--- Log opened Tue Feb 20 00:00:10 2018 | ||
-!- witness [uid10044@gateway/web/irccloud.com/x-gfzvbpobzihzfynp] has joined #shogun | 07:43 | |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4168 synchronized by syashakash | 08:36 |
---|---|---|
-!- travis-ci [~travis-ci@ec2-54-145-158-139.compute-1.amazonaws.com] has joined #shogun | 09:26 | |
travis-ci | it's syashakash's turn to pay the next round of drinks for the massacre he caused in syashakash/shogun: https://travis-ci.org/syashakash/shogun/builds/343710621 | 09:26 |
-!- travis-ci [~travis-ci@ec2-54-145-158-139.compute-1.amazonaws.com] has left #shogun [] | 09:26 | |
@wiking | fa | 09:30 |
@wiking | celitf :) | 09:30 |
-!- travis-ci [~travis-ci@ec2-54-205-49-211.compute-1.amazonaws.com] has joined #shogun | 09:41 | |
travis-ci | it's syashakash's turn to pay the next round of drinks for the massacre he caused in syashakash/shogun: https://travis-ci.org/syashakash/shogun/builds/343710621 | 09:41 |
-!- travis-ci [~travis-ci@ec2-54-205-49-211.compute-1.amazonaws.com] has left #shogun [] | 09:41 | |
-!- HeikoS [~heiko@p4FED57F8.dip0.t-ipconnect.de] has joined #shogun | 10:15 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 10:15 | |
@HeikoS | lisitsyn: you there? | 11:13 |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4181 synchronized by sorig | 11:14 |
@sukey | [https://github.com/shogun-toolbox/shogun] Issue https://github.com/shogun-toolbox/shogun/issues/4182 opened by goldturtle | 11:32 |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4183 opened by vinx13 | 11:57 |
lisitsyn | HeikoS: hey | 11:59 |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4181 synchronized by sorig | 12:01 |
-!- witness [uid10044@gateway/web/irccloud.com/x-gfzvbpobzihzfynp] has quit [Quit: Connection closed for inactivity] | 12:22 | |
@HeikoS | lisitsyn: hi | 12:29 |
lisitsyn | hey | 12:29 |
@HeikoS | lisitsyn: saw the grant thingi? | 12:29 |
lisitsyn | not really | 12:30 |
lisitsyn | what's this about? | 12:30 |
@HeikoS | about us getting $$$ | 12:32 |
@HeikoS | lisitsyn: and then you viktor and me spending a week or two together implementing stuff :D | 12:33 |
@HeikoS | I pm'ed you the link | 12:33 |
lisitsyn | oh I see | 12:33 |
@HeikoS | question is: are you up for that? | 12:33 |
lisitsyn | HeikoS: week yes, two not really | 12:33 |
lisitsyn | HeikoS: and most importantly when | 12:34 |
@HeikoS | thats fine as well | 12:34 |
lisitsyn | :) | 12:34 |
@HeikoS | thats why I ask you | 12:34 |
@HeikoS | discuss that | 12:34 |
@HeikoS | think it doesnt matter for the application | 12:34 |
@HeikoS | only that we do it | 12:34 |
@HeikoS | but since this is about the tags, would be good if you were around | 12:34 |
lisitsyn | may is no good for me | 12:34 |
lisitsyn | btw | 12:35 |
@HeikoS | you think it is feasible to do this stuff I wrote in a week or two? | 12:35 |
lisitsyn | if we have some powerful machine to compile yes | 12:35 |
lisitsyn | :) | 12:35 |
lisitsyn | because the most terrible slowdown I have in that | 12:35 |
lisitsyn | is compilation time | 12:35 |
@HeikoS | we have the bots | 12:36 |
@HeikoS | can also offer my work machine | 12:36 |
@HeikoS | has 24 cores | 12:36 |
@HeikoS | when would be good for you? | 12:37 |
lisitsyn | HeikoS: any time after may actually | 12:37 |
@HeikoS | june maybe? | 12:37 |
lisitsyn | uh | 12:38 |
lisitsyn | let me think | 12:38 |
lisitsyn | HeikoS: the problem is that I have 2-week holidays because you know I marry :D | 12:38 |
lisitsyn | in may | 12:38 |
@HeikoS | I didnt know | 12:38 |
@HeikoS | congrats :) | 12:38 |
lisitsyn | I am not sure it is good to take another one-week break just after that.. | 12:39 |
lisitsyn | this means I would be absent at my job for three weeks | 12:39 |
@HeikoS | sure | 12:39 |
@HeikoS | so when s good? | 12:39 |
lisitsyn | I guess some week in or after august.. | 12:40 |
@HeikoS | wiking: can we up swig dependency to 3.0.12? | 12:40 |
lisitsyn | if it blocks you | 12:40 |
lisitsyn | i think you can go ahead | 12:40 |
lisitsyn | I can spend some time remote trying to help you | 12:40 |
@HeikoS | i think would be good to have you around | 12:40 |
@HeikoS | otherwise it doesnt really make sense | 12:40 |
lisitsyn | this might work as well | 12:40 |
lisitsyn | sorry about that :P | 12:41 |
@HeikoS | there is no difference to now then, as I hang out with viktor already | 12:41 |
lisitsyn | I see | 12:41 |
@HeikoS | no worries, I am just thinking: when is good | 12:41 |
lisitsyn | I can easily do that in august, september, blabla | 12:41 |
@HeikoS | works for me as well | 12:42 |
lisitsyn | but june not sure, I am constrained by the time I would be absent | 12:42 |
@HeikoS | ok | 12:42 |
@HeikoS | I will ask NF if it is ok to spend the $$$ in summer | 12:43 |
@HeikoS | and then we plan for august | 12:43 |
@HeikoS | lisitsyn: and what about a shorter hackathon over a long weekend? | 12:43 |
@HeikoS | independent of the NF stor | 12:43 |
lisitsyn | HeikoS: should be possible but not in may | 12:45 |
lisitsyn | :) | 12:45 |
@HeikoS | april? :D | 12:45 |
lisitsyn | it might be good for me yes | 12:45 |
@HeikoS | like Friday-Monday | 12:45 |
@HeikoS | in Z?rich | 12:45 |
@HeikoS | lisitsyn: ok cool | 12:46 |
@HeikoS | I will talk to wiking and adjust | 12:46 |
@HeikoS | lisitsyn: now I have some other qs :) | 12:46 |
@HeikoS | lisitsyn: so I think "put" is kinda done now | 12:47 |
@HeikoS | lisitsyn: now there is "get" | 12:47 |
lisitsyn | ok | 12:47 |
@HeikoS | so that is get_real in Python & co | 12:47 |
@HeikoS | but I dont think get_real should exist in c++ | 12:47 |
@HeikoS | but get<float...> | 12:47 |
@HeikoS | you agree? | 12:47 |
@HeikoS | or not? | 12:47 |
lisitsyn | yes | 12:47 |
lisitsyn | C++ should provide get<T> | 12:48 |
@HeikoS | the "put" api, btw, is also inconsistent between c++ and swig | 12:48 |
lisitsyn | how? | 12:48 |
@HeikoS | namely. put is typed for c++, but dispatches types in swig | 12:48 |
lisitsyn | ah | 12:48 |
@HeikoS | for example for put<float> | 12:48 |
@HeikoS | so you can put any type from swig | 12:48 |
@HeikoS | and it will convert | 12:48 |
lisitsyn | I see | 12:48 |
@HeikoS | while in c++ you get error if you assign float to double | 12:49 |
lisitsyn | yeah sounds good | 12:49 |
@HeikoS | I couldnt make it work otherwise | 12:49 |
lisitsyn | no need to let C++ not use its features | 12:49 |
@HeikoS | due to overloading problems with swig | 12:49 |
@HeikoS | yeah | 12:49 |
@HeikoS | what didnt work unfortunately is | 12:49 |
@HeikoS | float64_t results = get("name") | 12:49 |
lisitsyn | yeah it can't work | 12:49 |
@HeikoS | okok | 12:50 |
@HeikoS | then I will go ahead with get now | 12:50 |
lisitsyn | HeikoS: it can work though | 12:50 |
@HeikoS | and there is some swig bug in travis swig 3.0.10 | 12:50 |
lisitsyn | :) | 12:50 |
lisitsyn | not this way | 12:50 |
lisitsyn | get could return some object that is castable to anything | 12:50 |
lisitsyn | and then runtime | 12:50 |
lisitsyn | :) | 12:50 |
@HeikoS | lets not do that :) | 12:50 |
lisitsyn | yes | 12:50 |
@HeikoS | lisitsyn: ok cool | 12:50 |
@HeikoS | lisitsyn: then there is still that clone thing :D | 12:50 |
lisitsyn | aha | 12:51 |
lisitsyn | clone arrays i guess? | 12:51 |
lisitsyn | let me be honest it is on the top of my todo list :D | 12:51 |
lisitsyn | I just finished my course so hopefully will sort out and get to it | 12:51 |
@HeikoS | ok cool! | 12:53 |
@HeikoS | would be good | 12:53 |
@HeikoS | yeah arrays! | 12:53 |
@HeikoS | and SGVector and stuff | 12:53 |
@HeikoS | and CSGObject | 12:54 |
@HeikoS | just not = operator :) | 12:54 |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4181 merged by karlnapf | 12:55 |
@sukey | [https://github.com/shogun-toolbox/shogun] karlnapf pushed 2 commits: | 12:55 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/6f108a0cf06f75bc3c98078240748191e5b63086 | 12:55 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/ec4930047c5c84152a2f41ad6da3242ea675c453 | 12:55 |
@sukey | [https://github.com/shogun-toolbox/shogun] karlnapf pushed 13 commits: | 12:57 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/c77b514fae5f21a9daac98833a663a40c8815fb4 | 12:57 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/99b8ddbaee4787a39718fff5fbf7a5e6b8881ed1 | 12:57 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/9bd758b19f2c55fd51020d3cfb5659b3e6b428fc | 12:57 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/8ff6b01e18935749aa0c1414dc0769ef00a6d772 | 12:57 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/1bc05807f136e5ad749a5f51a9194c5f8b05f563 | 12:57 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/1361986b20f6442476c328e7e153bea61233521a | 12:57 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/8e069905863a19fd6888338742781ae8cd66ea3d | 12:57 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/6532b57bfe1a19319fcf70cbdf0087f836a0f36d | 12:57 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/3d87d53c6e91b048432303e33b80e8a20a81b12c | 12:57 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/39f04c79425b35f7ace848ade0fb20b70af00ffc | 12:57 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/aa8cde6fb5755396fc42af39efd1815a54a6e064 | 12:57 |
-!- sukey [~nodebot@ks312251.kimsufi.com] has quit [Remote host closed the connection] | 12:57 | |
@HeikoS | wiking: https://github.com/swig/swig/commit/b538070016e46e463b3897fbdb75e1ae995f936e | 13:23 |
@HeikoS | lisitsyn: ^ | 13:23 |
lisitsyn | wow! | 13:23 |
@HeikoS | so if we move the swig interface stuff for put/get into the swig.i files, we need swig 3.0.12 | 13:23 |
lisitsyn | wsfulton did something for us | 13:23 |
lisitsyn | :P | 13:23 |
@HeikoS | lol :P | 13:23 |
@HeikoS | lisitsyn: suggestions? | 13:24 |
lisitsyn | HeikoS: what suggestions? | 13:25 |
@HeikoS | well we dont require 3.0.12 yet | 13:25 |
@HeikoS | so I cannot do that | 13:25 |
lisitsyn | ah lets wait for 3.0.13 and then require | 13:25 |
lisitsyn | D: | 13:25 |
@HeikoS | 3.0.12 is in sid | 13:25 |
lisitsyn | already? | 13:25 |
@HeikoS | yes | 13:26 |
@wiking | HeikoS, ? | 13:36 |
@wiking | HeikoS, sorry had to have a timeout after lunch | 13:36 |
@wiking | 3.0.12 if fine | 13:36 |
@wiking | afaik | 13:36 |
@wiking | i mean honestly | 13:37 |
@wiking | we got you a ppa for 3.0.12 | 13:37 |
@wiking | for ubuntu last week when u were here | 13:37 |
@wiking | and anyways the idea is that nobody should compile manually those .so-s | 13:37 |
@wiking | and whoever wants to develop | 13:38 |
@wiking | the worse case scenario is that you pull a docker image | 13:38 |
@wiking | the sdk | 13:38 |
@wiking | btw whats the swig version in conda? | 13:38 |
@wiking | https://anaconda.org/anaconda/swig | 13:38 |
@wiking | 3.0.12 | 13:39 |
@wiking | only linux-ppc64le has 3.0.10 :P | 13:39 |
@wiking | sid = debian unstable | 13:42 |
@wiking | it's always the latest release | 13:42 |
@wiking | and sid is always the unstable of debian | 13:42 |
@wiking | this might be a problem https://packages.debian.org/stretch/swig | 13:43 |
@wiking | but i reckon that is solveable as well | 13:46 |
@HeikoS | yeah stretch has 3.0.10 | 14:06 |
@HeikoS | wiking: I mean I can also work around this | 14:06 |
@wiking | i think we could take 3.0.12 as requirement | 14:07 |
@HeikoS | via not using a template for these methods (they are only instantiated for two template parameters) | 14:07 |
@HeikoS | ok cool | 14:07 |
@wiking | it has been reelased a year ago or so | 14:07 |
@wiking | but if you go with it | 14:07 |
@HeikoS | shall I add the ppa to the dockerfile? | 14:07 |
@wiking | i'll check out the stretch | 14:07 |
@wiking | mmm probably yeah | 14:07 |
@wiking | i'll hav eto fix some stuff | 14:07 |
@wiking | so dont merge it | 14:07 |
@wiking | just like afternoon | 14:07 |
@HeikoS | sure nw | 14:07 |
@wiking | till then i figure out all our buildbots | 14:07 |
@HeikoS | typedo should work once swig 3.0.12 is used | 14:07 |
@wiking | how we can require 3.0.12 | 14:07 |
@HeikoS | it all works on my machine | 14:08 |
@wiking | k | 14:08 |
-!- travis-ci [~travis-ci@ec2-54-205-49-211.compute-1.amazonaws.com] has joined #shogun | 14:08 | |
@HeikoS | so I will work on "get" then | 14:08 |
travis-ci | it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/343791849 | 14:08 |
-!- travis-ci [~travis-ci@ec2-54-205-49-211.compute-1.amazonaws.com] has left #shogun [] | 14:08 | |
@wiking | then yeah docker sdk needs to be fixed | 14:08 |
@wiking | etcet c | 14:08 |
@wiking | ok i'll dive into this | 14:08 |
@wiking | just wanna finish up some graph traversing with llvm | 14:08 |
@wiking | :P | 14:08 |
@HeikoS | lol | 14:08 |
@HeikoS | enjoy! :P | 14:08 |
@wiking | yeah not so straightforward | 14:08 |
@wiking | to travers the ast | 14:09 |
@wiking | from root elements | 14:09 |
@wiking | :P | 14:09 |
@wiking | *root = leaf | 14:09 |
@wiking | or i mean | 14:09 |
@wiking | the doxygen is just asdf | 14:09 |
@wiking | i would love that there's a better way to document c++ code | 14:09 |
@HeikoS | me too! | 14:09 |
@wiking | https://libgit2.github.com/libgit2/#HEAD | 14:10 |
@wiking | that's created with docurium | 14:11 |
@HeikoS | looks cute! | 14:12 |
@wiking | the interesting part is | 14:12 |
@wiking | that it support | 14:12 |
@wiking | git tags | 14:12 |
@wiking | so you know | 14:12 |
@wiking | multi version | 14:12 |
@HeikoS | wiking: btw you saw the discussion with sergey re the dates? | 14:12 |
@wiking | we dont need to hack ourselves | 14:12 |
@HeikoS | he can do april for a weekend hack | 14:12 |
@wiking | mmm no | 14:12 |
@wiking | for NF? | 14:12 |
@HeikoS | but not before august for the NF fund week | 14:12 |
@wiking | or the other as well? | 14:12 |
@HeikoS | two days (Friday,Monday) in April is fine | 14:12 |
@HeikoS | a week only in August | 14:12 |
@wiking | mmm | 14:12 |
@wiking | ok | 14:12 |
@wiking | august is a bit shacky for me | 14:12 |
@wiking | as i have already some holiday plans | 14:13 |
@wiking | (lasd and first week definitely of august) | 14:13 |
@wiking | so only 2 weeks is empty in that month | 14:13 |
@wiking | for me | 14:13 |
@wiking | or at least it looks like that | 14:13 |
@wiking | atm | 14:13 |
@HeikoS | I mean if we apply anyways | 14:13 |
@HeikoS | it days "must be completable in 2018" | 14:13 |
@HeikoS | so all good | 14:13 |
@wiking | sure i mean | 14:13 |
@wiking | for me other then those 2 weeks | 14:13 |
@wiking | are open | 14:13 |
@wiking | april is good for me | 14:13 |
@wiking | (not right after easter, or if you guys wanna come to serbia then actually that's a doable thing as well :P) | 14:14 |
@wiking | you cannot get cheaper than serbia in europe :D | 14:14 |
@wiking | just saying | 14:14 |
@wiking | :P | 14:14 |
@HeikoS | I know | 14:14 |
@HeikoS | I like the place | 14:14 |
@HeikoS | so happy to do that | 14:14 |
@wiking | ok so that's an option as well | 14:14 |
@wiking | sergey can come i know | 14:14 |
@HeikoS | so haclathon in zurich in April funded by gunnar? | 14:14 |
@wiking | april? you mean may? ?:) | 14:14 |
@wiking | or we move it to april as sergey cannot do may? | 14:14 |
@wiking | atm the doodle is for may | 14:15 |
@wiking | but yeah the idea for zurich is funded by G | 14:15 |
@HeikoS | i know | 14:15 |
@HeikoS | but sergey cannot do in may | 14:15 |
@wiking | i'll doublecheck but yeah | 14:15 |
@wiking | ok | 14:15 |
@HeikoS | so april maybe? | 14:15 |
@wiking | so we shift april? | 14:15 |
@HeikoS | I can do that | 14:15 |
@wiking | for me that's good as well | 14:15 |
@wiking | as said | 14:15 |
@HeikoS | and we get fernando and sergey over | 14:15 |
@HeikoS | and maybe italian? | 14:15 |
@wiking | zurich april = only the last 3 weeks | 14:16 |
@wiking | first week (after eater) i'm still in .rs | 14:16 |
@HeikoS | i am flexible I think | 14:16 |
@HeikoS | question | 14:16 |
@HeikoS | get_object or get_distance? | 14:16 |
@wiking | ok then i'll create/edit the doodle | 14:16 |
@HeikoS | in swig | 14:16 |
@wiking | mmmm | 14:16 |
@HeikoS | which translates to | 14:16 |
@wiking | i guess you want to have base interface getters? | 14:17 |
@HeikoS | get<CSGobject*> or get<CDistance*> | 14:17 |
@wiking | but you know | 14:17 |
@wiking | actually | 14:17 |
@wiking | that branch hasn't been merged | 14:17 |
@wiking | but just because of a stupid reason | 14:17 |
@wiking | feature/obtain_from | 14:17 |
@HeikoS | yeah | 14:17 |
@wiking | but that has .as<> | 14:17 |
@HeikoS | we can always offer the .as | 14:17 |
@wiking | so we can translate those | 14:17 |
@wiking | for base interface classes | 14:17 |
@HeikoS | yeah that is an option | 14:18 |
@HeikoS | and c++? | 14:18 |
@wiking | so it's get_object().as_distance() | 14:18 |
@wiking | i reckon get().as<CDistance> | 14:18 |
@wiking | is fine | 14:18 |
@wiking | :) | 14:18 |
@wiking | i mean i'm just writin llvm code | 14:18 |
@HeikoS | ok | 14:18 |
@HeikoS | I think that is ok | 14:18 |
@wiking | and i have things like this | 14:18 |
@wiking | auto record = base.getType()->getAsCXXRecordDecl(); | 14:18 |
@HeikoS | then we have all this casting code only in the "as" and not in two places | 14:18 |
@wiking | indeed | 14:18 |
@wiking | the only reason i didn't merge that branch | 14:20 |
@wiking | is because there's somerhing wrong with macros :( | 14:20 |
@wiking | somehow ifndef swig is not kicking in | 14:20 |
@wiking | :( | 14:20 |
@wiking | https://github.com/shogun-toolbox/shogun/commit/a3acc968c2b0d2c4adde3b03743b6e67d8df09e5#diff-da4bd0301230a7269627808a09755b8bR27 | 14:21 |
@wiking | i would suppose that this is an empty string (the macro) when you generate the SWIG code | 14:21 |
@wiking | but it's not :( | 14:21 |
@wiking | which i really dont get why | 14:21 |
@HeikoS | probably some nested macro stuff | 14:25 |
@wiking | but i mean why? :) | 14:25 |
@HeikoS | dont know | 14:25 |
@wiking | if i plainly put it there | 14:25 |
@wiking | the very same macro | 14:25 |
@wiking | it all works | 14:25 |
@HeikoS | I remember having to first expand one macro before using it in another one | 14:25 |
@HeikoS | stringification stuff | 14:25 |
@HeikoS | but no idea here | 14:26 |
@wiking | mmm | 14:28 |
@wiking | cool -E works for swig as well | 14:28 |
@wiking | lemme check | 14:28 |
@HeikoS | wiking: get_object has one downside | 14:30 |
@HeikoS | I need a dispatcher internally then | 14:30 |
@wiking | ? | 14:30 |
@wiking | you can always upcast | 14:31 |
@HeikoS | well | 14:31 |
@HeikoS | so get_object("distance") say | 14:31 |
@HeikoS | get<CSGObject*>("distance") -> type error | 14:31 |
@wiking | (CSGObject*)whateverSGObjectLike | 14:31 |
@HeikoS | so need to (CSGObject*)get<CDistance*>("distance") | 14:32 |
-!- travis-ci [~travis-ci@ec2-54-196-252-130.compute-1.amazonaws.com] has joined #shogun | 14:32 | |
travis-ci | it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/343791849 | 14:32 |
-!- travis-ci [~travis-ci@ec2-54-196-252-130.compute-1.amazonaws.com] has left #shogun [] | 14:32 | |
@HeikoS | but then user calls an .as afterwards | 14:32 |
@HeikoS | so kind of back and forth | 14:32 |
@wiking | well | 14:32 |
@wiking | welcome to what happens in case of dynamic languages | 14:32 |
@wiking | all the time | 14:32 |
@wiking | :) | 14:32 |
@HeikoS | haha | 14:32 |
@wiking | just hat you done see this | 14:32 |
@wiking | *that | 14:32 |
@wiking | but partly it's because of these type of stuff | 14:32 |
@HeikoS | ok Ill leave it as this then | 14:32 |
@wiking | that they are slow | 14:32 |
@wiking | all of them constantly doing this | 14:33 |
@wiking | as you really have no other way around this of course | 14:33 |
@HeikoS | would be doing this in c++ as well then (in the examples) | 14:33 |
@wiking | python has this intermediate language | 14:33 |
@wiking | you ought to check that stuff | 14:33 |
@wiking | for some of your code :P | 14:33 |
@HeikoS | yeah | 14:33 |
@wiking | it is crazy what happens underneath | 14:33 |
@HeikoS | I could also offer get_distance instead | 14:34 |
@HeikoS | no conversion necessary then | 14:34 |
@HeikoS | just have to offer it for every base class | 14:34 |
@wiking | https://docs.python.org/2/library/dis.html | 14:34 |
@wiking | :) | 14:34 |
@HeikoS | which defeats the purpose of .as | 14:34 |
@HeikoS | so I guess, I'll stay with the cast for now | 14:34 |
@wiking | yeah i mean | 14:34 |
@HeikoS | see where it brings us | 14:34 |
@wiking | actually we have not actually exposed | 14:35 |
@wiking | .as in swig :P | 14:35 |
@HeikoS | ah | 14:35 |
@HeikoS | haha | 14:35 |
@HeikoS | Ill just play a bit and see what happens | 14:35 |
@HeikoS | most annoying | 14:35 |
@HeikoS | get_distance -> name clash everywhere | 14:35 |
@wiking | ah because of shadowing ? | 14:35 |
@HeikoS | yes | 14:36 |
@wiking | :< | 14:36 |
@HeikoS | but ok | 14:36 |
@HeikoS | these other methods will be removed from swig soon | 14:36 |
@HeikoS | so for now I can name it get_generic_distance or so | 14:36 |
@wiking | ic | 14:36 |
@HeikoS | maybe .as stays in libshogun then and not swig | 14:36 |
@wiking | could be | 14:37 |
@wiking | atm its not exposed | 14:37 |
@HeikoS | kk | 14:37 |
@HeikoS | thx! | 14:37 |
@wiking | mmm | 14:37 |
@wiking | atm i dont get anything :D | 14:37 |
@wiking | just compiled r interface w/o problems | 14:37 |
@HeikoS | hahaha | 14:38 |
@HeikoS | ah joy | 14:38 |
@HeikoS | all these langs and builds :D | 14:38 |
@HeikoS | Ill go for a walk now, see you later | 14:38 |
@wiking | ah ok | 14:40 |
@wiking | stale .cxx | 14:40 |
-!- travis-ci [~travis-ci@ec2-54-167-169-20.compute-1.amazonaws.com] has joined #shogun | 17:25 | |
travis-ci | it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/343843668 | 17:25 |
-!- travis-ci [~travis-ci@ec2-54-167-169-20.compute-1.amazonaws.com] has left #shogun [] | 17:25 | |
@HeikoS | wiking: lisitsyn https://gist.github.com/karlnapf/1d369f5ad6fdf400802ee8ab7f9dcf27 | 17:45 |
@HeikoS | which way do you like more? | 17:45 |
lisitsyn | HeikoS: what do you mean by way? | 17:46 |
lisitsyn | that's the same thing done in diff langs, no? | 17:46 |
@HeikoS | there is two versions of get | 17:46 |
lisitsyn | ahh | 17:46 |
@HeikoS | one gives base pointer | 17:46 |
lisitsyn | I think simple get should suffice | 17:46 |
lisitsyn | and then as | 17:46 |
lisitsyn | HeikoS: this way it is symmetric to put | 17:47 |
@HeikoS | well | 17:47 |
lisitsyn | and simpler | 17:47 |
@HeikoS | get_real_vector | 17:47 |
lisitsyn | HeikoS: yes for vectors you have to | 17:47 |
@HeikoS | I tend to agree with you | 17:47 |
@HeikoS | there is just two little downsides | 17:47 |
@HeikoS | 1. the simple get involves one more type dispatch cast | 17:47 |
lisitsyn | HeikoS: I don't think it is often to get a kernel and use it somehow | 17:47 |
lisitsyn | you usually construct something not get its parts | 17:48 |
@HeikoS | yeah I mean one can still call get("kernel").get_real("log_width") | 17:48 |
lisitsyn | yeah | 17:48 |
lisitsyn | but can't call compute etc etc | 17:48 |
@HeikoS | the other thing is error msgs | 17:48 |
lisitsyn | what's with errors? | 17:49 |
@HeikoS | get_kernel("svm") | 17:49 |
lisitsyn | ah | 17:49 |
@HeikoS | will tell you : you asked to get an SVM as Kernel, which is not possible | 17:49 |
lisitsyn | yes but the same thing would be get("svm").as_kernel() | 17:49 |
@HeikoS | yeah true | 17:49 |
@HeikoS | get("log_width") | 17:49 |
@HeikoS | will say "I cannot convert that to object" | 17:49 |
@HeikoS | you think that is OK? | 17:50 |
lisitsyn | not that good but not terrible | 17:50 |
@HeikoS | kk | 17:50 |
@HeikoS | thx for the opinion | 17:50 |
lisitsyn | to my taste | 17:50 |
lisitsyn | :) | 17:50 |
@HeikoS | it is getting quite complete now | 17:50 |
@HeikoS | soon will be able to change all examples to this | 17:50 |
lisitsyn | HeikoS: we have one more opportunity this way | 17:50 |
@HeikoS | and also use the kwargs | 17:50 |
lisitsyn | which is | 17:50 |
lisitsyn | get("log_width") could return some wrapper then | 17:51 |
lisitsyn | and then asks you to cast itself to something | 17:51 |
@HeikoS | yeah sure | 17:51 |
@HeikoS | could do that later | 17:51 |
@HeikoS | btw | 17:51 |
lisitsyn | yes | 17:51 |
@HeikoS | how can I find out the type string of a parameter whose name I have? | 17:51 |
@HeikoS | is that possible? | 17:51 |
lisitsyn | yes sure | 17:51 |
lisitsyn | Any provides typename | 17:52 |
lisitsyn | lemme find | 17:52 |
lisitsyn | well | 17:52 |
lisitsyn | type() | 17:52 |
lisitsyn | :) | 17:52 |
@HeikoS | how does it work if I only hav ename? | 17:52 |
lisitsyn | HeikoS: you get it from parameter map | 17:52 |
@HeikoS | but I need tag for that no? | 17:53 |
lisitsyn | and then .type() | 17:53 |
lisitsyn | not really | 17:53 |
lisitsyn | parameter map is string-keyed | 17:53 |
lisitsyn | no? | 17:53 |
@HeikoS | ah wait | 17:53 |
lisitsyn | ah | 17:53 |
lisitsyn | use BaseTag | 17:53 |
lisitsyn | there is typed Tag and untyped BaseTag | 17:54 |
lisitsyn | :) | 17:54 |
lisitsyn | HeikoS: so get(BaseTag(name)).type() | 17:54 |
@HeikoS | self->map[BaseTag(name)].get_value().type(); | 17:54 |
lisitsyn | self->get(BaseTag(name)).get_value().type(); | 17:55 |
lisitsyn | yeah | 17:55 |
@HeikoS | thx | 17:55 |
@HeikoS | cool | 17:55 |
@HeikoS | so error msg is better now | 17:56 |
-!- travis-ci [~travis-ci@ec2-54-167-169-20.compute-1.amazonaws.com] has joined #shogun | 17:58 | |
travis-ci | it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/343843668 | 17:58 |
-!- travis-ci [~travis-ci@ec2-54-167-169-20.compute-1.amazonaws.com] has left #shogun [] | 17:58 | |
-!- HeikoS [~heiko@p4FED57F8.dip0.t-ipconnect.de] has quit [Ping timeout: 255 seconds] | 18:20 | |
-!- HeikoS [~heiko@p2003005BE30A2100BD85DE74700336FE.dip0.t-ipconnect.de] has joined #shogun | 18:29 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 18:29 | |
-!- HeikoS [~heiko@p2003005BE30A2100BD85DE74700336FE.dip0.t-ipconnect.de] has quit [Ping timeout: 265 seconds] | 18:51 | |
-!- siddhartha [95a29b32@gateway/web/freenode/ip.149.162.155.50] has joined #shogun | 19:09 | |
-!- siddhartha [95a29b32@gateway/web/freenode/ip.149.162.155.50] has quit [Client Quit] | 19:09 | |
-!- shubham808 [0e8bf0fb@gateway/web/freenode/ip.14.139.240.251] has joined #shogun | 21:28 | |
-!- shubham808 [0e8bf0fb@gateway/web/freenode/ip.14.139.240.251] has quit [Client Quit] | 21:32 | |
--- Log closed Wed Feb 21 00:00:11 2018 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!