogun-toolbox/shogun/commit/7dccde870407ccf04be17b61f588d7e6ea78c69110:07
shogun-notifier-shogun: Update bugfix NEWS for latest release10:07
shogun-notifier-shogun: Fernando Iglesias :develop * 609dcad / NEWS: https://github.com/shogun-toolbox/shogun/commit/609dcad8bd098bdb63c2449ac70d01521671f63c10:07
shogun-notifier-shogun: Merge pull request #1808 from iglesias/develop10:07
shogun-notifier-shogun:10:07
shogun-notifier-shogun: Update bugfix NEWS for latest release10:07
besser82sonne|work, wiking, lisitsyn:  G' Morning!!!10:36
@wikingmornin10:44
shogun-buildbotbuild #165 of clang34 - undefined behaviour analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/165  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>11:12
shogun-buildbotbuild #161 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/161  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>11:29
shogun-buildbotbuild #176 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/176  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>12:12
-!- new_lido [~walid@41.218.176.150] has joined #shogun12:32
-!- new_lido [~walid@41.218.176.150] has quit [Ping timeout: 246 seconds]12:46
-!- iglesiasg [~fircuser@213.143.60.190] has joined #shogun12:47
iglesiasgsonne|work: hey12:50
iglesiasgmaybe you remember I asked you a few days past about using Baum-Welch with real valued features12:51
iglesiasgI am not sure if that is currently possible sonce I just saw StringFeatures there12:52
iglesiasgerr, StringFeatures<uint16_t> actually13:03
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]13:07
-!- iglesiasg [~fircuser@213.143.60.190] has quit [Read error: Connection reset by peer]13:10
sonne|workiglesias - yes I remember - unfortunately not yet in shogun - I can mentor you though how to do it...13:53
-!- iglesiasg [~iglesiasg@211.Red-83-40-129.dynamicIP.rima-tde.net] has joined #shogun14:06
-!- mode/#shogun [+o iglesiasg] by ChanServ14:06
@iglesiasgsonne|work, I need a couple of extensions to that14:07
@iglesiasgsonne|work, handle real values as well as vector observations14:07
@iglesiasgI have some Matlab code that is able to handle that, so I think I will try first with it14:08
-!- iglesiasg [~iglesiasg@211.Red-83-40-129.dynamicIP.rima-tde.net] has quit [Ping timeout: 264 seconds]14:21
-!- iglesiasg [~iglesiasg@211.Red-83-40-129.dynamicIP.rima-tde.net] has joined #shogun14:23
-!- mode/#shogun [+o iglesiasg] by ChanServ14:23
sonne|workiglesiasg: yeah sure - you would need to write a new HMM class for that if you would want to do that in shogun...15:00
-!- lambday [67157f36@gateway/web/freenode/ip.103.21.127.54] has joined #shogun15:14
-!- besser82 [quassel@fedora/besser82] has quit [Remote host closed the connection]15:21
-!- besser82 [quassel@fedora/besser82] has joined #shogun15:40
-!- zxtx_ [~zv@c-98-223-196-32.hsd1.in.comcast.net] has quit [Ping timeout: 264 seconds]16:00
-!- zxtx_ [~zv@129-79-241-148.dhcp-bl.indiana.edu] has joined #shogun16:51
-!- pickle27 [d8bfe7da@gateway/web/freenode/ip.216.191.231.218] has joined #shogun17:34
pickle27sonney2k: what irc client do you use on osx?17:35
-!- iglesiasg [~iglesiasg@211.Red-83-40-129.dynamicIP.rima-tde.net] has quit [Quit: Leaving]17:36
-!- bb_ [458fc8de@gateway/web/freenode/ip.69.143.200.222] has joined #shogun17:43
@wikingpickle27: colloquy is quite ok17:51
pickle27@wiking, alight I saw that one17:52
pickle27osx has too many choices but I find it harder to make up my mind compared to linux17:52
pickle27anyways figure I should get irc on my work machine so I can at least idle in here even though I don't have much time to do any development atm17:53
@wikingpickle27: no worries most of us is busy as well with other stuff but still we idle here :P18:13
@sonney2kwiking, pickle27 but we will do a rlease this weekend :)18:48
* sonney2k afk18:48
pickle27awesome!18:48
lisitsynsonney2k: okay I am speeding up to get notebook ready before release ;)18:50
-!- bb_ [458fc8de@gateway/web/freenode/ip.69.143.200.222] has quit [Ping timeout: 272 seconds]19:39
-!- bb_ [458fc8de@gateway/web/freenode/ip.69.143.200.222] has joined #shogun20:05
-!- travis-ci [~travis-ci@ec2-54-205-49-151.compute-1.amazonaws.com] has joined #shogun20:39
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/1630567120:39
-!- travis-ci [~travis-ci@ec2-54-205-49-151.compute-1.amazonaws.com] has left #shogun []20:39
-!- pickle27 [d8bfe7da@gateway/web/freenode/ip.216.191.231.218] has quit [Ping timeout: 272 seconds]20:40
@wikingsonney2k: great! :)21:56
@wikingbesser82: ping21:57
bb_quit22:30
-!- bb_ [458fc8de@gateway/web/freenode/ip.69.143.200.222] has quit []22:30
-!- lambday [67157f36@gateway/web/freenode/ip.103.21.127.54] has quit [Ping timeout: 272 seconds]22:35
-!- wiking_ [~wiking@info2k1.hu] has joined #shogun22:57
-!- wiking [~wiking@huwico/staff/wiking] has quit [Read error: Connection reset by peer]22:57
-!- wiking_ is now known as wiking23:07
-!- wiking [~wiking@info2k1.hu] has quit [Changing host]23:07
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun23:07
-!- mode/#shogun [+o wiking] by ChanServ23:07
besser82wiking:  pong23:55
--- Log closed Sat Jan 04 00:00:36 2014