--- Log opened Thu Jul 25 00:00:46 2013 | ||
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 238ead3 / src/shogun/mathematics/ajd/FFDiag.h: https://github.com/shogun-toolbox/shogun/commit/238ead3c7616dbc86553a76be548004774f1c2f6 | 00:01 |
---|---|---|
shogun-notifier- | shogun: drop unicode character | 00:01 |
@sonney2k | wiking, could you please install jinja2 on the bsd bot? | 00:01 |
@iglesiasg | sonney2k, did you see clang in travis still complain about c++11 unrecognized option? | 00:02 |
@sonney2k | iglesiasg, no | 00:03 |
@sonney2k | iglesiasg, which one? | 00:03 |
@sonney2k | iglesiasg, btw did 3km today :) | 00:03 |
@iglesiasg | sonney2k, swimming? | 00:03 |
@iglesiasg | sonney2k, that's a good one in that case! | 00:03 |
@sonney2k | if the weather stays good I can go for 4 but I was afraid that kids are awake already | 00:04 |
@sonney2k | yeah | 00:04 |
@iglesiasg | sonney2k, here for instance -- pickle27's last PR https://travis-ci.org/shogun-toolbox/shogun/jobs/9445750 | 00:04 |
van51 | sonney2k: you 're a disgrace to programmers :p | 00:04 |
van51 | sonney2k: we 're supposed to be fat and lazy | 00:05 |
@iglesiasg | hahaha | 00:05 |
@iglesiasg | van51, that's just an stereotype! :P | 00:05 |
shogun-buildbot | build #1066 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1066 blamelist: Soeren Sonnenburg <sonne@debian.org> | 00:05 |
gsomix | good stereotype. I like it. :3 | 00:06 |
shogun-buildbot | build #753 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/753 blamelist: Soeren Sonnenburg <sonne@debian.org> | 00:06 |
@sonney2k | iglesiasg, arg kid crying | 00:06 |
@sonney2k | sec | 00:06 |
@sonney2k | van51, I am fat and lazy but I like swimming | 00:10 |
@sonney2k | ^ iglesiasg | 00:10 |
@sonney2k | iglesiasg, I don't understand that | 00:10 |
@sonney2k | iglesiasg, the error should not be there | 00:10 |
@iglesiasg | sonney2k, yes, I don't understand that error either | 00:11 |
@iglesiasg | sonney2k, I think HeikoS and lambday already reverted the change in the templates to fix that | 00:11 |
@iglesiasg | sonney2k, but it is still there | 00:11 |
@sonney2k | iglesiasg, I explicitly did -Werror -std=c++11 | 00:11 |
@iglesiasg | sonney2k, and someone else got that error in ubuntu 12.04 too | 00:11 |
@iglesiasg | sonney2k, why would -Werror -std=c+11 fix that? Or that is about clang travis? | 00:12 |
@sonney2k | iglesiasg, I would need the output of configure.log from sb with a recent clang version | 00:12 |
@sonney2k | no one seems to have clang 4.2 (as we have on the buildbot) | 00:12 |
@iglesiasg | yep, that looks too modern for my ubuntu box | 00:13 |
@iglesiasg | maybe somewhere else rather than the official repositories to get it? | 00:13 |
shogun-buildbot | build #1270 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1270 blamelist: Soeren Sonnenburg <sonne@debian.org> | 00:14 |
@sonney2k | iglesiasg, debian has 3.2 so I suspect no | 00:14 |
@iglesiasg | sonney2k, can't we get that can of info from travis machine somehow? | 00:15 |
@sonney2k | iglesiasg, well we could add to configure.log | 00:18 |
@sonney2k | iglesiasg, can you add that to the travis yml? | 00:18 |
@iglesiasg | sonney2k, I can have a quick look at it | 00:18 |
@sonney2k | ohh man what I am doing now is intrusive | 00:18 |
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has quit [Ping timeout: 256 seconds] | 00:19 | |
* sonney2k is getting rid of EFeatureType / EFeatureClass | 00:19 | |
@iglesiasg | nasty | 00:19 |
@iglesiasg | sonney2k, so a 'cat configure.log' after ./configure should give us what we want if I understood correctly | 00:20 |
@iglesiasg | I just issued a PR with that, let's don't merge it and see travis output when it fires for the PR | 00:23 |
@sonney2k | iglesiasg, maybe we should have this always turned on though | 00:24 |
@sonney2k | but lets see first | 00:24 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * e67e109 / src/configure: https://github.com/shogun-toolbox/shogun/commit/e67e1091ac2b7632b8d8488ae232e8edede95463 | 00:29 |
shogun-notifier- | shogun: add rtti test in configure | 00:29 |
@sonney2k | alright sleep time! | 00:35 |
@sonney2k | cu | 00:35 |
@iglesiasg | bye bye, good night | 00:36 |
gsomix | nite | 00:39 |
gsomix | today is not productive day :( | 00:39 |
shogun-buildbot | build #1542 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1542 blamelist: Soeren Sonnenburg <sonne@debian.org> | 00:39 |
* gsomix yawns | 00:40 | |
shogun-buildbot | build #1543 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1543 blamelist: Soeren Sonnenburg <sonne@debian.org> | 00:56 |
gsomix | sonney2k, ok, part by part | 01:00 |
gsomix | sonney2k, first part on github already | 01:00 |
gsomix | sonney2k, https://github.com/shogun-toolbox/shogun/pull/1275/files | 01:00 |
gsomix | good night people | 01:00 |
shogun-buildbot | build #1387 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1387 blamelist: Soeren Sonnenburg <sonne@debian.org> | 01:04 |
@iglesiasg | good night gsomix ! | 01:07 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has joined #shogun | 01:09 | |
travis-ci | [travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/9446508 | 01:09 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has left #shogun [] | 01:09 | |
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has joined #shogun | 01:29 | |
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/9449499 | 01:29 |
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has left #shogun [] | 01:29 | |
-!- hushell [~hushell@c-24-21-169-136.hsd1.or.comcast.net] has joined #shogun | 01:40 | |
@iglesiasg | hi hushell! | 01:41 |
@iglesiasg | hushell, are you available a few minutes? | 01:41 |
-!- van51 [~van51@athedsl-399972.home.otenet.gr] has left #shogun ["QUIT :Leaving."] | 01:50 | |
-!- travis-ci [~travis-ci@ec2-75-101-178-60.compute-1.amazonaws.com] has joined #shogun | 01:53 | |
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/9449640 | 01:53 |
-!- travis-ci [~travis-ci@ec2-75-101-178-60.compute-1.amazonaws.com] has left #shogun [] | 01:53 | |
@iglesiasg | hushell, never mind, I did a github issue instead | 01:56 |
@iglesiasg | good night! | 01:56 |
shogun-notifier- | shogun: Kevin :develop * 1a5e075 / src/shogun/converter/ica/Jade.cpp: https://github.com/shogun-toolbox/shogun/commit/1a5e075afb04d5e408e4af08f866ef4ea38a4b7f | 01:57 |
shogun-notifier- | shogun: bug fixes in Jade.cpp | 01:57 |
shogun-notifier- | shogun: Kevin :develop * d468914 / src/shogun/converter/ica/Jade.cpp,src/shogun/converter/ica/Jade.h: https://github.com/shogun-toolbox/shogun/commit/d468914f70d2c2f56e9c0638f4127788d0291802 | 01:57 |
shogun-notifier- | shogun: fixed Jade | 01:57 |
shogun-notifier- | shogun: Kevin Hughes :develop * f499939 / src/shogun/converter/ica/Jade.cpp: https://github.com/shogun-toolbox/shogun/commit/f49993991a55d9e83f172025e986cfd38440cc98 | 01:57 |
shogun-notifier- | shogun: Jade fixed looping code style | 01:57 |
shogun-notifier- | shogun: Kevin :develop * 41e938b / src/shogun/converter/ica/Jade.cpp: https://github.com/shogun-toolbox/shogun/commit/41e938b63c3f8b0f34ced472afa9f30eac5b554c | 01:57 |
shogun-notifier- | shogun: Jade, code clean up around the eigenvectors | 01:57 |
shogun-notifier- | shogun: Kevin :develop * bdada87 / src/shogun/converter/ica/Jade.cpp,src/shogun/converter/ica/Jade.h: https://github.com/shogun-toolbox/shogun/commit/bdada874070a851e9a4f2eab72ae252519d8504b | 01:57 |
shogun-notifier- | shogun: Jade cleanup | 01:57 |
shogun-notifier- | shogun: Kevin :develop * 4307eee / tests/unit/converter/ica/Jade_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/4307eee48e2738b6ff14cbe6ac9ca83dae71f4ae | 01:57 |
shogun-notifier- | shogun: fixed Jade unit test | 01:57 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 783117c / / (3 files): https://github.com/shogun-toolbox/shogun/commit/783117c6af1a8fc42801463f24ece5ec1131448e | 01:57 |
shogun-notifier- | shogun: Merge pull request #1296 from pickle27/develop | 01:57 |
shogun-notifier- | shogun: | 01:57 |
shogun-notifier- | shogun: Fixed Jade | 01:57 |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Leaving] | 01:57 | |
shogun-buildbot | build #1544 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1544 blamelist: Kevin <kevinhughes27@gmail.com> | 02:14 |
shogun-buildbot | build #1545 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1545 blamelist: Kevin <kevinhughes27@gmail.com>, Kevin Hughes <kevinhughes27@gmail.com> | 02:17 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has joined #shogun | 02:17 | |
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/9450327 | 02:17 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has left #shogun [] | 02:17 | |
shogun-buildbot | build #1546 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1546 | 02:29 |
shogun-buildbot | build #1067 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1067 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com> | 02:30 |
shogun-buildbot | build #754 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/754 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com> | 02:31 |
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has joined #shogun | 02:37 | |
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/9450535 | 02:37 |
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has left #shogun [] | 02:37 | |
shogun-buildbot | build #1271 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1271 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com> | 02:39 |
-!- travis-ci [~travis-ci@ec2-75-101-178-60.compute-1.amazonaws.com] has joined #shogun | 02:59 | |
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/9451106 | 02:59 |
-!- travis-ci [~travis-ci@ec2-75-101-178-60.compute-1.amazonaws.com] has left #shogun [] | 02:59 | |
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has quit [Ping timeout: 240 seconds] | 03:06 | |
shogun-buildbot | build #403 of nightly_all is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/403 | 03:08 |
shogun-buildbot | build #1388 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1388 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com> | 03:15 |
-!- travis-ci [~travis-ci@ec2-75-101-178-60.compute-1.amazonaws.com] has joined #shogun | 03:41 | |
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/9452059 | 03:41 |
-!- travis-ci [~travis-ci@ec2-75-101-178-60.compute-1.amazonaws.com] has left #shogun [] | 03:41 | |
shogun-buildbot | build #468 of nightly_default is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/468 | 04:15 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has joined #shogun | 04:24 | |
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/9453999 | 04:24 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has left #shogun [] | 04:24 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 04:57 | |
-!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has quit [Quit: Leaving] | 05:22 | |
-!- nube [~rho@116.90.239.3] has joined #shogun | 06:10 | |
-!- nube [~rho@116.90.239.3] has quit [Quit: Leaving.] | 06:20 | |
-!- nube [~rho@116.90.239.3] has joined #shogun | 06:21 | |
gsomix | good morning | 07:35 |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 08:28 | |
-!- gsomix [~gsomix@109.169.225.10] has quit [Ping timeout: 246 seconds] | 08:48 | |
-!- gsomix [~gsomix@109.169.242.187] has joined #shogun | 09:01 | |
-!- lambday [67157f4c@gateway/web/freenode/ip.103.21.127.76] has joined #shogun | 09:52 | |
lambday | HeikoS: sonney2k: sorry my network went down suddenly last night :( | 09:53 |
lambday | HeikoS: adding all template instantiations since before I removed them it was fine (I am not sure why it gives this error) | 09:55 |
gsomix | good morning | 10:11 |
gsomix | awesome! http://photojournal.jpl.nasa.gov/figures/PIA17171_fig1.jpg | 10:12 |
gsomix | "That's us... from Saturn." | 10:13 |
lambday | gsomix: wow! :-o | 10:15 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 10:48 | |
shogun-notifier- | shogun: lambday :develop * 7891ac3 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/7891ac321ffa7cc16257da3d01175c2dc2948d0c | 10:48 |
shogun-notifier- | shogun: template instantiation added, accuracy reduced for log-det unittest | 10:48 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 2dcf81d / / (3 files): https://github.com/shogun-toolbox/shogun/commit/2dcf81dd141d007a6facd92f27164ca7edb4592f | 10:48 |
shogun-notifier- | shogun: Merge pull request #1300 from lambday/feature/log_determinant | 10:48 |
shogun-notifier- | shogun: | 10:48 |
shogun-notifier- | shogun: template instantiation added, accuracy reduced for log-det unittest | 10:48 |
lambday | HeikoS: keeping fingers crossed! I am not sure about this! | 10:48 |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun | 10:51 | |
shogun-buildbot | build #755 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/755 blamelist: lambday <heavensdevil6909@gmail.com> | 10:56 |
lambday | nope! didn't work! | 10:57 |
shogun-buildbot | build #1068 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1068 blamelist: lambday <heavensdevil6909@gmail.com> | 11:01 |
shogun-buildbot | build #1272 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1272 blamelist: lambday <heavensdevil6909@gmail.com> | 11:06 |
shogun-buildbot | build #1548 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1548 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 11:07 |
-!- nube [~rho@116.90.239.3] has quit [Ping timeout: 264 seconds] | 11:07 | |
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:c149:9096:8977:c686] has joined #shogun | 11:07 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 11:08 | |
@iglesiasg | good morning | 11:08 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has joined #shogun | 11:11 | |
travis-ci | [travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/9465161 | 11:11 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has left #shogun [] | 11:11 | |
lisitsyn | iglesiasg: thanks for reviewing and merging pickle27's code! | 11:15 |
@iglesiasg | lisitsyn, no problem at all :) | 11:15 |
thoralf | Hey. | 11:16 |
lisitsyn | hey thoralf | 11:16 |
thoralf | sonney2k: "thoralf, currently the buildbot installs stuff somewhere and then runs the make check-examples etc" <-- I don't understand. With my changes, check-examples relies on the compiled libraries from the source directory. Not the target location. | 11:17 |
thoralf | sonney2k: So, calling "make install" does not hurt, but has no effect on check-examples either. | 11:17 |
thoralf | sonney2k: Did I miss something? | 11:17 |
thoralf | Copy-paste from yesterday. ;) | 11:18 |
-!- nube [~rho@36.253.103.184] has joined #shogun | 11:29 | |
@iglesiasg | lisitsyn, gsomix, can SGNDArray be currently used from interfaces? | 11:38 |
gsomix | iglesiasg, hm, there are typemaps for SGNDArray. | 11:39 |
@iglesiasg | gsomix, all right, then I guess it could be used | 11:40 |
gsomix | iglesiasg, yep | 11:41 |
-!- nube [~rho@36.253.103.184] has quit [Ping timeout: 268 seconds] | 11:46 | |
@sonney2k | thoralf, I meant that it would be good to know we installed everything to compile & run the examples! | 12:00 |
@iglesiasg | sonney2k, hi there! | 12:00 |
@iglesiasg | sonney2k, configure.log in travis clang https://travis-ci.org/shogun-toolbox/shogun/jobs/9451871 | 12:00 |
thoralf | sonney2k: So you're suggesting that we could run both: check-examples on system wide installed shogun as well as without install? | 12:02 |
thoralf | s/could/should be able to/ | 12:02 |
lambday | sonney2k: hi | 12:02 |
lambday | sonney2k: my network went off all of a sudden last night :( sorry | 12:03 |
lambday | sonney2k: if you have a min, could you please take a look at this? http://www.compileonline.com/compile_cpp_online.php | 12:03 |
lambday | sonney2k: this is the situation we have, I am not sure while it works on my machine or travis but not on buildbot :( | 12:03 |
@sonney2k | iglesiasg, see it happens with Generating dependencies! | 12:04 |
@sonney2k | iglesiasg, no warnings during build! | 12:04 |
@sonney2k | thoralf, yes exactly | 12:05 |
-!- nube [~rho@116.90.239.3] has joined #shogun | 12:05 | |
@sonney2k | lambday, the link is not proper... | 12:05 |
-!- nube [~rho@116.90.239.3] has quit [Client Quit] | 12:05 | |
lambday | sonney2k: okay.. pasting on codepad | 12:06 |
thoralf | sonney2k: I see that this would be nice, but it's also a lot of work. Not so say that I can do it right now, but may I suggest: | 12:06 |
@iglesiasg | sonney2k, what is this 'generating dependencies' step exactly doing? | 12:06 |
lambday | sonney2k: http://codepad.org/vSkxeMgD | 12:07 |
@sonney2k | iglesiasg, creating rules which .h / .cpp files are dependencies for a file to be build | 12:07 |
thoralf | sonney2k: My change makes it impossible to run check-examples on the system-wide installation. So we maybe make it possible to create a fake-installation in the shogun directories. | 12:07 |
thoralf | sonney2k: Lets say in shogun/build/ | 12:08 |
thoralf | sonney2k: So the code for "make install" can be used to install both - depending on the configure variables. | 12:08 |
@sonney2k | lambday, templates and virtual functions + inheritance *urgs* | 12:09 |
@sonney2k | lambday, you sure you want that? | 12:09 |
lambday | sonney2k: :( no other way :( | 12:09 |
@sonney2k | lambday, that cries for problems! | 12:09 |
thoralf | sonney2k: More complicated and more flexible setup. | 12:09 |
thoralf | sonney2k: My current changes are tested and I can overlook the side effects. | 12:10 |
@sonney2k | thoralf, that would actually be trivial - just do make install check-examples DESTDIR=../build | 12:10 |
lambday | sonney2k: I don't know what else to do :( everything else seems to work fine.. In fact I used this same for another class and buildbot didn't complain for that!! | 12:10 |
thoralf | sonney2k: While I cannot do this for my lastest idea. | 12:10 |
thoralf | sonney2k: If that's true, my time spent for the PR was wasted. ;) | 12:11 |
@sonney2k | thoralf, I thought you need that for eclipse? | 12:12 |
@sonney2k | I mean to run it in-place | 12:12 |
thoralf | sonney2k: Yes. | 12:12 |
@sonney2k | so the DESTDIR thing doesn't fix anything | 12:12 |
shogun-buildbot | build #1389 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1389 blamelist: lambday <heavensdevil6909@gmail.com> | 12:12 |
@sonney2k | lambday, which buildbot does complain btw? deb1? | 12:13 |
lambday | sonney2k: checking | 12:13 |
lambday | rpm1 | 12:13 |
lambday | sonney2k: | 12:14 |
@sonney2k | lambday, the example compiles on all machines / compilers I've found | 12:15 |
@sonney2k | including rpm1 | 12:15 |
lambday | sonney2k: then I must have missed something | 12:16 |
lambday | sonney2k: checking.. | 12:16 |
@sonney2k | lambday, in any case you should make sure that we don't expose these classes to the outside | 12:16 |
thoralf | sonney2k: Actually I only wanted to submit a small patch... then you encouraged me to change this and to do it for all interfaces. And now it's maybe been for the waste bin. ;) | 12:16 |
lambday | sonney2k: yes.. | 12:16 |
@sonney2k | lambday, ? | 12:17 |
@sonney2k | thoralf, yeah I am evil | 12:18 |
thoralf | sonney2k: Sure! | 12:18 |
lambday | sonney2k: we'd not expose these classes to any modular interfaces.. these are all internals | 12:18 |
@HeikoS | thoralf, sonney2k whats the issue? | 12:19 |
@sonney2k | lambday, ok but what fails then? | 12:21 |
thoralf | HeikoS: We just found out that my PR would be obsolete if "make install check-examples DESTDIR=../build" works. ;) | 12:21 |
@HeikoS | thoralf: ah I see | 12:21 |
@HeikoS | thoralf: damn :) sorry for that | 12:21 |
lambday | sonney2k: I am didn't figure out :( | 12:21 |
lambday | sonney2k: might be that I missed some headers.. checking | 12:22 |
@sonney2k | lambday, I have local access to the rpm1 machine if you want me to try things | 12:22 |
@sonney2k | iglesiasg, I get it now | 12:24 |
@sonney2k | iglesiasg, we use g++ to generate dependencies | 12:24 |
@sonney2k | iglesiasg, but it gets all the compile flags | 12:25 |
@sonney2k | from clang! | 12:25 |
@iglesiasg | sonney2k, oh | 12:25 |
@iglesiasg | so the cc1plus error line is from g++ | 12:25 |
lambday | sonney2k: I don't understand what does the error msg say.. similar type of code is there for CDenseMatrixOperator (except that it doesn't inherit with the same template arg for parent).. | 12:25 |
@sonney2k | iglesiasg, yes! | 12:25 |
@iglesiasg | sonney2k, all right. Then I guess we have to change that so the dependencies are also created with clang when compiling with ti | 12:26 |
@iglesiasg | it* | 12:26 |
@sonney2k | iglesiasg, I guess we can use clang also to generate dependencies | 12:26 |
@iglesiasg | sonney2k, it should be possible indeed | 12:26 |
@sonney2k | iglesiasg, that won't fix the situation with other compilers | 12:27 |
@sonney2k | some might not be able to do gcc -M stuff | 12:27 |
@iglesiasg | sonney2k, what other compilers? BSD cygwin and so? | 12:27 |
@sonney2k | iglesiasg, or we don't pass some flags to that one which might be wrong though if stuff is ifdef'd in the headers | 12:28 |
@iglesiasg | it is headache to make things valid among several compilers or platforms in general | 12:30 |
@sonney2k | iglesiasg, yeah maybe we should just limit support to gcc / clang | 12:32 |
lisitsyn | icc | 12:32 |
@iglesiasg | sonney2k, but which ones are the others? | 12:32 |
@sonney2k | lambday, who has that? | 12:32 |
@iglesiasg | lisitsyn, ^ | 12:32 |
@sonney2k | lisitsyn, did you ever try shogun with that? | 12:32 |
lisitsyn | sonney2k: icc? yes why not | 12:32 |
lisitsyn | it is free | 12:32 |
@sonney2k | and did it work? | 12:32 |
lisitsyn | yes iirc | 12:32 |
lisitsyn | ICC is quite standart compliant I think | 12:33 |
@iglesiasg | I am a ignorant, first time I heard of icc | 12:33 |
lisitsyn | its warnings are even better than clang ones | 12:33 |
@iglesiasg | lisitsyn, really? omg I must try it then! | 12:33 |
lisitsyn | iglesiasg: just download intel composer xe 2013 | 12:33 |
@iglesiasg | lisitsyn, apt-get? | 12:34 |
lisitsyn | iglesiasg: no intel website | 12:34 |
lambday | sonney2k sorry I didn't get :( | 12:34 |
@sonney2k | lambday, meant lisitsyn :) | 12:34 |
lisitsyn | lisitday | 12:34 |
lisitsyn | lamdsyn | 12:34 |
lambday | lol | 12:34 |
@sonney2k | lamdsyny | 12:34 |
lisitsyn | sonney2k: I thought you knew how to spell that but you failed at our workshop! | 12:35 |
@sonney2k | ?? | 12:35 |
lisitsyn | sonney2k: well you said you don't know how to spell my last name | 12:35 |
gsomix | sonney2k, hey | 12:35 |
gsomix | sonney2k, have a minute for review? | 12:35 |
lisitsyn | HeikoS: sorry didn't have time to fill in some discussion in that issue | 12:36 |
@sonney2k | lambday, which g++ version is it that fails then? | 12:36 |
@HeikoS | lisitsyn: dont worry, maybe just later | 12:36 |
thoralf | sonney2k: Okay, your idea does not work, since $(DESTDIR) is concattenated with other paths from src/.config. | 12:36 |
lambday | sonney2k: the one that rpm1 has.. it passes in mine and on all of travis's | 12:36 |
@HeikoS | lisitsyn, sonney2k I plan to start adding shogun development ugidelines to the github wiki, maybe we can even link them from the webpage | 12:36 |
thoralf | sonney2k: So if you have some complicated prefix path, it will be also contained in $DESTDIR | 12:37 |
@HeikoS | lisitsyn, sonney2k we can use that to document general ideas and also particular concepts like SGVector etc | 12:37 |
* sonney2k installs g++-4.4 | 12:37 | |
lambday | sonney2k: which one did it have earlier? | 12:37 |
lisitsyn | sonney2k: necromancy stuff | 12:38 |
@sonney2k | HeikoS, please do that in README.developer | 12:39 |
@HeikoS | sonney2k: I disagree | 12:39 |
@sonney2k | thoralf, it just preprends stuff to prefix. | 12:39 |
@sonney2k | HeikoS, ? | 12:39 |
@HeikoS | sonney2k: would like to have something public available colaborative easy access | 12:39 |
@HeikoS | nobody reads readme | 12:39 |
@sonney2k | HeikoS, nobody reads wikis | 12:39 |
@HeikoS | sonney2k: so they get outdated | 12:39 |
@HeikoS | sonney2k: we can synch the Readme by the wiki | 12:40 |
@HeikoS | also readmes is just text no markdown | 12:40 |
@sonney2k | sure | 12:40 |
@sonney2k | HeikoS, it is not the format or anything it is the tl;dr issue | 12:40 |
@HeikoS | sonney2k: yeah, wikis are easier to digest I think, also most projects use them | 12:41 |
@HeikoS | sonney2k: readme is all very old-school | 12:41 |
@HeikoS | sonney2k: but I think we should still have them, just in addition | 12:41 |
thoralf | sonney2k: But the way paths are concattenated in make check-examples does not work. The resulting paths do not exist. | 12:41 |
thoralf | sonney2k: DESTDIR need to be (at least) absolute. | 12:42 |
thoralf | sonney2k: But it really works. | 12:42 |
@sonney2k | thoralf, yeah could be | 12:42 |
@sonney2k | thoralf, but I still like the idea of having a make check-examples locally *and* for the installed thing | 12:43 |
thoralf | sonney2k: Yeah, but if "make check-examples-local" (working title ;)) just does a "make install check-examples DESTDIR=/foo", it's already done. | 12:44 |
@sonney2k | thoralf, true would be sufficient | 12:44 |
@sonney2k | didn't think of that | 12:45 |
@sonney2k | lambday, compiles with g++-4.4.7 here | 12:45 |
lambday | sonney2k: on rpm1? | 12:46 |
lambday | sonney2k: shogun or the example code? | 12:46 |
@sonney2k | lambday, on my machine compiling shogun | 12:47 |
@sonney2k | lambday, rpm1 has the same g++ version (but redhat...) | 12:47 |
@sonney2k | and it fails there | 12:47 |
lambday | sonney2k: mine is redhat (fedora) :( and it works fine.. mine is gcc 4.6.3 | 12:48 |
lambday | sonney2k: no clue what's wrong :( | 12:48 |
shogun-notifier- | shogun: Evgeniy Andreev :develop * b7e6d6f / / (7 files): https://github.com/shogun-toolbox/shogun/commit/b7e6d6f964a030695ef9cd16c4156e3d3cba25c6 | 12:51 |
shogun-notifier- | shogun: LineReader and CircularBuffer are modified | 12:51 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 74263b9 / / (7 files): https://github.com/shogun-toolbox/shogun/commit/74263b99cfbc374b4882d153ce94f1f7191ba1be | 12:51 |
shogun-notifier- | shogun: Merge pull request #1275 from gsomix/feature/linereader | 12:51 |
shogun-notifier- | shogun: | 12:51 |
shogun-notifier- | shogun: Bug fixes and re-factoring for LineReader and CircularBuffer | 12:51 |
gsomix | sonney2k, cool. thanks! | 12:53 |
gsomix | sonney2k, another part will be little later. | 12:53 |
lambday | HeikoS: we didn't notice any failing compilation on rpm1 when I added that class, right? | 12:56 |
thoralf | sonney2k: I could do another PR with this change, since it seems to work. It's just a new make target an will thus have no side effects, except for a build-dir that's not gonna be cleaned up. | 12:57 |
thoralf | sonney2k: would a "rm $(BUILD_DIR)" after successful build be okay? | 12:58 |
@HeikoS | lambday: not really | 13:00 |
lambday | HeikoS: can we access to past build logs? | 13:00 |
lambday | sonney2k: HeikoS only a few are there.. | 13:00 |
shogun-buildbot | build #756 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/756 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:01 |
@HeikoS | lambday: I think it is possible to access them | 13:01 |
lambday | HeikoS: how? that might be helpful | 13:01 |
@sonney2k | lambday, well tests were broken so the buildbot didn't run things | 13:02 |
lambday | sonney2k: oh | 13:02 |
@sonney2k | lambday, I managed to create a different error | 13:02 |
lambday | sonney2k: different error? | 13:02 |
@HeikoS | sonney2k: if I add another .i file to moduarl interfaces, is it used automatically, or is there a list? | 13:03 |
@sonney2k | lambday, http://dpaste.com/1316870/ | 13:03 |
lisitsyn | HeikoS: there is a list somewhere | 13:03 |
lisitsyn | modshogun or so | 13:03 |
@sonney2k | HeikoS, you then have to add it to modshogun.i etc | 13:03 |
@iglesiasg | HeikoS, in Makefile.template is this list I think | 13:04 |
lambday | sonney2k: checking... | 13:04 |
@iglesiasg | HeikoS, but I think you need to modify that if you add a new directory rather than just a file | 13:04 |
lambday | sonney2k: so its not taking default template arg | 13:05 |
shogun-buildbot | build #1069 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1069 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:06 |
lambday | sonney2k: could you please add an instantiation in CIterativeLinearSolver.cpp for <float64_t, float64_t>? | 13:06 |
lambday | sonney2k: I think this should work :( | 13:06 |
@sonney2k | lambday, did this already | 13:06 |
@sonney2k | now the buildbot removed my dir | 13:06 |
@sonney2k | retrying | 13:06 |
lambday | sonney2k: sorry for this trouble :( | 13:07 |
shogun-buildbot | build #1273 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1273 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:09 |
shogun-buildbot | build #1549 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1549 blamelist: Evgeniy Andreev <gsomix@gmail.com> | 13:12 |
@sonney2k | lambday, the error msg is weird IMHO struct shogun::CIterativeLinearSolver<double, double>’ | 13:14 |
@sonney2k | why struct? | 13:14 |
lambday | sonney2k: no idea! I've seen the same at many places... they call it struct :( | 13:16 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has joined #shogun | 13:16 | |
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/9468545 | 13:16 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has left #shogun [] | 13:16 | |
thoralf | sonney2k: Another try? ;) https://github.com/shogun-toolbox/shogun/pull/1301 | 13:20 |
@HeikoS | iglesiasg, sonney2k thanks! | 13:20 |
@sonney2k | lambday, found it | 13:20 |
-!- nube [~rho@116.90.239.3] has joined #shogun | 13:20 | |
lambday | sonney2k: what is it? | 13:21 |
@sonney2k | lambday, the rpm1 doesn't hav EIGEN3 | 13:21 |
@sonney2k | e | 13:21 |
@sonney2k | lambday, so some classes in the hierarchy are not available | 13:21 |
lambday | errr! and that fails? but I've put #ifdef | 13:21 |
@sonney2k | lambday, you can reproduce it locally when you do ./configure --disable-eigen3 | 13:21 |
@sonney2k | lambday, some class must be missing it | 13:21 |
lambday | sonney2k: thankssss I'm checking | 13:21 |
lambday | sonney2k: alright... :D | 13:21 |
lambday | phew! | 13:22 |
@HeikoS | iglesiasg, sonney2k thanks! | 13:22 |
@iglesiasg | HeikoS, you are welcome! :) | 13:23 |
shogun-notifier- | shogun: Thoralf Klein :develop * 9fe7d6b / src/Makefile: https://github.com/shogun-toolbox/shogun/commit/9fe7d6ba17a8b4ed4365e8322e47b6b2430451ea | 13:23 |
shogun-notifier- | shogun: Added special make targets to allow builds without installing them. | 13:23 |
shogun-notifier- | shogun: | 13:23 |
shogun-notifier- | shogun: Just call: $ make {install,tests,check-examples,run-testsuite}-local-build clean-local-build | 13:23 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 7d5f236 / src/Makefile: https://github.com/shogun-toolbox/shogun/commit/7d5f236225387e1b137d631c9c23cbbd9c98912b | 13:23 |
shogun-notifier- | shogun: Merge pull request #1301 from tklein23/allow_builds_in_src_tree | 13:23 |
shogun-notifier- | shogun: | 13:23 |
shogun-notifier- | shogun: Added special make targets to allow builds without installing them. | 13:23 |
thoralf | Yay. ;) | 13:23 |
@sonney2k | HeikoS, thanks for what? | 13:23 |
@HeikoS | sonney2k: question: if the apply() method of machine is marked to return new objects, is that applied to a GP when its defined in another .i file? | 13:23 |
lisitsyn | HeikoS: I think so | 13:24 |
lisitsyn | HeikoS: well order matters | 13:24 |
thoralf | HeikoS, iglesiasg: We now have new make targets. Just try "make check-examples-local-build" (and don't blame me on the long name ;)) | 13:24 |
@sonney2k | thoralf, well you are to be blamed could be named check-examples-local at least :) | 13:24 |
@HeikoS | lisitsyn: cool thanks! | 13:24 |
@HeikoS | lisitsyn: I put all GP things in a seperate file, its too messy otherwise | 13:25 |
@sonney2k | HeikoS, parse error | 13:25 |
@iglesiasg | thoralf, this is to check examples without installing, right? | 13:25 |
@iglesiasg | thoralf, so first we make normally and then that | 13:25 |
thoralf | iglesiasg: Right. | 13:25 |
@iglesiasg | thoralf, all right, thank you! | 13:25 |
thoralf | iglesiasg: Internally it calls "make install", but to a directory inside the source tree. | 13:25 |
thoralf | iglesiasg: Hidden in shogun/structured/... ;) | 13:25 |
thoralf | sonney2k: What about check-examples-l? ;) | 13:26 |
@sonney2k | thoralf, you should choose the least used directory | 13:36 |
@sonney2k | thoralf, or the one with the most classes :) | 13:36 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 308f52a / src/Makefile.template,src/configure: https://github.com/shogun-toolbox/shogun/commit/308f52af693edbf3c3623df6de60b10150d72e0e | 13:37 |
shogun-notifier- | shogun: generate dependencies with default c++ compiler or gcc or clang or | 13:37 |
shogun-notifier- | shogun: makedepend - whatever is available first | 13:37 |
thoralf | sonney2k: Trying to maximize depth-in-tree should do something similar ;) | 13:37 |
shogun-buildbot | build #1390 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1390 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:38 |
@sonney2k | thoralf, hehe | 13:41 |
@sonney2k | iglesiasg, that should fix it | 13:41 |
shogun-buildbot | build #757 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/757 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:43 |
shogun-buildbot | build #1551 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1551 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 13:43 |
@sonney2k | HeikoS, do you know when wiking is back? | 13:43 |
shogun-buildbot | build #758 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/758 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:45 |
shogun-buildbot | build #1070 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1070 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:47 |
lambday | sonney2k: HeikoS: sent the patch.. | 13:50 |
@sonney2k | lambday, excellent! | 13:50 |
lambday | HeikoS: few gp related unit-tests fail to compile in absense of eigen3 | 13:51 |
@sonney2k | lambday, you need to ifdef them too then | 13:51 |
shogun-buildbot | build #1274 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1274 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:51 |
lambday | I should always make shogun thrice before sending a PR.. eigen3, older eigen3, and without eigen3 | 13:51 |
shogun-notifier- | shogun: lambday :develop * 2f15df7 / src/shogun/mathematics/logdet/ (5 files): https://github.com/shogun-toolbox/shogun/commit/2f15df7eb753a21c6ec0a86d49a0a18a5f6fa076 | 13:52 |
shogun-notifier- | shogun: fixed compilation error in absense of eigen3 | 13:52 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * cf32c71 / src/shogun/mathematics/logdet/ (5 files): https://github.com/shogun-toolbox/shogun/commit/cf32c71e1db435274bb4225f3d0449bc7fb3eeb1 | 13:52 |
shogun-notifier- | shogun: Merge pull request #1302 from lambday/feature/log_determinant | 13:52 |
shogun-notifier- | shogun: | 13:52 |
shogun-notifier- | shogun: fixed compilation error in absense of eigen3 | 13:52 |
@sonney2k | lambday, impossible to take care of everything | 13:52 |
@sonney2k | alright | 13:52 |
@sonney2k | blue time! | 13:52 |
@sonney2k | cu | 13:52 |
lambday | sonney2k: ciao :) | 13:52 |
shogun-buildbot | build #1071 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1071 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:55 |
lisitsyn | lambday: http://eigen.tuxfamily.org/dox-devel/classEigen_1_1MappedSparseMatrix.html you may be interested | 14:02 |
shogun-buildbot | build #1275 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1275 blamelist: Soeren Sonnenburg <sonne@debian.org> | 14:02 |
lambday | lisitsyn: aha! thanks.. checking | 14:04 |
@iglesiasg | thoralf, trying make-check-examples-loca-build right now | 14:04 |
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has joined #shogun | 14:09 | |
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/9469527 | 14:09 |
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has left #shogun [] | 14:09 | |
shogun-buildbot | build #1072 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1072 blamelist: lambday <heavensdevil6909@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 14:09 |
shogun-buildbot | build #759 of rpm1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/759 | 14:10 |
shogun-buildbot | build #1276 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1276 blamelist: lambday <heavensdevil6909@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 14:13 |
-!- nube [~rho@116.90.239.3] has quit [Ping timeout: 268 seconds] | 14:18 | |
thoralf | iglesiasg: Did it work? | 14:30 |
@iglesiasg | thoralf, yes | 14:30 |
@iglesiasg | well it crashed because I am missing some dependencies :D | 14:30 |
@iglesiasg | but your thing worked indeed | 14:31 |
thoralf | Phew :) | 14:31 |
-!- van51 [~van51@athedsl-399972.home.otenet.gr] has joined #shogun | 14:34 | |
van51 | hello | 14:34 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 5f56476 / src/interfaces/modular/ (6 files): https://github.com/shogun-toolbox/shogun/commit/5f5647643d3586af67f33ea16461e96f05aeb334 | 14:37 |
shogun-notifier- | shogun: cleaned up GP interfaces and added an own file for them | 14:37 |
shogun-notifier- | shogun: Heiko Strathmann :develop * d71edc3 / src/interfaces/modular/GaussianProcess.i/ (2 files): https://github.com/shogun-toolbox/shogun/commit/d71edc34ca6a1fde4b9b2d747cd22afa55e24c74 | 14:37 |
shogun-notifier- | shogun: added new files | 14:37 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 6b80376 / src/interfaces/modular/ (8 files): https://github.com/shogun-toolbox/shogun/commit/6b80376a41cca7aabb03e4cfb8c7c3d3ceaeb250 | 14:37 |
shogun-notifier- | shogun: Merge pull request #1303 from karlnapf/develop | 14:37 |
shogun-notifier- | shogun: | 14:37 |
shogun-notifier- | shogun: moved GP modular inferfaces to its own file | 14:37 |
@HeikoS | gsomix, sonney2k lib/CircularBuffer.cpp: In member function 'bool shogun::CCircularBuffer::has_next()': | 14:39 |
@HeikoS | lib/CircularBuffer.cpp:183:10: warning: unused variable 'tail_length' [-Wunused-variable] | 14:39 |
shogun-buildbot | build #1392 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1392 blamelist: lambday <heavensdevil6909@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 14:41 |
thoralf | HeikoS: what about -Werror when building to force fixing warnings immediately? | 14:42 |
@HeikoS | thoralf: I am up for that | 14:42 |
@HeikoS | thoralf: but others dont like it | 14:42 |
@HeikoS | some external libs are also causing warnings | 14:42 |
@HeikoS | sonney2k: ^ what do you think on Werror for the main build? | 14:42 |
thoralf | HeikoS: Yeah, but at least the library itself does not have any warnings by now. | 14:43 |
@HeikoS | thoralf: I know, I would totally activate it | 14:43 |
@HeikoS | for everything in fact | 14:43 |
@HeikoS | also python swig | 14:43 |
thoralf | HeikoS: with -Werror we don't rist degrading. :) | 14:43 |
@HeikoS | (only warnings of docs are missing or variable names or python incompatible=) | 14:43 |
thoralf | s/rist/risk/ | 14:44 |
lisitsyn | I find most warnings useless | 14:44 |
lisitsyn | well some of them are crazy useful like errors but some of them not | 14:45 |
lisitsyn | I am unsure we can get back to 0 warnings that way | 14:45 |
gsomix | HeikoS, ok, tnx. | 14:48 |
@HeikoS | lisitsyn: still we want to keep the build clean right? | 14:48 |
@HeikoS | lisitsyn: we regularily have zero-warning hunts | 14:48 |
@HeikoS | which take hours | 14:48 |
@HeikoS | Werror would avoid that | 14:49 |
lisitsyn | HeikoS: ohh well we spent quite some time to fix real errors | 14:49 |
lisitsyn | with warnings I am afraid it could get permanently red | 14:49 |
thoralf | lisitsyn, Heiko: What about making -Werror default for developers, but disabling in travis? | 14:50 |
shogun-buildbot | build #1391 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1391 blamelist: Soeren Sonnenburg <sonne@debian.org> | 14:50 |
@HeikoS | lisitsyn: well if it IS green once | 14:50 |
@HeikoS | then we could activate and see | 14:51 |
@HeikoS | thoralf: I think it would be most useful in travis actually | 14:51 |
@HeikoS | not on the buildbots | 14:51 |
shogun-buildbot | build #1073 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1073 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 14:51 |
@HeikoS | lisitsyn: ^ | 14:51 |
@HeikoS | lisitsyn: do you use ccache? | 14:51 |
lisitsyn | HeikoS: yes | 14:52 |
@HeikoS | lisitsyn: I have the feeling it doesnt do anything here anymore | 14:53 |
@HeikoS | I installed it via linking ccache to /usr/local/bin/gcc, g++, cc and /usr/local/bin is before gcc path in path | 14:54 |
@HeikoS | which gcc | 14:54 |
@HeikoS | /usr/local/bin/gcc | 14:54 |
@HeikoS | lisitsyn: but still it feels slow (I remember before it was lightening fast) | 14:54 |
shogun-buildbot | build #1277 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1277 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 14:54 |
lisitsyn | HeikoS: which /usr/local/bin/gcc | 14:55 |
lisitsyn | ? | 14:55 |
lisitsyn | HeikoS: just ensure you really call ccache! | 14:56 |
@HeikoS | lisitsyn: what do you mean? | 14:56 |
lisitsyn | HeikoS: what is called with /usr/local/bin/gcc? | 14:56 |
@HeikoS | lisitsyn: see above | 14:57 |
@HeikoS | lisitsyn: ah | 14:57 |
@HeikoS | sorry | 14:57 |
@HeikoS | lisitsyn: its a soft link to ccache | 14:57 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 69ce15f / src/Makefile.template: https://github.com/shogun-toolbox/shogun/commit/69ce15f3cc8d1299b6dee33a1d081578bcc78939 | 14:59 |
shogun-notifier- | shogun: added GaussianProcess in modular interface structure | 14:59 |
shogun-notifier- | shogun: Heiko Strathmann :develop * e304c64 / src/Makefile.template: https://github.com/shogun-toolbox/shogun/commit/e304c64aa7e1360cf19cf01f71473157288a6333 | 14:59 |
shogun-notifier- | shogun: Merge pull request #1304 from karlnapf/develop | 14:59 |
lisitsyn | HeikoS: what you can check now | 14:59 |
shogun-notifier- | shogun: | 14:59 |
shogun-notifier- | shogun: added GaussianProcess in modular interface structure | 14:59 |
lisitsyn | HeikoS: is whether your ccache is filled | 15:00 |
lisitsyn | with something | 15:00 |
lisitsyn | there is a key that stands for stats | 15:00 |
lisitsyn | I don't remember exactly | 15:00 |
lisitsyn | HeikoS: if you see nothing cached something is obviously wrong | 15:00 |
@HeikoS | lisitsyn: I have another question | 15:01 |
@HeikoS | lisitsyn: swig related | 15:01 |
lisitsyn | yes | 15:01 |
@HeikoS | GaussianProcessBinaryClassification(ExactInferenceMethod() ) | 15:01 |
@HeikoS | NotImplementedError: Wrong number or type of arguments for overloaded function 'new_GaussianProcessBinaryClassification'. | 15:02 |
@HeikoS | Possible C/C++ prototypes are: | 15:02 |
@HeikoS | shogun::CGaussianProcessBinaryClassification::CGaussianProcessBinaryClassification() | 15:02 |
@HeikoS | shogun::CGaussianProcessBinaryClassification::CGaussianProcessBinaryClassification(CInferenceMethod *) | 15:02 |
@HeikoS | swig somehow doesnt realise that ExactInferenceMethod is a subclass of InferenceMethod | 15:02 |
@HeikoS | lisitsyn: do you know this problem? | 15:02 |
lisitsyn | HeikoS: try forward declaration | 15:02 |
@HeikoS | lisitsyn: in which file? | 15:03 |
@HeikoS | lisitsyn: or: what is that? | 15:03 |
lisitsyn | HeikoS: a lot of possible cases actually | 15:03 |
lisitsyn | HeikoS: you may try to forward declare CInferenceMethod in CGPBC | 15:04 |
@HeikoS | lisitsyn: really? | 15:04 |
@HeikoS | lisitsyn: that is usually not necessary | 15:04 |
shogun-buildbot | build #1074 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1074 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 15:04 |
@HeikoS | its a subclass | 15:04 |
@HeikoS | should have the same type | 15:04 |
lisitsyn | HeikoS: that's just swiggy-wiggy stuff | 15:04 |
@HeikoS | lisitsyn: but I never had that with other classes | 15:04 |
lisitsyn | I remember it can be solved that way sometimes | 15:05 |
lisitsyn | HeikoS: no guarantees but that's what I'd try | 15:05 |
@HeikoS | lisitsyn: ok I will try thanks | 15:06 |
shogun-buildbot | build #1075 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1075 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 15:06 |
lisitsyn | HeikoS: I'd say it is either some missed #include | 15:06 |
lisitsyn | or forward declaration | 15:07 |
@HeikoS | lisitsyn: ah its probably the order | 15:10 |
@HeikoS | I remember s?ren told me about that | 15:10 |
@HeikoS | if I only could remember those things ;) | 15:11 |
@HeikoS | lisitsyn: yeaaah worked :) | 15:13 |
@HeikoS | now ccahce is the only problem | 15:13 |
shogun-buildbot | build #1278 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1278 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 15:14 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 28b5644 / src/interfaces/modular/GaussianProcess.i/ (2 files): https://github.com/shogun-toolbox/shogun/commit/28b56449575dc9fa3611a36afbc2687e68db0daf | 15:14 |
shogun-notifier- | shogun: changed include order to get rid of some error in modular | 15:14 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 5a95118 / src/interfaces/modular/GaussianProcess.i/ (2 files): https://github.com/shogun-toolbox/shogun/commit/5a95118f37daa8f01b8fdc5b8f8e4e629cfa529b | 15:14 |
shogun-notifier- | shogun: Merge pull request #1305 from karlnapf/develop | 15:14 |
shogun-notifier- | shogun: | 15:14 |
shogun-notifier- | shogun: changed include order to get rid of some error in modular | 15:14 |
thoralf | HeikoS: https://github.com/shogun-toolbox/shogun/pull/1306 | 15:23 |
thoralf | HeikoS: Thanks ;) | 15:23 |
shogun-buildbot | build #1279 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1279 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 15:25 |
-!- travis-ci [~travis-ci@ec2-75-101-178-60.compute-1.amazonaws.com] has joined #shogun | 15:26 | |
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/9470299 | 15:26 |
-!- travis-ci [~travis-ci@ec2-75-101-178-60.compute-1.amazonaws.com] has left #shogun [] | 15:26 | |
lisitsyn | HeikoS: yes it is usually order | 15:27 |
@HeikoS | lisitsyn: another thing for the dev wiki :) | 15:27 |
lisitsyn | HeikoS: devil wiki | 15:28 |
@HeikoS | lisitsyn: ? | 15:28 |
lisitsyn | HeikoS: jk | 15:29 |
@iglesiasg | thoralf, should we add src/local-build-tmp to gitignore? | 15:29 |
@HeikoS | lisitsyn: ehm? whaat? :) I dont know any acronyms | 15:29 |
lisitsyn | HeikoS: I mean it could be hell | 15:29 |
@HeikoS | lisitsyn: what? | 15:29 |
thoralf | iglesiasg: Of course. I'll fix that. | 15:30 |
@iglesiasg | thoralf, nice! thank you | 15:30 |
lisitsyn | HeikoS: all the hidden things | 15:30 |
@HeikoS | lisitsyn: I dont follow | 15:30 |
lisitsyn | HeikoS: there are a lot of things like that to document | 15:30 |
@HeikoS | lisitsyn: yeah, so we are currently 8+5 people | 15:31 |
@HeikoS | if everyone does a little bit, this is easy | 15:31 |
@HeikoS | thats why I want to set up a a wiki | 15:31 |
gsomix | http://d24w6bsrhbeh9d.cloudfront.net/photo/axNGMxK_460s.jpg | 15:47 |
gsomix | 9gag time! | 15:47 |
shogun-buildbot | build #1394 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1394 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 15:55 |
shogun-buildbot | build #1393 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1393 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 15:57 |
shogun-buildbot | build #1558 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1558 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:14 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has joined #shogun | 16:16 | |
travis-ci | [travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/9471828 | 16:16 |
-!- travis-ci [~travis-ci@ec2-50-19-69-230.compute-1.amazonaws.com] has left #shogun [] | 16:16 | |
thoralf | Can anyone explain, why SGVector defines so many const-methods "display_vector()" and one non-const "SGVector<T>::display_vector()" which just calls const ones? | 16:19 |
-!- lambday [67157f4c@gateway/web/freenode/ip.103.21.127.76] has quit [Ping timeout: 250 seconds] | 16:20 | |
thoralf | Is there a difference to just create a bunch of non-const methods? | 16:20 |
@iglesiasg | thoralf, it is good to const things wherever you can | 16:29 |
thoralf | iglesiasg: As a rule-of-thumb: Okay, got it. | 16:30 |
@iglesiasg | thoralf, but that does not mean that it is properly done in SGVector :) | 16:31 |
@iglesiasg | thoralf, I think that in general our const-ness is not very good. We should refactor some parts related to it | 16:31 |
thoralf | But can you elaborate that? When only using the non-const version, which calls the const version... isn't that a bit overhead for "just try const everywhere"? ;) | 16:31 |
@iglesiasg | thoralf, according to what you have said, I see no reason why there is a non-const version | 16:32 |
@iglesiasg | thoralf, if a method doesn't modify the argument, then it should be const and that's all | 16:33 |
@iglesiasg | thoralf, the same method may be repeated in a non-const and const version for the return value | 16:33 |
thoralf | Hmm. Maybe there's a trick that the compiler already knows on compile time, which non-const method to take and doesn't need a vtable lookup? lisitsyn? | 16:33 |
@iglesiasg | thoralf, I don't know if it is about vtable lookup, but declaring const arguments allow the compiler to do optimizations (not copy things around and so) | 16:34 |
thoralf | iglesiasg: It's not about the arguments. It's about the methods which are declared const. | 16:35 |
thoralf | iglesiasg: members vs. functions | 16:35 |
@iglesiasg | thoralf, ah all right | 16:35 |
@iglesiasg | thoralf, is the one that is non-const static? | 16:35 |
van51 | there is a problem with this unit test : FactorGraph.evaluate_energy_param_data | 16:37 |
thoralf | iglesiasg: Yes. | 16:37 |
@iglesiasg | van51, yep. I already made an issue for hushell about that https://github.com/shogun-toolbox/shogun/issues/1298 | 16:37 |
van51 | iglesiasg: ok then :) | 16:38 |
@iglesiasg | let's see if we can fix it soon | 16:38 |
van51 | iglesiasg: weird thing is it crashes only when I call all the tests together | 16:38 |
-!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun | 16:38 | |
@iglesiasg | thoralf, then you got your answer in there. It makes no sense to have a static const method | 16:38 |
thoralf | iglesiasg: SGVector<T>::display_vector() is const, while the static versions SGVector<...>::display_vector() are declared for every type. | 16:38 |
@iglesiasg | van51, yeah, it has some jumps around depending on non-initialised memory and so -> undefined behaviour | 16:38 |
van51 | iglesiasg: ah I see! | 16:39 |
pickle27 | iglesiasg: saw your feedback and it totally makes sense | 16:39 |
thoralf | iglesiasg: const SGVector<T>::display_vector() calls static display_vector() | 16:40 |
pickle27 | iglesiasg: pass by ref made sense when I started but almost all the ajd algs immediatly make a copy as to not modifiy the original anyways | 16:40 |
thoralf | iglesiasg: Why not have many const and no static? | 16:40 |
@iglesiasg | pickle27, the thing is that in languages like python the difference pass by ref / pass by value do not exist | 16:41 |
pickle27 | iglesiasg: yes that too | 16:41 |
@iglesiasg | pickle27, so when you get a method with & in one of the arguments, it is unusable from interfaces | 16:41 |
@iglesiasg | pickle27, gsomix mentioned something about doing typemaps so it could work | 16:41 |
pickle27 | and correct me if I'm wrong but python usability seems to be a big focus for us | 16:41 |
@iglesiasg | pickle27, indeed it is :) | 16:41 |
@iglesiasg | everyone should be using python! :P | 16:41 |
pickle27 | iglesiasg: I'm a fan of python so I'm pretty happy about that | 16:42 |
@iglesiasg | pickle27, in any case, pass by value of SGVector or SGmatrix is not a crime; just copy one or two integers and a pointer | 16:42 |
pickle27 | ah okay, I hadn't looked at the behind the scenes to see what was happening | 16:43 |
@iglesiasg | thoralf, I agree with you in that the static display_vector(const SGVector<T>, ...) could be droppe | 16:45 |
@iglesiasg | dropped* | 16:46 |
-!- gsomix_ [~gsomix@95.67.134.30] has joined #shogun | 17:06 | |
-!- gsomix [~gsomix@109.169.242.187] has quit [Read error: Connection reset by peer] | 17:09 | |
pickle27 | lisitsyn: hey | 17:10 |
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has joined #shogun | 17:10 | |
travis-ci | [travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/9472899 | 17:10 |
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has left #shogun [] | 17:10 | |
lisitsyn | pickle27: hey! | 17:11 |
lisitsyn | pickle27: that's cool jade is finally here | 17:11 |
pickle27 | lisitsyn: I know, that bug was driving me nuts! | 17:11 |
lisitsyn | pickle27: I haven't checked andreas ipython thing yet - what about it? | 17:11 |
pickle27 | lisitsyn: thats what I was just going to message about, its pretty awesome | 17:11 |
pickle27 | lisitsyn: he basically did what I was planning to spend the rest of this week on lol | 17:12 |
@iglesiasg | can we see that ipython? :) | 17:12 |
pickle27 | iglesiasg: yeah for sure, how is best to share? | 17:12 |
pickle27 | lisitsyn: I made my own ipython after the meeting based on my python example | 17:13 |
@iglesiasg | pickle27, is it online somewhere? or do you have the source instead? | 17:13 |
@HeikoS | hushell: unit tests fail on factor graphs | 17:13 |
pickle27 | but then andreas made it work with some simple audio | 17:13 |
@HeikoS | hushell: could you fix? | 17:13 |
pickle27 | iglesiasg: I have the source | 17:13 |
@iglesiasg | I have made none ipython notepad yet :( | 17:13 |
pickle27 | iglesiasg: its so easy! and amazing | 17:13 |
@iglesiasg | pickle27, mail then maybe? | 17:13 |
@HeikoS | pickle27, iglesiasg the notebook should be checked into tutorial folder, wherer there are some already, please create a new folder gsoc_2013 for that and a subfolder with your topic | 17:14 |
@iglesiasg | HeikoS, I made an issue about that, I think hushell must be already on it :) | 17:14 |
@HeikoS | iglesiasg: I see, thanks! :) | 17:14 |
@HeikoS | iglesiasg: this should not be merged | 17:14 |
@HeikoS | iglesiasg: unit tests were green so travis tells us before | 17:14 |
@iglesiasg | HeikoS, yeah but hushell was testing it in his machine and didn't find anything wrong with it | 17:15 |
@HeikoS | iglesiasg: I see | 17:15 |
@iglesiasg | HeikoS, and that pull request had been there open for ages (2000 lines of code already :S ) | 17:15 |
@HeikoS | iglesiasg, hushell if its not solved by tonight, please remove the test temporarily | 17:15 |
@HeikoS | iglesiasg: wow, too much, but still, things should only be merged when they are green | 17:16 |
@iglesiasg | HeikoS, all right | 17:16 |
@iglesiasg | HeikoS, deactivate the test could be enough? | 17:16 |
@iglesiasg | we can do that right now in a moment if so | 17:16 |
@HeikoS | iglesiasg: yes, since otherwise new patches cannot be tested to be ok | 17:16 |
@HeikoS | iglesiasg: maybe a good idea | 17:16 |
@iglesiasg | indeed | 17:16 |
@HeikoS | iglesiasg: and the re-activate once it is solved | 17:16 |
@HeikoS | iglesiasg: there is a disabled test list somewhere | 17:17 |
@iglesiasg | let's do it then | 17:17 |
@HeikoS | I think at least :) | 17:17 |
@iglesiasg | has anyone already deactivated any test already? | 17:17 |
pickle27 | HeikoS: iglesiasg okay I'll try and get my notebook checked in soon, I want to do a bit more work on it first | 17:17 |
@iglesiasg | pickle27, sure! | 17:17 |
@HeikoS | pickle27: cool! | 17:17 |
@HeikoS | wiking: are you around? | 17:17 |
@HeikoS | wiking: any news on the automagic notebook stuff? | 17:17 |
@HeikoS | wiking: your automagic unit tests were already extremely useful | 17:17 |
thoralf | anyone want to merge some trivial cleanup? ;) https://github.com/shogun-toolbox/shogun/pull/1306 | 17:18 |
thoralf | unfortunately no travis status is displayed. | 17:18 |
@iglesiasg | oh changes in Makefile and so, I am probably not the right now to check that | 17:20 |
thoralf | HeikoS: About our automatic code generation: We should do more testing with types != float64_t/int32_t. Just wrote a shell-one-liner and found (only format string) bugs in sgvector. | 17:24 |
@HeikoS | thoralf: yep totally | 17:24 |
@HeikoS | thoralf: but still no good idea how to do this in a nice way | 17:24 |
@HeikoS | thoralf: I think the "standard instance" is still the easiest yet powerful method | 17:24 |
@HeikoS | thoralf: but, maybe this can be done better by people with some hacker skills ;) | 17:25 |
thoralf | HeikoS: No hackers around ;) | 17:26 |
@HeikoS | wiking is one! maybe he has ideas | 17:27 |
@HeikoS | thoralf: but other things first | 17:27 |
thoralf | another simple PR https://github.com/shogun-toolbox/shogun/pull/1307 | 17:27 |
thoralf | (No Makefiles involved ;)) | 17:27 |
-!- nube [~rho@49.244.36.218] has joined #shogun | 17:27 | |
shogun-notifier- | shogun: Thoralf Klein :develop * c540d72 / src/shogun/lib/SGVector.cpp: https://github.com/shogun-toolbox/shogun/commit/c540d723d1c31948853b7b583951a306308917e0 | 17:28 |
shogun-notifier- | shogun: Boy scout rule: Fixed format strings of uint/char SGVector::display_vector() methods. | 17:28 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 6751620 / src/shogun/lib/SGVector.cpp: https://github.com/shogun-toolbox/shogun/commit/675162040d8d37ae89b70150811b7023521fc547 | 17:28 |
shogun-notifier- | shogun: Merge pull request #1307 from tklein23/display_sgvector | 17:28 |
shogun-notifier- | shogun: | 17:28 |
shogun-notifier- | shogun: Boy scout rule: Fixed format strings of uint/char SGVector::display_vector() methods. | 17:28 |
pickle27 | iglesiasg: I updated my PR and re ran the relevant unit tests on my machine | 17:33 |
pickle27 | should be good to merge | 17:33 |
shogun-buildbot | build #1559 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1559 | 17:33 |
shogun-buildbot | build #1076 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1076 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 17:34 |
@iglesiasg | pickle27, let's do it then :) | 17:35 |
pickle27 | iglesiasg: awesome! | 17:36 |
shogun-notifier- | shogun: Kevin :develop * 93ac51b / src/shogun/mathematics/ajd/ (9 files): https://github.com/shogun-toolbox/shogun/commit/93ac51b73517c81bb563b0ef3effb73a04457e4f | 17:36 |
shogun-notifier- | shogun: changed ND Array input to const for the AJD class | 17:36 |
shogun-notifier- | shogun: Kevin :develop * d2dea43 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/d2dea4311592a27919437ae0750393b32d32bf71 | 17:36 |
shogun-notifier- | shogun: fixed a memory leak in the ICA unit tests and example | 17:36 |
shogun-notifier- | shogun: Kevin :develop * 27a4e9e / src/shogun/mathematics/ajd/ (9 files): https://github.com/shogun-toolbox/shogun/commit/27a4e9ec631dac742e66b831d452dcc143ad033e | 17:36 |
shogun-notifier- | shogun: changed AJD interface to pass by value | 17:36 |
shogun-notifier- | shogun: Fernando Iglesias :develop * da8857a / / (12 files): https://github.com/shogun-toolbox/shogun/commit/da8857a412ca4933dd510667cc17a7b32300cdb3 | 17:36 |
shogun-notifier- | shogun: Merge pull request #1299 from pickle27/develop | 17:36 |
shogun-notifier- | shogun: | 17:36 |
shogun-notifier- | shogun: Changed input NDArray to pass by value for AJD class and fixed a memory leak in ICA unit tests and examples | 17:36 |
@iglesiasg | it is a shame we don't get a mail when a pull request is updated with a commit | 17:37 |
@iglesiasg | or maybe there is something missing in my github conf... | 17:37 |
shogun-buildbot | build #1280 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1280 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 17:44 |
shogun-buildbot | build #1560 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1560 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 17:47 |
-!- nube [~rho@49.244.36.218] has quit [Quit: Leaving.] | 17:58 | |
@iglesiasg | thoralf, no idea about testing IO really | 18:00 |
@iglesiasg | all right, I am off for a while now. See you! | 18:01 |
@iglesiasg | HeikoS, do you know how to disable a test? | 18:02 |
@HeikoS | iglesiasg: I dont | 18:02 |
@HeikoS | iglesiasg: but some testa re disabled | 18:02 |
@HeikoS | maybe grep for it | 18:02 |
@iglesiasg | HeikoS, yeah, grep told me nothing but matching in object files | 18:02 |
shogun-buildbot | build #1561 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1561 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com> | 18:04 |
@iglesiasg | HeikoS, DISABLED in capitals :) | 18:04 |
@HeikoS | iglesiasg: ah, nice :) | 18:04 |
@HeikoS | cool that is useful to know | 18:04 |
@iglesiasg | I am going to make an alias for grep without case sensitiveness lol | 18:04 |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 18:07 | |
shogun-notifier- | shogun: Fernando Iglesias :develop * 4a27e1a / tests/unit/structure/FactorGraph_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/4a27e1a6da72e85044cf8513a415b3ee8bd54568 | 18:16 |
shogun-notifier- | shogun: Disable temporarily buggy FactorGraph tests | 18:16 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 8ffd7ea / tests/unit/structure/FactorGraph_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/8ffd7ea39ae760e4c6df40c0f7d7548133e10536 | 18:16 |
shogun-notifier- | shogun: Merge pull request #1311 from iglesias/develop | 18:16 |
shogun-notifier- | shogun: | 18:16 |
shogun-notifier- | shogun: Disable temporarily buggy FactorGraph tests | 18:16 |
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:c149:9096:8977:c686] has quit [Quit: Ex-Chat] | 18:17 | |
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has joined #shogun | 18:18 | |
travis-ci | [travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/9473282 | 18:18 |
-!- travis-ci [~travis-ci@ec2-184-72-165-99.compute-1.amazonaws.com] has left #shogun [] | 18:18 | |
shogun-buildbot | build #1395 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1395 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 18:22 |
shogun-buildbot | build #1077 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1077 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 18:28 |
shogun-buildbot | build #1562 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1562 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 18:37 |
shogun-buildbot | build #1281 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1281 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 18:37 |
shogun-buildbot | build #1396 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1396 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 19:06 |
hushell | HeikoS: Hi | 19:07 |
hushell | HeikoS: I am doing it right now | 19:07 |
-!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has quit [Quit: Leaving] | 19:09 | |
-!- lisitsyn [~lisitsyn@213.87.140.158] has joined #shogun | 19:37 | |
thoralf | Bye. | 19:39 |
-!- thoralf [~thoralf@enki.zib.de] has quit [Quit: Konversation terminated!] | 19:39 | |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has joined #shogun | 20:10 | |
travis-ci | [travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/9478867 | 20:10 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has left #shogun [] | 20:10 | |
hushell | HeikoS: Finally I fixed it, I'll send a PR :) | 20:11 |
@HeikoS | hushell: nice thanks a lot! | 20:11 |
hushell | HeikoS: it turns out I forgot to unref one thing | 20:12 |
@HeikoS | hushell: I see | 20:12 |
@HeikoS | hushell: it is really important to fix memory errors/leaks straight away. Its horrible to do this afterwards ;) | 20:12 |
hushell | I referenced your SubsetStack.cpp, helped a lot! | 20:12 |
@HeikoS | hushell: cool! :) | 20:14 |
@HeikoS | what for? | 20:14 |
hushell | HeikoS: for DynamicObjectArray | 20:24 |
-!- van51 [~van51@athedsl-399972.home.otenet.gr] has quit [Quit: Leaving.] | 20:24 | |
hushell | HeikoS: I did a wrong commit, how do I roll back after commit --amend? | 20:25 |
lisitsyn | hushell: I am afraid that's not really possible | 20:33 |
lisitsyn | but I don't know exactly | 20:33 |
hushell | well, I had created a new branch and compile from scratch :( | 20:34 |
shogun-notifier- | shogun: Thoralf Klein :develop * 4410c98 / src/shogun/lib/SGSparseVector.cpp,src/shogun/lib/SGSparseVector.h: https://github.com/shogun-toolbox/shogun/commit/4410c98e95564042870c00972ec18025407b336e | 20:48 |
shogun-notifier- | shogun: Added display_vector() method to SGSparseVector for easier debugging. | 20:48 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 389afef / src/shogun/lib/SGSparseVector.cpp,src/shogun/lib/SGSparseVector.h: https://github.com/shogun-toolbox/shogun/commit/389afef6526aeeedd0f3301cc3c9087a955ab1a2 | 20:48 |
shogun-notifier- | shogun: Merge pull request #1310 from tklein23/display_for_sparse_vector | 20:48 |
shogun-notifier- | shogun: | 20:48 |
shogun-notifier- | shogun: Added display_vector() method to SGSparseVector for easier debugging. | 20:48 |
shogun-notifier- | shogun: Thoralf Klein :develop * 254bfd6 / src/shogun/lib/SGVector.cpp: https://github.com/shogun-toolbox/shogun/commit/254bfd66aa4fda95fce7bc5732b41f88b9097373 | 20:48 |
shogun-notifier- | shogun: Undo change to SGVector<char>::display_vector. | 20:48 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 1d60d35 / src/shogun/lib/SGVector.cpp: https://github.com/shogun-toolbox/shogun/commit/1d60d35864eec22f15d8cb3264c0c0867367223e | 20:48 |
shogun-notifier- | shogun: Merge pull request #1308 from tklein23/display_vector_char_revert | 20:48 |
shogun-notifier- | shogun: | 20:48 |
shogun-notifier- | shogun: Undo change to SGVector<char>::display_vector. | 20:48 |
-!- lisitsyn [~lisitsyn@213.87.140.158] has quit [Read error: Connection reset by peer] | 20:54 | |
-!- HeikoS [~heiko@nat-185-60.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 21:00 | |
shogun-buildbot | build #1078 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1078 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 21:00 |
shogun-buildbot | build #1282 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1282 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 21:04 |
shogun-buildbot | build #1565 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1565 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Thoralf Klein <thoralf.klein@zib.de> | 21:07 |
-!- lisitsyn [~lisitsyn@213.87.140.158] has joined #shogun | 21:14 | |
-!- Netsplit *.net <-> *.split quits: shogun-buildbot | 21:24 | |
-!- Netsplit *.net <-> *.split quits: @wiking | 21:24 | |
-!- Netsplit *.net <-> *.split quits: lisitsyn | 21:25 | |
-!- Netsplit *.net <-> *.split quits: shogun-notifier- | 21:25 | |
-!- Netsplit *.net <-> *.split quits: gsomix_ | 21:26 | |
-!- Netsplit *.net <-> *.split quits: zxtx | 21:27 | |
--- Log closed Thu Jul 25 21:29:04 2013 | ||
--- Log opened Thu Jul 25 21:29:11 2013 | ||
-!- shogun-toolbox [~shogun@7nn.de] has joined #shogun | 21:29 | |
-!- Irssi: #shogun: Total of 8 nicks [1 ops, 0 halfops, 0 voices, 7 normal] | 21:29 | |
-!- Irssi: Join to #shogun was synced in 6 secs | 21:29 | |
shogun-buildbot | build #1397 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1397 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 21:31 |
-!- sonney2k [~shogun@7nn.de] has joined #shogun | 21:35 | |
-!- ServerMode/#shogun [+o sonney2k] by wolfe.freenode.net | 21:35 | |
-!- hushell [~hushell@c-24-21-169-136.hsd1.or.comcast.net] has quit [Ping timeout: 256 seconds] | 21:39 | |
-!- mmyros999 [889818eb@gateway/web/freenode/ip.136.152.24.235] has joined #shogun | 21:46 | |
mmyros999 | Hi, I am trying to install shogun matlab interface from source. A call of ./configure --interfaces=matlab_static results in an error. Shogun cannot find matlab. I created an alias alias matlab='~/matlab13a', so typing matlab into shell starts matlab. Any clues on the origin of the error? Thanks | 21:48 |
-!- lisitsyn [~lisitsyn@213.87.140.158] has quit [Read error: Connection reset by peer] | 21:54 | |
-!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has joined #shogun | 22:09 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 22:09 | |
@iglesiasg | hi mmyros999 | 22:10 |
@iglesiasg | I just read about your problem in the chat logs | 22:10 |
@iglesiasg | mmyros999: did you solve it already? | 22:10 |
@iglesiasg | mmyros999: anyway, I guess you should export that alias so it is also visible when the configure script runs | 22:25 |
mmyros999 | @iglesiasg I tried alias matlab="export MATLAB="~/matlab"" but that didn't work. Was this wrong? | 22:31 |
@iglesiasg | mmyros999: write in your ~/.bashrc or ~/.bash_profile (depending on your operative system) alias matlab=<command you use to run matlab> | 22:35 |
@iglesiasg | mmyros999: you could also write a symbolic link called matlab in /usr/local/bin pointing at the matlab executable | 22:36 |
@iglesiasg | mmyros999: what is the command | 22:36 |
@iglesiasg | sorry | 22:36 |
@iglesiasg | mmyros999: what is the command you normally use to run matlab? | 22:36 |
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has joined #shogun | 22:38 | |
mmyros999 | ~/matlab | 22:39 |
mmyros999 | I will try what you suggested. thanks | 22:39 |
@iglesiasg | mmyros999: you are welcome | 22:40 |
@sonney2k | iglesiasg, did it help? | 22:46 |
@iglesiasg | sonney2k: let me see | 22:50 |
@iglesiasg | sonney2k: yep, it worked indeed :) | 22:53 |
@iglesiasg | sonney2k: but I still find weird that c++11 doesn't appear in the list of C++ compiler flags | 22:54 |
@sonney2k | iglesiasg, what? | 22:55 |
@sonney2k | iglesiasg, how do you know? | 22:55 |
@iglesiasg | sonney2k: https://travis-ci.org/shogun-toolbox/shogun/jobs/9479220 | 22:55 |
@iglesiasg | for instance that one | 22:55 |
@sonney2k | iglesiasg, btw I have another issue with your matrixlist and my changes that cause a test failure but later | 22:56 |
@sonney2k | first this one | 22:56 |
@sonney2k | iglesiasg, where do you find this? | 22:56 |
@sonney2k | iglesiasg, I don't see any compiler flags? | 22:57 |
@iglesiasg | sonney2k: line 1040 | 22:57 |
@iglesiasg | it is quite on top | 22:57 |
@iglesiasg | damn it doesn't work crtl+f of the line number, at least in my browser | 22:57 |
@sonney2k | iglesiasg, ahh ok these are only warnings etc support | 22:57 |
@iglesiasg | -O0? | 22:58 |
@sonney2k | iglesiasg, ./configure --disable-optimization ... | 22:58 |
@iglesiasg | sonney2k: yeah I know, I meant that they are not only warnings :) | 22:58 |
@sonney2k | except for -O0 only warnings :D | 22:59 |
@iglesiasg | sonney2k: anyway, I just thought they all should appear there | 22:59 |
@sonney2k | compiler flags are speed tuning & warnign flags | 22:59 |
@sonney2k | iglesiasg, well there is a test trying to enable warning flags & compiler speed / tuning stuff | 22:59 |
@sonney2k | that is the one that displays the used flags | 22:59 |
@sonney2k | then a lot of other flags are added | 22:59 |
@sonney2k | needed for pthread/ python etc etc | 23:00 |
@iglesiasg | sonney2k: what test? | 23:00 |
@sonney2k | iglesiasg, configure test | 23:00 |
@sonney2k | iglesiasg, it runs all these parameters with g++ on some toy C++ program to see if it compiles with them | 23:00 |
@sonney2k | iglesiasg, are we good on this? | 23:02 |
@iglesiasg | sonney2k: yeah, everything all right | 23:02 |
@sonney2k | iglesiasg, ok then next issue | 23:02 |
@sonney2k | iglesiasg, we are having these crashers here: | 23:03 |
@sonney2k | iglesiasg, http://shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1396/steps/test%20python_modular/logs/stdio | 23:03 |
@sonney2k | iglesiasg, search for structure_discrete_hmsvm_bmrm | 23:03 |
@sonney2k | and scroll down to the gdb analysis | 23:04 |
@iglesiasg | let me see | 23:04 |
@iglesiasg | there was actually an user yesterday with this problem | 23:06 |
@iglesiasg | it seems to happen rarely | 23:06 |
shogun-notifier- | shogun: van51 :develop * 627d8cd / / (7 files): https://github.com/shogun-toolbox/shogun/commit/627d8cd56d34bed3a67eefbfd89f6974e22d75ec | 23:07 |
shogun-notifier- | shogun: Quadratic support for rest numerical hashed classes | 23:07 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 9e8ebe4 / / (7 files): https://github.com/shogun-toolbox/shogun/commit/9e8ebe4b60e5252a7c2c8026a3f366249664c2a8 | 23:07 |
shogun-notifier- | shogun: Merge pull request #1309 from van51/feature/quadratic | 23:07 |
shogun-notifier- | shogun: | 23:07 |
shogun-notifier- | shogun: Quadratic support for rest numerical hashed classes | 23:07 |
@sonney2k | iglesiasg, well I guess not | 23:07 |
@sonney2k | iglesiasg, I am actually pretty sure that my change caused this | 23:07 |
@sonney2k | iglesiasg, so I am trying to understand what is wrong | 23:08 |
@iglesiasg | sonney2k: ah all right, what change was then? | 23:08 |
@sonney2k | iglesiasg, basically you were calling the inplace constructor / destructor manually for SGMatrix | 23:08 |
@sonney2k | right? | 23:08 |
@iglesiasg | sonney2k: yes, I remember doing something like that | 23:08 |
@iglesiasg | sonney2k: let me look it up | 23:09 |
@sonney2k | iglesiasg, so I removed that and added support for that in SG_MALLOC | 23:09 |
@sonney2k | iglesiasg, like it is done for SGVector/SparseVector | 23:09 |
@iglesiasg | sonney2k: I am trying to find it, i don't see it right now in SGMatrixList | 23:10 |
@sonney2k | iglesiasg, now when I look at get_matrix and operator[] I see that you return a SGMatrix<T>& | 23:10 |
@sonney2k | iglesiasg, look above I removed it! | 23:10 |
@sonney2k | iglesiasg, so my guess is that returning a SGMatrix<T>& is the reason | 23:11 |
@sonney2k | iglesiasg, and I would like to change that to without & | 23:11 |
@iglesiasg | sonney2k: yes | 23:12 |
@iglesiasg | sonney2k: I think get_matrix without & should work right away | 23:12 |
@sonney2k | iglesiasg, is that ok? | 23:12 |
@sonney2k | ok then let me try it | 23:12 |
@iglesiasg | but what about the places that use it | 23:12 |
@sonney2k | iglesiasg, that is what I meant | 23:12 |
@iglesiasg | sonney2k: I am not sure now if somewhere I was changing the SGMatrixList using this | 23:13 |
@iglesiasg | let me check | 23:13 |
@sonney2k | iglesiasg, what did you expect there? | 23:13 |
@iglesiasg | sonney2k: how do you mean it? | 23:13 |
@sonney2k | iglesiasg, the only difference when using SGMAtrix& vs no & is that when you actually change pointers / sizes of the matrix | 23:14 |
@sonney2k | that this won't propagate back to your SGMatrixList | 23:14 |
@sonney2k | gsomix_, all good? | 23:15 |
gsomix_ | sonney2k, yep | 23:15 |
@iglesiasg | sonney2k: oh true, if I am changing the actual contents of the matrix it will be fine | 23:15 |
-!- gsomix_ is now known as gsomix | 23:15 | |
@sonney2k | iglesiasg, so the question is what do you do? | 23:16 |
@sonney2k | gsomix, btw no single thunderstorm here... I've seen a cloud today though | 23:16 |
@iglesiasg | sonney2k: in MatrixFeatures get_matrix is not even used | 23:16 |
gsomix | sonney2k, good. | 23:16 |
gsomix | sonney2k, it seems I need create my own tokenizer for support of quoting | 23:17 |
@sonney2k | it really is a crazily hot summer this year and I am very happy to not be in berlin (weather forecast 37C on saturday) | 23:17 |
-!- mmyros999 [889818eb@gateway/web/freenode/ip.136.152.24.235] has quit [Quit: Page closed] | 23:17 | |
@iglesiasg | sonney2k: and IIRC that is the only part where SGMatrixList is used | 23:17 |
@sonney2k | here it's nice 22-26 C :) | 23:17 |
@sonney2k | iglesiasg, and the operator? | 23:17 |
@sonney2k | gsomix, btw did you see the compiler warning: | 23:18 |
@sonney2k | lib/CircularBuffer.cpp: In member function ‘bool shogun::CCircularBuffer::has_next()’: | 23:18 |
@sonney2k | lib/CircularBuffer.cpp:183:10: warning: unused variable ‘tail_length’ [-Wunused-variable] | 23:18 |
@sonney2k | ^ | 23:18 |
gsomix | sonney2k, because I have no idea how properly parse something like this: "one", "two", "super, complex, bad case" | 23:18 |
@sonney2k | gsomix | 23:18 |
gsomix | sonney2k, yep | 23:18 |
@sonney2k | gsomix, then please fix | 23:18 |
gsomix | sonney2k, properly parse I mean with only DelimiterTokenizer | 23:18 |
@iglesiasg | sonney2k: CMatrixFeatures::set_feature_vector | 23:19 |
@sonney2k | gsomix, and more complex "" for a real " :) | 23:19 |
@sonney2k | iglesiasg, yeah kaboom! | 23:20 |
gsomix | sonney2k, so I'll just write class like QuotingTokenizer, then I can work with string directly and have standard Tokenizer interface | 23:21 |
@sonney2k | gsomix, please send a PR supporting no " first and then we can continue to improve on sth we have with cornercases we fetch from mldata.org | 23:21 |
@iglesiasg | sonney2k: we can put another setter in SGMatrixList, then the setter in MatrixFeatures simply call this one. Then we can remove the & | 23:21 |
@sonney2k | gsomix, that would be a good test too btw | 23:21 |
@iglesiasg | sonney2k: does that make sense? | 23:21 |
@sonney2k | iglesiasg, yes - what do we do with the operator? | 23:22 |
@iglesiasg | sonney2k: return by value instead | 23:23 |
@iglesiasg | remove inline and move it to cpp instead :) | 23:23 |
@sonney2k | iglesiasg, could you please do that change right away? | 23:24 |
@iglesiasg | sonney2k: yeah, sure | 23:24 |
@sonney2k | iglesiasg, thx :) | 23:25 |
shogun-buildbot | build #1566 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1566 blamelist: van51 <vangelis_51@hotmail.com> | 23:27 |
shogun-buildbot | build #1079 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1079 blamelist: Soeren Sonnenburg <sonne@debian.org> | 23:30 |
shogun-buildbot | build #1283 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1283 blamelist: Soeren Sonnenburg <sonne@debian.org> | 23:34 |
@iglesiasg | sonney2k: compiling | 23:37 |
@iglesiasg | sonney2k: it seems we are getting more users these days btw | 23:39 |
@iglesiasg | sonney2k: new people in the mailing list and posting, that's nice | 23:40 |
@sonney2k | iglesiasg, I guess I am so swamped in gsoc that I didn't notice | 23:41 |
@sonney2k | but surely nice thing if this is true! | 23:41 |
@sonney2k | iglesiasg, if we manage to do the ipython-notebook stuff for large parts of shogun | 23:41 |
@sonney2k | we will get plenty of users I can tell | 23:41 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * e94465b / tests/unit/base/clone_unittest.cc.py: https://github.com/shogun-toolbox/shogun/commit/e94465bf5527269d728804f17faaff6ac508633a | 23:43 |
shogun-notifier- | shogun: don't fail tests if jinja2 is not available | 23:43 |
@sonney2k | iglesiasg, btw I did about 3km today again but it is kind of stupid it could have been 4 but I started too late | 23:44 |
gsomix | sonney2k, btw what's status of R interface problem now? | 23:44 |
@sonney2k | and there were some currents and I was running out of time | 23:44 |
@iglesiasg | sonney2k: tomorrow is the day, 4km tom | 23:44 |
@sonney2k | heh | 23:44 |
@sonney2k | gsomix, which problem? | 23:45 |
gsomix | sonney2k, refcounting in R modular. | 23:45 |
@sonney2k | gsomix, no clue | 23:45 |
@sonney2k | gsomix, no idea if it is refcounting even | 23:45 |
@sonney2k | gsomix, I tried w/ a simple example and refcounting worked | 23:45 |
@sonney2k | gsomix, but with shogun it is all just kaboom | 23:46 |
@sonney2k | gsomix, I tried to create some minimal shogun class example (only stuff in modshogun.i that is absolutely necessary for 2 depending objects) | 23:46 |
@sonney2k | but ran out of time - you know some students wanting CR's | 23:46 |
@sonney2k | gsomix, could you give us a quick PR fix for the build warning? | 23:47 |
gsomix | sonney2k, warning about unused variable? yep, just a moment - compiling | 23:48 |
shogun-notifier- | shogun: Thoralf Klein :develop * 01a714c / / (3 files): https://github.com/shogun-toolbox/shogun/commit/01a714c24744e944ac938a8b30e99f9a160fff30 | 23:50 |
shogun-notifier- | shogun: Minor changes to build scripts: Bogus whitespaces, missing debug output, missing make dependecies | 23:50 |
shogun-notifier- | shogun: Thoralf Klein :develop * be8dac9 / .gitignore,src/Makefile: https://github.com/shogun-toolbox/shogun/commit/be8dac97a1d950d417edbc5ec51412aad96f5587 | 23:50 |
shogun-notifier- | shogun: Shorter names for "new" local make targets. Added build directory to .gitignore | 23:50 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 44e6099 / / (5 files): https://github.com/shogun-toolbox/shogun/commit/44e60998fd06925ad831c4b3c882153f0e6542be | 23:50 |
shogun-notifier- | shogun: Merge pull request #1306 from tklein23/develop | 23:50 |
shogun-notifier- | shogun: | 23:50 |
shogun-notifier- | shogun: Minor changes to build scripts | 23:50 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * f8ea22c / .gitignore: https://github.com/shogun-toolbox/shogun/commit/f8ea22c08edb1f8fd9fcc6ed40f34746580a8e4d | 23:51 |
shogun-notifier- | shogun: add GP swig file to ignores | 23:51 |
@sonney2k | iglesiasg, build done? | 23:52 |
@iglesiasg | sonney2k: getting there | 23:52 |
@sonney2k | iglesiasg, what is wrong with your superfast machine? | 23:52 |
@iglesiasg | sonney2k: I made a couple of silly mistakes | 23:52 |
@sonney2k | iglesiasg, ahh ok np | 23:52 |
gsomix | compiling is so slow... | 23:52 |
@iglesiasg | sonney2k: actually I am a bit disappointed, the computer I am using at uni is faster :( | 23:53 |
gsomix | even with -j4 and cchache | 23:53 |
gsomix | *ccache | 23:53 |
@iglesiasg | gsomix: oh I just saw this in my build lib/CircularBuffer.cpp:183:10: warning: unused variable ‘tail_length’ [-Wunused-variable] | 23:55 |
@iglesiasg | gsomix: :P | 23:55 |
gsomix | agrh :) | 23:55 |
shogun-buildbot | build #1080 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1080 blamelist: Soeren Sonnenburg <sonne@debian.org> | 23:56 |
shogun-buildbot | build #1284 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1284 blamelist: Soeren Sonnenburg <sonne@debian.org> | 23:59 |
--- Log closed Fri Jul 26 00:00:40 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!