[01:33:37] hal yes i can [01:33:44] halfak* [01:36:40] Amir1: hello [01:36:57] Zppix: hey [01:37:08] please don't self merge [01:37:38] also travis is still failing [01:38:23] Amir1: halfak asked me to [01:38:25] :) [01:38:51] no its not its havent done a build yet [01:39:06] its waiting for the next repo update [01:39:37] Zppix: did you squash your commits? [01:39:53] yes [01:40:42] now I'm happy [01:40:43] :D [01:41:04] I also moved ores to trusty from precise ci tests Amir1 [14:20:09] o/ Amir1 [14:20:11] sorry I'm late [14:20:43] Hmm. Looks like you're not here. I sort of half-remember you saying you had something to do. [15:04:46] Zppix, hey! [15:04:55] I see you merging your own PRs :/ [16:11:02] halfak: isnt that what you wanted? [16:11:13] if not then i misunderstood your message [16:12:05] No worries. I think it's all OK. [16:12:57] halfak: sorry next time ill just leave it be if i dont understand :D [16:13:19] If I said "just self-merge it", my bad. But in this case, call is OK :) [16:14:07] halfak: i dont think you did but your msg made me think that but whatever it shouldnt of broke anything regardless [16:15:17] halfak: the builds work but they error out because it doesnt get what it wants to make it pass ... i dont know anything to fix that. [16:15:32] i may turn off ci until we figure this all out [16:16:09] Oh. If the build fails, that's a good reason to not merge :/ [16:16:19] its not failing because of anything we did [16:16:43] its failing due to it not reciving what its expecting (because we dont have it setup to do that via the script you have me runing) [16:16:58] Did you rebase? [16:17:02] https://travis-ci.org/wiki-ai/wikilabels/builds/240664852 [16:17:11] halfak: it rebases automatically [16:17:26] What does? [16:17:29] its not the same errror halfak its not really an error [16:17:44] halfak: well ci and the github webui rebases when prs are merged [16:17:47] Oh! so the test isn't a good one. [16:18:20] halfak: well it is but travis doesnt like it, but the idea is [16:18:22] Need to start the server in the background and then kill it. [16:18:34] This is what should have been worked out before the merge. [16:19:07] halfak: i would of thought travis would of been smart enough to do that automatically [16:19:23] Important to check. [16:19:30] Let's fix this, OK? [16:19:56] yes, im unsure how considering the way i can think of doing it requires its pid which isnt easy to get considering its random each time [16:20:36] we may need to do a command line arg for the server script to allow an arg that has it run for x mins or so [16:20:50] pkill [16:21:10] pkill? [16:21:13] kills a process based on pattern matching [16:21:17] pkill wikilabels [16:21:25] oh ya i forgot that exists [16:21:26] or pkill dev_server [16:21:36] okay working on it [16:22:54] halfak: linux's sleep command is in ms right? [16:23:08] nvm [16:23:11] I wouldn't let the server run for a fixed amount of time. [16:23:21] I'd make some requests and then just manually shut it down. [16:23:28] halfak: but i dont want it killing it before it finishes [16:24:19] "finishes"? [16:24:32] the dev_server startup [16:26:11] let me try it just using pkill and see how it acts? [16:30:22] I think i need to run wikilabels dev_server && pkill dev_server cause pkill wont work otherwise [16:31:13] Zppix, maybe we should loop through status checks for a limited amount of time waiting for the server to come online. [16:31:17] Then we run some checks. [16:31:20] Then pkill [16:32:51] it wont allow pkill because dev_server stops commands from being sent it requires ctrl+c [16:38:26] Run it in the background [16:38:42] how? [16:39:28] travis_wait? [16:40:09] https://www.google.com/search?q=how+to+run+a+program+in+the+background+linux [16:40:10] :P [16:40:32] The short answer is append " &" on the command [16:41:57] according to travis docs that does the same thing [16:43:34] travis_wait does not do the same thing [16:43:43] worth a try [16:43:47] travis_wait signals to travis that it should not fail after 10 minutes [16:43:57] So it will wait 20 minutes [16:44:06] damn [16:44:08] Unless you give it an argument that says a specific amount of minutes. [16:44:47] so i do & wikilabels dev_server halfak ? [16:46:58] i got it [16:51:49] halfak: pkill exited with status 1 [16:51:54] causing build to fail [16:53:24] want me to ask releng? [16:56:47] Not sure releng supports travis [16:57:01] i dont know what to do im out of ideas [16:57:29] pkill is causing it to fail cause it sends signal 1 which travis sees as bad [16:59:26] it's not matching the process [17:02:29] so is dev_server killing itself? [17:02:47] nope [17:02:52] pattern isn't matching [17:03:02] try running locally and looking for it with pgrep [17:03:12] pgrep and pkill use the same type of pattern to match [17:04:45] i dont use linux i cannot do pgrep [17:05:32] Use '-f' [17:05:42] pkill -f "wikilabels dev_server" [17:05:53] ok [17:08:05] Zppix, "git commit --amend" is your friend :D [17:08:33] halfak: if only github webui had that [17:09:00] i have issues using command line with git [17:09:12] still fails [17:09:17] github* [17:09:51] its a typo extra quote [17:09:59] You're working with the web ui? [17:10:06] yes [17:10:15] wow [17:10:20] Whatever works I guess [17:10:21] ? [17:10:34] So you can only work on one file at a time? [17:10:44] no [17:11:05] i can work on a whole repo at a time if i wanted [17:11:55] i feel like dev_Server is killing itself [17:12:12] let me try without pkill [17:12:28] Seems unlikely given that it ran for > 10 mins before [17:16:27] halfak: it passes [17:18:10] Zppix, so now the dev_server shuts itself down? [17:18:24] halfak: i guess [17:18:33] seems broken [17:18:56] halfak: its not it waits for the dev_server to finish startup [17:21:57] PROBLEM - mem-check on Ores-Compute is UNKNOWN: NRPE: Unable to read output [17:21:58] ACKNOWLEDGEMENT - mem-check on Ores-Compute is UNKNOWN: NRPE: Unable to read output Ack [17:22:45] paladox: what was up with that ^? [17:22:53] Zppix nothing [17:23:04] that command wont exist on the servers any ways. [17:24:15] Zppix, what waits for startup to finish? [17:25:19] halfak: travis [17:26:04] travis shouldn't know when the server is ready [17:26:15] And if travis kills a process, that should be a failure. [17:27:48] i asked a buddy who works with CI somtimes and he knows travis (i didnt know that) hes says travis waits for processes in background to wait for a input if travis isnt told to do an input or nothing is expected to happen from the current state travis stops the procs and finishes the build [17:31:31] How would travis know when the process is waiting for input? [17:31:44] It still uses a little CPU while waiting for a request [17:31:58] Either way, this isn't an effective test. [17:33:44] halfak: i mean this all i have to work with feel fre to add on as you see fit [17:33:49] I bet it's just letting it run in the background and moving on. [17:34:02] Just left a note about how you messed up the pattern for pkill [17:34:48] Any more servers that need monotoring? [17:36:35] paladox, we have some wmflabs ORES monitoring done with prod icinga. I'm not that tuned into what's in place. [17:36:54] Seems like if we are doing some monitoring with your labs install of icinga2, we should do all of it. [17:37:29] yep [17:37:52] halfak: that and we can customize more considering we dont monitor a whole lot and we're newer version [17:46:39] wiki-ai/revscoring#1010 (md-from-rst - df90585 : Devin/Zppix): The build passed. https://travis-ci.org/wiki-ai/revscoring/builds/240886983 [17:48:09] wiki-ai/revscoring#1012 (md-from-rst - c42e0cd : Devin/Zppix): The build passed. https://travis-ci.org/wiki-ai/revscoring/builds/240887517 [17:50:09] Zppix, check that MD. Lots of bad rendering [17:50:29] looking [17:52:19] * halfak heads to lunch [17:53:05] * paladox had lunch 7 hours ago [18:26:15] halfak|Lunch: want me to add codecov.io to wikilabels? [18:57:21] Zppix, sure. [18:57:34] ok [18:57:42] I just cut down a tree in my yard. [18:57:51] lol [18:57:53] Planning to get rid of it and it's just starting to grow it's death apples [18:58:09] * paladox yard -> backgarden :) [18:58:21] "death apples"? [18:59:00] Yeah. ineddible hard berries that kill everything they touch. [18:59:23] wiki-ai/revscoring#1016 (ci-to-trusty - eaba308 : Devin/Zppix): The build passed. https://travis-ci.org/wiki-ai/revscoring/builds/240911989 [19:03:24] halfak: ill work on that later it will be just as fun as travis was :P [19:05:50] halfak: i fixed the formatting [19:08:31] * paladox waits 1:51 hours for the exit polls [19:29:28] who's watching the uk election? :) [19:40:31] halfak: o/ [19:41:12] o/ [19:41:17] paladox, any prelim results? [19:41:18] paladox: I would be curious to see the poll result tonight. But I think the conservative will have the most number of seats :) [19:41:51] halfak well the poll this mornning on the news said torries then another one said labour [19:42:00] * paladox hates torries :) [19:42:31] my area has andrea leadsome who tryed to win last year at being pm of the uk [19:42:48] halfak http://election.news.sky.com/general-election-results-24 [19:43:00] shortly live blogs will be starting [19:43:18] As it is currently illegal for news sites to talk about who to vote for. [19:43:35] until 10pm [19:43:38] halfak: I want to confirm something regarding to the item quality model. I think my research is done once we have developed a model based on the top performing classifier right? [19:43:59] I am trying to start writing my thesis now. So, this kind of information would be useful :D [19:44:35] glorian_wd, hmm... I thought you'd have a goal here. [19:44:37] or I wonder, if there's something else that comes after that [19:44:51] Yeah. My goal would be finding the best classifier for item quality model [19:44:53] But I think if you can show a substantial improvement on the baseline performance of the model, that'd be worth writing about. [19:45:17] But I don't know if you come up with a better goal :P [19:45:41] baseline performance of the model? do you mean the one that was done in Hackathon? [19:47:06] halfak: can you merge my pr in wikilabels so i can setup codecov? [19:49:51] Zppix, I just commented [19:50:47] halfak found a live blog on bbc http://www.bbc.co.uk/news/live/election-2017-40171454 [19:51:59] paladox, wish I had time. In a meeting, talking here, reviewing PRs, and eating lunch ATM. Will check it out in a bit :) [19:52:10] ok :) [19:52:13] Looks like the results will be in when I'm done with work (5PM CDT) [19:52:14] :D [19:52:23] well partial results [19:52:32] counting goes on all through the night [19:52:40] halfak: baseline performance of the model? do you mean the one that was done in Hackathon? [19:52:48] (counting starts at 10pm bst) :) [19:52:52] thought you missed that :P [19:54:29] glorian_wd, yes [19:54:38] halfak: fixed and build passed [19:55:12] Zppix, this will kill the server before it fully starts up. [19:55:50] so what do you want me to do? [19:57:00] Wait N seconds. Send a request to the server. [19:57:09] There were need to be a database running for the server to work. [19:57:37] See the README.md [19:58:58] i suck at dbs... ill lookinto how to do that in travis [20:00:02] Should just be a sequence of commands. [20:06:16] 54 minutes :). Also sky has it's front headlines as the elections now :) [20:40:34] sorry for spamming im trying to fix something :/ [20:42:39] halfak: would just sudo adduser wikilabels be good enough when i try to go to edit the db the dev_server comes back from background [20:45:25] 15 minutes left :) [20:48:04] Zppix, uh. no. Gotta install postgres and set it up [20:48:37] ok [20:57:51] 3 mins left :) [20:58:07] paladox: make me a bot that does travis-ci for me please [20:59:20] Zppix i doint know how to create bots that much :) [20:59:28] travis-ci? [20:59:50] use travis-ci yaml file that notifiys irc? [21:00:23] paladox: no i mean setup a bot that does the config for me [21:00:27] polls are closed now :) [21:00:28] uh [21:00:39] paladox: all day today has been dedicated to travis-ci config :P [21:00:39] I have never setup the travis-ci bot though [21:00:46] im joking [21:00:49] lol [21:01:52] "BREAKING NEWS: The broadcasters' exit poll is predicting that the Conservatives will be the largest party in the next parliament but could fall 12 seats short of an overall majority." [21:02:04] glorian_wd your right ^^ [21:04:01] Snap election was a good bet then? [21:04:24] halfak: i need help now it is saying line 1 is wrong but it hasnt changed [21:04:34] https://travis-ci.org/wiki-ai/wikilabels/jobs/240957482 [21:04:51] paladox: If the conservative end up with the majority seat, I'd guess that might be because the recent tragedies in Manchester and London [21:04:53] halfak though the 12 seats short of an overall majourity make me think they need a coallition. [21:05:15] maybe Britons support the notion to close their border :/ [21:05:34] Zppix, there isn't a rake file? [21:05:43] a what? [21:05:47] Rakefile [21:05:50] * paladox dosen't. [21:05:52] That's what the error says [21:06:33] halfak: it cannot parse the config so its going with default which is ruby [21:06:49] its saying it cannot parse line 1 of travis.yml [21:08:25] oh i see [21:08:45] The file is all messed up. [21:09:14] how so [21:11:00] Try to parse it as yaml and get a bunch of errors [21:11:08] http://yaml-online-parser.appspot.com/ [21:12:16] but all i did was add the apt-gets [21:12:20] what broke it? [21:12:46] formatting [21:13:04] Missing spaces. And a list without a key [21:13:09] well let me commit this (apt-gets have to be in before-install [21:13:21] after fixing ^ [21:14:46] i dont see a list without key [21:15:15] are you referring to the postgres stuff? [21:28:28] halfak: i need help again... i dont understand [21:59:18] Zppix, will be happy to take another look at this tomorrow [21:59:22] Or maybe Amir1 can help [22:00:00] Amir1, we're essentially trying to install and run postgres inside our travis instance so we can send a few requests to wikilabels API as a CI test [22:02:45] hmm, that doesn't look too complicated [22:02:56] but needs some time to read the manuals