[00:15:00] Ryan_Lane: Did the Jenkins <--> Gerrit link get broken in the restart somehow? Doesn't seem to be running tests automatically....no rush, but figured someone should know [00:17:14] The connection to Gerrit is down! Check your settings and the Gerrit server. [00:17:23] #Since the connection is down, Jenkins cannot determine the Gerrit version. So all Gerrit features have been enabled. [00:17:44] test connection works... [00:18:01] marktraceur: try that... [00:18:06] I pressed the start button [00:18:09] the error went away [00:19:34] Reedy: Is it working, though? Let me push a change. [00:20:04] Best way to find out ;) [00:20:14] It doesn't appear to be working [00:24:42]

Project WikipediaMobile - Nightly builds build #399: SUCCESS in 11 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/399/ [00:24:45] Seems to work for mobile.. [00:25:06] Reedy: Mobile is on GitHub, methinks [00:25:19] ah. true [00:26:30] Just restarted Jenkins... [00:27:15] marktraceur: looks more promising... [00:27:16] INFO: Project [Ext-MobileFrontend] triggered by Gerrit: [PatchsetCreated: Change: 26431 PatchSet: 1] [00:27:43] Awesome com.sonyericsson.hudson.plugins.gerrit.trigger.gerritnotifier.ToGerritRunListener [00:28:05] Sweet [00:28:13] Thanks, Reedy [00:28:27] Please wait while Jenkins is getting ready to work [00:28:32] Lazy butler [00:29:34] *nod* [00:30:21] it's using a far amount of cpu time (lol java), so it's seemingly still doing something [00:30:39] I hope it gets up soon :) [00:42:16] marktraceur: back up now [00:43:18] Reedy: I saw! Thanks! [00:49:11] Reedy: just tried to trigger a Jenkins test run on the Parsoid repo, but might be missing the rights to do so [00:52:27] could be the backlog too though [00:56:02] I'm not sure how the rights work.. It let me run one [00:56:10] https://integration.mediawiki.org/ci/job/Parsoid-parserTests/2/console [01:10:31] AaronSchulz: what is going on with, "Oct 3 01:06:16 10.0.2.199 apache2[25581]: PHP Warning: require() [function.require]: GC cache entry '/usr/local/apache/common-local/php-1.20wmf12/extensions/MobileFrontend/includes/MobileFrontend.body.php' (dev=2049 ino=662124) was on gc-list for 601 seconds in /usr/local/apache/common-local/php-1.20wmf12/includes/AutoLoader.php on line 1150" [01:12:42] Reedy: that run triggered, but apparently missed the gerrit changeset to test [01:12:46] AaronSchulz: GC cache entry '%s' (dev=%d ino=%d) was on gc-list for %d seconds" or "GC cache entry '%s'was on gc-list for %d seconds [01:13:11] preilly: yeah, those gc notices having been cropping up for weeks in random places [01:13:20] AaronSchulz: e.g., when process unexpectedly died, reference is not decreased. First apc.ttl seconds is active in APC cache, then is deleted (there isn't next hit on this item). Now item is on garbage collector list (GC) and apc.gc_ttl timeout is running. When apc.gc_ttl is less then (now - item_deletion_time), warning is written and item is completely flushed. [01:13:24] Reedy: when I try to schedule another run on a revision, it just tells me that "No jobs triggered for this event" [01:13:34] so it's web server segfault issues right? [01:14:52] gwicke: You'd probably be better asking Hashar when he's about [01:15:07] Reedy: ok, will try that next week [01:15:09] maybe it's php dying in libxml when parsing or something [01:15:51] TimStarling: are you still monitoring your node count limit change? [01:16:42] not for a while [01:16:50] but I don't think it could cause that [01:17:32] this is why we have gc_ttl, because APC is buggy and often leaks references [13:56:22] Change merged: Ottomata; [analytics/udp-filters] (master) - https://gerrit.wikimedia.org/r/26404 [14:46:35] hashar: Thanks for the mail, I sent one back, I'll be in and out as I commute into work :) [14:46:57] marktraceur: fine. Have a nice commute :-) [14:51:18] marktraceur: ahhhh [14:51:23] marktraceur: I will redo the job [15:03:20] Hey, MediaWiki experts. I have spans with a noprint class that don't show on the printable version (so far so good) but do show in its HTML source. Is this wanted behavior or a bug? [15:03:43] (For my use it's annoying, but I thought I'd ask.) [15:04:36] guillom: that's expected behavior. printable version is just regular version with the print stylesheet attached for screen as well as print [15:04:49] ok, thanks [15:05:04] I'll continue to remove those items manually, then. [15:36:40] New patchset: Diederik; "Resolve merge conflict" [analytics/udp-filters] (master) - https://gerrit.wikimedia.org/r/26459 [15:37:13] Change abandoned: Diederik; "See https://gerrit.wikimedia.org/r/26459" [analytics/udp-filters] (master) - https://gerrit.wikimedia.org/r/25408 [15:37:55] New review: Diederik; "Ok." [analytics/udp-filters] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/26459 [15:37:55] Change merged: Diederik; [analytics/udp-filters] (master) - https://gerrit.wikimedia.org/r/26459 [16:20:58] marktraceur: I have added a couple of sentences to my page: http://www.mediawiki.org/wiki/User:Zeljko.filipin(WMF) [16:21:14] thanks for letting me know I should do it [16:22:50] zeljkof: I added some crucial info, I hope it's correct [16:22:58] add more languages if you know more [16:29:44] zeljkof: Cool! Looks pretty nice. [16:38:40] marktraceur: I just wanted to add anything :) [16:38:48] will add more data later, if needed [16:38:52] zeljkof: Works for me! [16:46:41] chrismcmahon: did you see this? http://lists.wikimedia.org/pipermail/wikitech-l/2012-October/063644.html [16:47:30] zeljkof: I have both https://wikimediafoundation.org/wiki/User:MHolmquist and http://www.mediawiki.org/wiki/User:MarkTraceur, I enjoy both :) [16:47:57] Poll of the room: Does anyone have any idea about writing plugins in Jenkins? [16:49:51] ^demon: I know you've worked with Gerrit plugins / core, have you looked at Jenkins at all? [16:50:35] <^demon> I was the one who set up our jenkins before hashar took it over. [16:50:47] <^demon> Never did any internal-ly stuff with it though. [16:51:06] ^demon: Hrm. I may be in unexplored territory, then. [16:52:12] zeljkof, chrismcmahon, I don't suppose you have any insight re: Jenkins? I'd take either plugin help or help with tracking test results against master instead of "whatever ran just now" [16:53:02] marktraceur: I'm pretty innocent wrt modern Jenkins. if hashar or Krinkle|detached were around, that'd be who to ask for sure [16:55:04] hashar seems to have stepped out, which is fine, I'll catch him later, but he seems to be exploring as much as I am, so I'm looking to see if someone has immediate thoughts :) [16:55:33] Krinkle|detached (or anyone else), if you have thoughts on the scrollback, I'm definitely listening :) [17:00:18] marktraceur: I am not sure I understand what are you trying to do [17:01:08] I do not have any experience writing jenkins plugins, but I do have some experience with test results, if that is what you are trying to do [17:01:39] chrismcmahon: I think I have configured zoiper correctly, I hope it will work for the conference [17:01:53] zeljkof: Briefly, I have test results that come back in JUnit XML, and I want to diff them against the same results in master. If things that *were* passing are suddenly failing, I want the build to fail and Jenkins to report the failure on Gerrit. Else, Jenkins should +1 the change. [17:02:48] marktraceur: I still do not understand :) [17:02:54] zeljkof: That's OK :) [17:02:59] let my try to explain what I understood [17:03:13] you run the tests in jenkins and the tests produce xml file that jenkins understands [17:03:20] *nod* [17:03:35] where do you get the other xml, the one you want to compare? [17:04:09] zeljkof: By doing this exact build process, the one that's already configured, but for the master branch in git. [17:04:27] lost again [17:04:39] let's start from the beginning [17:04:59] K [17:05:03] you want to compare two xml files produced by two runs of the same jenkins job? [17:05:26] zeljkof: Essentially, yes. The problem is, these two runs aren't necessarily sequential. [17:05:33] I see [17:06:01] zeljkof: I want to compare between running the tests on origin/master and running the tests on the gerrit changeset that just got submitted, which triggered the tests. [17:06:11] zeljkof: The purpose being, basically, regression testing. [17:06:36] I guess I am confused because I have no idea how gerrit works [17:06:56] but that should be fixed later today, I have registered for git/gerrit conference [17:08:25] marktraceur: so jenkins sometimes runs the job using code from origin/master, sometimes from gerrit branch? [17:08:42] marktraceur: or did I misunderstood it? [17:09:00] * zeljkof is back in 20 minutes [17:09:34] zeljkof: I don't know if that's possible, this is my dilemma :) if I could get the Jenkins job to run once for master, then once for the new branch, from the same job....that would be perfect. [17:09:48] But I can't....wait. Brain blast. Hold on. [17:10:34] marktraceur: I think that should be possible to implement at least [17:11:03] chrismcmahon: I think I can do it, actually. I just need to run the right job twice, from a master job that's triggered on Gerrit changes. [17:16:50] marktraceur: in the worst case a simple solution would be to do our own diffing and archive the actual test output as a success message, if such a thing exists [17:19:40] marktraceur: potentially worth a look: https://github.com/pliljenberg/build-regression-checker [17:21:04] Not a lot of docs.... [17:21:19] And ugh, installing manuall. [17:21:21] y. [17:22:41] it uses the last successful build as a baseline, which sounds like the default Jenkins behavior [17:22:44] New patchset: Milimetric; "Merge branch 'release/August2012' into develop" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26474 [17:22:44] New patchset: Milimetric; "test 1 of git flow gerrit interaction" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26475 [17:22:45] New patchset: Milimetric; "test 2" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26476 [17:22:45] New patchset: Milimetric; "Merge branch 'feature/GerritTest1' into develop" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26477 [17:22:45] New patchset: Milimetric; "=test 3" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26478 [17:22:46] New patchset: Milimetric; "test 4" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26479 [17:22:46] New patchset: Milimetric; "Merge branch 'feature/GerritTest2' into develop" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26480 [17:23:00] Oh, wow. [17:23:06] marktraceur: see https://github.com/pliljenberg/build-regression-checker/blob/master/src/main/java/jenkins/plugins/regression_checker/AnalysisRegressionChecker.java#L100 [17:23:21] marktraceur: might still be a good starting point [17:23:53] gwicke: I want to try something without more plugins first, I think I might (might) be able to do it. [17:24:40] New review: Diederik; "Ok." [analytics/reportcard/data] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/26477 [17:24:53] marktraceur: to me it seems extending that method to also require that revision to be merged would pretty much do the trick [17:25:03] New review: Diederik; "Ok." [analytics/reportcard/data] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/26474 [17:25:04] Change merged: Diederik; [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26474 [17:25:27] New review: Diederik; "Ok." [analytics/reportcard/data] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/26478 [17:25:28] gwicke: Seems reasonable, but I'm not sure how to do that yet [17:25:40] I have another idea, it just might take more time for testing. [17:25:45] New review: Diederik; "Ok." [analytics/reportcard/data] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/26479 [17:25:48] (i.e. the builds will take longer) [17:26:02] New review: Diederik; "Ok." [analytics/reportcard/data] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/26480 [17:26:56] New review: Diederik; "Please improve." [analytics/reportcard/data] (master); V: 1 C: -1; - https://gerrit.wikimedia.org/r/26475 [17:27:24] New review: Diederik; "Ok." [analytics/reportcard/data] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/26476 [17:27:40] No, maybe not. OK, plugin it is! [17:30:14] * marktraceur re-learns Eclipse [17:33:39] where is this "Git+Gerrit with saper live" ? it was scheduled for today in, I think #wikimedia-dev [17:37:25] spagewmf: It's on SIP and in #git-gerrit [17:37:59] marktraceur, Yeah I see it now https://www.mediawiki.org/wiki/Git/Workshop [17:46:37] * marktraceur watches 500 java packages get installed on his laptop and is glad he doesn't need to do this very often [17:48:13] <^demon> marktraceur: Me too! I had to `apt-get install openjdk-6-jdk maven` yesterday. [17:48:16] <^demon> Took forever :( [17:48:51] Ugh [17:49:06] Wait, I think I'm just running a mvn command..... [17:49:09] Maybe not [18:06:37] gerrit-events/src/main/java/com/sonyericsson/hudson/plugins/gerrit/gerritevents [18:06:50] Java developers love to talk [18:40:22] siebrand: ping [18:40:46] siebrand: what is the name of the i18n room on freenode? [18:40:58] anyone else can feel free to respond too [18:41:25] preilly: #mediawiki-i18n [18:41:32] RoanKattouw: thanks! [19:40:00] <^demon> RoanKattouw: We never managed to dig up the missing objects for mediawiki-config, did we? [19:41:19] I don't think so no [19:41:25] I gotta run to a meeting sory [19:41:28] *sorry [19:42:05] <^demon> RoanKattouw_away: K, thnx. [19:46:36] ^demon, things went missing from mediawiki-config? [19:47:01] <^demon> When we had downtime. We've got 3 objects from a change missing. [19:47:35] <^demon> https://gerrit.wikimedia.org/r/#/c/12185 - PS 5/6/7 [19:50:16] Might someone involved with reviewing that change be able to find them? [19:50:55] <^demon> We tried, nobody had the old patches iirc. [19:51:07] Well I don't seem to have them [19:51:12] <^demon> I might just rewrite history to pretend they didn't happen. [19:51:14] <^demon> Might be easier. [19:52:37] Are you sure these are the only things that went missing? [19:53:49] <^demon> It's the only things git fsck is complaining about. [19:55:15] Have you talked to Hashar and Platonides? [19:56:16] <^demon> Yep, we did that. [20:03:31] ^demon: I asked Platonides but never heard back [20:03:47] <^demon> Ah, ok. [20:05:16] ^demon: Off the top of your head, do you think I could do the equivalent of pressing the rebase button, but automatically from Jenkins? [20:05:39] ^demon: Just a thought, not crucial [20:06:00] <^demon> `git rebase origin/BRANCH && git push origin HEAD:refs/for/BRANCH`? [20:06:11] Hm. K [20:06:28] <^demon> All gerrit does is try to trivially rebase. So if a `git rebase origin/BRANCH` failed, it shouldn't try to push it. [20:06:42] Noted! Will do, thanks [20:14:57] ^demon, I notice that there are no comments for the missing patchsets.. [20:15:24] <^demon> Unrelated. Comments are stored in the database. [20:15:33] <^demon> Just means no one left comments. [20:18:55] Should it not say "Uploaded patch set " for each missing one though? [20:19:07] Or are those not really stored as comments and are based on the patchset info? [21:11:30] ori-l: Have a bit to chat about diff classification? [21:13:10] New patchset: Milimetric; "=test 3" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26478 [21:13:10] New patchset: Milimetric; "test 4" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26479 [21:13:10] New patchset: Milimetric; "test 2" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26476 [21:13:10] New patchset: Milimetric; "test 1 of git flow gerrit interaction - fixed" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26475 [21:16:43] New review: Diederik; "Ok." [analytics/reportcard/data] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/26475 [21:16:43] Change merged: Diederik; [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26475 [21:21:55] marktraceur: not at the moment, 4pm ok? [21:22:19] ori-l: Anytime! IRC, or would you prefer a meeting? [21:24:29] marktraceur: can you swing by my desk on 6th? [21:26:19] I can [21:26:24] ori-l: See you then [21:29:02] Woa dabadaba doo http://codepen.io/joshnh/pen/ohbHl [21:29:23] 1 element. CSS3 "construction.GIF" animation [21:29:38] (not me, don't ask why) [21:39:05] New patchset: Hashar; "Update CodeSniffer git repository" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/26545 [21:39:05] New patchset: Hashar; "sh script to apply a Gerrit changeset" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/26546 [21:39:29] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/26546 [21:39:30] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/26545 [21:46:48] marktraceur: thanks [21:56:41] Change abandoned: Milimetric; "abandoning git flow / gerrit test" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26477 [21:56:59] Change abandoned: Milimetric; "abandoning git flow / gerrit test" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26480 [22:06:19] jQuery in the wild: http://i.imgur.com/eAlkZ.jpg [22:06:22] lol [22:07:57] Krinkle: Even rarer, because it's not written with $ :) [22:08:24] yeah, I guess $ might give too much linking with money for layman [22:08:31] Ha! [22:25:27] New patchset: Milimetric; "undoing gerrit test changes" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/26551 [22:47:30] New patchset: Stefan.petrea; "Added switch for dump directory to collector" [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/26554 [22:50:10] ori-l: Will I be interrupting if I come up nowish? [22:50:39] marktraceur: no-ish :) [22:50:40] come by [22:50:43] Hmmmkay [22:51:03] New review: Diederik; "Some minor issues." [analytics/webstatscollector] (master); V: 1 C: -1; - https://gerrit.wikimedia.org/r/26554 [23:49:36] ori-l: http://parsoid.wmflabs.org/dump/Cond%C3%A9_Nast_Publications [23:49:48] could that be an issue with urlencoding? [23:54:04] gwicke: probably -- does that happen with other articles with titles containing code points above the ascii range? [23:54:30] this is the first I tried, let me look for other ones.. [23:54:51] yeah, "Élan" doesn't work either [23:55:15] same for http://parsoid.wmflabs.org/dump/Enrique_Pe%C3%B1a_Nieto [23:56:15] there's a lesson here about premature optimization. i should've just used an xml parser. ok, i'll figure out where the encoding gets screwed up. [23:58:25] sorry about that :(