[05:18:27] 10MediaWiki-Core-Team, 6Discovery, 10Wikidata, 10Wikidata-Query-Service: Mourn Titan - https://phabricator.wikimedia.org/T88550#1395305 (10MZMcBride) [05:18:51] 10MediaWiki-Core-Team, 6Discovery, 10Wikidata, 10Wikidata-Query-Service: Mourn Titan - https://phabricator.wikimedia.org/T88550#1395307 (10MZMcBride) What replaced Titan? [05:19:46] 10MediaWiki-Core-Team, 6Discovery, 10Wikidata, 10Wikidata-Query-Service: Mourn Titan - https://phabricator.wikimedia.org/T88550#1395309 (10ori) >>! In T88550#1395307, @MZMcBride wrote: > What replaced Titan? [Blazegraph](http://www.blazegraph.com/) [05:20:03] 10MediaWiki-Core-Team, 6Discovery, 10Wikidata, 10Wikidata-Query-Service: Mourn Titan - https://phabricator.wikimedia.org/T88550#1395310 (10Smalyshev) @MZMcBride Blazegraph [05:20:13] SMalyshev: I was quicker! [05:20:37] *and* I included a link. [05:22:25] 10MediaWiki-Core-Team, 6Discovery, 10Wikidata, 10Wikidata-Query-Service: Mourn Titan - https://phabricator.wikimedia.org/T88550#1395311 (10MZMcBride) Hmmm, Blazegraph™, I think you mean. Thanks for the info. It looks like an interesting partnership. [06:13:06] ori: heh :) [06:25:22] why http://stats.grok.se/en/top has so many duplicates? [10:38:07] SMalyshev: seems a one time error in the rank generation, other languages are fine http://stats.grok.se/la/top [10:38:39] Probably all the fake requests to "undefined", data:// and so on don't help [15:50:31] * ori solicits +2 for https://gerrit.wikimedia.org/r/#/c/220398/ [15:56:52] * ostriches hands ori a freshly baked +2 [15:57:13] thank youuu [15:57:53] yw [19:00:11] bd808: Timestamps in the logs on fluorine are in the local time now, rather than UTC? [19:00:17] Is that a bug? [19:19:03] hoo: https://phabricator.wikimedia.org/T99581 [19:19:50] Thanks :) [19:39:21] legoktm: how far are we from setting $wgCentralAuthStrict = true? [19:40:38] bd808: it's blocked on rolling out global user merge, which may happen tomorrow, but more likely on monday [19:40:48] sweet [19:41:02] anything I can help with? [19:41:45] I was checking my team goals -- https://www.mediawiki.org/wiki/Wikimedia_Engineering/2014-15_Goals#Reading_Infrastructure -- and its the big success measure [19:41:53] which is funny as you know [19:42:42] * bd808 is tired of pretending that giant projects can be done on a calendar schedule [20:01:00] bd808: want to look into either https://phabricator.wikimedia.org/T103738 or https://phabricator.wikimedia.org/T103739 ? I'll do the other one [20:03:32] legoktm: I can look at https://phabricator.wikimedia.org/T103738. Sounds easy enough [20:04:43] ok, thanks [20:21:42] > As a security measure, you are only allowed to globally merge a user once per minute. Please try again after one minute. [20:21:44] * legoktm grumbles [20:24:03] legoktm: that message is missing for sure. Possibly others introduced in the same patch too. I'm going to check all of them [20:24:19] :/ [20:24:23] thansk [20:24:28] np [21:03:08] legoktm: poke? [21:03:44] Keegan: pong [21:09:16] lol @ https://gerrit.wikimedia.org/r/#/c/219584/ - turns out, I broke it [21:09:21] * MaxSem hides [23:30:57] bd808: so, how should I test https://github.com/wikimedia/composer-merge-plugin/pull/36 ? normally I just patch my copy of the plugin in vendor and run stuff, except in this case I need to test functionality without already having it installed... [23:31:49] legoktm: you can test it in the "example" directory with a local checkout [23:32:03] the compser.json there points back at the clone [23:32:36] I should really explain that in the README [23:33:07] so git clone; cd example; composer install [23:33:15] * legoktm tries [23:34:16] you can also test by just blowing away your vendor dir and making sure you have a composer.local.json that does something [23:35:02] using example worked, except your patch isn't [23:35:38] bd808: http://fpaste.org/236353/88924143/raw/ it merges at the end, but doesn't re-trigger the installer? [23:36:04] and I'm using composer alpha10 [23:37:02] legoktm: :/ [23:37:14] and you have the patch in the clone? [23:37:21] yes [23:37:34] 85a7920 [23:37:49] ohh [23:37:52] maybe not [23:38:04] I applied it to a branch called 36 [23:38:10] and composer is using dev-master [23:38:14] ah. it needs to be on master [23:38:27] or you need to adjust the compser.json magic [23:39:07] wooooo it works!!! [23:39:07] you [23:39:12] sweet [23:39:29] it was kind of too simple to seem sane once I got it coded up [23:39:54] I was really surprised that the plugins are imported into the runtime as soon as they are downloaded [23:40:42] how many people could you pwn if you got control of a popular github repo... [23:40:49] lots [23:40:52] lots and lots [23:41:23] which is one of the horrible things about how composer works [23:42:11] but really pear/pecl is no better [23:42:26] or most distribution systems [23:49:17] sweet [23:50:00] bd808: and https://github.com/wikimedia/composer-merge-plugin/pull/38 :) [23:50:49] legoktm: I'm going to track down that crash bug that is reported on github and fix it too. I'm pretty sure I know what causes that [23:51:01] ok and we'll do a release after that? [23:51:12] sounds good.