[08:59:32] New patchset: Yuvipanda; "Add new version of EL Table into the fray" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/55026 [09:03:16] New patchset: Yuvipanda; "Added .gitreview" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/55027 [09:03:17] New review: Ori.livneh; "The complexity of the SQL is a little alarming, but the execution plan looks OK in 'explain extended'." [analytics/limn-mobile-data] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/55026 [09:03:17] Change merged: Ori.livneh; [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/55026 [09:03:51] New review: Yuvipanda; "It's only going to get worse as we add more tables :)" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/55026 [09:05:04] hello [09:05:16] hi hashar [09:05:40] * YuviPanda waves at hashar [09:08:00] Change merged: Ori.livneh; [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/55027 [09:08:26] thanks ori-l [09:08:43] ahhh [09:08:47] we have toooo many projects :-D [09:09:52] yes, and then some [09:11:35] moooar tests [09:11:39] andre__: around? [09:11:46] YuviPanda, yes [09:11:55] hashar: apps have no tests. Analytics stuff has no tests. makes me a little sad [09:12:04] andre__: what exactly does the 'ASSIGNED' status mean? [09:12:05] go ahead and add some ? [09:12:18] and if we set that should we also add assignee? [09:12:39] hashar: apps should have some tests within the next few weeks. I'm going to carve out one week from next sprint for adding tests [09:13:03] YuviPanda, if you click on "assigned to" in a bug report, you will end up at https://bugzilla.wikimedia.org/page.cgi?id=fields.html#assigned_to [09:13:09] assignee = The person in charge of resolving the bug. [09:13:20] and you expressed that you're working on it, so I set you as the assignee. [09:13:29] ah, okay [09:13:31] got it [09:13:40] andre__: do we use that projectwide? [09:13:42] (assignee) [09:13:47] religiously, I mean? [09:14:39] YuviPanda: mostly but not extremely consistently. It is a nice way to show "somebody is working on this" to avoid duplicated work. Very unlikely in the case of Mobile App, I know, but might be good in the long run :) [09:14:53] yeah :) [09:15:23] so I don't expect everybody to religiously set it, but when I see indicators that somebody works on a report I sometimes do it. [09:16:12] andre__: okay [09:16:19] andre__: do you want to talk about the apps and bugzilla now? [09:16:32] andre__: considering my bandwidth situation (situation: miserable) I think IRC > Hangout [09:16:56] hehe... [09:17:10] YuviPanda, can we do this a little bit later today? I'm still trying to wake up :) [09:17:14] aha [09:17:15] okay [09:17:20] andre__: i'll be around, poke me whenever [09:17:27] like in two hours if that's not too late? ah, ok [09:17:53] sure [09:17:57] I woke up about an hour ago [09:18:02] I'm technically in IST [09:18:12] :D [09:18:15] but I think tz wise you can consider me as being in europe [09:18:24] ehehe [09:18:30] :D [09:20:17] some Russian regions surely have the same TZ as you [09:20:27] more or less [09:21:21] Nemo_bis: you mean IST or my actual timezone? [09:23:52] YuviPanda: any :D [09:24:12] by an approximation of half a hour [09:25:25] :D [09:25:34] MaxSem is probably not in too different a timezone than my real one [09:25:43] but he still is around for most of PST anyway [09:26:39] wassup YuviPanda [09:26:53] just talking about real and effective Timezones [09:26:57] MaxSem: what is your timezone? [09:27:03] +4 [09:27:23] MaxSem: but what is your effective timezone? [09:27:28] when do you wake up + when do you sleep? [09:27:49] do I sleep? [09:27:55] not that I know of [09:27:59] you're awake when I wake up [09:28:01] I get up at noon [09:28:02] you're awake when I go to sleep [09:28:25] heh, sounds about right [09:28:36] so it can be approximated to Azores time or something [09:29:14] heh [09:29:24] similar to mine too, wake up around 1:30 usually [11:11:51] hashar: I'd like to try to see whether I could fix the way we build gerrit plugins in Jenkins. I've been given acces already, but as job configurations are not yet in a repo ... Does "Be bold" also extend to Jenkins? [11:12:40] qchris: yes :-] [11:12:45] Ok. Thanks :-) [11:12:46] qchris: the jobs have been created by Chad [11:13:16] I just wanted to make sure I don't set hell loose. [11:24:49] * hell_ is loose [12:11:03] New patchset: Yuvipanda; "Rename dashboard to be more accurate" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/55045 [12:39:11] qchris: we have a Jenkins plugin to keep track of the modifications made to jobs :-D that might help [12:39:31] qchris: example https://integration.wikimedia.org/ci/view/Java/job/gerrit-core/jobConfigHistory/? [12:39:44] Oh. That's great \o/ [12:39:47] so we now you broke everything hehe https://integration.wikimedia.org/ci/view/Java/job/gerrit-core/jobConfigHistory/showDiffFiles?histDir1=%2Fvar%2Flib%2Fjenkins%2Fjobs%2Fgerrit-core%2Fconfig-history%2F2013-03-20_18-36-22&histDir2=%2Fvar%2Flib%2Fjenkins%2Fjobs%2Fgerrit-core%2Fconfig-history%2F2013-03-21_12-29-03 [12:40:12] now / know [12:40:13] grbmbl [12:41:21] So I was just about to send Email to demon about what I did behind his back... Now I can just point him to the above URL. Yay! [12:42:09] But it doesn't show names. So I can break stuff anonymously Hehehe :-) [12:42:22] it does in the history [12:42:27] https://integration.wikimedia.org/ci/view/Java/job/gerrit-core/jobConfigHistory/ [12:42:27] Ah it does have names in the overview. [12:42:37] I just saw that ;-) [12:42:39] I don't think we can use blame though [12:43:03] Anyway, we're all green in the Java section again. [12:54:15] <^demon> :) [12:56:19] * YuviPanda has just started using Gerrit, should try to not bitch too much [12:56:31] <^demon> qchris: Got your e-mail. Looks great! [12:56:43] :-) [12:57:12] <^demon> Do we need to do the same in hooks-its for hooks-bugzilla's benefit? [12:58:08] When I checked yesterday, some plugins already had the direct "mvn clean install", but I'll go and check all plugins to be sure. [12:58:51] <^demon> Ah, that's what I was doing in hooks-its. [12:58:56] <^demon> That's probably sufficient :) [13:04:39] <^demon> Ooh, I see the gc stuff went in. [13:05:20] That sounds like it's time to upgrade gerrit :-) [13:09:35] <^demon> Yeah, I'm gonna start trying to pick a point to use. [13:10:04] <^demon> Shawn's rumbling about 2.7, and Martin's looking at 2.6, so might want to see what settles in prep for that too. [13:13:55] <^demon> 2.6 is gonna be kind of a boring release though :p [13:14:12] <^demon> Since he wants to branch just-before-Schema74. All the fun stuff happened after that :p [13:19:14] :-) But he has a point when saying that too many commits happened since the last release. [13:20:33] <^demon> Yeah, I'm just glad we're already past that upgrade. [13:20:33] <^demon> :) [13:52:40] Change merged: Yuvipanda; [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/55045 [13:57:58] <^demon> Hmm, once https://gerrit-review.googlesource.com/#/c/43770/ goes in, we might be able to stop building with gwt.style=PRETTY and go back to serving the smaller js. [13:59:11] hey ^demon [13:59:20] <^demon> hiya [13:59:29] where do I file requests for wm-bot to spam changes to analytics/* to #wikimedia-analytics? [13:59:35] it currently just comes through here [14:06:04] <^demon> Ask petan, wm-bot is his :) [14:06:12] <^demon> Or do you mean gerrit-wm? [14:12:00] <^demon> YuviPanda: Like this? https://gerrit.wikimedia.org/r/#/c/55057/ :) [14:12:29] ^demon: exactly like that, yes ;) [14:12:33] ^demon: and yes, I meant gerrit-wm [14:12:48] ^demon: I'm sure I've mentioned how awesome you are before. [14:12:50] thank you! [14:13:34] <^demon> You're welcome :) [14:13:44] <^demon> Just need someone to review+merge it. [14:14:30] I just did a +1, though I doubt that's going to help [14:18:02] hi [15:50:10] New patchset: Hashar; "more MediaWiki extensions" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/55077 [15:53:03] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/55077 [15:55:29] I am out for now, will be back tonight [15:56:42] New patchset: Zfilipin; "Disable AFTv5 tests because of a known bug at beta cluster" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/55078 [15:58:47] New review: Cmcmahon; "Turn off AFT tests pending significant refactoring after deployment to dewiki yesterday." [qa/browsertests] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/55078 [15:58:48] Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/55078 [16:47:19] greg-g: https://gerrit.wikimedia.org/r/#/c/55017/ is merged [16:47:55] I can cherry-pick and deploy it if you like [16:48:04] Tim switched all wikis back to wmf11 so it wouldn't actually go live anywhere [16:48:20] RoanKattouw: one sec, lemme get us a window [16:48:31] they're doing the AFT stuff right now [16:48:39] Oh right OK [16:49:04] Hm there's no window until 4pm [16:49:09] Never mind then [16:49:10] RoanKattouw: how much time do you think you need to do it? [16:49:15] we can bump someone [16:49:21] and/or shorten their window [16:49:36] So, what I'm offering to do is the easy part and will take very little time [16:49:42] Like 5 minutes [16:49:55] The harder and more dangerous part is actually switching wmf12 back on on certain sites [16:50:03] gotcha [16:50:06] Which I'll leave up to Aaron or Reedy or whoever does that normally [16:50:21] any sense in doing the two parts separately, or should we just do them all at once when we have the window? [16:50:36] You could do both at the same time, it doesn't really matter [16:50:44] k, just checking [16:50:55] so, the db locking shouldn't be an issue in your mind? [16:51:05] Step 1 is just a matter of changing the wmf12 code which isn't actually live anywhere anyways, step 2 is putting it live places [16:51:07] What DB locking issue? [16:51:15] Oh, the thing that Aaron added? [16:51:19] yeah [16:52:45] ^demon|lunch: so, I guess you're eating now, but, when would be a good time for you to do a deploy? [16:52:57] It uses FOR UPDATE, I think that's locking in a different way, let me see [16:53:02] k [16:53:35] Yeah so FOR UPDATE has the effect of locking out *others* [16:53:47] Until such time as the row is updated [16:54:14] <^demon|lunch> greg-g: I haven't actually walked away to get food yet. [16:54:20] <^demon|lunch> So, what needs deploying? [16:54:38] ^demon|lunch: wmf12 broke all page moves, so Tim rolled everything back to wmf11. We now have a fix merged in master for that bug [16:54:41] wmf12 with the cherrypicked fix for https://bugzilla.wikimedia.org/show_bug.cgi?id=46397 [16:55:17] <^demon|lunch> Has it been cherry picked to wmf12 yet? [16:55:44] not yet, RoanKattouw was offering to do that part [16:56:46] Working on that now [16:57:35] <^demon|lunch> So, all Roan needs me to do is scap and flip the wikis back? [16:58:21] The change itself is one file, sync-file will do that [16:58:33] As for flipping the wikis back, I've never done that before and I'm about to walk to a different office [16:58:35] so, don't do it yet, AFT is doing stuff right now [16:59:17] ^demon|lunch: and, the wikis we'd switch on would be 1) test2, bang at it and test the fix, if ok then 2) the wikis that are a part of the phase1 and phase2 deploys (ie: to get back where we were as of yesterday) [16:59:29] OK, Gerrit is merging the cherry-pick now, very slowly [16:59:32] <^demon|lunch> Ok, so I'm going to run out and grab some food. Can do that when I'm back. [16:59:59] <^demon|lunch> RoanKattouw: Speaking of slowness, we should play with some gc/repacking. I want your help since you kept crap from exploding last time I tried it :p [17:00:06] <^demon|lunch> $someTimeSoonish [17:00:40] New patchset: Cmcmahon; "cuke for move-page browser test" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/55087 [17:00:56] ^demon|lunch: OK that's fine. I am on vacation most of next week though [17:01:10] <^demon|lunch> Yeah, doesn't have to be right now. [17:01:13] <^demon|lunch> Just...sometime soonish [17:01:35] chrismcmahon: ^^ awesome, thanks for the quick turn around on the test [17:01:53] ^demon|lunch: cool, ping me when you're back, think you'd be ready in like 45-1hr? [17:02:13] so, 11am Pacific/2pm Eastern [17:02:19] <^demon|lunch> Yeah, probably less. Just going to grab food and bring it home. [17:02:27] greg-g: that's a great browser test. unless something goes haywire it will be running tomorrow afternoon [17:02:29] cool [17:03:00] greg-g: what I checked in is just the "story", zeljko and I will pair on the executable code first thing tomorrow [17:03:02] chrismcmahon: k, we're going to deploy on test2 when ^demon|lunch is back, in about 30 minutes or so, mind helping test it? [17:03:06] cool [17:03:11] greg-g: sure, np [17:03:15] thanks man [17:05:05] matthiasmullie: heya, pinging you here, too. Just let me know when you're done with the AFT deploy this morning, please. We have to do a redeploy of wmf12 after a rollback last night [17:06:26] YuviPanda, apart from the manual,is there any documentation about extension dev? [17:07:08] Rahul_21: http://svn.wikimedia.org/doc was useful to me when I was doing it, unsure if it is still up to date [17:07:29] matthiasmullie: I don't see anything in the server log, have you started yet? [17:12:33] YuviPanda, thanx [17:14:15] greg-g: you can do so now [17:14:34] matthiasmullie: our deployer isn't back from lunch yet :) [17:14:37] I've not yet done anything yet; patches not yet merged [17:14:46] ok [17:14:57] so go ahead and steal the current window [17:15:34] gotcha, thanks matthiasmullie [17:17:24] greg-g: soo, how about those weekly deploys? :) [17:17:53] Aaron|home: :P [17:25:20] <^demon|lunch> nom nom nom. [17:25:53] ^demon|lunch: so, csteipp is doing the deploy to test2 right now [17:27:10] <^demon|lunch> So, only thing I need to do when you guys are done testing is roll it to phase1/2 wikis? [17:28:42] right [17:30:04] is sync-file relative to /common? [17:30:15] ^demon|lunch: ^^ [17:30:25] <^demon|lunch> Yep. [17:30:35] <^demon|lunch> So you'd do sync-file php-1.21wmf12/foo/bar/... [17:34:32] chrismcmahon: about ready for your testing help on test2 [17:35:07] greg-g: check :) [17:35:47] just playing "keeping all parties aware of the current situation" point :) [17:37:55] ^demon|lunch: So I updated wikiversion.dat. Do I need to run a full scap? [17:38:03] Or just sync that one file? [17:38:16] maybe sync-wikiversions? [17:38:22] sync-wikiversions SUMMARY [17:38:32] that will update the cdb file and sync the dat/cdb files [17:38:36] <^demon|lunch> Yeah, sync-wikiversions. [17:38:43] <^demon|lunch> Was gonna say update-wikiversions, but I knew that was wrong. [17:40:11] Alright, it's done. [17:40:38] Btw, the instructions on het-deploy say to edit the file in /home. Do I need to push those back into gerrit? [17:41:19] <^demon|lunch> Yes, should push them back to gerrit, otherwise the heads will diverge. [17:41:54] Cool. Just commit, and git push? [17:42:26] <^demon|lunch> I can't remember if we have direct pushing on that repo or not. [17:42:27] <^demon|lunch> Sec. [17:42:46] Or do you normally do it through gerrit? [17:43:05] <^demon|lunch> I usually go through gerrit. [17:43:12] <^demon|lunch> But you can push your change to gerrit. [17:43:25] <^demon|lunch> `git push [remote] HEAD:refs/for/master` [17:46:38] csteipp: so, test2 should be OK? I'm getting http://paste.lisp.org/display/136153 [17:47:58] greg-g: wfm [17:48:21] yeah, Infinity now redirects to Infinity Plus One, correctly, but I got a stack trace at the time of doing it [17:50:40] aude: I got http://paste.lisp.org/display/136154 when moving https://test2.wikipedia.org/wiki/External_link_templates_test [17:50:54] chrismcmahon: csteipp ^demon|lunch Aaron|home ^^ [17:50:55] http://pastebin.com/fKxkvE1k [17:51:10] https://test2.wikipedia.org/wiki/Cool_Hand_Luke is corrupt now [17:51:15] can't move it out of the way [17:51:30] ok, something's not right http://test2.wikipedia.org/w/index.php?title=0.15215567970984123 should have a redirect [17:51:39] checking something... [17:51:53] Aaron|home: ^^ help please, we have corruption here [17:52:11] anomie: any idea on the lua exception ? ^ [17:52:59] <^demon|lunch> At first glance, I don't think it's Lua's fault, I think it has to do with the incomplete page move. [17:53:20] moved to http://test2.wikipedia.org/wiki/0.15215567970984123_Moved which should not have a redirect to itself [17:53:21] https://bugzilla.wikimedia.org/show_bug.cgi?id=46427 is what i got the first time around when moving [17:53:33] i had a script error broken lua script there (my fault) [17:53:48] i removed the invoke and could move hte page fine [17:53:56] * anomie look [17:53:58] * anomie looks [17:54:03] but the old title is corrupt from the moves [17:54:29] <^demon|lunch> Do we have a page_id for the old title? [17:54:33] <^demon|lunch> Curious what it looks like. [17:54:46] we must have had a page id [17:54:51] now, no idea [17:54:52] I wonder if that exception was the one on line 2950 of Title.php [17:55:37] * anomie git pulls [17:55:41] line 2965 now... [17:56:07] https://test2.wikipedia.org/wiki/Cool_Hand_Luke2 is not good either [17:56:36] https://test2.wikipedia.org/w/index.php?title=Testpage1&redirect=no [17:57:05] <^demon|lunch> http://p.defau.lt/?8oELRJH6DEz5Qz8Eib1MIw - 'Cool_Hand_Luke%' page entries. [17:59:52] YuviPanda, as i was telling yesterday,i have an idea of extension that marks all the depreciated tags in HTML5 by color red ,so the user knows he has to change it [18:00:09] ^demon|lunch: strange [18:01:23] hmm, move deleted the entire contents of this page after saving under the previous name: http://test2.wikipedia.org/w/index.php?title=0.8878279934637614_move_has_lua&action=history [18:01:40] yuck yuck [18:01:45] no errors shown [18:02:16] <^demon|lunch> Maybe we didn't fix the bug, we just hid it ;-) [18:03:08] Aaron|home: sorry to ping again, just making it explicit that you're advice/insight here would be greatly appreciated :) [18:03:18] <^demon|lunch> greg-g: So, I'm fine with leaving it on test2 so people can keep testing, but let's forego any other rollouts of 1.21wmf12 until we've verified a fix. [18:03:21] <^demon|lunch> This bug is *nasty* [18:03:27] ^demon|lunch: heh [18:03:40] ^demon|lunch: 1 million percent agree [18:04:18] yep, deleted normal text in page upon move also http://test2.wikipedia.org/w/index.php?title=0.7984418109470285_move_normal_text_in_page&action=history [18:04:23] <^demon|lunch> Maybe an easy fix eventually, but data loss is Very Bad(tm) [18:04:48] yeah, data is still in history, but it's Not Nice(tm) [18:05:53] <^demon|lunch> Well, we're losing the ability for page to keep track of its latest rev :) [18:06:09] true [18:06:18] * aude fine with waiting on wmf12 until fix (for wikidata) [18:06:24] fine with whatever on test2 [18:06:36] thanks aude [18:06:51] sad it's not fixed, but understand and appreciate the attempts [18:06:53] so no other deployments to prod today, but one to mw.o for Echo. [18:07:02] $redirectArticle->updateRevisionOn( $dbw, $redirectRevision, 0 ) must do nothing [18:07:20] (just want to make sure this does not escape from test2 to prod, which has happened on occasion) [18:07:21] <^demon|lunch> greg-g: We should send an e-mail to engineering@, so people known for any deploys they do. [18:07:24] a valid revision is in the redirect history, page_latest needs to actually point to it [18:07:54] ^demon|lunch: will do, good point, thanks [18:11:25] email sent [18:13:26] Aaron|home: Since Tim identified the commit that broke this, is there any reason we can't just revert that commit? [18:17:43] <^demon|lunch> Aaron|home: http://p.defau.lt/?yzW_2M5M5AzA2xovnkhh_Q :p [18:23:57] marktraceur: is this the right channel to ask questions about parsoid? [18:24:37] hey Rahul_21 [18:24:47] I think that idea would be better implemented as a gadget or as an userscript [18:24:50] than an extension [18:25:37] Rahul_21: if you're looking for some extensions feature work to do, I can be greedy and suggest you add an API to Extension:ShortURL :) [18:26:01] Rahul_21: it is going to be a good learning experience, and more importantly - it'll get deployed on the cluster! [18:26:27] YuviPanda, Ill take a look at it for sure :) [18:28:01] YuviPanda: #mediawiki-parsoid [18:28:49] YuviPanda, i found this cool new thing http://discover-devtools.codeschool.com [18:36:01] Hello everyone, I"m setting up for today's Tech Chat. The stream URL is: http://youtu.be/ju8suSJY0zQ [18:36:07] we will begin in about a half hour [19:03:00] Hello, yurik & anomie :) I bet you remember me, from $last_days->gerrit()->ApiFeedWatchlist(). What do you think, as we're talking about commas, should I restore the commas I removed or should I additionally put commas wherever they can be? ;) [19:03:26] Kaligula- Just restore the ones you removed. No need to add them everywhere, either. [19:03:44] hi Kaligula, anomie is a faster typist [19:03:55] yurik :D [19:04:11] anomie: ok [19:04:14] I'm at the SMWCon and might be able to get some Postgres testers, which is nice [19:04:38] cool! sumanah, I will drop by to see how it is. Will present tomorrow [19:04:40] Yippie, build fixed! [19:04:40] Project browsertests-linux-chrome build #226: FIXED in 5 min 26 sec: https://wmf.ci.cloudbees.com/job/browsertests-linux-chrome/226/ [19:04:41] zeljko.filipin: Disable AFTv5 tests because of a known bug at beta cluster [19:04:48] for some reason I read that as 'portugese testers' instead of 'postgres testers' [19:05:00] That would be nice too [19:05:03] YuviPanda, i'm sure we could use them too [19:05:07] although arguably less of a bottleneck that needs clearing right now [19:05:17] yeah, I'm sure that the i18n team has that covered [19:05:31] We have Waldir :) [19:05:37] one of the things I'm looking forward to is data collection by the i18n team on language usage. [19:05:52] at least just to check if I'm the only guy who can't do jack in his native language [19:05:58] postgress is overrated. I say we should start using GIT as the storage engine [19:06:01] YuviPanda, http://discover-devtools.codeschool.com its great [19:06:33] Rahul_21: wooo, that's quite nice! [19:06:47] does the HPHP talk take place now? [19:07:19] sumanah: that link Rahul_21 found is quite wonderful, I'd reccomend it to people who are doing Gadgets [19:07:23] MaxSem: Yes [19:07:31] linkie? [19:08:14] yay audio! [19:08:21] YuviPanda, yeaaa!i shared it with my friends right away [19:08:26] oh linkieeeee? [19:08:26] Rahul_21: nice find! [19:09:19] aha, should be at https://www.youtube.com/watch?v=ju8suSJY0zQ&feature=plcp but not yet [19:09:41] ahha that explains why i was looking at ryan speaking about git-deploy from january.. [19:10:24] cndiv, will there be a video feed? [19:10:26] YuviPanda: have you tried any of the videos yet? [19:10:44] sumanah: I just hit one [19:10:57] going through it, but I've been looking for this exact thing for a while now [19:11:28] sumanah, seems like they are compeeting with codacademy.com [19:11:38] yeah. [19:11:39] sorry, http://www.codecademy.com [19:11:55] I'm curious how actual newbies find it [19:11:59] btw, sumanah, i think i will use that site to demo tomorrow :) [19:12:19] Rahul_21: have you gone through the videos? [19:12:31] yurik: you'll use discover-devtools to demo the MediaWiki API? [19:12:39] * sumanah may be missing referent [19:12:39] http://www.codecademy.com [19:12:50] * YuviPanda uses chrome dev tools to demo mw api / gadgets every workshop he does [19:12:52] YuviPanda, a few yes! [19:12:58] Rahul_21: how many? [19:13:02] Rahul_21: how are you finding it? [19:13:27] sumanah, its easy there because i have a tiny javascript that shows results in realtime [19:13:31] as i type [19:13:38] yurik: oooo, link? [19:13:39] Right. That sounds good [19:13:55] YuviPanda, haven't published it yet, but i can add you as a contributor if you want to see it [19:13:56] yurik: it is amusing to me that you're actually going to present at the exact same lectern as you did at Wikipedia Day NYC [19:13:58] YuviPanda, Amazing!!its very good for people new to debugging [19:14:00] sure! [19:14:09] sumanah, hehe :) [19:14:11] Rahul_21: go through the entire thing :) [19:14:26] sumanah: there was talk of adding mw.org api to codeacademy, any idea what happened to it? [19:14:38] YuviPanda, that's what i am working on :) [19:14:52] aaah [19:14:55] right! [19:15:00] yurik: is there an activity page somewhere, where people can track progress, see what needs help, etc? [19:15:01] YuviPanda, codeacademy has videos related to git too!! [19:15:12] https://www.mediawiki.org/wiki/Wikimedia_Engineering/Project_documentation_howto can help you if you want yurik [19:15:22] sumanah, hmm, not really, is there a good place for these things? [19:15:40] I think mediawiki.org + use the templates in the HOWTO [19:16:12] YuviPanda, http://www.codeschool.com/paths/electives [19:16:41] sumanah, thanks, will take a look [19:16:43] http://openhatch.org/missions/git is what I point people to to give them interactive git bootstrapping [19:16:56] YuviPanda: Have I talked with you about developer personas? [19:17:20] no! [19:17:24] sumanah, i remember you suggested me to peep into openhatch! [19:17:29] http://ur1.ca/d299c [19:18:26] Basically, we have opportunistic and systematic developers in the communities we want to engage with, and people who learn in different ways, and some people want lectures and references so they can code defensively, and some people want interactive consoles and rapid iteration [19:18:27] https://blogs.msdn.com/b/stevencl/archive/2010/07/21/do-developers-have-culture.aspx [19:19:17] * YuviPanda reads up followup links [19:19:31] TimStarling would you please mute your mic in the hangout? getting some odd noises [19:20:15] right... [19:20:19] sumanah: interesting, I've not heard that before [19:20:33] no mic problem then [19:20:43] YuviPanda: Very few people had. (I learned of this via the "Are your APIs usable?" chapter of the _Making Software_ book and it made many things make sense) [19:21:16] TimStarling sounds better, not sure if it was on your end or mine :-) [19:21:32] I didn't change anything [19:21:44] must have been mine then, thanks though [19:22:00] sumanah: hmm, most of the people I interact with in workshops are perhaps not developed enough to fall anywhere on that spectrum yet [19:22:56] YuviPanda: you think so? Hmm. I wouldn't be surprised if you quickly discovered what learning methods they gravitated to -- but only if all the paths were equally available and high-quality [19:23:13] YuviPanda, Just listen to this! http://gitreal.codeschool.com/levels/1 [19:23:43] sumanah: true, and that is hard, since I'm usually the only person leading those, my biases will bias which ones are available [19:24:13] however, unfortunately, in my experience, many developers don't think "my learning style is one where I prefer sandboxes & using libraries to reading reference material" -- instead they look at available paths and think "god, who reads these? why do they even make them? useless" [19:24:40] if they spare a single thought for the path they're not taking [19:24:50] or they get frustrated and they misattribute the source of the frustration [19:24:56] well, yes. *yesterday* i ran into someone who didn't want to read ESR's Hacker-Howto because it was too long [19:25:23] "x language sucks" instead of "x language has affordances that don't jibe with the way I think" [19:25:36] there was a post on slashdot a week ago, that apparently 50% of documentation devs get are from stackoverflow.com [19:25:46] heh, at the very beginner level, it gets more to 'programming sucks' [19:26:08] so who needs to read those boring manuals anyway :) [19:26:12] and the way that programming is explained, taught ("Memorize this large set of documentation-ish things that you can lookup in Google anyway!"), most people come with a negative bias [19:26:22] indeed, manuals are for lookup, not afternoon reading [19:26:28] for you they are [19:26:34] for some other people, they really are afternoon reading [19:26:42] because that is how they learn [19:27:09] that was my point about bias, I guess. [19:27:42] it *is* hard for me to see the other ways, on either side of the spectrum, and not go 'eeek! that is so inefficient!' [19:27:56] Right [19:28:38] batch-processing versus line-by-line.... [19:28:56] http://opensourcebridge.org/proposals/991 "Search-first writing makes you think about the structure of your document and product as a series of topics, instead of a big book. The days of linear documentation are over, or at least numbered. Users are much more likely to come to documentation through searches." I hope this proposal gets accepted because I want to see it [19:29:16] http://hiphop-php.com/wiki/index.php/HHVM [19:29:47] sumanah, YuviPanda ,People see ugly code base [19:29:57] Rahul_21: ? [19:29:58] * anomie wonders if they have Scribunto on their MediaWiki install [19:30:14] very vanilla: http://hiphop-php.com/wiki/index.php/Special:Version :) [19:30:33] sumanah: pretty nice! I should go through videos from past OSBs at some point [19:30:53] :-) [19:30:55] YuviPanda, people see ugly code base's and they infer that its not their cup of tea! [19:31:30] TimStarling, if you have hangout connection issues but can get the stream, happy to relay comments/questions from irc [19:31:30] YuviPanda, Thats what happens in many Indian colleges sadly! [19:32:42] Rahul_21: more than ugly, it is 'large' [19:32:55] Rahul_21: they think 'i have at best worked on 100-200 lines of code, this thing is thousands!' [19:33:01] hasharAW: JSDuck is awesome, automatic @since :) https://github.com/senchalabs/jsduck/wiki/@since (also, what's AW?) [19:34:11] YuviPanda, Yes!Dont You think Open Source Programming should be introduced at college level?Cause the first time i ventured into it ,i felt like i was in some parallel universe [19:34:11] * sumanah leaves the channel alone so people can talk about HipHopVM :) [19:34:28] * sumanah leaves http://teachingopensource.org/ behind as she goes :) [19:36:13] Krinkle: AW == away [19:37:06] Krinkle: ahwah automatic @since ? :-D [19:37:48] yeah, we might not want to use it in favour of having the information in the source code (which also has an advantage), but it's nice to know it can do this. [19:38:19] so once we have it generated beyond 1.21 we could start using it [19:38:29] we'll see [19:39:04] python time :-D [19:40:19] hashar: What does that mean? [19:40:46] I am putting my python hat and going to fix my patches that have been reviewed during US morning [19:40:51] also packaging some modules for debian [19:40:56] k [19:44:16] Q. what prompted them to write a trace-based jit vs. a method-based jit? [19:52:03] and are there performance differences between "sandbox" and "production" modes? [19:55:44] New patchset: Hashar; "triggers for some more extensions" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/55122 [19:56:02] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/55122 [19:57:14] paravoid: I'll ask yours next [20:00:02] no worries, I'm sure we'll eventually find out :) [20:00:07] spagewmf1: is E3 done with their deployemnt? [20:00:20] I'm just wondering if it'd make sense fo rus to run in sandbox, since we don't particularly care about source code leakage [20:02:23] paravoid: yeah, we'll have to benchmark it and see what the tradeoff looks like [20:02:51] if it's a few percent who cares, if it's a 50% savings…. maybe it's worth slower deploys [20:02:58] hell it can't be slower than scap is these days :P [20:03:38] RobH: did we ask tim's question about how long global analysis for production mode takes? [20:03:41] er [20:03:42] robla: ^ [20:03:44] too many robs :D [20:03:56] * RobH smites brion for summoning him [20:04:03] * brion ow [20:04:27] <^demon> Hmm, I wonder if we could make jenkins faster by running unit tests with it. [20:04:48] <^demon> parser tests on hhvm :p [20:05:11] ^demon: I got a patch floating around to use the old parsertest script :D [20:05:23] <^demon> I don't like the old script :p [20:05:39] i love the old script! but i'm biased, i wrote it [20:05:41] ^demon: it is faster ! I made it to output JUnit (still have to test though) [20:06:04] the --record was a nice thing [20:06:14] :D [20:06:17] <^demon> I'd rather see us improve phpunit's performance there. [20:06:23] <^demon> I like being able to run it as part of all my other tests. [20:06:32] *nod* [20:06:56] one thing that would help is to mock the databases access [20:07:19] kaldari, superm401 and I are done. Ori wants to update NavigationTiming in, he's going to coordinate with bsitu and matthias [20:07:26] also I think we have MW caches disabled which does not help [20:09:13] why is test running wmf11 and test2 running wmf12? [20:09:28] how is that even possible? [20:10:16] oh no gwicke is willing to move everything to JS I knew it [20:10:54] Yippie, build fixed! [20:10:54] Project browsertests-linux-firefox build #212: FIXED in 8 min 57 sec: https://wmf.ci.cloudbees.com/job/browsertests-linux-firefox/212/ [20:10:55]