IRC logs of #shogun for Wednesday, 2016-08-10

« Tuesday, 2016-08-09 Index Thursday, 2016-08-11 »
--- Log opened Wed Aug 10 00:00:59 2016
shogun-buildbotbuild #1183 of nightly_default is complete: Failure [failed test notebooks]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/1183  blamelist: Esben Sorig <esben@sorig.eu>, Viktor Gal <vigsterkr@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>03:55
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]04:26
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun04:26
-!- mizari [~mizari@95-174-213-100.nts.su] has joined #shogun05:03
@wikingSaurabh7, ping?06:58
Saurabh7hey wiking06:58
@wikingmmm06:58
@wikingdo you inlcude the falconn headers in other shogun headers?06:58
Saurabh7no only in KNN.cpp06:59
@wikingbecause this sounds like that the includes06:59
@wikingmmm06:59
@wikinglemme go trhough06:59
@wikingbut looks like the header is not only included in the cpp somewhere06:59
@wikingand because of that06:59
@wikingwhen you compile example (that doesn't have anymore the -isystem or -I for eigen headers) it fails07:00
Saurabh7i see that in my hierarchical clustering generated meta example07:00
Saurabh7the generated cpp07:00
Saurabh7there is a stray #include <shogun/lib/external/falconn/wrapper/cpp_wrapper_impl.h>07:00
Saurabh7oh i see07:00
Saurabh7lemme double check again then07:00
shogun-buildbotbuild #56 of memleak - valgrind is complete: Failure [failed memory check]  Build details are at http://buildbot.shogun-toolbox.org/builders/memleak%20-%20valgrind/builds/56  blamelist: Esben Sorig <esben@sorig.eu>, Viktor Gal <vigsterkr@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>07:11
Saurabh7wiking: even in KNN meta example I have #include <shogun/lib/external/falconn/wrapper/cpp_wrapper_impl.h>07:11
Saurabh7there is no need for that as KNN.h is already included07:12
@wikingmmm how did that got there?07:12
@wikingi mean into the meta07:13
Saurabh7no idea really, its generated right07:13
@wikinglemme check how it is generated07:13
Saurabh7even in KMeans meta07:13
@wikingbecause imo it should only include KNN.h07:13
Saurabh7so its everywhere where DistanceMachine is there07:14
shogun-buildbotbuild #1054 of nightly_none is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_none/builds/1054  blamelist: Esben Sorig <esben@sorig.eu>, Viktor Gal <vigsterkr@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>07:19
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]08:36
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun08:36
@wikingSaurabh7, /usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/tests/unit/multiclass/KNN_unittest.cc:15:25: error: no member named 'generate_gaussians' in 'shogun::CDataGenerator'09:19
@wiking        feat = CDataGenerator::generate_gaussians(num,classes,feats);09:19
@wiking               ~~~~~~~~~~~~~~~~^09:19
@wiking1 error generated.09:19
@wikingplease be more careful with macros09:19
-!- mizari [~mizari@95-174-213-100.nts.su] has quit [Ping timeout: 250 seconds]10:25
-!- c4goldsw [5da420e6@gateway/web/cgi-irc/kiwiirc.com/ip.93.164.32.230] has joined #shogun11:37
c4goldswzing wing ping wiking11:38
@wikinghey11:38
c4goldswI managed to get my code to compile11:38
c4goldswBut11:38
c4goldswIs there an easy way to debug the code in Shogun?11:38
c4goldswWithout having to make a seperate copy?  I don't know where the .o files are11:39
@wiking?11:39
c4goldswLet me rephrase that11:39
c4goldswif you're using gdb for instance11:39
@wikingi mean the only thing you need to be able to debug shogun11:39
@wikingis to have the library compiled with -g flag11:40
@wikingthat is if you for example to run cmake with -DCMAKE_BUILD_TYPE=Debug flag and then compile it11:40
c4goldswOh - okay.11:41
c4goldswSo you can also refer to lines within files you specify?11:41
@wikingyeah11:41
@wikingafaik you just need to give gdb the head of the path where the src is11:41
@wikingbut yes11:41
@wikingdefinitely you can do that11:41
c4goldswCool, I'll give that a try then.11:41
c4goldswAlso wiking, is that the src directory where all the code is or the compiled code in build/src ?11:42
c4goldswwhere all the source code is *11:42
@wikingc4goldsw, the source is in src/11:43
@wikingbuild/src is just where the built objects are11:43
c4goldswRight, okay :)11:43
-!- sanuj [~sanuj@117.203.23.66] has joined #shogun12:12
sanujlisitsyn, there?12:59
-!- c4goldsw [5da420e6@gateway/web/cgi-irc/kiwiirc.com/ip.93.164.32.230] has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]13:11
-!- c4goldsw [5da420e6@gateway/web/cgi-irc/kiwiirc.com/ip.93.164.32.230] has joined #shogun13:11
-!- sorig [8056b1a0@gateway/web/freenode/ip.128.86.177.160] has joined #shogun14:17
@wikingsorig: what should i merge? :)14:17
sanujsorig, hello14:17
sanujwiking, wazzup14:18
sorigHey!14:18
sanujsorig, what shall i update to make C# work?14:19
sorigYou just need to update the float literal14:19
sorig1.0f instead of 1.014:19
sanujokay14:19
sorigTo get java working we need to merge #341614:19
sorigBut travis hasn't run yet, so maybe you'd rather wait?14:20
sanujsorig, do i also need to update the float literals in other examples in a separate PR?14:20
sanujyeah14:20
sorigI did that already14:20
sanujokay, cool!14:20
sorigSee #341314:20
sanujsorig, shall i also send a separate PR for data?14:20
sanujenums.sg produces no data14:21
sorigFor the meta_api tests? Yeah14:21
sanujbut the literals example does14:21
sorig@wiking  #3413 needs to be merged simultaneously with my PR in shogun-data14:21
@wikingsorig, ok i'll wait for travis14:22
@wikingand then i merge those14:22
sorigCool, thanks!14:22
sanujsorig, how shall i structure the enums tests14:22
sanujnow there is only 114:22
sanujand it is not producing any data14:23
sorigWhat you have is fine, but you can add a few more if you want. You could reassign a new value to an already instantiated variable (you could go through all values of the enum for example)14:23
sanuji see14:24
sorigBut getting some comments in to document the float literals vs double literals would be great. Also just mention that real is 64bit and float is 32 etc.14:24
sorigFeel free to add comments you think would be useful for example authors to understand what's possible and what's not :)14:25
sanujsorig, okay, i'll add comments14:25
sanujso the enums.sg would look something like this14:25
sanujLIBLINEAR_SOLVER_TYPE myVar = enum LIBLINEAR_SOLVER_TYPE.L2R_L2LOSS_SVC14:25
sanujmyVar = enum LIBLINEAR_SOLVER_TYPE.L2R_LR14:26
sanujmyVar = enum LIBLINEAR_SOLVER_TYPE.L2R_L2LOSS_SVC_DUAL14:26
sanujmyVar = enum LIBLINEAR_SOLVER_TYPE.L2R_L1LOSS_SVC_DUAL14:27
sorigYeah, exactly. If you do that, you don't really need to check other enum types14:27
sanujokay14:27
sanujsorig, and after literals and enums, what all shall we write tests for?14:28
Saurabh7hey sorig14:31
sorigWe need to test: initializing shogun type variables (1. by using the constructor 2. by assigning them to some object returned by a method on another object), static calls and method calls (make sure to test ones with different number of arguments)14:31
sorighey14:32
Saurabh7sorig: how are includes generated for cpp, I am having a problem with an meta example due to an unnecessary include14:32
Saurabh7I understand it was based on classes used14:33
sorigCan you send a link to a gist?14:33
sorigFor C++ we include all class names and enum types ever mentioned in an example14:34
Saurabh7umm in the example , CKNN is used14:36
Saurabh7I have added some external headers to KNN class14:36
sanujsorig, so can i use any shogun class for constructor and assignment tests?14:36
sorigYeah, you choose14:36
sanujand i would use multiple classes in each test]14:37
Saurabh7sorig: the example is including external headers too, when it should include only KNN.h14:37
sorigAh right, yeah we are using a heuristic to look up the header paths. It sounds like it needs some tweaking14:38
sorigSend the specific example, I'll have a look at it14:38
Saurabh7https://gist.github.com/Saurabh7/cb66deb288de81c19599af465082562f#file-genreated-cpp-L614:39
sorigsanuj: yeah, put them in the same file, that's fine14:39
Saurabh7this is the generated one14:39
sanujsorig, okay14:39
sanujsorig, thanks a lot!14:39
sanujleaving for now.14:39
sorigSaurabh7, Can you give me the meta-language version also?14:40
Saurabh7um its in develop14:40
Saurabh7but my pr changes the class14:40
sorigAh sorry!14:40
Saurabh7lemme send link14:40
@wikingsorig, around?14:40
@wikingah cool Saurabh7 already asked you :)14:41
Saurabh7sorig: https://github.com/shogun-toolbox/shogun/blob/develop/examples/meta/src/multiclass_classifier/knn.sg14:41
Saurabh7and this is the branch https://github.com/shogun-toolbox/shogun/pull/341514:41
sorigCool, thanks. Let me take a look14:42
Saurabh7sorig: Thanks!14:43
-!- sorig [8056b1a0@gateway/web/freenode/ip.128.86.177.160] has quit [Quit: Page closed]14:43
-!- sorig [~textual@128.86.177.160] has joined #shogun14:44
-!- sanuj [~sanuj@117.203.23.66] has quit [Ping timeout: 240 seconds]14:44
c4goldswwiking: yo, I'm running into another issue with GDB.  I'm able to debug the code, but any changes I make to Shogun aren't reflected in the code GDB.  My process is as follows: edit a source file, compile (I've also tried to reinstall the library which did nothing).  Any suggestions?14:45
-!- sorig [~textual@128.86.177.160] has quit [Client Quit]14:45
c4goldswaren't reflected in the code in GDB*14:45
@wikingso i would run gdb like this14:50
@wikinggdb --directory ~/shogun/src14:50
@wikingor i mean wherever your shogun is14:50
c4goldswAlright, I'll try that then.14:50
c4goldswBy the way - the source code changes, but the compiled code doesn't.14:51
@wiking?14:52
c4goldsw(i.e. GDB is aware of changes to the source, but not of changes to the byte code)14:52
c4goldswIf i do (gdb) list, I can see the new source code.  But, for instance, if I add a new variable, it isn't aware of it.14:53
lisitsynit just reads the fiels14:53
lisitsynfiles*14:53
lisitsynsources14:53
c4goldswi.e. print new_var doesn't work.14:53
@wikingc4goldsw: the source is just trying to map the gdb instructions back to the src14:54
@wikingif you change the src but not the binary14:54
@wikingthen this mapping might even be foobar14:54
c4goldswBut I do change the binary (by building shogun).14:54
@wikingthen if you use that binary with the src14:55
@wikingthen it should be all good14:55
@wiking:)14:55
@wikingif not then please try to explain your case a bit more in details14:55
c4goldswSo, I edit and compile shogun14:55
c4goldswThese changes are not reflected in what I run in GDB, but they are still in the compiled code (I've confirmed this through the mock tests)14:56
-!- sanuj [~sanuj@117.203.23.66] has joined #shogun14:56
@wikingare you sure that the gdb-ed binary14:56
@wikingis actually using the newly compiled shared lib?14:57
@wikingsay if you do ldd <binary>14:57
@wikingwhat is the output14:57
c4goldswInside of GDB or in bash?14:57
@wikingbash14:57
@wikingso instead of gdb <binary>14:57
@wikingrun ldd <binary>14:57
c4goldswI have to go to where the library is actually located to have it return annything14:58
@wikingc4goldsw: ok not really14:58
@wikingyou could use14:58
c4goldswfrom a random directory, it just returns as no such file or directory14:59
@wikingLD_LIBRARY_PATH=<path to the library>14:59
@wikingso14:59
@wikingLD_LIBRARY_PATH=<path to the library> ldd <binary>14:59
@wikingshould then work14:59
@wikingbut the thing is then14:59
@wikinghow did gdb worked till now? :)14:59
c4goldswDon't know, I'm wondering that myself - but it has worked.14:59
@wikingweird14:59
c4goldswI just compiled my code along with the library using g++15:00
@wikingthen i suppose it has definitely not using the right libshogun :)15:00
@wikingok but did you do15:00
@wiking-lshogun15:00
@wikingor libshogun.a15:00
@wiking?15:00
c4goldswlshogun15:00
c4goldswthe shared one15:00
c4goldsw-lshogun *15:00
@wikingok so then i really wonder :P15:00
@wikinganyhow15:00
c4goldswHere: g++ -g -std=gnu++11 LDATest.cpp -o LDATest -lshogun15:01
@wikingok15:01
@wikingand where's your libshogun?15:01
@wikinglibshogun.so?15:01
c4goldswd /usr/local/lib15:01
c4goldsw(just prepending with d so my client doesn't think it's a command)15:01
@wikingyep15:02
@wikingyou can though add a space15:02
@wiking:P15:02
c4goldswlol, okay15:02
@wikinganyhow so the /usr/local/lib/libshogun.so is the one15:02
@wikingthe latest15:02
@wikingyou wanna use15:02
@wiking?15:02
c4goldswYes - so I just set LD_LIBRARY_PATH=</usr/local/lib> ?15:02
@wikingLD_LIBRARY_PATH=/usr/local/lib15:03
@wikingbut yes15:03
c4goldswAh, yes15:03
c4goldswGreat, I'll just add that to my .bashrc.  Will report back.15:03
@wikingbut no15:03
@wikingi mean15:03
@wikingbefore that15:03
@wikingdo15:03
@wikingsudo ldconfig15:03
@wikingand then do the ldd15:03
@wikingwithout LD_LIBRARY_PATH15:03
c4goldswI also did that beforehand15:03
c4goldswBut, I'll do it again.15:04
sanujlisitsyn, i'm getting crazy merge conflicts after i rebased the plugins PR with feature/tags15:07
sanujhaha15:07
lisitsynoh yeah I'd expect that15:08
sanujlisitsyn, i need to write a readme for plugins15:09
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun15:09
shogun-notifier-shogun: Esben Sorig :develop * a4be802 / examples/meta/generator/targets/java.json: https://github.com/shogun-toolbox/shogun/commit/a4be80206ae6a08bb3e5af3be1b50842ebfd4db415:09
shogun-notifier-shogun: Fix java enum imports and literals again15:09
shogun-notifier-shogun: Viktor Gal :develop * 68f8710 / examples/meta/generator/targets/java.json: https://github.com/shogun-toolbox/shogun/commit/68f8710af62d97f7a9085760ad8fe10689b3cbd715:09
shogun-notifier-shogun: Merge pull request #3416 from sorig/fix_java_enums_again15:09
shogun-notifier-shogun:15:09
shogun-notifier-shogun: Fix java enum imports and literals again15:09
sanujbut i'm not sure what is our future plan with plugins15:10
shogun-notifier-shogun-data: Esben Sorig :master * 5375dd8 / testsuite/meta/ (6 files): https://github.com/shogun-toolbox/shogun-data/commit/5375dd86f9baafc1d5fca4552f86dd360594639315:10
shogun-notifier-shogun-data: Change incorrect int types to float64 types15:10
shogun-notifier-shogun-data: Viktor Gal :master * 97f5012 / testsuite/meta/ (6 files): https://github.com/shogun-toolbox/shogun-data/commit/97f5012e1e402999c2bdaf44bd83bf8475ac8fda15:10
shogun-notifier-shogun-data: Merge pull request #121 from sorig/fix_wrong_int_types15:10
shogun-notifier-shogun-data:15:10
shogun-notifier-shogun-data: Change incorrect int types to float64 types15:10
shogun-buildbotbuild #1322 of precise - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/1322  blamelist: Esben Sorig <esben@sorig.eu>, Viktor Gal <vigsterkr@gmail.com>15:12
shogun-buildbotbuild #2964 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2964  blamelist: Esben Sorig <esben@sorig.eu>, Viktor Gal <vigsterkr@gmail.com>15:17
-!- sorig [~textual@128.86.177.160] has joined #shogun15:23
-!- mizari [~mizari@95-174-213-100.nts.su] has joined #shogun15:24
-!- sanuj [~sanuj@117.203.23.66] has quit [Ping timeout: 276 seconds]15:52
shogun-buildbotbuild #50 of FC23 - libshogun - aarch64 is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC23%20-%20libshogun%20-%20aarch64/builds/50  blamelist: Esben Sorig <esben@sorig.eu>, Viktor Gal <vigsterkr@gmail.com>15:55
shogun-buildbotbuild #103 of xenial - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun/builds/103  blamelist: Esben Sorig <esben@sorig.eu>, Viktor Gal <vigsterkr@gmail.com>15:55
-!- c4goldsw [5da420e6@gateway/web/cgi-irc/kiwiirc.com/ip.93.164.32.230] has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]16:32
-!- sanuj [~sanuj@117.203.23.66] has joined #shogun16:58
-!- sorig [~textual@128.86.177.160] has quit [Quit: My Mac has gone to sleep. ZZZzzz…]17:18
-!- sorig [~sorig@128.86.177.160] has joined #shogun17:56
sanujlisitsyn, shall i do SFINAE thing in NonOwningValueAnyPolicy?18:04
sanujfor now, it's only on PointerValueAnyPolicy18:04
@wikingsorig, so shall i merge 341718:04
@wikingas it's green18:04
sorig@wiking yeah go ahead18:05
shogun-notifier-shogun: Esben Sorig :develop * 44a65aa / examples/meta/generator/translate.py: https://github.com/shogun-toolbox/shogun/commit/44a65aa329fda630aab0133152bc410030887a0018:05
shogun-notifier-shogun: Fix heuristic for include path search. Warn users if multiple candidates are found18:05
shogun-notifier-shogun: Viktor Gal :develop * c45f8b0 / examples/meta/generator/translate.py: https://github.com/shogun-toolbox/shogun/commit/c45f8b01e0f2a0c2706d2723dac1960c0f88499918:05
shogun-notifier-shogun: Merge pull request #3417 from sorig/include_path_multiple18:05
shogun-notifier-shogun:18:05
shogun-notifier-shogun: Fix heuristic for include path search. Warn users if multiple candida…18:05
@wikingdone18:05
sorig@Saurabh7 it should work now18:06
shogun-buildbotbuild #1323 of precise - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/1323  blamelist: Esben Sorig <esben@sorig.eu>18:08
sorig@wiking I keep getting blamelisted but the errors all look unrelated to my changes. Should I just ignore?18:09
shogun-buildbotbuild #2965 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2965  blamelist: Esben Sorig <esben@sorig.eu>, Viktor Gal <vigsterkr@gmail.com>18:13
shogun-buildbotbuild #1324 of precise - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/1324  blamelist: Viktor Gal <vigsterkr@gmail.com>18:14
@wikingyes18:14
@wikingignore :)18:14
shogun-buildbotbuild #104 of xenial - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun/builds/104  blamelist: Esben Sorig <esben@sorig.eu>, Viktor Gal <vigsterkr@gmail.com>18:19
-!- sorig [~sorig@128.86.177.160] has quit [Quit: My Mac has gone to sleep. ZZZzzz…]18:41
shogun-buildbotbuild #51 of FC23 - libshogun - aarch64 is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC23%20-%20libshogun%20-%20aarch64/builds/51  blamelist: Esben Sorig <esben@sorig.eu>, Viktor Gal <vigsterkr@gmail.com>18:51
-!- sanuj [~sanuj@117.203.23.66] has quit [Ping timeout: 244 seconds]18:58
-!- sanuj [~sanuj@117.203.23.66] has joined #shogun19:00
-!- sorig [~sorig@nat-180-135.internal.eduroam.ucl.ac.uk] has joined #shogun19:12
-!- sorig [~sorig@nat-180-135.internal.eduroam.ucl.ac.uk] has quit [Quit: My Mac has gone to sleep. ZZZzzz…]20:17
-!- sanuj [~sanuj@117.203.23.66] has quit [Ping timeout: 265 seconds]20:37
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]21:05
-!- sorig [~sorig@5.80.215.61] has joined #shogun22:09
-!- sorig [~sorig@5.80.215.61] has quit [Quit: My Mac has gone to sleep. ZZZzzz…]22:44
-!- sorig [~sorig@5.80.215.61] has joined #shogun23:04
--- Log closed Thu Aug 11 00:00:01 2016
« Tuesday, 2016-08-09 Index Thursday, 2016-08-11 »

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