[00:46:37] kaldari, is it possible to get test.wikipedia.org updated with ori-livneh's patch? Or does that not happen until deployment time? [00:51:10] not until deployment time, but I can get them set up on ee-prototype [00:51:18] in the meantime if you want [01:02:02] kaldari: it's already 6pm, it's ok [01:03:44] StevenW: He still has a small change to make before I merge the code anyway [01:04:10] Yeah [12:17:10] Hihi. Question: are rows in the externallinks table ever deleted? [14:10:39] * werdna skulks in [14:11:47] hey chrismcmahon [14:11:59] hi werdna [14:12:04] how goes? [14:12:13] how was berlin? did you make it? [14:12:30] very well, actually. no, I didn't go to Berlin unfortunately [14:14:20] that's alright, I didn't make it either [14:14:28] particularly well? [14:16:42] werdna: we have two open reqs for QA staff and we have some great candidates who have been doing browser test automation out in the wild. One of them found the labs beta cluster without any help and managed to get himself blocked there. :-) we got him unblocked and he's done some very neat experiments and posted them to github. [14:18:30] heh! [14:26:55] chrismcmahon: heh, nice! [14:27:00] how did he get blocked? [14:27:01] wow [14:27:03] this is awkward [14:42:32] maybe he submitted too many requests? [14:54:54] Platonides: yes, he made a lot of random edits :-). now he's just making random edits to his user page. I think it's great that someone went out and used beta labs for serious work, that's exactly what it's for. [16:23:32] anybody know how to make the resource loader give me an error message when it fails to load a module? i've got some javascript that's silently failing to load in ie8. [17:13:55] Greetings, rsterbin, yoni_omniti, mtlin! I look forward to our AFT5 deployment today! RoanKattouw and I are sitting in a conference room now, he is reviewing the code as we speak. We have mtlin on Skype, happy to bring any of you guys into that voice backchannel if you like. [17:14:14] Sorry for the CR backlog everyone [17:14:18] I just got back from Europe yesterday [17:18:54] RoanKattouw: I've got a fairly large back log too, although I didn't consider it a back log since its mostly commits that were submitted during and from the hackathon [17:19:06] I should stop spreading my name like that next time :P [17:19:14] haha [17:19:23] Welcome to my life dude ;) [17:19:50] Now testing on new article feedback form option 6 on prototype: http://prototype.wikimedia.org/release-en/Golden-crowned_Sparrow?aftv5_form=6 This is one of our big new features today, which Roan is reviewing today. [17:21:07] got 24 left [17:21:25] of which 7 already have a review [17:21:31] but not closed yet [17:23:32] RoanKattouw: Well, this looks promising when accessing the bug tracker for tracker: http://geshi.org/developers/bugs/ [17:23:53] mysql_connect system denied and a bunch of php deprecation warnings [17:24:03] lol [17:24:11] for geshi* [17:24:16] Fatal error: Call to undefined function access_denied() in /var/www/virtual/geshi.org/htdocs/developers/bugs/core/authentication_api.php on line 479 [17:24:18] whahaha [17:24:41] yeah, and there is that [17:25:18] so ahm.. I see that geshi is not svn:externals anymore (since we're on git), but I didn't know we had a hardcopy in master as well (rather than just in the wmf branches) [17:25:37] is that all kosher ? [17:25:53] I don't really know [17:26:08] OTOH we have hard copies of jquery and a lot of other stuff too [17:26:15] yeah, true [17:26:28] but we tend to not change/add files to those upstream libs [17:26:32] rsterbin, yoni_omniti : We are not able to see the URL changes which Matthias just made to Gerrit and prototype (for the terms link and the help link). Do you know why that is? I refreshed my local caches, but still no dice. Could it be related to some server caching issues? (e.g. from international team). [17:26:44] https://gerrit.wikimedia.org/r/#/c/10286/ [17:27:16] Hm.. not new in that change.. when was html5.php added.. [17:28:10] https://bugzilla.wikimedia.org/show_bug.cgi?id=37044 [17:28:44] Hm.. maybe it is part of geshi [17:29:20] https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/extensions/SyntaxHighlight_GeSHi.git;a=history;f=geshi [17:29:23] alright, whatever [17:33:15] fabriceflorin: are you still seeing that issue? it looks okay to me now. [17:35:58] Yes, I am experiencing the issue with both links on Firefox, Chrome and Safari, after clearing my cache for all 3 browsers. http://prototype.wikimedia.org/release-en/Golden-crowned_Sparrow?aftv5_form=6 [17:38:34] rsterbin: all code on prototype is ok [17:38:45] yeah, everything i see here is right [17:38:51] however, i18n.php edits don't seem to be picked up [17:39:21] articlefeedbackv5-help-tooltip-linkurl & articlefeedbackv5-help-transparency-terms were edited, but still have their old value [17:39:34] do you know if these are cached in a way (and how to clear it)? [17:56:35] rsterbin, yoni_omniti : FYI, mlitn and RoanKattouw are discussing some of the recent revisions over Skype, mostly about his code. Let us know if you want to be included in that Skype call. [17:57:04] ok [17:57:08] fabriceflorin: we're also in our irc chat with mlitn [18:06:34] * chrismcmahon follows along [18:28:54] rmoen: hi; I'm around for ~30 more minutes if you need anything besides my e-mail. [18:30:28] guillom: reading, I'll let you know [18:30:35] great :) [18:35:02] rsterbin: wfWaitForSlaves() [18:35:33] mlitn, rsterbin : The CTA 3 on prototype no longer has the link to your last post. I don't know if that's related to the caching issue or not, but it should say "Thanks! Your feedback has been " (with permalink, after you click on 'Post your review'). [That is supposed to replace what I see now on prototype: 'Thanks! Your feedback has been saved.'] [18:39:41] guillom, this sounds reasonable. I can begin without having the template fully created. Though, please let me know when you have it. Enjoy your dinner :) [18:40:04] rmoen: awesome, thank you so much :) [18:40:18] guillom: np [18:41:44] rmoen: are the two of you discussing the proj status tool? [18:41:51] Krinkle: https://gerrit.wikimedia.org/r/#/c/4694 has been awaiting re-review since May 21 [18:42:17] there's at least a couple of bugs that Krinkle and I introduced that I'm guessing you hit [18:42:52] if you aren't discussing that, then nevermind :) [18:43:22] I have a new version that I worked on on the plane [18:43:42] RoanKattouw: I see over half a dozen reviewers on that commit though, none have reviewed the latest patch set yet [18:44:28] I gave an earlier (static) review, I don't have that extension installed nor used it. So I'm fine doing another static review, but I wouldn't merge it [18:44:37] robla: we were talking about a wikitech gadget for helping cleanup of some ops pages [18:44:58] oh, you've moved on! xlnt, more gadgets :) [18:45:25] robla: I still need to sync your code into my dev copy, not sure what you changed yet. [18:45:42] rmoen: Krinkle and I just decided to be bold over the weekend [18:45:57] oh, you have a separate dev copy? [18:46:31] rmoen: robla Do both of you have sysop rights on mediawiki.org [18:46:37] RoanKattouw has completed code review for AFT5 and is now ready to coach mlitn on how to deploy the merged code, following his excellent deployment guide: https://wikitech.wikimedia.org/view/How_to_deploy_code Wish us luck! [18:46:46] Krinkle: I do [18:46:51] Krinkle, I do not. [18:47:05] robla: yes i have a secret dev copy [18:47:16] if so, I'd like to move it to a gadget right away to optimize loading and cache busting (I hate to keep clearing my browser cache) [18:47:17] rmoen: mw.org user name again [18:47:25] Robmoen [18:47:36] Krinkle: lemme save the current version first [18:47:40] sure [18:47:54] hm.. I my mw.org magic wand has no dust for +sysop [18:48:37] Reedy: can you make that happen? [18:48:39] Krinkle: done [18:48:42] thx [18:48:51] (And thank you all for your help with this!) [18:48:53] I noticed Reedy added +sysop for robla [18:48:57] Krinkle: I backed out the promise-based callbacks. They busted saving [18:49:06] interesting [18:49:23] Krinkle: I've got the staff bit, so presumably I could have added myself [18:49:53] robla: I suppose. although that might have required a self +(local steward) first [18:50:15] anyway, sounds like it's a moot point now :) [18:50:16] maybe it is possible from meta as well, I don't know [18:51:01] robla: Do you see a non-disabled checkbox for administrator on https://www.mediawiki.org/wiki/Special:UserRights/Robmoen ? [18:51:02] new version saved https://www.mediawiki.org/wiki/User:Robmoen/projectStatusHelper.js [18:51:34] Krinkle: yeah, looks like I've got the ability to fiddle with all of the checkboxes [18:51:42] alright, good to know [18:51:51] robla: cool ty [18:52:17] lots of changes, and I've got a few more that I was working on this morning on the train that aren't quite done yet [18:52:43] moved [18:53:13] and done; it is now an opt-in gadget on mw.org [18:53:15] preserved history [18:56:05] rmoen: one way to see what's different now: visit https://www.mediawiki.org/wiki/OAuth and click on Project Status. it will automatically select the right project now [18:56:23] second change: replacement of status works now [18:57:01] robla: cool, thanks [18:57:13] third change: if you put in a date associated with a status that already exists, the edit dialog will populate with the text of that status [18:57:38] if you've made edits prior to changing the date, it will ask for confirmation before smashing your edits [18:58:36] robla: oh well, it looks like you've taken care of everything :) [18:58:58] the change I'm working on now is to make it so that edit links on the "Wikimedia * Engineering" pages (and other places) open up the dialog prepopulated with the right stuff [18:59:18] I may need some template work to pull that off, though [18:59:33] * Krinkle fixing bug in ProjectStatusHelper where clicking the Project Status links jumps to the top ('#'), thus undoing the current scroll position and hash [18:59:40] anyway, I've gotta run off to lunch before an interview [18:59:45] 1 line, easy :) nice. fixed [19:00:34] ty Krinkle [19:10:06] mlitn: mwscript rebuildLocalisationCache.php --wiki=testwiki --outdir=/home/wikipedia/common/php-1.20wmf4/cache/l10n [19:20:34] roblaAFK: http://amo-probos.org/post/14 [19:20:45] since it seems everyone goes away for lunch. heh [19:20:54] Krinkle: http://amo-probos.org/post/14 [19:21:00] maybe you'd be interested in that [19:21:17] I have no clue who's doing our jenkins stuff right now :) [19:22:15] Ryan_Lane: hashar and chrismcmahon are the primary folks interested in Jenkins-y things [19:22:36] this is a way of specifying which kinds of jobs are what [19:22:42] so that you can run tests in parallel [19:22:46] reading... [19:23:22] and this is why I like having the same toolkits as openstack :) [19:29:58] rsterbin, yoni_omniti, chrismcmahon: wanted to give you an update on how Roan and mlitn are coming along with the AFT5 deployment. We have moved the code over to testing, but are experiencing some issues with Option 6: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_form=6 Nothing ready to test yet, just an FYI. [19:30:13] ok [19:30:15] thanks fabriceflorin [19:31:08] You're welcome, we'll keep you posted. We thought it was related to a local caching issue, but Roan thinks it's something else. Stay tuned ... [19:36:49] fabriceflorin, mlitn, RoanKattouw: just noticed something really obvious -- the feedback page only gathers feedback for option1 [19:36:55] option 6 is excluded [19:37:14] i'm going to take a look and see how hard it would be to fix [19:37:30] OK, guys, we're ready for testing of AFT 5: Option 1 is http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_form=1 -- and Option 6 is http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_form=6 [19:38:00] Before you try these links, check that the 50% - 50% bucketing works on a fresh browser. [19:38:57] So you would only use this short URL to test that the bucketing works: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow Try it on different browsers, to see if you get 50% option 1 vs 50% option 6 [19:41:13] rsterbin: indeed; weird, didn't do that on my local setup [19:41:13] right sidebar only contains "Feature" now, not Hide/Oversight/view? [19:43:03] that is, on http://test.wikipedia.org/w/index.php?title=Special:ArticleFeedbackv5 and http://test.wikipedia.org/w/index.php?title=Special:ArticleFeedbackv5/Golden-crowned_Sparrow&ref=cta [19:43:59] RoanKattouw: remember how I said those two insert lines didn't matter? They do. [19:44:18] option 6 comments won't save without them, [19:45:08] Hi guys, could this be related to the fact that the feedback page is programmed to only list Option 1 -- not Option 6? But then why did this seem to work properly on prototype? [19:47:33] it should show both 1 & 6, but 2 lines seem to prevent that from happening [19:48:27] rsterbin: Didn't I merge that commit? [19:48:31] Ooooh /those/ [19:48:35] Wait no [19:48:40] Which lines are these again? [19:48:41] I'm confused now [19:48:48] rsterbin: would it be helpful if we put you on Skype now, so we can troubleshoot together? Just turn your Skype on and I can reconnect you. [19:48:58] mlitn, RoanKattouw, fabriceflorin -- working on a fix now [19:49:01] OK thanks [19:51:49] RoanKattouw: did you run those two inserts? [19:52:07] Which ones? [19:54:03] (I guess that means no) [19:55:14] any idea how to force the regeneration of the message cache? (in labs) [19:57:03] RoanKattouw: the two added lines toward the bottom of sql/ArticleFeedbackv5.sql [19:57:22] inserts the field names for option 6 [19:57:36] Aha, no I did not [19:57:37] Doing that now [19:58:26] Done [20:07:54] rsterbin: $wgMemc->set( $key, $value, $timeout ) [20:08:25] rsterbin: php /path/to/maintenance/eval.php then type $wgMemc->delete( wfMemcKey( 'blahblah' ) ); [20:08:41] rsterbin: (and please tell me what the key is so I can do this on prod too) [20:09:00] $key = wfMemcKey( 'articlefeedbackv5', 'getFields' ); [20:13:58] mlitn: Alright, let's deploy then. To do this you simply run 'scap' [20:14:14] After that we'll have to purge the memc key and update the filter counts, but scap takes ~15 minutes usually [20:14:19] RoanKattouw: no schema changes to be done? [20:14:24] I did those already [20:14:27] cool :) [20:14:43] running scap [20:16:00] OK, we seem to have solved the problem on testing, which was preventing Option 6 comments from appearing on the feedback page: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_form=6 [20:16:48] So I think we are good to go to production, right? [20:17:43] mlitn has started deploying to production, should have the code on en-wiki in about 15 mins. [20:41:36] rsterbin: mw.loader.using( 'modulename', function() { console.log( 'success' ); }, function() { window.failureData = arguments; console.log( 'failure' ); } ); [20:41:47] i'm looking for an example of a really crazy complex template on WP [20:41:49] any ideas? [20:42:12] TrevorParscal: With or without parser functions? [20:43:09] fabriceflorin: it's not on all servers already, but I'm already seeing everything: http://en.wikipedia.org/wiki/Golden-crowned_Sparrow [20:43:43] Hmm, the cite templates? [20:43:44] Some common ones with varying degrees of complexity: {{IPA}}, {{Infobox musical artist}}, {{Infobox album}} [20:43:53] Or those [20:44:03] Yeah, and of course the whole "stack trace" of {{cite web}} and friends [20:44:19] aye [20:44:22] which did (or still does) include {{str len}} etc. [20:45:17] OK, chrismcmahon, it would be great if you could do a bit of testing on the new feedback page and feedback form on en-wiki. Here's the feedback page: http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow [20:45:28] TrevorParscal: I assume none of these can be edited in the VisualEditor, though.. [20:45:34] TrevorParscal: is this for Parsoid performance? [20:45:48] fabriceflorin: the SHOUT warning in prod refers to edits not comments, different than the test warning, which seems more correct [20:45:54] no, it's for a VE presentation, trying to show particularly offensive use of wikitext [20:46:00] ah [20:46:33] like crazy curly brace soup and whatnot [20:46:39] qif? :D [20:46:57] And as mlitn mentioned, here's the feedback form URL - If you try it on different browsers, it should alternate between option 1 and option 6: http://en.wikipedia.org/wiki/Golden-crowned_Sparrow [20:47:16] TrevorParscal: Not terribly large but maybe complex enough to be ironic? https://en.wikipedia.org/w/index.php?title=Template:Welcome&action=edit [20:47:25] TrevorParscal: Have you ever read [[Template:qif]]? Should be required reading for all WMF engineers [20:47:27] Just "safesubst:BASEPAGENAME" is crazy enough for me [20:47:53] for a beginner at least [20:48:06] hehe [20:48:32] um, qif is marked for speedy deletion... [20:48:46] wee, deprecated as of March 28, 2012 [20:48:47] fabriceflorin: sort order on http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow seems random, much like that bug I reported last time seen only in incognito/private mode [20:50:25] "Error loading feedback" and all results disappear from http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow. not sure how I did it, will try to repro [20:50:39] TrevorParscal: haha, don't think that'll be deleted, it's kept because of its historical significance [20:51:00] Qif is only (indirectly) used on 337 pages now [20:51:29] chrismcmahon: works fine for me - the default view is showing all comments, sorted by newest first [20:52:25] yoni_omniti: you could not repro https://bugzilla.wikimedia.org/show_bug.cgi?id=37102 last time either. but apparently some can. [20:58:43] chrismcmahon: Thanks for reporting a sort order issue on http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow -- I am not able to reproduce this issue on Chrome in Incognito mode: I keep getting the same order, Most Relevant, as it should be. Are you using a different browser? [20:58:58] From rsterbin : http://prototype.wikimedia.org/release-en/Special:ArticleFeedbackv5/Golden-crowned_Sparrow?debug=true [21:00:24] fabriceflorin: I'm seeing the longer selections (oversight requested for instance) in the filter selectbox on the feedback page truncated [21:00:48] ie. the box isn't long enough [21:02:01] I seem to have made my feedback post with all the Russian characters un-findable after oversighting it, trying to repro that. [21:02:02] chrismcmahon: I guess it's time to submit some bugs for all these issues, so we can get to the bottom of this. In the meantime, are you seeing any major show-stoppers that would require us to abort this deployment? [21:02:43] chrismcmahon: I can reproduce your bug about the shouting abuse filter -- I am going to look to see if anyone has changed that filter now. [21:04:48] fabriceflorin: I can also still repro https://bugzilla.wikimedia.org/show_bug.cgi?id=37104 [21:06:01] actually that one might be worse than last week because of the persistence now working [21:08:40] fabriceflorin: is is the case that an un-oversighted entry should still also require un-hiding to be visible? [21:09:32] fabriceflorin: put another way: oversight a normal post; to return that post to normal requires both un-oversight and un-hide. [21:09:46] which seems not intuitive to me [21:14:19] wow, checking FF there are no hide/oversight etc. flags on the http://en.wikipedia.org/w/index.php?title=Special:ArticleFeedbackv5/Golden-crowned_Sparrow at all, anyone can confirm that? [21:15:12] I see one of my issues is that the 'Sort by' selectbox is shown very far to the right. [21:16:08] fabriceflorin: if action links are really not visible in FF, that would be a blocker I think [21:17:27] fabriceflorin: also seeing all indented choices in "More filters" not visible in Firefox. [21:18:17] chrismcmahon: We are aware there are issues with IE8 -- I just posted a bug for this on Bugzilla: https://bugzilla.wikimedia.org/show_bug.cgi?id=37351 [21:18:30] fabriceflorin - just wanted to remind you before i forget -- could you close out all the fixed issues in bugzilla, now that we've deployed. i'd like to open one ticket for each new bug instead of leaving open the feature requests [21:18:52] and with that... i'm off to get food :-P [21:20:44] fabriceflorin: FF seems to have issues similar to IE8 [21:21:26] fabriceflorin: seems to me that likely Chrome is the only browser in which http://en.wikipedia.org/w/index.php?title=Special:ArticleFeedbackv5/Golden-crowned_Sparrow will function correctly, with all features visible and available. [21:24:40] fabriceflorin: summary: major problems with Firefox, minor problems (display) in Chrome, minor problems (one action to oversight, two actions to return to normal) in workflow. [21:24:54] I'm not even going to look at IE7 [21:28:57] fabriceflorin: I'm going to take a short break and then write up what I've found so far, will send you that list in email. hopefully we can get fixes before Saturday [21:29:19] chrismcmahon: Thanks for the update on our current status. I am not thrilled to hear about the FF and IE issues, but don't think we should revert the deployment at this point. I would prefer to keep it up, and try to find solutions for FF and IE in coming days, so we can deploy them again on Tuesday, which is our next deployment. The feedback page is not very visible yet (a small link on the talk page), so I think we'll be OK. Does that work [21:29:20] for you? [21:29:56] Thanks, chrismcmahon, that sounds good to me. [21:42:52] fabriceflorin: OK. would be nice to have that fixed by Saturday, but if it's not possible it's not [21:45:10] fabriceflorin: btw, https://bugzilla.wikimedia.org/show_bug.cgi?id=37104 does have the keyword. is it because of the comma that it didn't get found? [21:58:53] TrevorParscal: Wow, this must have been the craziest @import in Monobook.css I've removed so far [21:58:58] https://fur.wikipedia.org/w/index.php?title=MediaWiki%3AMonobook.css&diff=139168&oldid=132826 [21:59:01] Common.css [22:00:10] since 2006 [22:13:33] rsterbin_away: So I've been investigating this ResourceLoader bug of yours but I don't think it's a bug on our end [22:13:45] rsterbin_away: Because I see 'msie 8' in your browser blocklist in both ext.articleFeedbackv5.startup and .dashboard [22:33:57] RoanKattouw: huh. now that you mention it, I think that might have been put in a few weeks ago. [22:34:19] Well we just deployed 2 weeks worth of backlog [22:37:55] RoanKattouw: I just commented out that check on prototype, and still nothing [22:38:08] in any case, dashboard.js doesn't seem to be loading, either. [22:38:21] dashboard also has that same check [22:38:32] I put some console.log and debugger stuff in those files [22:38:47] And it doesn't trigger on the initial page load (which is weird) but it does trigger with the .using() line [22:43:19] RoanKattouw: I just dumped the user agent string... it's msie 7. [22:44:07] or, well, it has "msie 7" in the user agent string [22:46:00] RoanKattouw: got it. it was in "compatibility mode" [22:46:18] prototype is working on ie8 now, at least in the filtering/sorting area