[00:07:34] New review: Bsitu; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74547 [00:12:14] Change merged: Mwalker; [wikimedia/fundraising/crm] (master) - https://gerrit.wikimedia.org/r/74557 [00:12:34] New review: Mattflaschen; "Thanks! Confirmed locally." [mediawiki/extensions/GettingStarted] (master) C: 2; - https://gerrit.wikimedia.org/r/74559 [00:12:35] Change merged: Mwalker; [wikimedia/fundraising/crm] (master) - https://gerrit.wikimedia.org/r/74558 [00:12:36] Change merged: jenkins-bot; [mediawiki/extensions/GettingStarted] (master) - https://gerrit.wikimedia.org/r/74559 [00:15:20] awight: Yo. [00:15:41] New patchset: Cmcmahon; "Several refactorings, see comment" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/74564 [00:17:27] New patchset: Cmcmahon; "Several refactorings, see comment" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/74564 [00:17:50] Elsie: Hi! glad to be of service [00:18:24] MatmaRex: Oh, well it already got merged. [00:19:23] awight: :-) I run reports on Gerrit and you've currently (and for the past few months) held the record for the most open changesets. [00:19:41] awight: I'm wondering if we have a plan for how to reduce this figure. [00:19:43] legoktm: worst-case, l10n-bot will overwrite these changes on next update, then it'll overwrite them again after content on twn is updated [00:19:44] hehe thx! I was just in the process of addressing that... [00:19:58] Elsie: One issue is that I cannot abandon patchsets on closed repos [00:19:58] Oh, then I have good timing. ;-) [00:20:06] closed meaning, read-only [00:20:15] MatmaRex: Heh. I guess that works. [00:20:17] We have closed repos already? [00:20:22] New review: Cmcmahon; "* @browser.text makes test flaky, use a page element instead" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/74564 [00:20:54] Elsie: Yeah, I suppose that only accounts for three of my pet patchsets... [00:21:02] awight: Well, let me know if you need any help. In the worst case, you can just leave a comment saying "trying to abandon but can't" and Chad H. or Christian can likely help. [00:21:16] You've got another 48 without that issue, then. ;-) [00:21:18] Change merged: jenkins-bot; [mediawiki/extensions/Vector] (master) - https://gerrit.wikimedia.org/r/74315 [00:21:43] I think it's mostly fundraising-related code. If we need more reviewers in that area, that would be good to know. [00:21:46] Elsie: I'm interested in making the system work better for what I'm doing. A handful are meant to be demonstrative, not sure how I should present that sort of thing. [00:22:24] awight: It depends if you mean demonstrative in the sense that you only want to demonstrate or if you're interested in pursuing them further. [00:22:36] Abandoning the changes doesn't make them disappear, of course. [00:22:55] Elsie: well... I'm putting them out there to determine whether there is enough interest to pursue what I'm doing [00:23:22] Abandoning them is a decent workaround, but then people might get the idea that I've given up ;) [00:23:25] awight: An example changeset? [00:23:36] mmm... https://gerrit.wikimedia.org/r/24142 [00:24:12] Elsie: another example, https://gerrit.wikimedia.org/r/29527 [00:24:37] Right... code from 1.19.2 that hasn't been touched since September 2012 is ... unlikely to merged. [00:25:12] to be * [00:25:17] Verbs overrated. [00:25:29] Elsie: exactly, but it isn't quite "abandon" material, since it would be a simple enough port to update to HEAD. It's just a question of whether anyone else is interested. [00:25:58] Respectfully, I think we have to consider what Gerrit's purpose is. [00:25:59] I don't want to open the specific can of worms, but in general I am totally open to suggestions for a better way to present this sort of material... [00:26:16] * Elsie nods. [00:27:18] Elsie: I suppose I could create a User: subpage of Crackpot Ideas, and tarball stuff there, but I like the accessibility and readability of gerrit [00:27:21] awight: "Abandon" just means "don't show as a commit people should worry about merging" - it doesn't delete it. [00:27:36] James_F: well, I've -1 or -2 in various cases [00:27:40] awight: We abandon code in VisualEditor that we end up rescuing every now and then. [00:27:54] awight: Or stealing for use in other cases, more normally. [00:28:10] generally, I actually intend to pursue these things, but only once they appear on my coworker's plates [00:28:24] So yes, abandonment would work just fine for me ;) [00:30:46] Change merged: jenkins-bot; [mediawiki/extensions/Cite] (master) - https://gerrit.wikimedia.org/r/73096 [00:36:54] New review: Krinkle; "I'm not sure what those stats are meant to mean, but for what its worth, the latter does not include..." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74079 [00:37:59] New review: Krinkle; "So depending on whether one looks at it as a grepability or 'current usage' stat, it may not be repr..." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74079 [00:39:04] awight: Sorry, I got stuck on a call. I certainly don't want to discourage you or have you throw away work. I was just hoping to get you out of the #1 open changesets spot. :-) [00:39:14] awight: If there's anything I can do to help, please let me know. [00:40:16] Elsie: hehe, I'll reclaim my rankings soon enough ;) Thank you for encouraging me to clean up... The abandonment thing actually works well, cos it still allows people to review and post comments. [00:47:30] New patchset: Alex Monk; "mediawiki.notification: Return Notification object from notify()" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74407 [01:05:43] New review: Alex Monk; "This causes issues for LQT and likely other extensions if the user does not have viewmywatchlist:" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/67875 [01:21:04] New review: Kaldari; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74547 [01:23:45] New review: CSteipp; "code looks good. I'll test and merge if everything passes." [mediawiki/extensions/OAuth] (master) C: 1; - https://gerrit.wikimedia.org/r/74423 [01:27:22] New patchset: Alex Monk; "Fix breakage from Ie33446a2" [mediawiki/extensions/LiquidThreads] (master) - https://gerrit.wikimedia.org/r/74569 [01:27:38] New patchset: Hoo man; "Suppress notices while deleting output buffers" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74570 [01:32:58] New review: Kaldari; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74547 [01:35:39] Change merged: jenkins-bot; [mediawiki/extensions/LiquidThreads] (master) - https://gerrit.wikimedia.org/r/74569 [01:37:57] New review: Alex Monk; "(fixed in Icb81016d)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/67875 [01:42:32] New review: Hoo man; "Mostly style comments, didn't test." [mediawiki/core] (master) C: -1; - https://gerrit.wikimedia.org/r/36330 [01:49:42] New patchset: CSteipp; "Update test clients to call Special:MWOAUth" [mediawiki/extensions/OAuth] (master) - https://gerrit.wikimedia.org/r/74572 [01:53:10] New patchset: Adamw; "WIP (FR #986) prepare mailing job to apologize for May 2013 hole." [wikimedia/fundraising/crm] (master) - https://gerrit.wikimedia.org/r/74573 [01:57:35] Change merged: Aaron Schulz; [mediawiki/extensions/OAuth] (master) - https://gerrit.wikimedia.org/r/74572 [02:00:43] hoo: What part of the world do you live in? [02:01:05] Elsie: Germany, but I tend to stay up long :P [02:01:32] Fair enough. [02:07:28] but I think it's bed time now :D Good night ;) [02:07:31] New patchset: Kaldari; "More accureate function descriptions" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74575 [02:08:00] Good night, hoo. [02:08:34] New patchset: MZMcBride; "More accurate function descriptions" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74575 [02:16:37] New review: CSteipp; "merge conflict..." [mediawiki/extensions/OAuth] (master); V: 2 - https://gerrit.wikimedia.org/r/74423 [02:21:21] New review: CSteipp; "(1 comment)" [mediawiki/extensions/OAuth] (master) - https://gerrit.wikimedia.org/r/73977 [02:22:12] New patchset: Alex Monk; "(bug 16214) Create GlobalBlocking API module." [mediawiki/extensions/GlobalBlocking] (master) - https://gerrit.wikimedia.org/r/47539 [02:23:14] New patchset: Alex Monk; "(bug 16214) Create GlobalBlocking API module." [mediawiki/extensions/GlobalBlocking] (master) - https://gerrit.wikimedia.org/r/47539 [02:23:24] New patchset: Reedy; "Actually use $ignoreInvalidDB in UserRightProxy::getDB()" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74576 [02:26:48] New patchset: Reedy; "Add archive, externallinks PK" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/51675 [02:30:01] New patchset: Aaron Schulz; "Cleaned up MWOAuthDataStore object construction" [mediawiki/extensions/OAuth] (master) - https://gerrit.wikimedia.org/r/74577 [02:31:42] New patchset: Krinkle; "User: Perform ignoreHidden check before searching for name" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74578 [02:35:39] New patchset: Aaron Schulz; "Added conflict detection to some special pages" [mediawiki/extensions/OAuth] (master) - https://gerrit.wikimedia.org/r/74423 [02:36:27] New review: Krinkle; "external_user is added?" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/51675 [02:36:39] New review: Aaron Schulz; "Trivial rebase" [mediawiki/extensions/OAuth] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/74423 [02:36:39] Change merged: Aaron Schulz; [mediawiki/extensions/OAuth] (master) - https://gerrit.wikimedia.org/r/74423 [02:42:13] TimStarling: I'm curious what https://gerrit.wikimedia.org/r/#/c/70923/ would log [02:45:06] "Note: [02:45:06] It is usually not possible to match more than pcre.backtrack_limit characters in ungreedy mode. " [02:45:41] so if you tried to insert a string more than 1MB in length, generalize would fail to match it and the whole string would go into the profiling section name [02:46:40] also, using greedy regexes has the accidental benefit of collapsing large IN() lists [02:47:26] instead of IN('X','X','X'), one profile section name per IN item count, you get IN('X') [02:52:57] TimStarling: the current way seemed like a bug, though some more deliberate way to shorten it would be more complex [02:53:06] maybe that could at least have a comment [02:53:33] it was a kind of accidental feature [02:53:50] would be nice if ... IN(lots of stuff) AND (a=b) ORDER BY X would not get botched horribly [02:53:52] I only noticed what a nice benefit it was after it was committed and deployed [02:53:55] maybe just collapse down the INs [02:54:51] ok sounding like that should be a separate patch though [02:55:15] it's a difficult part of the code to patch, it's really quite hot in terms of performance [02:56:28] New patchset: Aaron Schulz; "profiler: log slow methods that worsen DB locks in transactions" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/70923 [02:56:28] maybe it should be done in C [02:58:01] we have this in report.py at the moment: http://paste.tstarling.com/p/pCLLil.html [02:58:16] this is what happens when you have IN() without quoting [02:59:09] :) [03:04:14] TimStarling: is categorylinks the only thing with this problem now? [03:05:58] waiting for it to load again... [03:10:57] something (probably CentralAuth) is causing a lot of profile spam with queries like: [03:11:05] query-m: SELECT user_id FROM `abwiki`.`user` WHERE user_name = 'X' LIMIT N [03:11:09] query-m: SELECT user_id FROM `acewiki`.`user` WHERE user_name = 'X' LIMIT N [03:14:11] categorylinks is the only one with the IN() problem [03:14:50] but probably that greedy regex is hiding all sorts of thing [03:14:51] s [03:26:22] New review: Tim Starling; "(1 comment)" [mediawiki/core] (master) C: -1; - https://gerrit.wikimedia.org/r/70923 [03:29:13] Yippie, build fixed! [03:29:13] Project browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_6 build #119: FIXED in 12 min: https://wmf.ci.cloudbees.com/job/browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_6/119/ [03:34:12] New patchset: Aaron Schulz; "profiler: log slow methods that worsen DB locks in transactions" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/70923 [03:35:55] AaronSchulz: did you miss the inline comment? [03:36:20] you commented out "if( $wgDebugProfiling ) {", apparently accidentally [03:37:07] oh, never mind [03:37:53] or uncommented out, odd, I don't remember why off hand [03:38:13] oh, just deleted [03:38:14] actually you removed commented-out code [03:38:20] right [03:38:39] https://git.wikimedia.org appears to be down. [03:39:38] it's still in DefaultSettings.php though [03:41:23] you know, this is the third issue of this type I've dealt with this week [03:41:57] Yippie, build fixed! [03:41:57] Project browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_8 build #157: FIXED in 12 min: https://wmf.ci.cloudbees.com/job/browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_8/157/ [03:43:23] it was commented out accidentally, by me, in July 2005 [03:43:56] eh, maybe accidentally [03:44:30] either $wgDebugProfiling should have been removed properly or not at all, right? [03:46:56] New patchset: Tim Starling; "profiler: log slow methods that worsen DB locks in transactions" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/70923 [03:48:08] the fourth issue, if you count me getting annoyed about wikipedians deleting historical pages [03:48:46] https://en.wikipedia.org/wiki/Wikipedia:Redirects_for_discussion/Log/2007_August_4#CNRs_to_the_category_namespace [03:49:19] nobody there wonders whether maybe it is not a good idea to delete pages with hundreds of edits [03:49:29] it's a redirect across namespaces, therefore it should be deleted [03:50:28] there tends not to be a good distinction between deleting and archiving [03:50:58] or at least for certain people [03:51:10] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/70923 [03:51:52] 590 edits actually [03:51:59] "Hundreds of edits" in a vacuum means little. It's about the value of the edits, unless $wgEditcountitis is set to true. [03:52:04] Yippie, build fixed! [03:52:05] Project browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_9 build #168: FIXED in 19 min: https://wmf.ci.cloudbees.com/job/browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_9/168/ [03:52:09] including https://en.wikipedia.org/w/index.php?title=Special:Undelete&target=Wikipedia%3AWikipedians%2FAustralia×tamp=20021018015718&diff=prev [03:53:43] I was reminding myself about this guy: https://en.wikipedia.org/w/index.php?title=Special:Undelete&target=Wikipedia%3AWikipedians%2FAustralia×tamp=20030319035258&diff=prev [03:53:51] who has been talking to Erik this week [03:55:25] the discussion at the bottom of that page is pre-phase3 [03:56:48] https://en.wikipedia.org/w/index.php?title=Special:Undelete&target=Wikipedia%3AWikipedians%2FAustralia×tamp=20011026013236 [03:57:43] I should have taken Mav's advice and implemented a blank page = red link feature [03:58:05] There's still an open bug about that, I think. [03:58:08] Pure wiki deletion. [03:58:12] he was very keen, very early about stopping the deletions [03:59:23] this is the page, for the benefit of non-sysops: https://nostalgia.wikipedia.org/wiki/Wikipedians/Australia [03:59:33] of course you can't see anything after December 2001 there [04:02:30] https://meta.wikimedia.org/wiki/Pure_wiki_deletion_system_(proposal) [04:03:29] http://meatballwiki.org/wiki/PlayingDead [04:05:10] Of course, while Wikimedians dearly cherish page histories, many pieces of wiki software had no issue deleting all but the most recent revisions. [04:06:03] meatball deletes old revisions [04:06:14] but despite that, they are better at preserving old discussion than wikipedia [04:06:53] How would you know if they're better? ;-) [04:07:56] yeah, you know I once tried to write there about something that had happened in the past [04:08:02] and to compare it against what was happening now [04:08:20] and I tried to do that by stupidly putting a timestamp on my comment [04:08:29] that's Not Allowed [04:08:57] Hah, a comment from Aryeh in 2008: "Something like HotCat deserves to be in core, IMO." [04:09:02] but you can date the comments to some extent by who writes them [04:10:14] http://meatballwiki.org/wiki/TimeStamp [04:11:53] And http://meatballwiki.org/wiki/TemporalContext [04:12:02] MeatballWiki is so trippy. I love it. [04:17:54] New patchset: Legoktm; "If a user does not have permission to view a log type, throw a PermissionsError" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74580 [04:18:01] Elsie: 4 years later ^ :D [04:19:05] legoktm: :-) [04:19:21] New review: Yury Katkov; "Yaron, how is the review process of this MARVELOUS feature going?" [mediawiki/extensions/SemanticForms] (master) - https://gerrit.wikimedia.org/r/57170 [04:19:45] "If your statements have no temporal content, don't put a date in at all." [04:19:55] * Aaron|home waits for Postmodern critique [04:20:22] I thought the better line was "If your statements can only make sense today (e.g. the news qua the news), it's probably best to not write them." [05:04:26] Project browsertests-en.wikipedia.org-linux-chrome build #195: FAILURE in 3 min 48 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-linux-chrome/195/ [05:11:52] New patchset: Santhosh; "Naming convention fix." [mediawiki/extensions/UniversalLanguageSelector] (master) - https://gerrit.wikimedia.org/r/74582 [05:11:56] New patchset: Santhosh; "Update the previous UI and content font whenever font value changes" [mediawiki/extensions/UniversalLanguageSelector] (master) - https://gerrit.wikimedia.org/r/74583 [05:16:50] Yippie, build fixed! [05:16:51] Project browsertests-en.wikipedia.org-windows-internet_explorer_6 build #155: FIXED in 5 min 2 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-windows-internet_explorer_6/155/ [05:16:59] Yippie, build fixed! [05:16:59] Project browsertests-en.wikipedia.org-windows-internet_explorer_7 build #155: FIXED in 4 min 33 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-windows-internet_explorer_7/155/ [05:21:50] Yippie, build fixed! [05:21:51] Project browsertests-en.wikipedia.org-windows-internet_explorer_8 build #159: FIXED in 4 min 59 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-windows-internet_explorer_8/159/ [05:23:02] Yippie, build fixed! [05:23:02] Project browsertests-en.wikipedia.org-windows-internet_explorer_9 build #147: FIXED in 6 min 2 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-windows-internet_explorer_9/147/ [05:38:30] New review: Mattflaschen; "You should test db code like this by running update.php . See https://www.mediawiki.org/wiki/Manual..." [mediawiki/extensions/Annotator] (master) C: -1; - https://gerrit.wikimedia.org/r/72361 [05:41:50] Project browsertests-test2.wikipedia.org-linux-firefox build #422: FAILURE in 18 min: https://wmf.ci.cloudbees.com/job/browsertests-test2.wikipedia.org-linux-firefox/422/ [06:02:04] New patchset: Santhosh; "Better blacklisting of browsers using $.client" [mediawiki/extensions/UniversalLanguageSelector] (master) - https://gerrit.wikimedia.org/r/74584 [06:14:32] New patchset: Santhosh; "Remove unused test html file" [mediawiki/extensions/UniversalLanguageSelector] (master) - https://gerrit.wikimedia.org/r/74587 [08:43:39] ohai Reedy [09:10:57] * YuviPanda pokes Reedy some more [09:11:06] help with testing that patch on testwiki? [09:33:56] YuviPanda: I should be able to. Just a few more things I need to catch up on first [09:34:03] Reedy: super. ty [09:35:02] New patchset: Netbrain; "Added support for ctrl+a in the export code window." [mediawiki/extensions/Maps] (master) - https://gerrit.wikimedia.org/r/74594 [09:36:38] New patchset: Netbrain; "Added support for ctrl+a in the export code window." [mediawiki/extensions/Maps] (master) - https://gerrit.wikimedia.org/r/74594 [09:47:53] New patchset: Reedy; "Force re-cache of Canonical Namespaces" [mediawiki/core] (wmf/1.22wmf11) - https://gerrit.wikimedia.org/r/74598 [09:58:50] New patchset: Pastakhov; "Refactoring, compiler is used instead of the interpreter (version 1.0.0)" [mediawiki/extensions/Foxway] (master) - https://gerrit.wikimedia.org/r/72073 [10:04:02] New review: Pastakhov; "Add increment operators." [mediawiki/extensions/Foxway] (master) - https://gerrit.wikimedia.org/r/72073 [10:06:26] New review: Reedy; "Merging for testing in production" [mediawiki/core] (wmf/1.22wmf11) C: 2; - https://gerrit.wikimedia.org/r/74598 [10:09:45] Change merged: jenkins-bot; [mediawiki/core] (wmf/1.22wmf11) - https://gerrit.wikimedia.org/r/74598 [10:09:56] YuviPanda: ^ [10:10:06] Reedy: can you run the maint script on testwiki again? [10:10:17] hello [10:10:18] Yeah, need to push the change first [10:10:22] alright [10:11:09] hello hashar! :) ^demon has asked me to rewrite gerrit-wm (and I am), and am thinking of relaying jenkins-bot changes to IRC only when they have an approval on them, and discarding others. [10:11:13] does that sound ok? [10:11:29] Done... [10:11:29] so V+1, -1 will come through, others won't [10:12:32] YuviPanda: bring that on wikitech-l :-] [10:12:34] Reedy: did you run the script as well? [10:12:49] hashar: ah, and get bikeshedded to death. I'll think about that then :) [10:12:51] YuviPanda: it has the potential of some evil bike shedding. I myself would only mention new patchset on IRC. [10:13:19] Yes [10:13:38] YuviPanda: gerrit-wm is a python hook isn't it ? [10:14:38] Reedy: hmm, that doesn't seem to help, since https://test.wikipedia.org/wiki/Campaign:wlm-test still redirects to https://test.wikipedia.org/wiki/Wlm-test [10:14:43] YuviPanda: there is a very lame integration test under files/gerrit/hook also [10:15:07] YuviPanda: and if you know java, maybe that could be handled as a Gerrit plugin *evil laugh* [10:15:17] hashar: I don't think it is a python hook, I was told it is something rather lame-ish, involving file systems and ircecho [10:15:27] hashar: I do know Java but I'm not touching anything GWT :) [10:15:46] YuviPanda: you don't need GWT :-] but do need java hehe [10:15:49] https://test.wikipedia.org/w/api.php?action=query&meta=siteinfo&siprop=namespaces|namespacealiases [10:15:58] Campaign [10:15:58] Campaign talk [10:16:04] hmm [10:16:08] why the redirecting then [10:16:22] Reedy: and what did the migration script output? [10:16:32] reedy@tin:/a/common$ mwscript extensions/UploadWizard/maintenance/migrateCampaigns.php --wiki=testwiki [10:16:32] Migrated wlm-test [10:16:32] Migrated wlm-nl [10:16:32] Migrated test [10:16:32] Migrated foobar [10:16:33] Migrated triage-test [10:16:35] Migrated wlm-pl [10:16:38] Migrated wlm-it [10:16:40] Migrated wlm-us [10:16:42] reedy@tin:/a/common$ [10:16:51] hmm, weird since I don't see any of that in the RecentChanges [10:17:12] It doesn't appear in the drop down on Special:AllPages [10:17:22] YuviPanda: all repos have some hooks which are in operations/puppet.git under files/gerrit/hooks . The hookhelper.py let yuoo decide in which file to write the message. Then we have ircecho reading those files and relaying the content on IRC. Note that ori has written yet another bot: tcpircbot which is in a puppet module. [10:17:50] Reedy: yeah, but https://test.wikipedia.org/w/index.php?title=Campaign:wlm-pl works, and thea ppropriate hooks are called [10:18:02] Reedy: can you disable ShortUrl from testwiki and see if this goes away? [10:19:28] hashar: hmm, will take a look. Idea is to make it more compact, less spammy, and have more info. Perhaps even use colors [10:20:03] YuviPanda: colors are evil :-] Don't waste your time on that hehe [10:20:10] yeah, probably not. [10:20:43] YuviPanda: what would be nice is to get rid of ircecho in favor of tcpircbot [10:20:54] YuviPanda: and make the hooks state of the art python :-] [10:20:59] hashar: ^demon wanted me to get rid of hooks themselves. [10:21:02] ahhh [10:21:10] so this is on stream events. [10:21:16] +30 [10:21:28] we could even build that inside Zuul hehe [10:21:32] with a layer of reliability on top, to protect against lost messages due to freenode being bad [10:21:46] hashar: oh noes :P [10:21:47] (its python, has all the logic about stream/events/jenkins/projects already :-) [10:22:07] this has a layer of redis for added reliability. no losing messages [10:22:11] YuviPanda: https://test.wikipedia.org/wiki/Campaign:wlm-pl [10:22:12] then one could do something like: [10:22:19] Yup, looks to be short url.. [10:22:29] Or weird caching [10:22:33] !zuul approve 12345 This change is fine [10:22:51] !zuul retrigger 4567 rerunning tests to see what is happening [10:23:06] yeah, this works too https://test.wikipedia.org/wiki/Special:UploadWizard?campaign=wlm-pl [10:23:30] could we be missing purge of short urls ? :( [10:23:49] hashar: well, this is affecting namespace registration, so doubt it [10:24:23] hashar: I already have the redis infrastructure running fine on toollabs, so I'm going to use that. we can move this to production at some point if needed. [10:25:39] YuviPanda: so you stream events in red is ? :D [10:26:05] Reedy: I guess we can leave ShortUrl disabled for now, but that's not a good Idea, I guess - since it is deployed elsewhere [10:26:32] hashar: yes. streamevents into redis, and 'clients' can register to read from the redis queue. so if your client crashes, you don't lose messages you haven't popped from the queue [10:27:22] YuviPanda: that sounds similar to the EventLogging system. [10:27:26] hashar: :D [10:27:30] YuviPanda: I am wondering whether you could get the events submitted there [10:27:38] hashar: indeed, except that is 0mq and not redis [10:27:39] and write a zmq subscriber to relay the irc notification [10:27:51] this way we get all events at the same place :-] [10:27:58] 0mq doesn't give us reliability [10:28:05] I guess that would ease maintenance/debugging since more people would know about EventLogging infrastructure [10:28:06] it's just... faster [10:28:07] YuviPanda: Well, only testwiki and test2wiki are the UW/SU overlap wikis [10:28:22] Reedy: true, but this has already bitten 3 extensions now [10:28:27] FR, EL and TW [10:28:36] err [10:28:36] UW [10:28:37] not TW [10:29:25] hashar: EL infrastructure is built to be fast, and it is okay if it drops messages now and then [10:29:25] YuviPanda: zmq can buffer messages while the client is busy reconnecting :-D [10:29:46] hashar: true, but Redis also keeps them when both the server or the client die [10:31:06] Reedy: let's leave ShortUrl disabled for now so people can test UW, and figure out the core problem later (anomie can help again, I guess). Can you leave comments on the bug report? [10:31:39] YuviPanda: I got your point, but really consider reusing something we already know before implementing yet another infrastructure :-] [10:33:09] hashar: they are different things, really. this is a 'hub' of sorts with redis at the center. [10:33:48] hashar: we can also use this for the IRC RecentChanges Feed at... some point [10:34:15] YuviPanda: another advantage of having Gerrit stream events in EventLogging is that we could have the data in mongodb/redis/whatever and reuse any dashboard / alarm system already existing [10:34:24] IRC ++ [10:34:41] the irc notification are a big huge tech debt nobody wants to maintain anymore [10:34:42] hashar: we don't really have any functioning dashboard systems that plug directly into EL :) [10:34:56] we will! :) [10:35:11] hashar: nobody in analytics is actively working on it and it isn't on any roadmap either :P [10:35:20] hashar: and the only used EL backend now is to mysql [10:35:26] which doesn't make sense for Gerrit [10:35:34] New review: Hashar; "Thank you Chad!" [integration/docroot] (master) C: 2; - https://gerrit.wikimedia.org/r/74553 [10:43:20] New patchset: Siebrand; "Have no-repeat follow url for background images" [mediawiki/extensions/UniversalLanguageSelector] (master) - https://gerrit.wikimedia.org/r/74602 [10:45:30] bahh [10:47:48] New review: Hashar; "again" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/74553 [10:48:17] gerrit is broken again :( [10:48:59] Gerrit? [10:49:02] Git? [10:50:09] New review: Hashar; ":(" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/74553 [10:51:26] Change merged: Hashar; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/74553 [10:51:40] na I am not sure what went wrong [10:51:42] anyway change is in [10:52:55] Reedy: on another topic mediawiki/extensions.git is no more updating a whole bunch of extensions :( [10:54:22] :( [10:54:49] maybe I should push them manually hehe [10:54:58] YuviPanda: where is your github/gerrit sync bot ? :) [10:55:06] hashar: where is it running? [10:55:07] hashar: toollabs [10:55:12] na source :-) [10:55:13] hashar: want me to enable it for some repo? :D [10:55:15] hashar: ah [10:55:31] hashar: github. github.com/yuvipanda/SuchABot [10:55:37] needs a little bit more work, though. [10:57:24] New patchset: Reedy; "Update a few extensions to master" [mediawiki/extensions] (master) - https://gerrit.wikimedia.org/r/74605 [10:57:25] New patchset: Reedy; "Add HotCat" [mediawiki/extensions] (master) - https://gerrit.wikimedia.org/r/74606 [10:58:02] Change merged: Reedy; [mediawiki/extensions] (master) - https://gerrit.wikimedia.org/r/74605 [10:58:35] Change merged: Reedy; [mediawiki/extensions] (master) - https://gerrit.wikimedia.org/r/74606 [10:59:19] hashar: I've 7 repos showing as dirty [11:00:36] YuviPanda: I am correct that SuchABot relay pull requests back to Gerrit ? [11:00:41] hashar: yes [11:00:43] Reedy: ahh I got some more dirty ones hmm [11:01:00] hashar: it also relays comments from Gerrit back to GitHub. Will eventually relay comments from GitHub to Gerrit too [11:01:36] YuviPanda: that is crazy :-] Thank you! [11:01:54] hashar: :D I've been using it for 99% of my patches for the last month or so. [11:02:13] hashar: it needs to be manually enabled for repos though. Currently enabled in ~15. Let me know if you want me to add more [11:02:28] YuviPanda: just wanted to let OpenStack folks know about your bot :) [11:02:35] fatal: Needed a single revision [11:02:36] Unable to find current revision in submodule path 'DataTypes' [11:02:37] bah [11:02:48] hashar: ah :) Let me know if you want me to turn up on some channel [11:05:59] Reedy: ah I might just be choking on DataTypes :D [11:06:30] hashar: is there any case where jenkins-bot will add V+2 but not immediately merge? [11:09:23] YuviPanda: any project not having a gate-and-submit pipeline [11:10:27] I am heading out for lunch [11:10:43] hashar: okay, will poke when you are back [11:31:06] New patchset: Matmarex; "(bug 43282) Add per-namespace restrictions and rights to Special:ListGroupRights" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/40096 [11:53:49] Change merged: jenkins-bot; [mediawiki/extensions/UniversalLanguageSelector] (master) - https://gerrit.wikimedia.org/r/74602 [12:12:47] re [12:13:31] hashar: the new bot is running and relaying things to ##legoktm-bots-chatter now [12:13:39] hop by if you want a look :) [12:31:55] New review: Manybubbles; "Oh my empty isn't what I expected! I'm not sure how I missed that when reading documentation but wh..." [mediawiki/extensions/CirrusSearch] (master) C: 2; - https://gerrit.wikimedia.org/r/74560 [12:31:58] Change merged: jenkins-bot; [mediawiki/extensions/CirrusSearch] (master) - https://gerrit.wikimedia.org/r/74560 [12:36:05] manybubbles: rule of thumb : do not use php.empty() :D [12:36:35] you usually want: count( $array ) === 0 [12:37:18] although count( false ) yield int(1) [12:44:03] New review: TheDJ; "In terms of consistency, this shows it I guess:" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/54484 [12:44:19] php seems a lot like javascript - useful but has all kinds of wacky nasty stuff you should just pretend doeesn't exist [12:47:08] ughhhh. i'd say it's the other way around ;) [12:47:13] New patchset: Matthias Mullie; "Split the code to generate a proper URL/link from setTitleLink()" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/74617 [12:49:22] New patchset: Matthias Mullie; "Send Echo notification upon submitting/moderating feedback" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/74166 [12:59:27] New review: TheDJ; "Shouldn't we return mw.message objects instead of the .text() values ? Seems cleaner if this is part..." [mediawiki/core] (master) C: -1; - https://gerrit.wikimedia.org/r/36391 [13:02:24] New patchset: Fantasticfears; "add a anonymous function in appendNamespaceAliases for pretty output" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74618 [13:30:34] New review: Matthias Mullie; "(1 comment)" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/72979 [13:41:07] New patchset: Hashar; "experimental breadcrumb display" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/68343 [13:42:06] New review: Hashar; "Made it a bit safer by expanding $_GET['path'] to a real path based on __DIR__ and making sure the r..." [integration/docroot] (master) - https://gerrit.wikimedia.org/r/68343 [13:48:31] New review: Anomie; "Apparently there's a second cache of namespace names in Language. Sigh. Fortunately the cache of lan..." [mediawiki/core] (master) C: -1; - https://gerrit.wikimedia.org/r/74523 [14:01:03] New patchset: Raimond Spekking; "Consistency tweaks for translatewiki.net" [mediawiki/extensions/BetaFeatures] (master) - https://gerrit.wikimedia.org/r/74621 [14:22:30] New review: Umherirrender; "Would be nice to get response to avoid javascript errors in production, which breaks ULS and maybe s..." [mediawiki/extensions/UniversalLanguageSelector] (master) - https://gerrit.wikimedia.org/r/73448 [14:25:21] New patchset: Anomie; "Hooks to tie OAuth in with User rights and the API" [mediawiki/extensions/OAuth] (master) - https://gerrit.wikimedia.org/r/73977 [14:26:05] New review: Anomie; "(1 comment)" [mediawiki/extensions/OAuth] (master) - https://gerrit.wikimedia.org/r/73977 [14:36:35] New patchset: Hashar; "move footer stylesheet to its own file" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/74624 [14:36:35] New patchset: Hashar; "global top navbar" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/74625 [14:36:35] New patchset: Hashar; "slightly smaller WMF logo" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/74626 [14:37:28] Change merged: Hashar; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/74624 [14:38:10] Change merged: Hashar; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/74625 [14:38:39] Change merged: jenkins-bot; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/74626 [14:39:07] yumm navbar on top of https://integration.wikimedia.org [15:03:10] New review: Yaron Koren; "Hi guys," [mediawiki/extensions/SemanticForms] (master) - https://gerrit.wikimedia.org/r/57170 [15:23:29] New patchset: Rachel99; "check wikisource text to ensure source editor is opened" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/74631 [15:30:18] New review: Cmcmahon; "A small suggestion, I think you should use a word other than "wikisource" so as to avoid any confusi..." [qa/browsertests] (master) C: -1; - https://gerrit.wikimedia.org/r/74631 [15:30:38] New patchset: Rjain; "Added the search class" [mediawiki/extensions/Annotator] (master) - https://gerrit.wikimedia.org/r/72361 [15:33:55] New patchset: Anomie; "API: Allow extensions to specify "useless" query pages" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/73034 [15:34:01] New patchset: Anomie; "API: Allow extensions to specify "useless" query pages" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/73034 [15:34:27] New patchset: Anomie; "API: Allow extensions to specify "useless" query pages" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/73034 [15:35:39] New review: Anomie; "PS2: Rebase" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/73034 [15:41:32] New patchset: Rjain; "Added the search class" [mediawiki/extensions/Annotator] (master) - https://gerrit.wikimedia.org/r/72361 [15:56:08] New patchset: Cmcmahon; "Maintenance update for References test" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/74640 [15:58:09] New patchset: Hashar; "basic CI dashboard, list build status of important jobs" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/74641 [16:00:41] Change merged: jenkins-bot; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/74641 [16:04:26] Change merged: Zfilipin; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/74640 [16:10:24] New patchset: Alex Monk; "Replace remaining sajax use" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/65641 [16:10:47] New patchset: CSteipp; "Prevent empty secrets" [mediawiki/extensions/OAuth] (master) - https://gerrit.wikimedia.org/r/74643 [16:14:50] New review: Alex Monk; "(This also sucks of course, but significantly less than that entire module.)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/65641 [16:23:30] New patchset: Rjain; "Added the destroy class" [mediawiki/extensions/Annotator] (master) - https://gerrit.wikimedia.org/r/74150 [16:33:24] New patchset: Rjain; "Added the destroy class" [mediawiki/extensions/Annotator] (master) - https://gerrit.wikimedia.org/r/74150 [16:38:31] New patchset: Alex Monk; "Replace remaining sajax use" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/65641 [16:42:22] New patchset: Alex Monk; "Return Notification object from mw.notification.notify" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74407 [16:44:47] New review: Alex Monk; "Accidentally made changes based on an old version of the commit, ignore this patch set..." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74407 [16:48:21] New patchset: Alex Monk; "mediawiki.notification: Return Notification object from notify()" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/74407 [16:48:47] New review: Krinkle; "See bug comment." [mediawiki/core] (master) C: -1; - https://gerrit.wikimedia.org/r/74570 [16:50:34] New patchset: Rjain; "Added the destroy class" [mediawiki/extensions/Annotator] (master) - https://gerrit.wikimedia.org/r/74150 [16:53:32] New patchset: Rachel99; "Change the name wikisource to source code so its less confusing" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/74647 [16:56:40]