[04:17:04] anyone knows the fast way to unify ur account in all wikis? [04:17:59] pathoschild had a tool that did that but it didn't crossover to wmflabs or was never kept.. [04:20:30] comets: https://meta.wikimedia.org/wiki/User:Krinkle/Tools/Global_SUL ? [04:21:35] * comets tests [04:23:19] hmm 720 wikis, i thot we had more than 750? [04:24:07] private or fishbowl wikis maybe? [04:26:31] probably, tis a good way to make sure people don't take ur name in some other language wiki ;) [04:27:11] kinda slow though, pathos tool did it like in 20 seconds :) [04:27:51] if you have the global account, no one can take your name unless a crat is stupid and renames someone into your name [04:29:46] yeah saw this and thot its not worth taking that risk .. http://en.wikipedia.org/wiki/Wikipedia:Village_pump_%28technical%29#SUL_question [04:33:22] o.O [04:45:49] greg-g: Poked around https://bugzilla.wikimedia.org/show_bug.cgi?id=47566#c10 a bit. [04:45:53] Hope that helps. [04:49:49] Carmela: neat, thanks [04:50:28] I should chat with g.wicke about it tomorrow [04:52:57] I dug up http://lists.wikimedia.org/pipermail/wikitech-l/2010-January/046367.html as well. [04:53:13] I was wondering about the reference to the discussion [04:53:45] thanks for all the digging [04:54:16] No problem. [04:55:07] alright, early morning tomorrow. g'night. [04:55:56] Night. [08:44:15] uh, no jayvdb [13:17:55] is CodeEditor an opt-in feature? I don't see it enabled on jawiki. [13:19:48] you may have disabled it [13:20:35] hm, I don't see it at ja:special:mypage/common.css either [13:22:09] some of the site js could be intefering, but I don't know which one . [13:23:26] I no longer remember where I'm supposed to be seeing codeeditor even on other wikis, I'd better shut up :P [13:34:01] @seen guillom [13:34:01] odder: Last time I saw guillom they were changing the nickname to guillom, but guillom is no longer in channel #wikisource at 7/21/2014 6:59:39 PM (18h34m22s ago) [13:58:10] whym: you need to have enhanced toolbar enabled [14:03:09] thedj: that did the trick, thanks a lot [14:04:30] whym: i intend to decouple that again at some point in the future... [14:04:49] perhaps during the hackathon. i've been too busy with work lately to work on it. [17:20:44] i'm a bit confused by the following: [17:21:07] these are the backlinks from the page "Einstein" which redirects to Albert Einstein: http://en.wikipedia.org/w/api.php?action=query&format=json&list=backlinks&bltitle=Einstein [17:21:51] meaning that i expected to find some Link on "Albert Einstein" linking to "Einstein" [17:22:13] Try adding &bllimit=max [17:22:18] i can't find that link... [17:22:27] Oh [17:23:44] i searched in the rendered html with $('a[href$="/Einstein"]') [17:23:50] (chrome dev tools) [17:24:39] it'd be /wiki/Einstein [17:25:18] ok, but it still ends in /Einstein which i don't find (only one is to wiktionary) [17:29:10] so am i wrong to expect a link to "/wiki/Einstein" on the "Albert Einstein" page, is it a cache problem or is it a bug? [18:28:08] hey, anyone knows why the dump process is idle and yet there is no enwikivoyage dump for almost a month now (used to be twice a month before)? could be that something is broken, or are the dumps supposed to happen less often now ? [18:28:41] apergos: ^^ :) [18:35:24] http://dumps.wikimedia.org/ seems to be the same as http://dumps.wikimedia.org/backup-index.html which "Database backup dumps" links to.. [18:40:29] yeah, according to http://dumps.wikimedia.org/backup-index.html there is nothing happening now, the dump process is idle.. [18:40:54] That's why I pinged ariel [18:41:03] and there's a few abortions yesterday, which is also weird [18:41:09] (at the bottom of the page) [18:41:21] that's why i am a bit worried [18:51:31] it was already asked in xmldatadumps-l [19:43:53] oh was it? :) that was me, but i didn't think the message made it, because i am not on the list, and i got a auto-reply saying the message needs approval [20:34:57] Reedy: [20:34:59] datasets@snapshot1002:/srv/dumps$ /usr/bin/php /apache/common/multiversion/MWScript.php getSlaveServer.php --wiki='dewiki' --group=dump -globals [20:35:00] ERROR: l parameter given twice [20:35:18] anything change recently? [20:38:17] apergos: Not that I know of.. [20:38:27] Looks to be the script itself though (doesn't work locally) [20:38:52] apergos: Though, looks like it should be --globals [20:39:51] Your branch is behind 'origin/production' by 132 commits, and can be fast-forwarded. [20:40:23] ha [20:42:06] oh that's my typo [20:42:19] anyways the actual comand is with two -- and in fact not that error message [20:42:30] lol [20:42:33] thanks :P [20:42:40] but it does have an error return code of 139 so that's new [20:43:30] mwscript getSlaveServer.php --wiki='dewiki' --group=dump --globals works on tin [20:44:49] we check the return code and expect it to be 0 [20:45:02] what exit code does it give though? [20:46:27] I see 139 over there too [20:47:01] 139 [20:47:12] it looks like it dies though [20:47:13] [static] => Array [20:47:13] ( [20:47:13] [keys] => reedy@tin:~$ echo $? [20:49:59] reedy@tin:~$ mwscript eval.php dewiki [20:49:59] > echo strlen( print_r( $GLOBALS, true ) ); [20:49:59] reedy@tin:~$ [20:50:22] Yeah, there's something weird going on there [20:52:42] apergos: Any idea when it broke? [20:52:50] let;s see [20:53:37] could have been as long ago as the 15th, once dumps get past that early step they keep running [20:53:58] so I see some completed well after that but the small ones look like that's the last date o them [20:54:19] 139 is a segfault? [20:54:36] php is segfaulting printing out all our globals? [20:55:04] great isn't it [20:55:09] to gdb! [20:55:22] thanks for being awake :-D [20:55:37] (it's almost midnight here so I'm definitely on the fuzzy side) [20:57:45] I got the segfault in the same place as you [20:58:18] [MessageCache::get] => Array [20:58:18] [0] => Closure Object [20:58:18] [static] => Array [20:58:18] [keys] => Segmentation fault (removing the parens) [20:58:45] string too long or something? [20:59:38] no idea and it' going to be very annoying if true [21:01:50] I dig something out of the globals which is why I ask for them [21:04:51] Program received signal SIGSEGV, Segmentation fault. [21:04:51] 0x000000000072d88e in ?? () [21:05:24] http://p.defau.lt/?SwTNmxA5Fr6oZHztDv_n7Q [21:05:40] sudo -u apache bash; gdb --args php /usr/local/apache/common-local/multiversion/MWScript.php getSlaveServer.php --wiki='dewiki' --group=dump --globals [21:05:40] run [21:06:23] Oh, need php debug stuff loaded.. [21:06:55] I was going to say, that stacktrace wasn't very informative [21:07:22] Can we get php5-dbg installed on tin? [21:08:03] or on snapshot1002 (seems I have access) [21:10:25] apergos: ^ [21:11:17] um sure [21:11:24] sec [21:11:49] which would... actually I'm thinking tin makes more sense as you'll prolly want it again at some point [21:11:54] and not just you [21:12:30] Possibly should be puppetised [21:13:09] likely but I am not doing that at midnight o clock [21:13:13] :) [21:13:28] can be easily done in retrospect [21:13:38] installed [21:13:48] lemme know if there are any libraries you need too [21:14:46] sapi_cli_ub_write () at /tmp/buildd/php5-5.3.10/sapi/cli/php_cli.c:284 [21:14:46] 284 /tmp/buildd/php5-5.3.10/sapi/cli/php_cli.c: No such file or directory. [21:14:54] bt now gives [21:14:55] #0 sapi_cli_ub_write () at /tmp/buildd/php5-5.3.10/sapi/cli/php_cli.c:284 [21:19:11] No bugs obviously logged [21:19:39] groan [21:20:13] So I could log a bug... [21:20:35] actually, I wonder [21:22:16] if you can't find anything tonight I'll try to fin a workaround for my situation tomorrow, get the value out using eval and something horrid instead of wading through the globals [21:22:28] it will be more fragile than this but that's the way it goes [21:22:57] I wonder if just getting the variables you want via script that echos $wgConf->get() might work [21:23:29] I'm just wondering if it's specifically relatd to MessageCache::get [21:24:05] haha [21:24:10] var_dump works [21:24:55] I was expecting it to be (relates to message cache), yes [21:25:01] oh wunnerful [21:25:30] just trying to get print_r to just dump the MessageCache::get bit.. [21:25:59] I don't of course give two figs for the message cache, globals could skip that and I would be perfectly happy [21:27:08] print_r( $GLOBALS['wgHooks']['MessageCache::get'] ); fails [21:27:33] array(1) { [21:27:33] ["keys"]=> [21:27:33] UNKNOWN:0 [21:27:33] } [21:28:39] So it's some weird subscriber that PHP is barfing at [21:29:12] extensions/WikimediaMessages/WikimediaMessages.php:$wgHooks['MessageCache::get'][] = function( &$lcKey ) { [21:29:56] huh [21:29:57] Yeah, it's that [21:30:14] boooo [21:30:42] We've other closures... [21:31:11] Yup [21:31:17] Comment that out and it works [21:32:42] apergos: You can blame https://github.com/wikimedia/mediawiki-extensions-WikimediaMessages/commit/37585a58c68c9f343a094d148438384962b404fd [21:33:08] ah so it would be the switch to the branch with that in it [21:33:09] nice [21:33:11] grrrr [21:33:27] Yup [21:34:37] so in the meantime the --globals option to maintenance scripts is broken then [21:34:41] pphhht [21:38:16] apergos: So swapping it from a closure fixes it [21:38:19] I'll commit and deploy [21:38:30] ah you will? bless you [21:38:34] Stupid PHP [21:41:09] oh right, there's a PHP bug with global closures [21:41:11] * legoktm finds [21:41:46] legoktm: thanks [21:42:00] https://bugs.php.net/bug.php?id=52144 [21:42:37] https://gerrit.wikimedia.org/r/#/c/86154/ says "Intermittent error in production from top-level lambda functions" [21:44:48] Presumably they didn't fix it in 5.3 [21:46:16] if they fix all the bugs in PHP there'll be no language left... [21:48:14] * bd808 shushes YuviPanda [21:49:49] Reedy: I'm not going to make it more than aboutanother 10 minutes [21:49:58] already having the whiplash syndrome effect in my chair [21:50:28] :( [21:50:32] I'll get the fix deployed etc [21:50:43] Should be good to set things off again when you get online in the morning [21:52:12] apergos: actually, fix deploying now [21:55:37] it made it around to snapshot1002 but not snapshot1004 (or maybe not to the branch that the smaller wikis are on) [21:55:53] yeah, just doing wmf14 now [21:55:59] ah ha [21:56:12] thanks a million for doing this [21:57:12] Should be done aswell now [21:58:03] and they are all running [21:58:10] whee [21:58:15] what a horrid little bug [21:58:37] ok off to drink juice and sleep asap, thanks again! [21:58:44] np [22:29:31] https://en.wikipedia.org/w/index.php?title=Module%3APing2&diff=618042544&oldid=590771379 this diff is retarded [22:29:43] the viewer of diff is using one of the least intuitive interpretations ever [22:36:26] viewer of diff? [22:36:44] diff2 (or whatever we use) [22:36:45] yes, I linked to a diff, it's viewed by a something [22:36:51] this something is being a bit silly [22:37:05] ... mildly to say; I'd be happy to curse at it for quite a while [22:37:25] never heard the differ called a viewer [22:37:59] we have https://bugzilla.wikimedia.org/show_bug.cgi?id=13462 but in your example there's also the problem of indenting [22:41:30] as usual dwdiff does better