--- Log opened Fri Sep 13 00:00:57 2013 | ||
@sonney2k | gsomix, btw you will need to read the whole protobuf message into memory anyway so you can aswell read it with fread() | 00:12 |
---|---|---|
shogun-buildbot | build #1791 of deb3 - modular_interfaces is complete: Failure [failed test python_modular test ruby_modular] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1791 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, hushell <hushell@hushell-U510.(none)> | 00:15 |
shogun-buildbot | build #1790 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1790 | 00:21 |
shogun-notifier- | shogun-data: Soeren Sonnenburg :master * 5e90268 / testsuite/tests/ (9 files): https://github.com/shogun-toolbox/shogun-data/commit/5e90268c3c1a3947747ce5d33ccf032dba662e43 | 00:27 |
shogun-notifier- | shogun-data: add tests for new examples | 00:27 |
shogun-buildbot | build #1447 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1447 | 00:28 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 77e6cf4 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/77e6cf49b1fba302316a3683d618c390b36ebe61 | 00:28 |
shogun-notifier- | shogun: require new data version and muten examples | 00:28 |
-!- HeikoS [~heiko@213.190.120.70] has joined #shogun | 00:37 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 00:37 | |
sonne|osx | HeikoS: doing the night shift? | 00:40 |
@HeikoS | sonne|osx: its 1040 here :) | 00:40 |
@HeikoS | and you=? | 00:40 |
@HeikoS | just came home | 00:40 |
sonne|osx | 00:40 :) | 00:40 |
@HeikoS | I see | 00:40 |
@HeikoS | so its you who is doing this :D | 00:40 |
@HeikoS | how are things? | 00:41 |
sonne|osx | HeikoS: we have a buildbot x-mas tree now | 00:41 |
sonne|osx | HeikoS: seen that armada http://buildbot.shogun-toolbox.org/waterfall | 00:41 |
sonne|osx | HeikoS: wikings new botnet | 00:41 |
@HeikoS | wow! | 00:42 |
@HeikoS | thats a lot | 00:42 |
@HeikoS | pretty cool | 00:42 |
@HeikoS | mmh the most important one fails | 00:42 |
@HeikoS | debian modular | 00:42 |
@HeikoS | but very nice | 00:42 |
@HeikoS | are those all running on the new server? | 00:42 |
sonne|osx | yeah and we've come quite far almost 50% are good | 00:42 |
sonne|osx | HeikoS: no | 00:42 |
sonne|osx | but most | 00:43 |
sonne|osx | actually modular is ok | 00:43 |
sonne|osx | I am changing examples and by doing so I am creating issues | 00:43 |
sonne|osx | trying to get rid of the tools dir | 00:43 |
sonne|osx | and AsciiFile | 00:43 |
sonne|osx | I can tell that I am happy to have tests for the examples | 00:44 |
@HeikoS | tests for examples? | 00:45 |
sonne|osx | yes | 00:45 |
sonne|osx | the integration tests are exactly that | 00:45 |
@HeikoS | ah | 00:45 |
@HeikoS | I see | 00:45 |
@HeikoS | so they are helping with things? | 00:45 |
@HeikoS | that could not otherwise be catched? | 00:46 |
@HeikoS | I am still wondering what that would be | 00:46 |
sonne|osx | well they ensure that the examples are doing the same thing even when I refactor them | 00:46 |
@HeikoS | API errors are catched by running them, result changes are detected by units | 00:46 |
@HeikoS | I dont know | 00:46 |
@HeikoS | there were again so many errors in the tests recently that were all just not really anything | 00:47 |
@HeikoS | where we replaced the test files | 00:47 |
@HeikoS | but | 00:47 |
@HeikoS | ok, thats a discussion for later | 00:47 |
sonne|osx | I know you don't like them. I am just happy to have them. Otherwise I would have broken examples. | 00:48 |
sonne|osx | anyway | 00:48 |
sonne|osx | we still have lots and lots on the todo | 00:48 |
@HeikoS | yeah I mean I want to replace them to not loose anything | 00:48 |
sonne|osx | before we can release | 00:48 |
@HeikoS | yes I know | 00:48 |
@HeikoS | I will join next week | 00:49 |
@HeikoS | for one week at least | 00:49 |
sonne|osx | but how would you check if an example returns the correct thing? | 00:49 |
@HeikoS | units | 00:49 |
sonne|osx | but that is C++ code | 00:49 |
@HeikoS | because then we also know *what* is wrong rather than *that* something is wrong | 00:49 |
sonne|osx | when you refactor an example I mean | 00:49 |
@HeikoS | I dont know what could sneak through | 00:50 |
sonne|osx | ohh with python examples you also know | 00:50 |
@HeikoS | as said, I haven't seen an example of an integration test that could not been otherwise catched | 00:50 |
@HeikoS | I still would run them | 00:50 |
@HeikoS | to get API/include errors | 00:50 |
sonne|osx | but they might return results that you didn't intend | 00:50 |
@HeikoS | we should discuss this in a quiet moment at some point | 00:51 |
@HeikoS | I am not in general against python tests | 00:51 |
@HeikoS | but these integration tests where we serialize, I dont like that | 00:51 |
@HeikoS | since it changes when you touch the class | 00:51 |
@HeikoS | would rather just store certain things or so | 00:51 |
@HeikoS | (or have migration working) | 00:51 |
sonne|osx | or then consciously update the test data | 00:52 |
sonne|osx | I don't like us breaking serialization | 00:52 |
sonne|osx | if we break it | 00:52 |
sonne|osx | some test should fail | 00:52 |
sonne|osx | and we should make a conscious decision | 00:52 |
@HeikoS | I spent probably about 10hrs "fixing" tests, where I did not have a single error but only changes | 00:52 |
@HeikoS | but it should not break if you add/remove a variable | 00:53 |
sonne|osx | if we had examples in all languages doing exactly the same | 00:53 |
sonne|osx | we could generate a reference output | 00:53 |
sonne|osx | and then compare | 00:53 |
@HeikoS | mmh | 00:53 |
@HeikoS | yeah thats good actually | 00:53 |
sonne|osx | so we know that java, octave, python ,... all return the same result | 00:53 |
sonne|osx | I didn't get that far as you can see | 00:53 |
@HeikoS | still my question: what errors dont we detect otherwise? | 00:53 |
@HeikoS | what is an example of a problem that only is catched like this? | 00:54 |
sonne|osx | how do you ensure that a python example gives the same result like an octave example? | 00:54 |
@HeikoS | did we ever have a problem with this? | 00:54 |
@HeikoS | for these inter-laguage things, I agree, thats nice to have | 00:55 |
sonne|osx | well I have. we tried to do examples to be exactly the same in all languages | 00:55 |
@HeikoS | but I would not do it with fixed files | 00:55 |
sonne|osx | I mean we have the same documentation for all | 00:55 |
@HeikoS | rather run code, dump, and check whether its the same | 00:55 |
sonne|osx | so they should do the same | 00:55 |
@HeikoS | can be done automatically | 00:55 |
@HeikoS | thats actually a nice thing to have | 00:55 |
@HeikoS | but then we would need the *same* examples | 00:56 |
sonne|osx | yeah I do that too but then when sty is changed it might break | 00:56 |
sonne|osx | we had | 00:56 |
sonne|osx | we are deviating from that | 00:56 |
@HeikoS | its impossible | 00:56 |
sonne|osx | why? | 00:56 |
@HeikoS | we rather should have automatically generated tests for all modular classes | 00:56 |
@HeikoS | a test that is defined once and then carried our automatically | 00:57 |
sonne|osx | you cannot test everything | 00:57 |
sonne|osx | and you don't test the example | 00:57 |
@HeikoS | true, but how would you get preople to write the very same example in all languages? | 00:57 |
@wiking | yey | 00:58 |
sonne|osx | it is very little work | 00:58 |
@wiking | streaming feature fixed | 00:58 |
@HeikoS | so, maybe let me phrase it this way: we should change the fact that the integration tests break all the time without something being wrong | 00:58 |
@HeikoS | wiking: hi! nice! | 00:58 |
sonne|osx | it is basically just copying into the different dirs | 00:58 |
sonne|osx | with very minimal adjustments | 00:58 |
sonne|osx | maybe we can automate that even | 00:58 |
@HeikoS | sonne|osx: writing is fine, but maintaining is not | 00:58 |
@HeikoS | sonne|osx: the only way to do it is automation | 00:58 |
@HeikoS | imagine the costs of changing one example | 00:59 |
@HeikoS | you would have to change 7 files | 00:59 |
@HeikoS | + integration test files | 00:59 |
sonne|osx | 1 integration test file which is just generator.py example.py | 00:59 |
sonne|osx | that is really cheap | 00:59 |
@HeikoS | no its like 3 comits | 01:00 |
@HeikoS | and it should not happen | 01:00 |
@HeikoS | or it need not to happen in my eyes | 01:00 |
@HeikoS | but maybe the better way to go is to to the migration for new varialbes | 01:00 |
@HeikoS | or something | 01:00 |
@HeikoS | or just even ignore | 01:00 |
@HeikoS | dont know | 01:00 |
@HeikoS | sonne|osx: but lets discuss that later | 01:01 |
@HeikoS | not the time now | 01:01 |
@wiking | oh man ascii reader has quite some error or i dont know | 01:02 |
sonne|osx | wiking: I am killing buildbot 3 | 01:02 |
sonne|osx | wiking: that can be removed soon | 01:02 |
@wiking | /Users/wiking/shogun/tests/unit/features/StreamingDenseFeatures_unittest.cc:48: Failure | 01:02 |
sonne|osx | wiking: it is too slow | 01:02 |
@wiking | Value of: example.vector[j] Actual: -0.80468112230300903 | 01:02 |
@wiking | Expected: expected.vector[j] | 01:02 |
@wiking | Which is: -0.80468114524960255 | 01:02 |
@wiking | -0.8046811452496025 | 01:02 |
@wiking | is actually in the file | 01:02 |
@wiking | is this error normal or i'm doing something wrong? | 01:02 |
sonne|osx | wiking: please pretend AsciiFile doesn't exist | 01:03 |
@wiking | sonne|osx: what should i use for writing then? | 01:03 |
sonne|osx | wiking: CSVFile | 01:03 |
@HeikoS | sonne|osx, wiking I will write a mail to the students/mentors now that people should stop coding very soon | 01:03 |
sonne|osx | HeikoS: yes very good | 01:04 |
sonne|osx | HeikoS: we should have a quick meeting again so they can all help with the release | 01:04 |
sonne|osx | otherwise we can hardly make it | 01:04 |
@HeikoS | yeah sergey scheduled it | 01:04 |
sonne|osx | woking has been working for 3 but it is still a looong todo list | 01:04 |
sonne|osx | wiking ^ | 01:04 |
@wiking | 3? | 01:05 |
@wiking | ah lol | 01:05 |
@sonney2k | shogun-buildbot, force build --branch=develop 'deb1 - libshogun' | 01:05 |
shogun-buildbot | build forced [ETA 9m15s] | 01:05 |
shogun-buildbot | I'll give a shout when the build finishes | 01:05 |
@sonney2k | shogun-buildbot, force build --branch=develop 'deb3 - shogun_modular' | 01:05 |
shogun-buildbot | no such builder 'deb3 - shogun_modular' | 01:05 |
@wiking | sonne|osx: hahaha | 01:05 |
@sonney2k | shogun-buildbot, force build --branch=develop 'deb3 - modular_interfaces' | 01:05 |
shogun-buildbot | The build has been queued, I'll give a shout when it starts | 01:05 |
@wiking | sonne|osx: CSVFile is buggy | 01:05 |
sonne|osx | wiking: bug gsomix! | 01:05 |
@wiking | sonne|osx: 2 dim vector | 01:05 |
@wiking | -1.69268,0.477893, | 01:05 |
@wiking | -1.56754,0.684399, | 01:05 |
@wiking | ^ first two line | 01:06 |
sonne|osx | wiking: but all examples work fine with it | 01:06 |
sonne|osx | wiking: what's the issue? | 01:06 |
sonne|osx | the trailing , ? | 01:06 |
@wiking | yeah | 01:06 |
@wiking | why is there? :D | 01:06 |
sonne|osx | wiking: create an issue and assign it to gsomix | 01:06 |
sonne|osx | we need as much testing as possible :D | 01:07 |
shogun-buildbot | build forced [ETA 34m11s] | 01:07 |
shogun-buildbot | I'll give a shout when the build finishes | 01:07 |
@wiking | CDenseFeatures<float64_t>* orig_feats=new CDenseFeatures<float64_t>(data); CCSVFile* saved_features = new CCSVFile(fname, 'w'); orig_feats->save(saved_features); saved_features->close(); SG_UNREF(saved_features); | 01:07 |
sonne|osx | alight I should sleep now | 01:07 |
@wiking | i did this | 01:07 |
sonne|osx | night! | 01:07 |
shogun-buildbot | Hey! build deb1 - libshogun #2115 is complete: Success [build successful] | 01:09 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2115 | 01:09 |
@wiking | gsomix: https://github.com/shogun-toolbox/shogun/issues/1579 | 01:10 |
shogun-buildbot | build #71 of FC19 - modular_interfaces is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/71 blamelist: Soeren Sonnenburg <sonne@debian.org> | 01:11 |
@HeikoS | wiking: you are really on a run :) | 01:21 |
shogun-buildbot | build #1792 of deb3 - modular_interfaces is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1792 | 01:34 |
shogun-buildbot | build #1448 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1448 blamelist: Soeren Sonnenburg <sonne@debian.org> | 01:40 |
shogun-notifier- | shogun: Viktor Gal :feature/FixStreaming * 1874222 / tests/unit/features/StreamingDenseFeatures_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/187422275532dad4983ff6777df5ac2ee5fc471b | 01:55 |
shogun-notifier- | shogun: Add basic unit test for StreamingDenseFeatures reading | 01:55 |
shogun-notifier- | shogun: Viktor Gal :feature/FixStreaming * 9f21f87 / src/shogun/io/streaming/ (3 files): https://github.com/shogun-toolbox/shogun/commit/9f21f879ec9c9baa7f270d267563f314cfa7649d | 01:55 |
shogun-notifier- | shogun: Fix Example<T> in ParseBuffer.h | 01:55 |
shogun-notifier- | shogun: instead of trying to store the example as SGVector<T> in an Example<T> | 01:55 |
shogun-notifier- | shogun: rather store it as T* as T can be SGSparseVectorEntry. | 01:55 |
shogun-notifier- | shogun: Use SG_REALLOC instead of SG_MALLOC in StreamingAsciiFile when reading | 01:55 |
shogun-notifier- | shogun: sparse vectors, this way enabling true buffering in ParseBuffer | 01:55 |
shogun-notifier- | shogun: Viktor Gal :feature/FixStreaming * b9a6012 / tests/unit/features/StreamingDenseFeatures_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/b9a60125f5fc9f242752d8dbbbc4a6fecce24e59 | 01:55 |
shogun-notifier- | shogun: Add example_reading_from_features to StreamingDenseFeatures unit test | 01:55 |
@wiking | let's see what travis says about this... | 01:55 |
@wiking | ok i'm off to sleep soonish | 01:55 |
@iglesiasg | good night! | 02:56 |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Ex-Chat] | 02:56 | |
-!- HeikoS [~heiko@213.190.120.70] has quit [Quit: Leaving.] | 02:57 | |
-!- sonne|osx_ [~sonne@f053047034.adsl.alicedsl.de] has joined #shogun | 03:17 | |
-!- sonne|osx [~sonne@f053046093.adsl.alicedsl.de] has quit [Ping timeout: 260 seconds] | 03:19 | |
-!- sonne|osx_ is now known as sonne|osx | 03:19 | |
-!- pickle27 [~Kevin@199.119.128.114] has quit [Quit: Leaving] | 03:38 | |
-!- besser82 [~besser82@fedora/besser82] has quit [Read error: Operation timed out] | 04:42 | |
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has quit [Read error: Operation timed out] | 04:48 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 04:55 | |
-!- besser82 [~besser82@fedora/besser82] has joined #shogun | 04:56 | |
gsomix | wiking, ok, I'll fix. | 05:42 |
gsomix | sonney2k, ok, got it. | 05:42 |
gsomix | sonney2k, but there are many possibilities in how we can read messages. | 05:57 |
gsomix | sonney2k, one of | 05:57 |
gsomix | https://developers.google.com/protocol-buffers/docs/reference/cpp/google.protobuf.io.coded_stream#CodedInputStream | 05:57 |
gsomix | need go to classes | 05:58 |
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has left #shogun [] | 07:25 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 08:01 | |
shogun-notifier- | shogun: Viktor Gal :develop * 057c667 / src/shogun/io/streaming/InputParser.h,src/shogun/io/streaming/ParseBuffer.h: https://github.com/shogun-toolbox/shogun/commit/057c667214ff6093c07d3dc12a88a76e6c06bae8 | 08:01 |
shogun-notifier- | shogun: InputParser and ParseBuffer requires pthread | 08:01 |
shogun-notifier- | shogun: check if it's available | 08:01 |
shogun-notifier- | shogun: Viktor Gal :develop * 1874222 / tests/unit/features/StreamingDenseFeatures_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/187422275532dad4983ff6777df5ac2ee5fc471b | 08:01 |
shogun-notifier- | shogun: Add basic unit test for StreamingDenseFeatures reading | 08:01 |
shogun-notifier- | shogun: Viktor Gal :develop * 9f21f87 / src/shogun/io/streaming/ (3 files): https://github.com/shogun-toolbox/shogun/commit/9f21f879ec9c9baa7f270d267563f314cfa7649d | 08:01 |
shogun-notifier- | shogun: Fix Example<T> in ParseBuffer.h | 08:01 |
shogun-notifier- | shogun: instead of trying to store the example as SGVector<T> in an Example<T> | 08:01 |
shogun-notifier- | shogun: rather store it as T* as T can be SGSparseVectorEntry. | 08:01 |
shogun-notifier- | shogun: Use SG_REALLOC instead of SG_MALLOC in StreamingAsciiFile when reading | 08:01 |
shogun-notifier- | shogun: sparse vectors, this way enabling true buffering in ParseBuffer | 08:01 |
shogun-notifier- | shogun: Viktor Gal :develop * b9a6012 / tests/unit/features/StreamingDenseFeatures_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/b9a60125f5fc9f242752d8dbbbc4a6fecce24e59 | 08:01 |
shogun-notifier- | shogun: Add example_reading_from_features to StreamingDenseFeatures unit test | 08:01 |
shogun-notifier- | shogun: Viktor Gal :develop * 625d74e / / (4 files): https://github.com/shogun-toolbox/shogun/commit/625d74e3198eb73423e0b912ba391a36c7df1f70 | 08:01 |
shogun-buildbot | build #72 of FC19 - modular_interfaces is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/72 blamelist: Viktor Gal <viktor.gal@maeth.com> | 08:12 |
-!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has joined #shogun | 08:19 | |
shogun-buildbot | build #124 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/124 blamelist: Viktor Gal <viktor.gal@maeth.com> | 08:20 |
shogun-buildbot | build #1449 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1449 blamelist: Viktor Gal <viktor.gal@maeth.com> | 08:37 |
shogun-buildbot | build #1450 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1450 blamelist: Viktor Gal <viktor.gal@maeth.com> | 09:10 |
shogun-buildbot | build #1794 of deb3 - modular_interfaces is complete: Failure [failed test python_modular test ruby_modular] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1794 blamelist: Viktor Gal <viktor.gal@maeth.com> | 09:15 |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 10:08 | |
zombie|sonnewrk | moin wiking! | 10:18 |
zombie|sonnewrk | look deb3 modular is back happy! | 10:18 |
zombie|sonnewrk | wiking: and cyg1 almost too! http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1450/steps/test/logs/stdio | 10:18 |
zombie|sonnewrk | wiking: however could do the following: when a test fails we rerun it with gdb and create a full backtrace at least or even w/ valgrind? | 10:19 |
zombie|sonnewrk | wiking: so we have something to work with? | 10:19 |
zombie|sonnewrk | besser82: btw did the version thing w/o git work? | 10:25 |
besser82 | zombie|sonnewrk: yes, works :) | 10:30 |
zombie|sonnewrk | perfect | 10:30 |
-!- lisitsyn1 [~lisitsin@mxs.kg.ru] has joined #shogun | 10:30 | |
-!- lisitsyn1 [~lisitsin@mxs.kg.ru] has left #shogun [] | 10:30 | |
-!- lisitsyn1 [~lisitsin@mxs.kg.ru] has joined #shogun | 10:30 | |
besser82 | zombie|sonnewrk: is there any closer targeted release date for 3.0.0 than Oct. 13 ??? | 10:31 |
besser82 | zombie|sonnewrk: like 2nd half of Oct or some KW? | 10:31 |
besser82 | zombie|sonnewrk: I'm asking, because beta freeze for F20 is target ~15.10.13 | 10:32 |
zombie|sonnewrk | besser82: I can't tell | 10:41 |
besser82 | zombie|sonnewrk: kk, thanks :) | 10:41 |
zombie|sonnewrk | we are making good progress but there are tons of issues stil | 10:41 |
zombie|sonnewrk | l | 10:41 |
zombie|sonnewrk | and gsoc is still running so we cannot do a hard freeze just yet | 10:41 |
zombie|sonnewrk | however it looks already pretty good http://buildbot.shogun-toolbox.org/waterfall | 10:42 |
zombie|sonnewrk | man buildbots became green | 10:42 |
zombie|sonnewrk | osx used to be green no idea what changed | 10:42 |
zombie|sonnewrk | cyg has only 3 tests failing | 10:42 |
zombie|sonnewrk | besser82: so there is only FC19 modular that needs fixing | 10:43 |
zombie|sonnewrk | the rest is rather optional | 10:43 |
zombie|sonnewrk | (various code quality / analysis stuff) | 10:43 |
zombie|sonnewrk | besser82: btw how did you manage to get it to configure on FC19 with http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/72/steps/configure/logs/stdio | 10:44 |
besser82 | zombie|sonnewrk: is this error on FCRH, too? | 10:45 |
zombie|sonnewrk | besser82: we don't have a buildbot for modular on FCRH | 10:45 |
zombie|sonnewrk | so no idea | 10:45 |
besser82 | zombie|sonnewrk: there should be an update for rubygem-narray on F19, which will fix this problem | 10:50 |
besser82 | zombie|sonnewrk: within next day or so | 10:50 |
zombie|sonnewrk | besser82: ahh cool! | 10:54 |
zombie|sonnewrk | then we sit it out | 10:54 |
besser82 | zombie|sonnewrk: I'll give you links to updated builds, when I'm done with it :) | 10:54 |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 11:01 | |
thoralf | Hello everybody | 11:20 |
zombie|sonnewrk | thoralf: good morning berlin! | 11:20 |
thoralf | Good morning Zombie ;) | 11:21 |
thoralf | Lets hope that wiking did not break any of the streaming stuff... | 11:24 |
thoralf | No break :) | 11:25 |
thoralf | But a leak. | 11:26 |
-!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has quit [Quit: WeeChat 0.3.7] | 11:30 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 11:39 | |
shogun-notifier- | shogun: ZhengyangL :develop * cd9c28e / / (8 files): https://github.com/shogun-toolbox/shogun/commit/cd9c28e592a8b660d48f8120f3ce6aa2456e2dcf | 11:39 |
shogun-notifier- | shogun: A Simple MLData Downloader with unittest | 11:39 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 1b219d5 / / (8 files): https://github.com/shogun-toolbox/shogun/commit/1b219d56d873cabe880ce63e6b5ee4f8704220c6 | 11:39 |
shogun-notifier- | shogun: Merge pull request #1580 from ZhengyangL/feature/mldatafile | 11:39 |
shogun-notifier- | shogun: | 11:39 |
shogun-notifier- | shogun: add unit-test to simple mldata downloader / fix the indent / make url_prefix configurable | 11:39 |
shogun-buildbot | build #2120 of deb1 - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2120 blamelist: Soeren Sonnenburg <sonne@debian.org> | 11:40 |
shogun-buildbot | build #2119 of deb1 - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2119 blamelist: ZhengyangL <liuzhengyang@outlook.com> | 11:42 |
zombie|sonnewrk | foulwall: around? Seen the build failures? | 11:50 |
zombie|sonnewrk | foulwall: you need to ifdef the examples/tests with the HAVE_CURL too | 11:50 |
thoralf | zombie|sonnewrk: About the suppressions file. But I think it doesn't make sense (yet) to have many suppression files - it's okay to collect a few suppressions first in one file and split them on demand. | 11:58 |
thoralf | zombie|sonnewrk: It's less convenient, btw., since you have to pass every suppression file on the valgrind command line. | 11:58 |
thoralf | Any suggestions for the name (or directory) are welcome. | 11:59 |
thoralf | As far as I checked there's no way to pass a directory or include suppressions into some "master suppression file". | 12:01 |
zombie|sonnewrk | thoralf: no but for example it woud be nice to store proper vimrc settings emacs settings too | 12:04 |
zombie|sonnewrk | thoralf: so sth like configs / settings ? | 12:04 |
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has joined #shogun | 12:07 | |
-!- HeikoS [~heiko@213.190.120.70] has joined #shogun | 12:09 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 12:09 | |
thoralf | zombie|sonnewrk: I'm no vi guy. Do you mean these comments to tell the editor tab width, indentation, character set? | 12:45 |
thoralf | zombie|sonnewrk: Oh, now I got it. | 12:46 |
thoralf | zombie|sonnewrk: You want a settings directory and the valgrind.supp inside? | 12:47 |
thoralf | zombie|sonnewrk: configs/valgrind.supp | 12:55 |
-!- lambday [67157e37@gateway/web/freenode/ip.103.21.126.55] has joined #shogun | 12:58 | |
lambday | HeikoS: hi :) | 12:58 |
lambday | zombie|sonnewrk: hi... you've turned into a zombie :-o | 12:58 |
-!- HeikoS [~heiko@213.190.120.70] has quit [Quit: Leaving.] | 12:59 | |
thoralf | unit tests are broken: tests/unit/io/MLDataHDF5File_unittest.cc:29:2: error: unknown type name 'CMLDataHDF5File' | 13:04 |
zombie|sonnewrk | thoralf: yes | 13:16 |
zombie|sonnewrk | foulwall: ^ that is yours | 13:16 |
zombie|sonnewrk | or I fix it | 13:16 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 0d3efc0 / tests/unit/io/MLDataHDF5File_unittest.cc,tests/unit/metric/LMNN_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/0d3efc068a59fb5fc758645fedcfce8dd79e6ca0 | 13:20 |
shogun-notifier- | shogun: don't fail when curl is not available | 13:20 |
foulwall | Thanks zombie|sonnewrk | 13:26 |
foulwall | merging van51's language detection demo | 13:28 |
shogun-buildbot | build #2121 of deb1 - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2121 | 13:28 |
shogun-buildbot | build #73 of FC19 - modular_interfaces is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/73 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:32 |
shogun-buildbot | build #125 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/125 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:34 |
-!- HeikoS [~androirc@217.28.182.64] has joined #shogun | 13:35 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 13:35 | |
@HeikoS | lambday, around? | 13:35 |
-!- lambday_ [67157f37@gateway/web/freenode/ip.103.21.127.55] has joined #shogun | 13:36 | |
lambday_ | HeikoS: hi | 13:36 |
lambday_ | HeikoS: sorry I was away | 13:36 |
@HeikoS | No worries | 13:36 |
lambday_ | HeikoS: how did your presentation go? | 13:37 |
gsomix | RE from classes | 13:37 |
gsomix | hello | 13:37 |
@HeikoS | Good so far, another one tonight | 13:37 |
lambday_ | great! | 13:37 |
@HeikoS | How are things gong for you? | 13:37 |
-!- lambday [67157e37@gateway/web/freenode/ip.103.21.126.55] has quit [Ping timeout: 250 seconds] | 13:37 | |
lambday_ | Should I use the ozone for the ipython or the new matrix that you said? 473x473 one? | 13:38 |
lambday_ | I had some trouble converting that one into ascii | 13:38 |
@HeikoS | Maybe both. Problem is that it takes very long | 13:38 |
@HeikoS | Leave the ozone maybe | 13:39 |
lisitsyn1 | HeikoS: your mail! | 13:39 |
lisitsyn1 | I am confused | 13:39 |
lambday_ | alirght... | 13:39 |
@HeikoS | lisitsyn1, ignore it | 13:39 |
lisitsyn1 | I confess I don't know what meeting I am trying to schedule | 13:39 |
lisitsyn1 | :D | 13:39 |
lisitsyn1 | HeikoS: really, what should we do | 13:39 |
@HeikoS | lisitsyn1, well you should be clear about that | 13:40 |
@HeikoS | lambday_, so for the notebook, it would be good to use the smaller matrix to illustrate things | 13:40 |
lambday_ | HeikoS: so we are postponing preconditioning stuff after gsoc? | 13:40 |
@HeikoS | lambday_, you can nicely explain the fill in effect etc | 13:40 |
lambday_ | HeikoS: yeah | 13:40 |
@HeikoS | lambday_, after | 13:40 |
lambday_ | alright | 13:41 |
@HeikoS | Now it's time to wrap up, takes time | 13:41 |
lambday_ | okay | 13:41 |
@HeikoS | Let me sen double my poster....you might take things from there | 13:41 |
lambday_ | double? | 13:41 |
lambday_ | lisitsyn1: how are you? :) | 13:42 |
lisitsyn1 | HeikoS: I know | 13:42 |
lisitsyn1 | lambday_: fine, what about you? | 13:42 |
@HeikoS | lambday_, sorry tablet. I meant let me send you my poster | 13:42 |
lisitsyn1 | HeikoS: so help me, what have I scheduled? :D | 13:42 |
lambday_ | HeikoS: hehe... ya checking | 13:42 |
lambday_ | lisitsyn1: good.. :) | 13:43 |
@HeikoS | lisitsyn1, I have no idea. I thought this was a Dev only meeting, ask wiking Nd Soren maybe | 13:43 |
@HeikoS | Gotta go, see you in a bit | 13:43 |
lisitsyn1 | haha okay this is getting worse | 13:43 |
lambday_ | HeikoS: alright see you.. poster looks amazing btw | 13:43 |
-!- lisitsyn1 is now known as lisitsyn | 13:44 | |
-!- lambday_ is now known as lambday | 13:46 | |
-!- HeikoS [~androirc@217.28.182.64] has quit [Ping timeout: 246 seconds] | 13:48 | |
zombie|sonnewrk | foulwall: cool | 13:48 |
zombie|sonnewrk | foulwall: does it work? | 13:49 |
shogun-buildbot | build #1451 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1451 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:58 |
zombie|sonnewrk | van51: now that you are `free' again would you want to start the kitchen duties :) | 14:01 |
van51 | zombie|sonnewrk: haha I have already ;) | 14:02 |
van51 | zombie|sonnewrk: but I would like some implementation instructions | 14:02 |
zombie|sonnewrk | van51: excellent | 14:02 |
van51 | zombie|sonnewrk: first of all I am trying to implement it as shown here in alg 1 : http://books.nips.cc/papers/files/nips21/NIPS2008_0885.pdf | 14:03 |
zombie|sonnewrk | van51: I assume it is the same principle like the random fourier features? | 14:04 |
thoralf | van51: Did you get some feedback about the hashing thing? | 14:04 |
van51 | zombie|sonnewrk: yeah, from my understanding actually the RFFeatures are a special case of kitchensinks | 14:04 |
zombie|sonnewrk | yes | 14:05 |
van51 | thoralf: zombie|sonnewrk said it was ok the way it is. the others haven't replied yet | 14:05 |
zombie|sonnewrk | thoralf: yes all ok | 14:05 |
zombie|sonnewrk | thoralf: we don't care too much about collisions | 14:05 |
thoralf | zombie|sonnewrk: hashing "feature" with "feature" as seed? | 14:05 |
zombie|sonnewrk | not with feature but the index | 14:06 |
thoralf | zombie|sonnewrk: Doesn't seem right. | 14:06 |
thoralf | s/feature/index/ yes | 14:06 |
thoralf | zombie|sonnewrk: You get a completely different hashing function with different collision properties. | 14:06 |
zombie|sonnewrk | thoralf: what do you propose instead? | 14:07 |
thoralf | zombie|sonnewrk: Choosing a seed (fixed, maybe const in the source) and then use this seed for hashing all indices. | 14:08 |
thoralf | Otherwise you're changing the hashing function every time. ;) | 14:08 |
thoralf | And maybe another seed for the quadratic features (I don't really care) to make sure both functions are independent. | 14:09 |
thoralf | (or using virtual indices from a different range than the linear features) | 14:09 |
van51 | thoralf: in a paper they had sent me to read it says that each feature was supposed to have a different hash function | 14:09 |
thoralf | van51: Really? Which one? | 14:10 |
van51 | thoralf: http://olivier.chapelle.cc/pub/ngdstone.pdf | 14:11 |
van51 | section 4.3 | 14:11 |
thoralf | van51: No, thats different: | 14:12 |
thoralf | No, sorry. It's not. ;) | 14:13 |
thoralf | van51: Okay, then ignore my first objection. What about the second one? About computing your "virtual" indices? | 14:16 |
van51 | thoralf: to use num_feats + index instead of index+index? | 14:17 |
thoralf | van51: Yes, because this way you're sure producing more collisions than neccessary. | 14:17 |
thoralf | I'm checking the code again, one second. | 14:18 |
van51 | thoralf: the only definite collision that it incurs is when the indexes are 0 I think | 14:18 |
van51 | thoralf: I think I had considered at that time to use num_feats + idx and that I had a reason for doing it the way I did but I can't recall it atm | 14:19 |
van51 | thoralf: it would also work the way you are saying and I can't find a reason against it but I don't see the difference it would make | 14:20 |
thoralf | I found two different ways of computing the hashes in the code. shogun/features/HashedSparseFeatures.cpp:142 and shogun/features/HashedDenseFeatures.cpp:198 | 14:22 |
thoralf | Let me check. | 14:22 |
thoralf | linear "CHash::MurmurHash3((uint8_t* ) &i, sizeof (index_t), i);" and quadratic "CHash::MurmurHash3((uint8_t* ) &n_idx, sizeof (index_t), n_idx) " <-- You get collisions for every "j = i * vec.size() + i" | 14:23 |
thoralf | This was in shogun/features/HashedDenseFeatures.cpp:198 | 14:23 |
thoralf | Agree? | 14:23 |
thoralf | Okay, but since i < vec.size(), this should not happen. Okay. ;) | 14:24 |
van51 | thoralf: hehe I like how you ask and then answer it yourself :p | 14:26 |
van51 | thoralf: I may not even have an answer, but it doesn't matter :p | 14:27 |
thoralf | In shogun/features/HashedSparseFeatures.cpp:142: hash_i(i) vs hash_{2*i}(i) | 14:27 |
thoralf | Should be okay, too. | 14:27 |
thoralf | LOL. | 14:27 |
thoralf | Yes, my first look was quite short. But now I had to double check my claim. ;) | 14:28 |
thoralf | And now I see why you're using two different ways of hashing (sparse vector without known length vs. dense vector with known length). | 14:29 |
van51 | thoralf: yeah it did have something to do with that | 14:30 |
thoralf | Hope you don't mind that the results using sparse are not the same as with dense vectors. ;) | 14:31 |
van51 | thoralf: hmm I hadn't thought of that but I don't think that it matters | 14:33 |
van51 | thoralf: you 're not after a specific representation but after the compactness | 14:34 |
thoralf | van51: I like the idea having two concurrent (but compatible) versions: You can use the learned model with dense and sparse features. | 14:36 |
thoralf | But that really doesn't matter. | 14:36 |
thoralf | Btw., I checked the implementation from scikits-learn. | 14:36 |
thoralf | They indeed hash all features with the same seed. | 14:36 |
van51 | thoralf: well tbh we also use the same seed in the HashedDocDotFeatures version | 14:38 |
van51 | thoralf: maybe it's a matter of preference | 14:39 |
thoralf | Yes, seems like. | 14:39 |
thoralf | tbh? | 14:39 |
van51 | to be honest | 14:40 |
thoralf | Thanks for the debate. ;) | 14:44 |
van51 | thoralf: anytime :) | 14:46 |
-!- HeikoS [~androirc@217.28.182.64] has joined #shogun | 15:01 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 15:01 | |
thoralf | wiking: Your changes broke StreamingAsciiFile - I'm getting segfaults at shogun/io/streaming/StreamingAsciiFile.cpp:573 | 15:06 |
thoralf | CStreamingAsciiFile::get_sparse_vector_and_label (this=0x6103b0, vector=@0x628798: 0x0, len=@0x6287a8: 47, label=@0x6287a0: 0) | 15:06 |
thoralf | wiking: The problem disappears when replacing SG_REALLOC my SG_MALLOC. | 15:12 |
thoralf | I found out that the vector returned by realloc is NULL. | 15:19 |
@HeikoS | lambday, you can also have the latex code if you want. How is the current state Btw. What are you working on, what is planned for the next days? | 15:24 |
lambday | HeikoS: I have been a bit slow due to insti related stuffs... I'll try to finish off the refactoring and examples as planned | 15:25 |
lambday | HeikoS: you told that there is something in the log-det that could have been done in a better way and you'll tell me... | 15:26 |
@HeikoS | What about the individual CG solvers for LogDetEstimator? | 15:26 |
lambday | HeikoS: COCG - still to be fixed | 15:26 |
@HeikoS | Ehm,.... :) | 15:26 |
@HeikoS | Can't remember | 15:26 |
lambday | :( | 15:27 |
@HeikoS | Ok, then COCG, refactoring is most important | 15:27 |
-!- thoralf [~thoralf@enki.zib.de] has quit [Quit: Konversation terminated!] | 15:27 | |
@HeikoS | And then examples for the most important classes, really important | 15:28 |
lambday | HeikoS: yeah.. but that won't take long... | 15:28 |
lambday | HeikoS: I can't convert the matrix into an ascii one | 15:28 |
@HeikoS | I noticed the class dox should also be updated... | 15:29 |
@HeikoS | Which matrix? | 15:29 |
lambday | HeikoS: checked out their dataset by the way, many many sparse matrices there to try out | 15:29 |
lambday | HeikoS: the sparse matrix 473x473 with max eig ~10^11 | 15:29 |
@HeikoS | Ah, what's the problem there? You can even store it densely | 15:30 |
lambday | the one you used for the probe sampler plot | 15:30 |
lambday | HeikoS: its in .mat format, my octave code that I used for ozone is not working for it | 15:31 |
@HeikoS | Ah... | 15:31 |
@HeikoS | I can send it as ascii | 15:31 |
@HeikoS | Will do tonight | 15:31 |
lambday | HeikoS: regarding class docs, I'll review them carefully after I am done with examples | 15:31 |
lambday | HeikoS: thanks man :) | 15:31 |
@HeikoS | Yeah cool, in particular you should explain how things work together | 15:32 |
@HeikoS | So many classes, so should help people through this jungle :) | 15:32 |
lambday | HeikoS: yeah... the LogDetEstimator unit-tests already do that, so you think I'll write similar examples, with that real matrix, right? | 15:32 |
lambday | hahaha yeah its a jungle alright | 15:33 |
lambday | now even I seem to forget which class was for what | 15:33 |
lambday | HeikoS: could you please explain a bit regarding the tutorial thing that you wrote in the mail? | 15:33 |
-!- HeikoS [~androirc@217.28.182.64] has quit [Read error: Connection reset by peer] | 15:33 | |
-!- heiko [~androirc@217.28.182.64] has joined #shogun | 15:33 | |
heiko | lambday, actually | 15:34 |
heiko | Just use python | 15:34 |
heiko | It CN load the file | 15:34 |
heiko | Can | 15:34 |
heiko | And then export as ascii | 15:34 |
heiko | That's easy | 15:34 |
heiko | See my notebokk | 15:34 |
lambday | alright | 15:34 |
heiko | Just two lines | 15:35 |
heiko | And you need to get into python anyways this week :) | 15:35 |
lambday | yeah! | 15:35 |
lambday | heiko: so, csc_matrix works compatibly with our sparse matrix? | 15:36 |
heiko | Yep | 15:37 |
heiko | That's how we store them | 15:37 |
lambday | cool! | 15:37 |
heiko | It's pretty easy to exchange data using python | 15:37 |
lambday | heiko: I actually had some trouble installing these packages... I'll check today only | 15:37 |
heiko | Everything is there | 15:37 |
heiko | Lambday just don't use cholmod | 15:38 |
heiko | That was only needed for the ozone | 15:38 |
lambday | heiko: alright.. | 15:38 |
heiko | Also note you addendum the sparse let to shogun | 15:38 |
heiko | So no need | 15:38 |
heiko | For cholmod | 15:39 |
heiko | Just a bit slower in Eigen3 | 15:39 |
lambday | okay | 15:39 |
heiko | But the other notebook doesn't need it | 15:40 |
lambday | heiko: what does this loadmat is taking? loadmat("S")["S"] ?? S.mat is the filename? | 15:40 |
heiko | It returns a dictionary | 15:40 |
heiko | With keys being strings and values being the data | 15:41 |
lambday | okay | 15:41 |
heiko | It's all well documented Btw Google for function names | 15:42 |
heiko | Python dictionaries are very cool btw | 15:42 |
lambday | yeah! | 15:43 |
lambday | was checking the documentation already... never worked in scipy... gonna be fun | 15:43 |
heiko | Yeah it's very cool stuff | 15:44 |
van51 | heiko: hey! quick question.. did you have a reason for postponing to register parameters in MulticlassStrategy? | 15:44 |
lambday | heiko: so, by this monday, all examples+ipython+refactoring has to be complete? | 15:45 |
heiko | lambday, as said in the mail, last week should be used for cleaning up, so try to finish these things as soon as possible | 15:46 |
lambday | okay.. yeah that sounds good | 15:46 |
heiko | van51, yes, integration tests were down. Feel free to change. But make sure to run the tests before and after. Might have to change quite a few test files | 15:46 |
lambday | I couldn't get pip-install and stuffs working.. as soon as I get that, things won't take long | 15:47 |
van51 | heiko: ok, I'll have a quick look :) | 15:47 |
zombie|sonnewrk | gsomix: do you have some WIP PR ready? | 15:49 |
heiko | van51, would be helpful if you did that and removed warnings :) we have to do this for the release anyways. ?.. | 15:50 |
gsomix | zombie|sonnewrk, nope. I'll send today at evening. need go to home now (100+ km) | 15:50 |
zombie|sonnewrk | gsomix: ok | 15:51 |
-!- heiko [~androirc@217.28.182.64] has quit [Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )] | 15:52 | |
-!- gsomix [~gsomix@185.6.239.51] has quit [Remote host closed the connection] | 15:52 | |
-!- heiko [~androirc@217.28.182.64] has joined #shogun | 16:02 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 16:20 | |
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has quit [Quit: Leaving.] | 16:24 | |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun | 16:29 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 16:29 | |
-!- heiko [~androirc@217.28.182.64] has quit [Read error: Connection reset by peer] | 16:41 | |
-!- heiko [~androirc@217.28.182.64] has joined #shogun | 16:41 | |
-!- zombie|sonnewrk [~sonnenbu@91-64-72-127-dynip.superkabel.de] has left #shogun [] | 16:53 | |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has left #shogun [] | 17:03 | |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Ex-Chat] | 17:27 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 17:27 | |
shogun-notifier- | shogun: hushell :develop * c64978c / / (7 files): https://github.com/shogun-toolbox/shogun/commit/c64978c1cb0c45652907d3f93fc7c804d70d5bed | 17:27 |
shogun-notifier- | shogun: refactor a bit MAPInference and add more unit-tests | 17:27 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 862f445 / / (7 files): https://github.com/shogun-toolbox/shogun/commit/862f4451018a4c41049c2b0bd7f7e441cc13d22e | 17:27 |
shogun-notifier- | shogun: Merge pull request #1581 from hushell/develop | 17:27 |
shogun-notifier- | shogun: | 17:27 |
shogun-notifier- | shogun: refactor a bit MAPInference and add more unit-tests | 17:27 |
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has joined #shogun | 17:31 | |
-!- heiko [~androirc@217.28.182.64] has quit [Read error: Connection reset by peer] | 17:34 | |
shogun-buildbot | build #74 of FC19 - modular_interfaces is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/74 blamelist: hushell <dom343@gmail.com>, Fernando Iglesias <fernando.iglesiasg@gmail.com> | 17:36 |
shogun-buildbot | build #126 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/126 blamelist: hushell <dom343@gmail.com> | 17:37 |
shogun-buildbot | build #127 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/127 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 17:43 |
shogun-buildbot | build #1452 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1452 blamelist: hushell <dom343@gmail.com> | 18:03 |
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has joined #shogun | 18:29 | |
shogun-buildbot | build #1453 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1453 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 18:38 |
shogun-notifier- | shogun: ZhengyangL :develop * b5ab66b / / (3 files): https://github.com/shogun-toolbox/shogun/commit/b5ab66bb905273e59cd364e2c93d65c745b0c85b | 18:56 |
shogun-notifier- | shogun: combine #IFDEFs and fix little ident mistakes | 18:56 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 123f403 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/123f403c09b6cbf05ea615bb060499bf988ad103 | 18:56 |
shogun-notifier- | shogun: Merge pull request #1582 from ZhengyangL/feature/mldatafile | 18:56 |
shogun-notifier- | shogun: | 18:56 |
shogun-notifier- | shogun: combine #ifdefs / fix little indent mistakes | 18:56 |
shogun-buildbot | build #75 of FC19 - modular_interfaces is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/75 blamelist: Soeren Sonnenburg <sonne@debian.org>, ZhengyangL <liuzhengyang@outlook.com> | 19:04 |
shogun-buildbot | build #128 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/128 blamelist: Soeren Sonnenburg <sonne@debian.org> | 19:07 |
shogun-buildbot | build #129 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/129 blamelist: ZhengyangL <liuzhengyang@outlook.com> | 19:13 |
shogun-buildbot | build #1454 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1454 blamelist: Soeren Sonnenburg <sonne@debian.org> | 19:36 |
@sonney2k | foulwall, do you have van51's demo ready? | 19:50 |
@sonney2k | or what is still missing | 19:50 |
lisitsyn | sonney2k: I can shogunize a bit if you guide me a bit | 20:05 |
lisitsyn | any suggestions? | 20:05 |
shogun-buildbot | build #1455 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1455 blamelist: ZhengyangL <liuzhengyang@outlook.com> | 20:11 |
lisitsyn | okay checking tests | 20:11 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 8539417 / / (27 files): https://github.com/shogun-toolbox/shogun/commit/85394179f0599e18e7df84b7582eebc10eb29816 | 20:18 |
shogun-notifier- | shogun: drop AsciiFile | 20:18 |
@sonney2k | lisitsyn, look at this http://buildbot.shogun-toolbox.org/static_analysis/2013-09-11-1/ | 20:19 |
lisitsyn | sonney2k: haha nice | 20:19 |
@sonney2k | lisitsyn, there are errors in slep, transfer, tapkee etc | 20:19 |
lisitsyn | good ok | 20:19 |
@sonney2k | not easy to fix and not easy to parse so have a look | 20:20 |
shogun-buildbot | build #2126 of deb1 - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2126 blamelist: Soeren Sonnenburg <sonne@debian.org> | 20:20 |
@sonney2k | lisitsyn, or any of https://github.com/shogun-toolbox/shogun/issues?labels=&milestone=3&page=1&state=open | 20:22 |
@sonney2k | lisitsyn, or sth cmake related | 20:22 |
@sonney2k | lisitsyn, whenever a test fails we should run the test with gdb to get a backtrace and/or valgrind | 20:23 |
@sonney2k | to know how to fix it | 20:23 |
lisitsyn | sonney2k: uh that's not really trivial | 20:23 |
lisitsyn | okay let me check what can I do | 20:24 |
@sonney2k | lisitsyn, we had that before | 20:24 |
shogun-notifier- | shogun: Thoralf Klein :develop * 29803fa / .valgrind.supp: https://github.com/shogun-toolbox/shogun/commit/29803faf703c5f08b40288809d7bc5180671e47a | 20:32 |
shogun-notifier- | shogun: Valgrind suppression file for pthread/openmp reachable-memory warnings. | 20:32 |
shogun-notifier- | shogun: Thoralf Klein :develop * 262f476 / configs/valgrind.supp,.valgrind.supp: https://github.com/shogun-toolbox/shogun/commit/262f4767fb4d9da0fdccc9803e649b3b267b0bb9 | 20:32 |
shogun-notifier- | shogun: Moved valgrind suppressions into new configs/ directory. | 20:32 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 7401d1a / configs/valgrind.supp: https://github.com/shogun-toolbox/shogun/commit/7401d1a4dfc49518fda3ae733d0503c3af0531c8 | 20:32 |
shogun-notifier- | shogun: Merge pull request #1570 from tklein23/valgrind_suppressions | 20:32 |
shogun-notifier- | shogun: | 20:32 |
shogun-notifier- | shogun: Valgrind suppression file for pthread/openmp reachable-memory warnings. | 20:32 |
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has quit [Quit: Leaving.] | 20:32 | |
shogun-buildbot | build #2127 of deb1 - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2127 | 20:36 |
shogun-buildbot | build #2129 of deb1 - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2129 blamelist: Soeren Sonnenburg <sonne@debian.org> | 20:37 |
shogun-buildbot | build #76 of FC19 - modular_interfaces is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20modular_interfaces/builds/76 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 20:40 |
shogun-buildbot | build #130 of osx1 - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/130 | 20:41 |
shogun-buildbot | build #1456 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1456 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 21:06 |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 246 seconds] | 21:07 | |
shogun-buildbot | build #1457 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1457 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 21:34 |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * 4cef68a / src/shogun/transfer/domain_adaptation/DomainAdaptationSVM.cpp: https://github.com/shogun-toolbox/shogun/commit/4cef68a61eeb620639e31d487c72b15cb4d394a9 | 21:52 |
shogun-notifier- | shogun: Added pre_svm check in DA SVM | 21:52 |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * 78130a2 / src/shogun/io/CSVFile.h: https://github.com/shogun-toolbox/shogun/commit/78130a24dd16b7c9807760d70d6e73cfa3ca2f0f | 21:52 |
shogun-notifier- | shogun: Added missed include in csv file | 21:52 |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * ae29935 / src/shogun/features/SparseFeatures.cpp: https://github.com/shogun-toolbox/shogun/commit/ae29935f09ea8086699f2656f704e7fc2b4a1f06 | 21:52 |
shogun-notifier- | shogun: Fixed possible nullptr in sparse features | 21:52 |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * 88061d1 / src/shogun/features/DenseFeatures.cpp: https://github.com/shogun-toolbox/shogun/commit/88061d1a006422b2248cc2c2395e0c15a066da0c | 21:52 |
shogun-notifier- | shogun: Fixed possible nullptr in dense features | 21:52 |
shogun-buildbot | build #2130 of deb1 - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2130 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 21:55 |
shogun-buildbot | build #2131 of deb1 - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2131 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 21:57 |
-!- lambday [67157f37@gateway/web/freenode/ip.103.21.127.55] has quit [Ping timeout: 250 seconds] | 22:06 | |
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 260 seconds] | 22:21 | |
@wiking | lisitsyn http://www.liveleak.com/view?i=37b_1378633407 | 22:37 |
@wiking | ROTFLMAO | 22:37 |
lisitsyn | wiking: ah yeah old joke | 22:37 |
@wiking | :P | 22:38 |
lisitsyn | wiking: very american though! | 22:38 |
lisitsyn | :D | 22:38 |
lisitsyn | wiking: do you know who's on asciifile removal? | 22:42 |
lisitsyn | wiking: I am not really sure what should be done about compilation failure now | 22:42 |
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has joined #shogun | 22:52 | |
--- Log closed Sat Sep 14 00:00:58 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!