[08:50:53] mors [09:00:21] mors? :) [09:16:46] = morning [10:13:47] petan|wk: morning [10:14:03] anybody here that could shed some light on shared databases? [10:14:26] Nikerabbit: depends [10:14:29] what exactly you mean [10:15:05] petan|wk: 1) how do I request access to shared database 2) how can I make it configurable which database my extension accesses for certain table [10:15:57] database where? are you talking about labs, production? or some other wiki [10:18:09] petan|wk: hypothetical for now... I'm implementing a translation memory, and its database should be shared by all wikis [10:18:18] I'm looking at CentralAuth now how it does it [10:18:24] ok [10:18:30] is it permanent memory? [10:18:38] if it's only a cache, memcached would be great [10:19:00] it's permanent, innodb... mayben even fulltext table in the future [10:19:04] you can also check global block code is simpler [10:19:16] global blocking [10:19:30] central auth is probably too complicated [10:20:16] I have no idea how it actually work, but probably you will need to create a new instance of database and connect it to shared db, no idea what you mean by "requesting access" [10:20:36] the connection will need to be done on every load [10:21:30] the rest of functions will be same as in wgDB, just you won't use the global database object, but your one [10:23:10] I think I got it [10:23:52] I just need new $wgFoo and do wfGetDB( DB_SLAVE/MASTER, 'foo', $wgFoo ); [10:24:26] exactly [10:46:06] Nikerabbit: Look into what was done for parser cache [10:46:59] It's a db server that's used to cache parser results long-term, with 1TB of space. [10:48:14] You'd have to ask ops about using that though, but I doubt using the centralauth db for that is a good idea. [11:01:57] johnduhart: I'm just looking how it is done [11:02:08] johnduhart: will probably try things out in labs if possible [13:39:46] what do you think about creating a global edit interface group for devs, so that people withou staff bit can assist on wikis with set up of various Mediawiki space issues? [13:45:09] petan|wk, there is already [13:45:33] petan|wk, https://meta.wikimedia.org/wiki/SRGP#Requests_for_global_editinterface_permissions [14:34:28] hexmode: guillom: Reedy: take a look at https://www.mediawiki.org/wiki/MediaWiki_1.19/Roadmap#Communications and see if you want to add anything [14:35:33] sumanah, btw, you know about https://www.mediawiki.org/wiki/Development_process_improvement/Communications_recommendations#Tech_announcements ; right? [14:35:47] Perhaps we could consolidate the content somewhere for future reference. [14:35:59] guillom: I knew there must be something like that somewhere but had missed it. Thanks. [14:36:44] guillom: you saw my draft, right? [14:36:46] The content, or rather its context, is a bit dated, but there's probably stuff to be salvaged. [14:37:00] I assume so, b/c you didn't ping me yesterday [14:37:16] hexmode, I saw it, but I didn't know if you were done [14:37:42] k, I do need to look at it again and maybe do some more [14:37:45] (I didn't ping you yesterday because I was out for dinner, so we didn't overlap much online :) [14:37:48] ok [14:37:56] just shoot me a quick e-mail when you think it's ready [14:39:47] sumanah: only real comment I have right now is that Saibo from commons really thought the RL changes being discussed were a show stopper and didn't think he could do more testing till the issues were addressed [14:44:54] sumanah: I should add that helderwiki contacted me last night about the (still un-resolved) issue here: https://bugzilla.wikimedia.org/show_bug.cgi?id=31511#c17 [14:46:26] hexmode: looking now. [14:52:14] Nemo_bis: that isn't really permanent one :o neither something what can be usually requested [14:52:25] there is it's only for a short time [14:57:42] hexmode: so, yes, RL is an issue [15:01:38] petan|wk, actually people do request it [15:01:47] and I think it's permanent for someone like Krinkle [15:07:47] hi hashar [15:07:58] hi :) [15:43:27] sumanah, might seem obvious but shouldn't wikitech-l on that list? [15:44:20] Also maybe a mail out to all wmf staff as a heads up too would be good [15:50:10] Reedy: good point. sorry. yes. add 'em? [15:50:24] I was assuming someone would do that but that is a bad assumption! [15:51:05] heh [15:51:05] right [15:51:34] The staff came to mind as I seem to recall us starting to do stuff then other staff members being quite confused [15:53:09] done [15:53:23] Reedy: cheers [16:22:20] sumanah, silly question again: I guess you're not using a centralnotice as with 1.17 because it shouldn't be as traumatic? [16:22:45] although now it can be filtered by user group and such [16:22:54] Nemo_bis: I think a CentralNotice at some point is worth doing, but not this early in the process [16:23:02] ok [16:23:02] Nemo_bis: please do add it to the plan though [16:23:07] Nemo_bis: or to the talk page [19:14:04] Hello [21:03:14] siebrand: just waiting for Trevor to show up. Waiting in the room that still smells of paint fumes.... [21:03:38] neilk_: I denied the meeting invitation. I have another meeting. [21:03:47] oh? [21:04:07] neilk_: Did that 14 hours ago... [21:06:06] <[Haekchen]> Hello! [21:06:07] <[Haekchen]> Is there no FeaturedFeeds-Component in Bugzillas MediaWikiExtension-Category or can't I find it? [21:08:00] siebrand: sorry, somehow Trevor and I missed that. I assume 10pm UTC is too late [21:08:03] there's probably no category [21:09:00] just add the bug in [other] [21:09:28] [Haekchen], it's new [21:09:36] give me a few minutes and I'll create it [21:09:47] MaxSem, ^ Do you want to be assignee or cc? [21:10:42] what's our current default assignee policy? [21:10:49] I don't mind assignee [21:11:28] Setting it to wikibugs is fine if people don't want to be set on it, but have enough interest to be cc'd [21:11:46] neilk_: yes. I've been working for the past 14 hours. Time for a break after this. [21:12:15] siebrand: Is 5pm UTC tomorrow okay? [21:12:43] [Haekchen], created [21:13:07] neilk_: what about 06:00 UTC tomorrow? (today evening for you, 22:00PDT, IIRC)? [21:13:46] siebrand: it's not easy for Trevor [21:14:34] neilk_: I wanna stry and start the weekend at 19:00 tomorrow evening... [21:14:50] I'm completely wasted at the moment... [21:14:52] siebrand: understood. Okay let's wrap around to Monday [21:15:44] siebrand: between what hours are you generally available [21:16:22] siebrand: nm you are in a meeting already, we'll send you an email [21:16:33] neilk_: 08:00-16:00 UTC is preferred, but it aligns badly with PDT. [21:16:45] (That's midnight-8am PST) [21:17:01] <[Haekchen]> Reedy: Thanks! [21:17:13] something like that ;) [21:17:16] siebrand: yeaaaah, I'm gonna need you to come in at midnight [21:17:31] siebrand: late nights aren't easy for Trevor at all [21:18:32] So usually I stretch that to 2pm PDT, but not all nights of the week, because I already *do* work 50+ hours a week. [21:18:38] Anyway??? lemme check.. [21:18:51] siebrand: you know, considering this meeting is just about design critiques, maybe we should just write up a wiki page and send that to you [21:19:16] if you get jorm in too, he knows what we've discussed so far. [21:19:28] I need deigns, and can then discuss those. [21:19:43] siebrand: sure -- we'll do something here and submit it for your review. [21:19:58] once I have final designs, my team can build them, and we can user test them. [21:36:24] Reedy: ping [21:36:40] ohai [21:36:56] Reedy: about the indentation of: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/110040#c30070 [21:37:09] Reedy: I ran it through stylize.php and that is what I get [21:38:08] You've got 8 if statements, and a loop, making 9 levels of nesting after the functions [21:38:38] or 8 ifs and 2 loops even [21:39:17] Reedy: so? [21:40:30] holy staircase code that [21:41:49] the most indented line is 49 characters in [21:42:13] Reedy: so, what do you think I should do? [21:54:09] home sweet home.. with a fire warning thing warning with a loud beep about its nearly empty battery each 15 seconds [21:59:48] siebrand: we're talking with a couple people about doing more full designs on this feature. [21:59:57] in fact, it's actually a "test" given to a couple designers so far. [22:00:55] jorm: I know. [22:01:22] jorm: I also know that only one candidate has handed in something worth sharing :) [22:01:29] yes. [22:03:28] <[Haekchen]> jorm, siebrand: What feature are you talking about? [22:03:36] A universal language selector. [22:05:54] <[Haekchen]> jorm: for interwikis, the interface or multi-lang wikis? [22:06:06] all things of the above. [22:06:40] <[Haekchen]> thanks [22:12:58] preilly, the code feels like it's re-inventing the wheel... [22:15:44] Reedy: yeah, I understand [22:15:53] It might not be [22:15:55] bleh [22:16:03] I've unindented it a few layers [22:16:31] Reedy: how? [22:18:46] where code doesn't have an alternate path, and doesn't need to hit the code at hte bottom of the loop, invert the condition, continue [22:19:04] have the "else" without the else {} and have it after the if ( !$foo ) { continue; } [23:19:26] rmoen: You'll need both MB and MAH updated, right? [23:19:54] RoanKattouw: we haven't made any changes to it, but there might be some language updates [23:20:08] "it" = ? [23:20:10] MAH? [23:20:15] Yes, sorry [23:20:32] We haven't updated MAH [23:20:37] OK [23:20:40] I'll leave it alone then [23:20:45] MB is all reviewed, gonna prep the update now [23:21:24] RoanKattouw: Sounds good [23:26:36] Reedy: Did you do this? 1 Fatal error: Call to undefined method HTMLForm::getLanguage() in /usr/local/apache/common-local/php-1.18/includes/specials/SpecialBlockList.php on line 435 [23:26:44] *RoanKattouw looks at svn log [23:27:42] Wouldn't suprise me [23:27:54] Differences between 1.18 and 1.19 are getting more blurry [23:28:49] looks fine to me [23:28:51] When was it? [23:29:07] The error? [23:29:12] yeah [23:29:39] Jan 26 23:21:04 10.0.2.202 apache2[24816]: PHP Fatal error: Call to undefined method HTMLForm::getLanguage() in /usr/local/apache/common-local/php-1.18/includes/specials/SpecialBlockList.php on line 435 [23:29:42] I pushed both getLanguage into 1.18wmf1 context stuff (I think) to try and fix these errors [23:29:44] Seems to be rare [23:29:56] As in, one occurrence in the past 1000 log lines [23:29:59] lol [23:30:06] Although we've probably still got ocwiki OOM errors polluting that log [23:30:08] *10,000 [23:30:19] Not sure when anyone last ran scapped [23:30:34] With apaches running out of free disk space... [23:30:38] Don't do it now, I'm in there [23:30:42] i wasn't going to [23:30:45] i was just loading SAL [23:30:50] Trying for 100,000 [23:30:58] Well, actually [23:31:02] Let me just find you a backtrace [23:31:14] No one this month, apparently [23:31:18] else the log bots were MIA [23:31:30] morebots was MIA yesterday [23:31:34] or the day before [23:31:46] there's been a few issues in the last couple of weeks [23:32:09] I swear to God I'm just gonna put log rotation for fatal.log in Puppet if no one else does it [23:32:20] It's still got all errors since Oct 5, 2011 [23:32:22] lol [23:32:56] rmoen, bsitu: MB changes should now be live on test [23:33:39] 2.6G /var/log/mw/fatal.log [23:33:41] *sigh [23:34:09] http://pastebin.com/wvcp6bEU [23:34:09] RoanKattouw: thx [23:36:04] HtmlForm in 1.18wmf1 doesn't have contexty crap [23:38:00] $wgLang and be done me thinks [23:39:23] r110095 [23:40:35] RoanKattouw: it seems that we need to be sure to add $wgMoodBarConfig['feedbackDashboardUrl'] = '//en.wikipedia.org/wiki/Special:FeedbackDashboard'; in the LocalSettings.php [23:40:51] Sorry, for failing to mention that [23:42:44] Ah, OK [23:42:47] Putting that in then [23:43:10] done [23:44:44] RoanKattouw: the email copy is still the old one, I guess it would be updated once the code gets deployed to prod? [23:46:01] That goes through the job queue, right? [23:46:08] yeah [23:46:13] The job queue is processed on separate machines, I didn't update those [23:46:16] so yeah [23:46:49] (the hack for updating test only basically entails all test.wp.o requests being routed to one box; this falls apart for job queue stuff, as you guys have exposed quite nicely) [23:52:11] bsitu: Let me know when I can push to production [23:52:43] RoanKattouw: will do, thank you [23:55:24] RoanKattouw: We are good to go [23:55:52] Yay [23:56:31] :-D [23:58:35] OK, it's live [23:59:12] yay!