[00:07:30] TimStarling: I've updated https://gerrit.wikimedia.org/r/#/c/50138/ [00:08:00] I'm looking at this lua timeout issue at the moment [00:08:15] there's a lot of discussion to catch up on [00:08:27] I read it yesterday, got a bit longer today [00:08:47] hello, is gerrit down? I can't access it [00:10:04] WFM [00:11:53] ? [00:11:53] wfm? [00:13:24] works for me [00:13:27] What problem are you having? [00:19:06] hashar: Hm..why are we running test and check pipelines on changes by whitelisted users? [00:19:09] https://gerrit.wikimedia.org/r/#/c/52308/ [00:19:13] V+1, V+2 [00:19:14] V+1, V+2 [00:19:15] V+1, V+2 [00:19:16] etc. [00:21:41] Krinkle: yeah that is a side effect. https://bugzilla.wikimedia.org/show_bug.cgi?id=43391 [00:22:00] Krinkle: need to update Zuul to support pipeline overriding (like 'test' would takeover 'check' [00:22:11] or finally isolate jenkins tests in vagrant boxes :-] [00:22:24] I don't see how the latter helps? [00:22:54] but yeah, we need a test to be of higher priority and make it never run more than 1 pipeline (only the most important one) [00:23:02] which I think is not a feature but a bug fi [00:23:04] fix8 [00:23:06] becuase... [00:23:12] there can't be more than 1 score [00:23:20] one will override the other, there is no point in having more than one [00:23:23] for nobody [00:23:42] to make sure it picks the right one, we'd need to either have a priority system or an inclusion system [00:23:58] e.g. say 'test' includes 'check'. Then we'd also have to duplicate less. [00:24:16] I'd prefer the latter [00:24:35] btw, memory usage on gallium. Getting closer to its limit [00:24:42] http://ganglia.wikimedia.org/latest/?r=year&cs=&ce=&c=Miscellaneous+eqiad&h=gallium.wikimedia.org&tab=m&vn=&mc=2&z=medium&metric_group=ALLGROUPS [00:24:58] nothing out of the ordinary, looking at the past year, but an issue? [00:39:48] anomie: I'm wondering about mw.title.getContent and invalidation [00:40:21] AaronSchulz- How so? [00:40:29] maybe the templatelinks table should reflect those calls [00:41:38] AaronSchulz- Hmm. I think you're right. Which makes me wonder, does expandTemplate add a templatelinks entry? [00:41:40] otherwise it's up to the parser cache ttl [00:42:47] csteipp: is there a central repository of CVE notification ? Sam proposed to auto link CVE links in Gerrit by using https://cve.mitre.org/ as a base [00:42:55] ( https://gerrit.wikimedia.org/r/#/c/52253/1 ) [00:43:21] the mitre database is pretty much the authority for them [00:43:26] great [00:43:27] +1 [01:14:06] Is the HTML of special pages ever cached by Squid/Varnish? [01:21:52] Reedy: superm401: I am off for some climbing this evening. [01:21:59] cool [01:22:00] Reedy: superm401 so don't look for me :-] [01:26:29] hashar, have fun. :) [01:32:22] superm401: I think it's a no [01:32:23] X-Cache-Lookup MISS from cp1006.eqiad.wmnet:3128, MISS from cp1017.eqiad.wmnet:80 [01:33:07] Reedy, alright, thanks. [01:33:12] That was my impression earlier too. [01:35:15] kaldari: How do I get echo to let me clear my notifications on wikitech? [01:35:48] if you hover over any of them, you can dismiss that type of notification [01:36:11] although there's a bug currently where it isn't reloading to fill in the missing slots in the flyout [01:36:12] Confusing [01:36:17] I pressed more and that cleared it :/ [01:36:45] but if you dismiss them, that means you don't get any more [01:37:24] I don't actually see anything when hovering [01:37:25] silly thing [01:37:35] it's supposed to be more of a log, i.e. you don't really clear it like an inbox [01:37:40] some of them are not dismissable [01:38:13] dismissing = unsubscribing [01:38:16] Right [01:38:24] I just wanted to get rid of the "unread" notifications [01:38:30] hmm [01:38:43] hashar mentioned the same desire [01:38:55] I'll pass that along to the designers [01:40:49] heh [01:40:53] LQT has something similar [01:41:03] Being in red makes it look like there's something to be done [01:41:39] If it's red that means you still have notifications you haven't viewed [01:42:18] But viewing the page didn't fix it.. [01:42:40] you might have to click on 'More' a few times if you have a backlog [02:07:44] hello [02:08:34] I need a little help with Auto-Wiki-Browser modification. [02:09:12] anyone here? [03:13:16] Krinkle (or anyone), jshint claims https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/extensions/GettingStarted.git;a=blob;f=resources/ext.gettingstarted.openTask.js;h=ab8f5dc39a907165ac3095001c42345e9283ed7a;hb=4aad7407373576b1380815a7d73b37bd13c07874#l144 has "bad escapement" [sic] [03:13:23] I agree the code is a little wonky. [03:13:33] But I don't think it's escaped wrong. [03:13:36] What do you think? [03:13:45] superm401: exact error? [03:13:54] "Bad escapement" on that line. [03:14:08] I get it locally to. [03:14:12] k [03:14:26] it means something is escaped that shouldn't be or isn;t needed to be [03:14:28] probably the $ [03:15:05] though I'd understand escaping that, see if that's what triggers it [03:15:14] I see. [03:15:20] Looks like '\\$1' is identical. [03:15:34] Not sure that should be an error, though it is a bit more verbose. [03:15:56] Thanks, Krinkle, that resolves it. [03:16:00] yw [03:16:15] It's really not "bad". [03:16:19] superm401: I understand the thought behind it though, it is redundant escaping. [03:16:24] I'll submit a patch to change the message. [03:16:39] i.e. it wants you to know that it is not needed [03:16:58] I would interpret bad escaping like ending a string with a backslash or something. [03:17:05] yeah [03:17:28] ... and so I lost my clipboard data. [03:17:43] I'll get it back. [03:18:07] I love Klipper. [03:18:19] I can keep a whole day's clipboard history in my status bar. [03:23:02] On a semi-related note, the replace table in the https://github.com/jshint/jshint/commit/d8f5cf4be33d3bffa9800b568b576e5f1487e723 message is a cracker's dream. [03:23:08] The second two literally do nothing at all. [11:48:44] New patchset: Zfilipin; "Updated Ruby gems" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/52403 [12:24:07] when I do code review in Gerrit, what's the difference between the "Publish" and "Publish and Submit" buttons? [12:25:08] You should only use Publish [12:25:36] except if you are working on a repo which doesn't run test and autosubmit automatically [12:28:08] ok, so I've done it right by guessing [17:01:17] hello [17:01:44] hashar: Hi! [17:02:01] the short climb yesterday basically killed me off :-] [17:02:06] slept 9hours straight [17:02:24] That'll happen [17:17:11] New patchset: Hashar; "JsonData-testextensions now non voting" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/52424 [17:17:30] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/52424 [18:41:28] xyzram: WARN org.wikimedia.lsearch.interoperability.RMIMessengerImpl - Error getting snapshot for index {file [18:41:29] java.lang.RuntimeException: Index {file doesn't exist [18:41:30] :)] [18:41:32] I am doomed [18:43:11] This is on the search box ? [18:43:17] the search idx [18:43:27] I guess my configuration file is not correct :-] [18:44:01] deployment-searchidx01.pmtpa.wmflabs /a/search/conf/lsearch-global-2.1.conf [18:44:07] it is generated by puppet though [18:44:59] Snapshots are generated via the lucene.jobs.sh script. [18:50:23] The MediaWiki script file "/data/project/apache/common-local/php-1.21wmf6/maintenance/dumpBackup.php" does not exist. [18:50:23] lovely [18:50:38] xyzram: I think I managed to trigger the snapshots creations [18:51:00] wmf6? lol [18:53:56] yeah I guess that is from wikiversion [18:54:19] win [18:55:55] After updating an index, the new version is not made available to searchers until a snapshot is created. [18:56:27] hashar: could you make the jslint checkstyleResult a non voting failure? [18:56:46] mwalker: which repo ? [18:56:47] I can't really control all the jquery stuff it just failed :) [18:56:54] fundraisings branch of core [18:56:59] ahh yeah that branch [18:57:00] :-] [18:57:09] I know; the evil black sheep [18:57:33] so yeah that would potentially be possible :-] [18:57:51] whoo [18:57:59] also -- totally awesome that we have these new checks [18:58:09] pgehres opened a bug about having fundraising branch to skip the tests antirely [18:58:29] right -- because a lot of them fail -- it would be nice to know [18:58:42] but if they didn't vote; that would be nice [19:04:09] xyzram: so I think I got all databases imported via lucene.jobs.sh :-] [19:05:16] That's a lot of data -- how many MB might that be roughly and how long did it take ? [19:05:59] don't know, it ran in a screen overnight [19:06:17] xyzram: on search idx, the dumps are 1.1GB :-] [19:06:58] /a only has 3.3GB used [19:07:06] so I guess that is small [19:53:29] how do I add names to the cc list of a new bugzilla ticket ? The dropdown is not populated and when I enter an explicit email address, it is not accepted. [19:54:36] <^demon> comma-delimited? [19:55:02] <^demon> Should be e-mail, e-mail, e-mail, if memory serves. [19:55:43] The dropdown suggestions should be working. [19:56:52] I just created 45795; the dropdown only has rainman. [19:58:20] When I enter chorohoe@wikimedia.org explicitly and click save changes, I get an error page with "Bugzilla was unable to make any match at all for one or more of the names and/or email addresses you entered on the previous page ...." [20:00:12] Is that user registered in Bugzilla? [20:05:26] xyzram: notpeter: any idea how I could get the lucene index/prefix on a search box? I apparently got them on the indexer but they are not on the search box :( [20:07:13] ahh /a/search/lucene.jobs.sh indexer-cron [20:07:14] [20:07:17] seems to publish some stuff [20:09:09] parent5446: That's Chad, I assume he is. [20:17:04] xyzram: turns out the search indexer on beta does not have enough memory apparently :/ [20:17:30] [ 5937.919767] Out of memory: Kill process 23377 (java) score 400 or sacrifice child [20:17:31] [ 5937.920497] Killed process 23377 (java) total-vm:9332356kB, anon-rss:1619648kB, file-rss:696kB [20:17:31] :( [20:17:44] Is that easy to raise ? [20:18:14] xyzram: Ah, OK. It's weird that he doesn't show up. [20:18:43] he will be using another email on bugzilla [20:19:19] use the gmail address with name innocentkiller [20:20:16] (another trick is to put the name and hope the user you wanted appears in the list bugzilla offers you) [20:20:44] email autocompletion is the single best feature in recent bugzilla releases :) [20:22:30] If I put in Chad, several names show but none is Horohoe; same with Tim -- no Starling [20:23:17] it's not always ideal if the email address doesn't match the name very well [20:23:19] he will appear as Chad H. [20:23:27] works great if you search for 'brion' tho :) [20:23:28] anyway, I gave you his bugzilla email above [20:23:36] so just put that [20:30:25] Ok, that works; 'tstar' also offers up the right email for Tim; Platonides: thanks. [20:55:57] xyzram: do you happen to know whether org.wikimedia.lsearch.config.StartupManager could be started with debug mode ? [20:56:15] seems there are a lot of log.debug("some useful message") but I can't find out how to enable that :-) [20:58:35] lsearch.log4j has settings [20:59:07] Tim recently changed things so that that file is dynamically monitored for changes. [20:59:33] ahh [20:59:59] log4j.appender.A1.Threshold=WARN [21:00:30] Changing that from WARN to DEBUG should cause debug messages to show up. [21:01:34] yeah that works :-] [21:02:51] java.lang.RuntimeException: Index enwiki doesn't exist [21:02:52] ah [21:02:52] ; [21:02:53] ) [21:03:11] yeah the search box does not have them grmrmgmbl [21:03:37] mw-titles.tspart1 sv-titles.tspart1 sv-titles.tspart2 [21:03:42] Is the incremental updater running ? [21:03:48] ah no that is bugged [21:04:02] I though somehow the search box would grab the index from the search indexer [21:05:34] I'm not sure how that happens initially --- in the middle of something right now, I'll look at it later this afternoon; is that OK ? [21:05:46] yeah sure =} [21:08:14] I think that is the global configuration which is wrong [21:35:45] Thehelpfulone: why are you accessing my bugzilla account? [21:35:58] Danny_B, did you not see the reason in the email? [21:36:04] I sent two emails before too! [21:36:15] bug https://bugzilla.wikimedia.org/show_bug.cgi?id=43610 [21:37:48] greg-g I think you're the only person that I didn't send that email to - if you're around atm, please can you go to https://bugzilla.wikimedia.org/userprefs.cgi?tab=saved-searches and change "editbugs" to "bz_canusewhines" for Share with a Group? [21:37:56] https://bugzilla.wikimedia.org/show_bug.cgi?id=43610 explains why we're doing that [21:38:56] Thehelpfulone: my mailbox for bugs allows only bugzilla mails to be delivered, everything else is treated as spam and thrown away immediately [21:38:59] Thehelpfulone: done [21:39:10] greg-g, thanks [21:39:23] Danny_B, hmm sorry about that, I didn't know [21:39:40] i use other email addresses for the communication [21:39:49] besides you know i'm 24/7 here ;-) [21:40:02] :P [21:40:17] greg-g, out of interest why did you decide to pick editbugs? did someone recommend that? [21:40:47] hashar, ^^ same thing for you as I said to greg please [21:41:13] Thehelpfulone: don't remember, maybe it was after I had a meeting with andre__ and I might have interpreted editbugs as a group I should share with (short answer: I don't remember anymore ;) ) [21:41:22] heh okay [21:42:49] Thehelpfulone: no idea what it is :-] [21:43:05] hashar, go to https://bugzilla.wikimedia.org/userprefs.cgi?tab=saved-searches and change the shared to permission to bz_canusewhines please [21:43:29] from canconfirm that is [21:43:41] I have no idea what it is for and can't really have a look at it right now, sorry :( [21:43:53] hashar, okay I'll do it for you then :) [21:43:58] :-] [21:43:59] heh [21:44:14] "either you share, or I'll share for you!" [21:44:35] I'm not that evil, just fixing permissions for bugs they've already shared :-) [21:46:13] fixing permissions sounds like euphemism... [21:48:43] so what is the issue with editbugs? [21:49:38] we had some vandalism problems with bugzilla a while back, so now we don't give editbugs by default to everyone [21:49:45] which means all the newer bugzilla accounts don't have it, so couldn't see a number of the shared searches [21:50:27] we add editbugs to these accounts when issues come up, but it's not really scalable to add it for each and every account - and searches don't really cause much harm, so we're making them accessible for everyone [21:51:24] 2013-03-06 21:51:09,372 [main] INFO org.wikimedia.lsearch.oai.IncrementalUpdater - Resuming update of aawiki from 2001-01-01 [21:51:25] 2013-03-06 21:51:09,372 [main] INFO org.wikimedia.lsearch.oai.OAIHarvester - Reading records from http://aa.wikipedia.beta.wmflabs.org/w/index.php?title=Special:OAIRepository&verb=ListRecords&metadataPrefix=mediawiki&from=2001-01-01 [21:51:33] oh yyyeahh xyzram ^^^ I got something working :-] [21:51:57] Krenair, you around? [21:52:02] yep [21:52:30] there's one bug left that needs the sharedto: permission fixing, it's yours [21:52:38] "Unconfirmed bugs " [21:53:26] please could you change that to bz_canusewhines at https://bugzilla.wikimedia.org/userprefs.cgi?tab=saved-searches ? [21:53:30] It's shared with canconfirm. [21:53:41] I could, but why? [21:53:49] Can't confirm bugs without canconfirm. [21:54:09] does that mean they shouldn't be able to see the search? [21:54:51] Thehelpfulone: is there any description of these groups? [21:54:52] meh. I'll just set everything (even unshared) to bz_canusewhines [21:55:14] Krenair, well you don't need to do that if there are already shared bugs for it [21:55:32] Danny_B, I don't think so, basically everyone is in the "bz_canusewhines" group [21:55:39] all these groups are bugzilla defaults [21:56:28] hmm, why there are same shared searches... [21:56:50] Day old patches (mah) vs. Day-old patches (Sumana) [21:56:53] etc. [21:57:00] needs some cleanup [21:57:33] yeah, feel free to file a bug for it! [22:01:42] hashar: nice [22:01:56] xyzram: will setup OAI tonight :-] [22:02:11] Ok, how'd you fix the index problem ? [22:03:48] why %user% doesn't work in fields? [22:22:30] Thehelpfulone: any idea? [22:22:49] I don't quite understand the question? [22:26:00] you should be able to put %user% to fields of search which substitutes to the current user (thus you can share bugs like "my opened bugs" etc.), however, it does not work in advanced fields [22:29:21] Reedy: https://gerrit.wikimedia.org/r/#/c/51794/ is just a bit of refactoring [23:18:54] New patchset: Stefan.petrea; "Fixed segfault" [analytics/webstatscollector] (time_travel) - https://gerrit.wikimedia.org/r/52549 [23:20:40] New review: Diederik; "Ok." [analytics/webstatscollector] (time_travel); V: 2 C: 2; - https://gerrit.wikimedia.org/r/52549 [23:20:40] Change merged: Diederik; [analytics/webstatscollector] (time_travel) - https://gerrit.wikimedia.org/r/52549 [23:29:02] qgil_, hi [23:29:20] Krenair, hi there [23:31:10] I was reading https://www.mediawiki.org/wiki/User:Qgil/Contributors [23:31:21] Krenair, yeah, working on it right now [23:31:33] The bit about one contributor = many usernames, lack of semantic data etc. [23:32:11] I was reminded that SMW is installed on wikitechwiki [23:33:02] Krenair, first I want to define a problem and agree on the solution, and then find the tools to implement that solution [23:33:15] So maybe that would be better for such things than mediawiki.org [23:34:31] Krenair, if SMW is part of the solution then why not having it in mediawiki.org - but as said I don't want to start a debate on tools & domains before we agree on the basics of the plan :) [23:35:11] Krenair, I'm happy to see that someone is reading that page so soon. Your feedback is welcome! [23:35:46] Krenair, see also the connected http://www.mediawiki.org/wiki/User:Qgil/Contribution_path :) [23:36:09] Krenair, as you can see our pre-discussion wikitech-mediawiki is reflected there already [23:36:13] That sort of ties in to the point I was about to make - Actually many of the WMF-specific pages on mediawiki.org could fit better on wikitech (once it's been tidied up a bit), but in the past wikitech was more closed which I assume is why it didn't go there instead [23:36:58] Krenair, that is another discussion. [23:40:52] Krenair, need to go now. As said, your earliest feedback is welcome. I hope you understand my focus on what we need to develop first. Also because a contractor budget might come from this. The domain discussion is easier & cheaper. ;) [23:41:18] okay. [23:41:24] thanks [23:44:13] New review: Reedy; "PHP Notice: Undefined index: message in /home/reedy/git/mediawiki/core/make-deploy-notes2 on line 81" [mediawiki/tools/release] (master) C: -1; - https://gerrit.wikimedia.org/r/52086 [23:44:53] New review: Reedy; "array(1) {" [mediawiki/tools/release] (master) - https://gerrit.wikimedia.org/r/52086 [23:48:21] New review: Reedy; "Looks like we've got multiple duplicates, with the first of each apparently having no message." [mediawiki/tools/release] (master) - https://gerrit.wikimedia.org/r/52086 [23:49:45] New patchset: Hashar; "submodule integration/junitdiff" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/52554 [23:50:28] Krinkle: marktraceur : some shinny new submodule for jenkins :-] https://gerrit.wikimedia.org/r/#/c/52554/ [23:50:45] Yaaay [23:50:48] Krinkle: marktraceur: tools/junitdiff being added to integration/jenkins. That is a Junit diff utility for parsoid team [23:50:55] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/52554 [23:51:04] I'll try not to eff it up, thanks hashar [23:51:14] * marktraceur sets about writing the thing [23:51:40] marktraceur: whenever you import your stuff in that repo, you will have to update the gitmodule in integration/jenkins , submit for review, get it merged and then have gallium local copy to be updated. [23:51:52] marktraceur: maybe I should just add you to the integration group :-] [23:52:04] OK. [23:52:16] hashar: If you think that's merited, I wouldn't mind it [23:52:18] New patchset: Reedy; "Rewrite most of make-deploy-notes so it handles bugs in footers" [mediawiki/tools/release] (master) - https://gerrit.wikimedia.org/r/52086 [23:52:37] marktraceur: get yourself an access on gallium. Would be happy to approve it [23:52:56] junitdiff is on gallium now, albeit empty [23:52:57] Submodule path 'tools/junitdiff': checked out '51ea1cf1a0524a885b3c500c22a40da2c1b07681' [23:54:25] hashar: What's the process for that? RT ticket or something equally arcane? :) [23:54:58] marktraceur: I have no idea :-] [23:55:09] ^demon: Krinkle: I have added marktraceur to the gerrit "integration" group [23:55:13] HRmm. [23:55:29] <^demon> k. [23:55:49] * marktraceur wonders who would know the process [23:55:55] <^demon> For shell to a server? [23:56:01] <^demon> RT + puppet change. [23:56:03] Ah. [23:56:05] ah yeah RT [23:56:37] Anyone know what class is responsible for the History view? [23:56:47] * marktraceur ... doesn't think he has his RT password available right now [23:57:01] marktraceur: you can just send an email I think [23:57:04] kaldari: HistoryPage? [23:57:07] anyway, I am out for now [23:57:08] Yes, I figure that's true [23:57:15] Bye, hashar! Thanks! [23:57:48] kaldari: When you find out, maybe add it to http://www.mediawiki.org/wiki/Developers/Maintainers ? [23:57:57] (the History box is empty now :( ) [23:57:59] kaldari: er, history/actions/HistoryAction.php [23:58:19] Er, wow [23:58:32] * marktraceur cleans glasses, didn't read "what class", read "who" [23:58:53] ori-l: It doesn't look like HistoryPage exists anymore [23:59:16] HistoryBlob?