[00:01:41] New review: Yuvipanda; "Cats can get annoying when you try to sleep" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [00:02:07] * PandaWithACat heads to sleep [00:02:16] Good luck! [00:14:00] Yippie, build fixed! [00:14:01] Project browsertests-en.wikipedia.beta.wmflabs.org-windows-internet_explorer_7 build #59: FIXED in 17 min: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.beta.wmflabs.org-windows-internet_explorer_7/59/ [00:54:30] Project browsertests-en.wikipedia.org-windows-internet_explorer_8 build #95: FAILURE in 5 min 13 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-windows-internet_explorer_8/95/ [01:11:10] New patchset: Brian Wolff; "(Optionally) Use {{DISPLAYTITLE:...}} whenever linking to a page" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/68936 [01:15:32] ugh, maybe I should have split that up [01:25:53] Yippie, build fixed! [01:25:54] Project browsertests-test2.wikipedia.org-windows-internet_explorer_8 build #330: FIXED in 9 min 53 sec: https://wmf.ci.cloudbees.com/job/browsertests-test2.wikipedia.org-windows-internet_explorer_8/330/ [03:00:52] greg-g/anyone: When do image scalars get wmf7? are they in the all other wikipedias group? [03:15:00] New patchset: Brian Wolff; "(Optionally) Use {{DISPLAYTITLE:...}} whenever linking to a page" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/68936 [03:15:05] bawolff: i don't think the imagescalers have mw installs? [03:15:31] New review: Brian Wolff; "PS2: Clarify comment and fix $id thing" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/68936 [03:16:22] p858snake|l: They have to since mediawiki controls which commands to run to make the thumbnail [03:17:40] Trying to figure out if https://upload.wikimedia.org/wikipedia/commons/thumb/4/45/langde-Gerrit_patchset_25838_test.svg/langde-345px-Gerrit_patchset_25838_test.svg.png is broken, or if they're just not updated yet [03:18:25] * bawolff guesses just not updated [03:20:23] Although thumb.php has a bunch of weird code in it, that doesn't seem to listen to the file handler [03:20:35] but I'm not sure if WM uses thumb.php [03:49:36] New review: Brian Wolff; "btw, I think this might not work with thumb.php, which might cause it to not work on Wikimedia's set..." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/25838 [04:35:13] New patchset: Ori.livneh; "Add API module for retrieving JSON Schema" [mediawiki/extensions/EventLogging] (master) - https://gerrit.wikimedia.org/r/68617 [05:36:55] New patchset: Aaron Schulz; "Removed redundant doc comments." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/68939 [06:01:59] New patchset: Aaron Schulz; "filebackend: throw exceptions during file iteration." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/68940 [07:15:15] Reedy: Hi!. [07:17:19] New review: Mattflaschen; "Upstream fixed the XSS bug by just removing SVG rendering." [mediawiki/extensions/UIFeedback] (master) - https://gerrit.wikimedia.org/r/50004 [07:23:59] New patchset: Rjain; "changed the way namespaces are checked" [mediawiki/extensions/Annotator] (master) - https://gerrit.wikimedia.org/r/68857 [07:36:12] New patchset: Ori.livneh; "Add API module for retrieving JSON Schema" [mediawiki/extensions/EventLogging] (master) - https://gerrit.wikimedia.org/r/68617 [07:36:12] New patchset: Ori.livneh; "Support inter-page JSON references" [mediawiki/extensions/EventLogging] (master) - https://gerrit.wikimedia.org/r/68941 [08:22:08] New review: Mattflaschen; "Code looks fine (much easier to read). I'm not sure about allowing it on talk pages for a couple re..." [mediawiki/extensions/Annotator] (master) - https://gerrit.wikimedia.org/r/68857 [08:24:20] New patchset: Nischayn22; "Adds signature to talk page messages" [mediawiki/extensions/PageTriage] (master) - https://gerrit.wikimedia.org/r/68942 [08:29:57] qchris: so I also sortof implemented some form of reliable message delivery for gerrit stream-events with Redis :) [08:30:07] deployed on tools labs now, should start the consumer shortly [08:30:20] YuviPanda: I read that yesterday. [08:30:37] I am very interesting in seeing how it works and taking a look [08:30:41] :) so I just have 1 super simple process that needs to be reliable [08:31:08] qchris: https://github.com/yuvipanda/SuchABot/blob/master/suchabot/stream-receiver.py [08:31:20] But I'll have to polish our gerrit<->bugzilla plugin a bit, so we get customizable notifications, can distinguish between "montions bug" and "fixed bug" and what not. [08:31:23] qchris: multiple clients can register with a secret and will get things added to a redis list which they can dequue [08:31:28] * qchris takes a look [08:31:34] i'm writing the client now [08:32:25] qchris: the 'secret' stuff is needed because this runs on a sortof-untrusted tools-labs environment. [08:34:07] YuviPanda: Gee! That looks nice and simple. I like that :-D [08:34:19] qchris: the client is even more simple :) [08:34:39] *gg* [08:34:39] qchris: I'm going to limit the queue size to like a couple of thousands, so we don't fill up memory badly. still reliable [08:34:58] i'm just going to have it close the GitHub PR when patch is merged or abandoned [08:35:04] commenting should come easily after that [08:35:46] New review: Nikerabbit; "(1 comment)" [qa/browsertests] (master) C: -1; - https://gerrit.wikimedia.org/r/68934 [08:36:23] YuviPanda: still up? [08:36:25] I guess you mean commenting on gerrit and then bringing the comment over to github? [08:36:33] Or the other direction as well? [08:36:35] Nikerabbit: yup [08:36:52] qchris: yes, commenting on gerrit and having it put up on github [08:37:05] qchris: other direction is well handled via github webhooks, I just need to write a gerrit api wrapper [08:37:43] Yes, sure. [08:38:03] But having aseparate bot for every task is getting us lots of software to maintain :-) [08:38:13] But you're of course right. [08:38:23] qchris: this is just two, one to do GitHub -> Gerrit and one to do Gerrit -> GitHub [08:38:29] qchris: the server can be used by anyone, of course [08:38:43] qchris: this also opens it up to more people who do not want to deal with DI or GWT [08:39:12] YuviPanda: gerrit wants to rip out GWT :-( So that problem will be gone sooner or later. [08:39:20] qchris: and replace it with? [08:39:38] qchris: I'm also told that there is a 'GitHub plugin' 'coming next week' for the last 2 months or so :P [08:39:44] YuviPanda: They do not know yet. I'll guess they take anything with a shiny new name. [08:39:49] heh [08:40:15] YuviPanda: Yes, that's true. I wrote several emails to the guy who is responsible for that. [08:40:34] YuviPanda: He showed me (confidential :-( ) videos of his plugin at work. [08:41:06] is 'why was it confidential' also confidential? [08:41:11] qchris: Apparently it covers quite everything github offers. Up to syncing accounts, teams, ... [08:41:25] so it isn't going to be Open SOurce? [08:41:50] YuviPanda: I do not know the exact reason. I suppose they wanted to sell it in a different market half a year ago. [08:41:52] or is it 'we will decide if we want to open source this or if we can make money off this' things? [08:42:34] YuviPanda: First they said that they might open source it. But the company now said that they'll open source it. [08:42:47] hmm [08:43:04] it's a bit 'chilling' because then if it were good it might make all my work go to waste :( [08:43:05] YuviPanda: But I am still waiting to see the code. [08:43:26] but then when I started working on the bot the plugin was 'going to be open sourced soon' for a few months already [08:43:31] YuviPanda: Your stuff works right now :-D [08:43:38] heh, indeed. [08:44:07] eventually it'd be nice to build some sort of system that figures out who are the most likely people to review a change based on past reviews and auto-assign to them [08:44:19] that way I can also practice a little bit of AI/ML [08:44:26] :-) [08:45:32] Wasn't there some "assign reviewer" thing before for wm's gerrit? [08:46:25] qchris: there is a bot [08:46:28] that depends on a page on mw.org [08:47:08] qchris: https://www.mediawiki.org/wiki/Git/Reviewers [08:47:16] YuviPanda: Thanks :-) [08:49:29] New review: Mattflaschen; "(1 comment)" [mediawiki/extensions/Annotator] (master) C: -1; - https://gerrit.wikimedia.org/r/68857 [08:49:29] New patchset: Mwjames; "Replace is_array/is_string with a more performant approach" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/68944 [08:57:01] New patchset: Rjain; "changed the way namespaces are checked" [mediawiki/extensions/Annotator] (master) - https://gerrit.wikimedia.org/r/68857 [08:57:36] New review: Yuvipanda; "Testing the client script for Redis" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [08:58:42] New review: Yuvipanda; "Another test. Python's famous 'backtrace oriented programming' :)" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [08:59:31] New review: Yuvipanda; "More backtraces!" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:00:55] New review: Yuvipanda; "Cat" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:01:32] qchris: https://dpaste.de/0OYMA/ is client code [09:01:37] works and has very low latency [09:01:51] (I put in a test comment and it is there, printed on my local machine before I could tab back to my terminal) [09:02:22] Nice! [09:03:03] Gonna have to setup Redis to play around with that :-) [09:03:22] qchris: yeah, it's pretty neat [09:03:37] qchris: or you could use tools labs / anything-on-labs and subscribe to the pre-existing redis instance :) [09:04:17] Mhmm. But I guess it'll have to wait for tomorrow. [09:04:25] I'll have to leave for now. [09:04:32] Enjoy the Sunday! [09:04:37] qchris: good night :) [09:04:38] * qchris waves [09:24:00] Change abandoned: Yuvipanda; "Testing" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:25:33] Change restored: Yuvipanda; "be" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:26:32] Change abandoned: Yuvipanda; "Testing abandonment" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:26:55] Change restored: Yuvipanda; "b" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:27:02] Change abandoned: Yuvipanda; "Testing abandonment?" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:32:59] Change restored: Yuvipanda; "Beer, anyone?" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:33:02] Change abandoned: Yuvipanda; "No beer" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:33:43] Change restored: Yuvipanda; "(no reason)" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:33:51] New review: Yuvipanda; "Python doesn't autoreload code." [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:34:36] Change abandoned: Yuvipanda; "I can't differentiate review and Abandon, apparently" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:35:34] Change restored: Yuvipanda; "(no reason)" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:35:41] Change abandoned: Yuvipanda; "No reason" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [09:48:44] New patchset: Mwjames; "Preparation to support SMW\JsonParameterFormatter" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/68945 [09:53:15] New review: Nikerabbit; "(1 comment)" [mediawiki/extensions/UniversalLanguageSelector] (master) - https://gerrit.wikimedia.org/r/68862 [10:03:52] Change merged: Nikerabbit; [mediawiki/extensions/InviteSignup] (master) - https://gerrit.wikimedia.org/r/67636 [10:08:08] Change merged: jenkins-bot; [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/68945 [10:14:19] Change abandoned: Mwjames; "??????????????????????????????????????" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/61769 [11:08:19] Change restored: Yuvipanda; "(no reason)" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [11:08:26] Change abandoned: Yuvipanda; "Testing" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [11:29:09] excuse me, is VisualEditor enabled for anonymous users on mw.org now?! [11:30:02] has been for awhile [11:30:46] well, it is not working [11:31:36] Nikerabbit: It's VE, not surprising. It has a pile of major bugs apparently [11:36:31] ther eis no bugzilla component for visual editor? [11:37:19] there should be... [11:37:29] its under extensions iirc [11:37:41] I don't see any [11:37:56] just filing under core then [11:39:49] Nikerabbit: its a root product [12:09:33] New patchset: Hiong3-eng5; "SQLite Compatibility" [mediawiki/extensions/WikiLexicalData] (master) - https://gerrit.wikimedia.org/r/68946 [12:14:08] Change restored: Yuvipanda; "Test" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [12:14:52] New patchset: Matmarex; "jquery.makeCollapsible: cleanup" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/55466 [12:32:21] New review: Yuvipanda; "Testing comments" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [12:33:46] Change abandoned: Yuvipanda; "Testing abandonment" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [12:34:15] Change restored: Yuvipanda; "Testing restore" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [12:35:05] Change abandoned: Yuvipanda; "Abandoning and stuff" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [12:35:13] Change restored: Yuvipanda; "Restoring the change" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [12:36:09] Change abandoned: Yuvipanda; "(no reason)" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [12:36:19] Change restored: Yuvipanda; "Testing Restore" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [12:37:00] New review: Yuvipanda; "Testing comment added plus CR" [qa/browsertests] (master) C: -2; - https://gerrit.wikimedia.org/r/65598 [12:38:30] w00t [12:41:53] Change merged: Yuvipanda; [apps/android/commons] (master) - https://gerrit.wikimedia.org/r/68545 [12:44:28] New review: Kipcool; "some comments. I'll send you my modifications per e-mail." [mediawiki/extensions/WikiLexicalData] (master) - https://gerrit.wikimedia.org/r/68946 [12:44:47] New review: Yuvipanda; "(1 comment)" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [12:50:13] New patchset: Bakert; "Mark required fields under "Release Rights"" [mediawiki/extensions/UploadWizard] (master) - https://gerrit.wikimedia.org/r/68948 [12:59:07] New patchset: Matmarex; "jquery.makeCollapsible: cleanup" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/55466 [13:00:34] New patchset: Matmarex; "jquery.makeCollapsible: cleanup" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/55466 [13:04:30] New review: Matmarex; "Rechecked, tested with Krinkle's suite, works perfectly now." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/55466 [13:10:55] New patchset: Matmarex; "jquery.makeCollapsible: fix jQuery memory leak" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/68949 [14:12:48] New patchset: Mwjames; "SMW\JsonParameterFormatter to be used in #set and #subobject" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/68950 [14:24:34] New review: TheDJ; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/68601 [14:26:07] New review: TheDJ; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/68601 [14:31:28] New review: Matmarex; "See inline. I hope this gets merged soon, the prefs do need cleanup :)" [mediawiki/core] (master) C: -1; - https://gerrit.wikimedia.org/r/65478 [15:08:39] New patchset: Matmarex; "Refactor watchlist token handling" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/64565 [15:08:53] New review: Matmarex; "Addressed Daniel's comments." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/64565 [15:20:39] New patchset: Amire80; "Add a test to check the appearance of the IME menu" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/68934 [16:22:59] New review: Yuvipanda; "Testing more?" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [16:29:46] New review: Yuvipanda; "ACtually testing?" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/65598 [16:34:22] New patchset: Brian Wolff; "Fix handling of height parameter in TMH." [mediawiki/extensions/TimedMediaHandler] (master) - https://gerrit.wikimedia.org/r/68953 [17:17:34] New patchset: Brian Wolff; "Make error in transcode one big link to the error details." [mediawiki/extensions/TimedMediaHandler] (master) - https://gerrit.wikimedia.org/r/68814 [17:24:26] made this mrconfig but I realize some "projects" seems to overlap http://paste.debian.net/10735/ [17:25:04] We still have USERINFO in gerrit? [17:25:10] that seems kind of pointless [17:25:56] bawolff: https://git.wikimedia.org/summary/USERINFO [17:26:24] I don't know either why it's not used any more or why it was used before [17:26:53] lol, that the last commit was svn:eol-style native, in true svn fashion [17:27:15] Historically it was used in the svn days so we had somewhere to link real names and emails to commit names [17:27:23] but that's less of an issue with git [17:27:37] ah I see [17:27:42] It was used to generate https://svn.wikimedia.org/users.php [17:27:53] except when people changes names and email all the time [17:28:40] bawolff: you know a quick way to fix https://wikitech.wikimedia.org/w/index.php?title=Special:LinkSearch&limit=500&offset=0&target=https%3A%2F%2Fgerrit.wikimedia.org%2Fr%2Fgitweb [17:28:51] they need to repoint to gitblit [17:29:13] Convince ^demon to make better rewrite rules for gitweb [17:29:22] heh [17:29:51] otherwise nothing other then the obvious of make a bot, or convince people to use a template to link to git so it could easily be changed [17:30:08] But I think the rewrite rules were actually planned, so I'd wait to see if that materializes first [18:19:44] New review: Parent5446; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/65478 [18:28:09] New review: TheDJ; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/64565 [18:42:34] New review: Brian Wolff; "After this change I get a "Please use an HTML5 browser" error on firefox 3.5, where before this chan..." [mediawiki/extensions/TimedMediaHandler] (master) C: -1; - https://gerrit.wikimedia.org/r/64932 [18:47:49] New patchset: Hoo man; "ResourceLoader: Serve stylesheets with "Access-Control-Allow-Origin: *"" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/60200 [18:48:06] thedj: ^ Do you want to add Release notes to that? I'm about to merge it [18:51:13] New review: Brian Wolff; "Looks fine from a code standpoint. Not giving a +2 since I'm not really qualified to judge if increa..." [mediawiki/extensions/TimedMediaHandler] (master) C: 1; - https://gerrit.wikimedia.org/r/63418 [18:57:18] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/60200 [19:06:01] New review: Parent5446; "(1 comment)" [mediawiki/core] (master) C: -1; - https://gerrit.wikimedia.org/r/64565 [19:20:35] hoo: i guess i'm too late for that :D [19:20:52] Somebody else stole my merge :D [19:21:11] besides, if it takes so much time to explain to devs, i'm not sure it's useful to explain in release notes :D [19:21:53] Maybe... our users are stupid, don't tell 'em :D [19:21:56] hey, hoo's doingreviews? [19:22:05] MatmaRex: No longer [19:22:10] boo. [19:22:29] hoo|away: The user is always wrong :P [19:34:46] New review: Parent5446; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/64565 [19:49:50] New review: Brian Wolff; "(1 comment)" [mediawiki/extensions/TimedMediaHandler] (master) - https://gerrit.wikimedia.org/r/68642 [20:11:12] Yippie, build fixed! [20:11:13] Project browsertests-commons.wikimedia.beta.wmflabs.org-linux-chrome build #124: FIXED in 9 min 59 sec: https://wmf.ci.cloudbees.com/job/browsertests-commons.wikimedia.beta.wmflabs.org-linux-chrome/124/ [20:12:31] New review: TheDJ; "ext.math.mathjax.enabler.js is supposed to contain configuration that is basically mediawiki specifi..." [mediawiki/extensions/Math] (master) - https://gerrit.wikimedia.org/r/64523 [20:16:39] New patchset: Matmarex; "Refactor watchlist token handling" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/64565 [20:17:40] New review: Matmarex; "Redone using FormSpecialPage per Tyler's suggestion." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/64565 [20:18:04] New review: Brian Wolff; "(1 comment)" [mediawiki/extensions/TimedMediaHandler] (master) - https://gerrit.wikimedia.org/r/68029 [20:19:24] Change abandoned: TheDJ; "Someone else please pick this up, if the need ever arises." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/60440 [20:19:32] where is wmf-selenium-bot defined? [20:20:09] AzaToth: ask zeljko :P [20:20:17] or chris [20:20:21] heh [20:20:59] it's not wikimedia/bots at least [20:21:55] Change abandoned: TheDJ; "I'm not pushing for this change myself, and there doesn't seem much interest in it. Whoever wants to..." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/27043 [20:23:10] thedj: the problem was that you quoted Verdy in the commit summary. Everyone's eyes glazed over... [20:23:42] nah. [20:23:55] it's just one of those things that not many ppl want to invest time into. [20:24:29] Probably a bit of "I don't want to be responsible for some tiny change in the layout of some obscure usecase on wikipedia" mixed into that too [20:24:55] exactly [20:25:12] bawolff: that did cause some layout changes [20:25:23] if you get en.wp on your ass then you are the bad guy [20:25:23] or maybe someone else's change which did the same did [20:26:10] but that's something which should bedone carefully if at all [20:26:21] anyway, some of that stuff i really don't care too much about [20:26:36] the new mantra is "put it in gerrit" [20:26:53] well my mantra is "unabondon it from gerrit" :D [20:27:06] :) [20:27:38] I sometimes worry that good patches get lost due to how old patches usually don't get abandoned [20:28:12] If you look at oldest patches in gerrit, its all -1 stuff that no one cares about [20:29:14] New review: TheDJ; "This should not be merged in this state, but i still care about it in the long term." [mediawiki/extensions/Math] (master) C: -2; - https://gerrit.wikimedia.org/r/4060 [20:32:17] Project browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_8 build #93: FAILURE in 11 min: https://wmf.ci.cloudbees.com/job/browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_8/93/ [20:32:38] New review: TheDJ; "I asked Nageh for help on this" [mediawiki/extensions/Math] (master) - https://gerrit.wikimedia.org/r/61253 [20:35:39] Change abandoned: TheDJ; "Mark forwarded this from my bugzilla patch submission, but it is demonstrably unsafe. I'm therefore ..." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/44379 [20:35:58] New patchset: Matmarex; "jquery.checkboxShiftClick: don't toggle if disabled or readonly" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/68961 [20:37:25] New patchset: Matmarex; "jquery.checkboxShiftClick: don't toggle if disabled or readonly" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/68961 [20:41:14] Change merged: TheDJ; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/64571 [20:42:34] New review: Alex Monk; "(1 comment)" [mediawiki/extensions/PageTriage] (master) - https://gerrit.wikimedia.org/r/68942 [20:43:00] a change of mine got merged. [20:43:06] *during weekend*! :o [20:43:55] zOMG! [20:44:31] wow, you have 2 pages of pending changes... [20:45:19] New review: TheDJ; "I'm not a fan of this change. I'm wondering why we set tabIndex on those collapsibles to begin with...." [mediawiki/extensions/Vector] (master) - https://gerrit.wikimedia.org/r/65115 [20:45:26] bawolff: haha [20:45:39] bawolff: a bunch of them has issues, though [20:45:44] two or three ~ [20:46:08] bawolff: yeah matma has been busy :D [20:47:57] New patchset: Matmarex; "Reorder SpecialRecentChanges::webOutput" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/65325 [20:53:29] New patchset: Matmarex; "AJAXify watchlist editor" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63274 [20:56:07] New review: Matmarex; "Bada-bump." [mediawiki/extensions/timeline] (master) - https://gerrit.wikimedia.org/r/47336 [20:57:03] New patchset: Jeroen De Dauw; "Fix dependency inclusion path" [mediawiki/extensions/WikibaseQueryEngine] (master) - https://gerrit.wikimedia.org/r/69007 [20:57:15] Change merged: Jeroen De Dauw; [mediawiki/extensions/WikibaseQueryEngine] (master) - https://gerrit.wikimedia.org/r/69007 [20:59:29] New patchset: Hiong3-eng5; "SQLite Compatibility" [mediawiki/extensions/WikiLexicalData] (master) - https://gerrit.wikimedia.org/r/68946 [21:01:42] New patchset: Jeroen De Dauw; "Fix dependency inclusion path" [mediawiki/extensions/WikibaseQueryEngine] (master) - https://gerrit.wikimedia.org/r/69008 [21:03:39] Change merged: Jeroen De Dauw; [mediawiki/extensions/WikibaseQueryEngine] (master) - https://gerrit.wikimedia.org/r/69008 [21:04:19] New review: Alex Monk; "(1 comment)" [mediawiki/extensions/PageTriage] (master) C: -1; - https://gerrit.wikimedia.org/r/68942 [21:07:40] New patchset: Jeroen De Dauw; "Fix the path on Database component inclusion" [mediawiki/extensions/WikibaseQueryEngine] (master) - https://gerrit.wikimedia.org/r/69009 [21:08:00] Change merged: Jeroen De Dauw; [mediawiki/extensions/WikibaseQueryEngine] (master) - https://gerrit.wikimedia.org/r/69009 [21:08:35] New patchset: Matmarex; "Don't use a for the TOC" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/39792 [21:08:46] Project browsertests-en.wikipedia.beta.wmflabs.org-windows-internet_explorer_6 build #61: FAILURE in 12 min: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.beta.wmflabs.org-windows-internet_explorer_6/61/ [21:08:53] New review: Matmarex; "Manually rebased." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/39792 [21:17:51] MatmaRex: the collapse code... [21:18:18] perhaps we should add new messages for the two states of the templates and hidden categories being collapsed. [21:18:45] it's bad to have buttons that don't really tell you what they do. [21:27:05] thedj: there are list headers, imo they're okay this way [21:27:18] thedj: the collapsible navigation sidebar doesn;t use different messages for this [21:27:23] New patchset: Nikerabbit; "Remove non-existing class from autoloader" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/69011 [21:27:38] thedj: also, this is harder than it looks [21:27:51] the templates... message has three variants [21:28:11] (and it really should have four if this was done right) [21:28:25] hiddencats has one, but similarly should have two [21:28:30] now double that, for show and hide [21:28:54] you'd also have to pull in jquerymsg for {{plural: parser [21:29:02] nasty stuff, basically [21:29:20] New patchset: Nikerabbit; "Remove non-existing class from autoloader" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/69012 [21:30:04] and very little gain, imo [21:30:48] New patchset: Nikerabbit; "Remove non-existing class from autoloader" [mediawiki/extensions/SemanticForms] (master) - https://gerrit.wikimedia.org/r/69013 [21:33:04] New patchset: Nikerabbit; "Remove non-existing class from autoloader" [mediawiki/extensions/SemanticMaps] (master) - https://gerrit.wikimedia.org/r/69014 [21:37:42] MatmaRex: yeah, but the sidebar is a terrible example [21:38:16] MatmaRex: perhaps we should offload that into a separate patch [21:39:26] or just append '[collapse]' '[expand]' to the end [21:40:48] still it fails my accessibility criteria for blind users this way, /me not like [21:45:10] New patchset: Parent5446; "Combine text and ID update fields" [mediawiki/extensions/UserMerge] (master) - https://gerrit.wikimedia.org/r/69017 [21:49:37] New review: TheDJ; "Still fails accessibility because the button text does not represent the action of the button making..." [mediawiki/core] (master) C: -1; - https://gerrit.wikimedia.org/r/54089 [21:52:54] New patchset: Parent5446; "Updated and added default configuration" [mediawiki/extensions/AWS] (master) - https://gerrit.wikimedia.org/r/69018 [21:53:23] Change merged: Parent5446; [mediawiki/extensions/AWS] (master) - https://gerrit.wikimedia.org/r/69018 [21:55:09] New review: Ori.livneh; "The value in having an object-oriented abstraction underlie your interface is that it allows you to ..." [mediawiki/extensions/OpenStackManager] (master) - https://gerrit.wikimedia.org/r/65421 [21:55:32] Yippie, build fixed! [21:55:33] Project browsertests-en.wikipedia.org-windows-internet_explorer_8 build #96: FIXED in 5 min 35 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-windows-internet_explorer_8/96/ [22:13:52] ori-l: around? [22:14:19] YuviPanda: partially :) what's up? [22:14:44] ori-l: no, just wanted to bounce some ideas about memcached + security/confidentiality in tools-labs/shared-hosting-environment [22:15:02] (also, your Redis idea was totally kickass, the code is now up and runnning - comments / actions on gerrit are replicated to GitHub now :) ) [22:15:03] shoot [22:15:11] oh, that's awesome! [22:15:29] ori-l: yeah, so adding github -> gerrit is a matter of writing a gerrit api wrapper. should be done next weekend, I hope [22:15:47] ori-l: anyway, memcached - so tools-labs has multiple tools running independently, sharing a single memcached server [22:16:10] ori-l: assuming everyone uses a long-enough+random-enough prefix [22:16:22] ori-l: would disabling STAT and FLUSH_ALL be enough to keep people from snooping on other's keys? [22:16:35] from my limited understanding of memcached, the answer to me seems to be 'yes' [22:16:43] (not counting DoS attacks, etc) [22:17:38] what do you mean by snooping on other's keys? [22:17:52] ori-l: imagine if one tool uses memcached to store local session keys [22:18:00] ori-l: you don't want another tool to be able to read it [22:18:05] the code for tools is open, so presumably it's not too hard to generate keys according to a pattern [22:18:15] you mentioned "random-enough", but don't forget that memcached is *fast* [22:18:20] ori-l: code is open, but not all data. [22:18:32] ori-l: well, 128chars of randomness generated out of random.org? [22:18:38] (that is what I am using for my redis prefix) [22:19:21] Project browsertests-test2.wikipedia.org-windows-internet_explorer_7 build #311: FAILURE in 10 min: https://wmf.ci.cloudbees.com/job/browsertests-test2.wikipedia.org-windows-internet_explorer_7/311/ [22:21:25] well, redis has a better solution, which is logical databases [22:21:32] see https://raw.github.com/antirez/redis/2.6/redis.conf [22:21:39] # Set the number of databases. The default database is DB 0, you can select [22:21:39] # a different one on a per-connection basis using SELECT where [22:21:39] # dbid is a number between 0 and 'databases'-1 [22:21:41] databases 16 [22:22:05] ori-l: well, yeah. but there's nothing to prevent you from just listing all the keys in someone else's database [22:22:36] no, you need something to constrain applications to a single database. i'm poking around the twemproxy docs to see if it makes that possible. [22:22:42] ah [22:22:53] yes, writing a proxy layer is one way to do it [22:23:09] the proxy layer could even simply transform keys in some application-specific way [22:23:09] Change merged: Jeroen De Dauw; [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/69012 [22:23:12] but then you'll have to proxy the entire protocol [22:23:13] ori-l: true [22:23:14] like it could actually add the prefix itself [22:23:21] Change merged: Jeroen De Dauw; [mediawiki/extensions/SemanticMaps] (master) - https://gerrit.wikimedia.org/r/69014 [22:23:45] ori-l: that also requires that we pass through the entire protocol without fucking up [22:23:54] and also be at least as fast as redis / memcached itself [22:24:01] twemproxy does support some automatic key transformations but it seems to be oriented toward sharding rather than application security [22:25:42] New patchset: Jarry1250; "Update thumb.php to handle "langXX"-type thumbnail names" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/69019 [22:26:08] ori-l: yes, that was my impression too [22:26:14] (I looked at it a while ago) [22:26:19] we could patch it... [22:27:09] ori-l: plus that's gonna be even more complex for Redis [22:27:26] twemproxy supports both [22:27:31] oh [22:27:33] i didn't know that [22:27:35] and has an expansive test suite to ensure it implements the protocol precisely [22:28:09] i'm also unsure how it can identify a connection as coming from an 'application' [22:29:14] ori-l: that would require some form of authentication that is per-application, at least [22:29:23] (this is what petan is doing, but that involves breaking the protocol) [22:29:34] New review: Jarry1250; "Indeed, bawolff (Reedy and I investigated this a fair bit after the patch appeared to not be working)." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/25838 [22:30:17] bawolff: ^ patched thumb.php (not to add handler support though) [22:30:36] jarry1250: Cool. [22:30:48] apergos confirmed that redirection isn't an issue, so the non use of a proper handler can only be a memory thing (or laziness I suppose) [22:31:11] jarry1250: But surely in order to generate the thumb we need the handler [22:31:28] I may put in a patch to change the behaviour and see what people say [22:31:34] YuviPanda: OK, stupid question. What about just having 1 memcached or redis instance per application? [22:31:43] ori-l: overhead? [22:31:52] minimal [22:31:56] * bawolff should probably try and find some docs for how wmf does thumb handling [22:31:57] can you add me as a reviewer if you do? (I don't plan to review it, I just need to keep track of it for something I'm working on) [22:31:59] bawolff: thumb.php breaks these things up into different functions though [22:32:07] ori-l: with hundreds of memcached instances? [22:32:08] apergos: Sure [22:32:12] Meant to, actually. [22:32:14] thanks [22:32:20] apergos: yes of course [22:32:38] jarry1250: But its not like we make a new handler class each time. We only make the class once [22:32:45] ori-l: memcached already supports sasl, and IIRC redis does something similar too, so one per tool works well. except for overhead and setup (how do you setup memcached for a tool, some won't need it, and setting up monitoring) [22:33:00] YuviPanda: it strikes me as a bit crazy too, but maybe that's the simplest solution. [22:33:15] possibly, yeah... [22:33:27] we can just have tools request for a memcached instance and set them up with memory limits [22:33:39] apergos: What are you on Gerrit as? [22:33:59] Ah, found you [22:34:03] ori-l: but, what *is* wrong with just having one big one + prefixes + no stat/flush? [22:34:12] they can't really dump it and do an offline attack... [22:34:28] I mean, if the prefix is say, 512bits... bruteforcing that is gonna take a while [22:35:08] ori-l: memcached development also seems dead :( [22:35:20] looking at https://github.com/memcached/memcached/pulls [22:35:30] plus I taled on #memcached, and none of the maintainers actually have time [22:35:37] (due to working at various different startups) [22:37:56] well, it works, except you can't implement it once and be done with it -- you have to hound each application developer and remind them to generate & use a prefix correctly [22:38:52] right [22:39:34] ori-l: so it is either that, and blame developers when something goes wrong, or do per-tool instances [22:39:43] hmm, alright [22:40:24] ori-l: either way, do check out the Gerrit -> GitHub features :) merging the patchset / abandoning it closes the Pull Request automatically, comments are replicated, so are merge / abandon reasons, etc [22:40:40] it can be a bit noisy though, since jenkins' comments are also replicated [22:41:01] ori-l: and I don't have support for inline comments yet, but other than that it's pretty neat [22:41:06] and looks reliable enough :) [22:43:20] YuviPanda: holy crap, that's awesome [22:43:34] ori-l: CR and V scores are replicated too [22:43:48] and it's pretty instantaneous [22:44:18] ori-l: see https://github.com/wikimedia/apps-android-commons/pull/6 (there was a bug at that time, so I had to close it myself - but it autocloses now) [22:44:39] ori-l: or see https://github.com/wikimedia/qa-browsertests/pull/2 which is my testbed pull rquest :) [22:45:04] ori-l: need to write up an email to wikitech-l tomorrow, when I'm sober :) [22:46:22] YuviPanda: i'm disappointed by your failure to leverage github emoticons [22:46:51] ori-l: blame it on the fact that I've not really been writing any software collaboratively :P [22:46:55] http://www.emoji-cheat-sheet.com/ -- there are actually ':+1:' and ':-1:' emoji [22:46:58] directly pushing everywhere gets old soon [22:47:01] hmm :D [22:47:23] +2 can be :v: [22:47:24] ori-l: should be a simple matter [22:47:33] of doing transformations from gerrit -> github [22:47:42] ori-l: the copywritng could definitely use an improvement [22:47:45] it currently is very... bland [22:48:16] -2 should be :poop:, obviously [22:48:18] ori-l: and it is only this big https://github.com/yuvipanda/SuchABot/blob/master/suchabot/github-sync.py [22:48:19] :D [22:48:37] I've Jinja2 templates there, so... [22:50:59] If i suggested that Special:Uploads be renamed Special:ThirdTimesTheCharmWhenItComesToUploadForms, that would probably be trollish... [22:51:36] bawolff: Special:MobileUploads would probably be not [22:51:48] Yeah, that was my first suggestion [22:52:08] we're apparently avoiding prefixing things with the word Mobile [22:52:24] oh [22:52:29] New patchset: Aaron Schulz; "lockmanager: QuorumLockManager subclasses can now EX/SH locks at once" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/69021 [22:52:34] I see [22:52:42] * bawolff goes reads the source code to see what it actually does, so I can make intelligent suggestions, instead of just annoying ones ;) [22:52:51] :) [22:52:55] * YuviPanda goes to sleep [22:53:02] nite folks [22:53:10] ori-l: try out the stuff :) [22:57:06] New patchset: Aaron Schulz; "Fixed mysql warning about expecting a resource." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/69022 [22:58:31] YuviPanda: night [22:58:35] YuviPanda: btw, why jinja2? [22:58:42] 'Change merged by {event.submitter.name}\nReason:\n{event.reason}'.format(event=myevent) works just as well [23:29:59] Yippie, build fixed! [23:29:59] Project browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_8 build #94: FIXED in 8 min 46 sec: https://wmf.ci.cloudbees.com/job/browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_8/94/ [23:51:04] Project browsertests-en.wikipedia.beta.wmflabs.org-linux-chrome build #133: FAILURE in 21 min: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.beta.wmflabs.org-linux-chrome/133/