[00:00:43] * ori takes Noam out for a bike-ride [00:00:45] \o [00:01:34] have fun [03:11:15] wow....I just caught up on the backscroll here [03:13:23] I generally agree no one should rush into a decision here, but wow, Daniel has his work cut out for him. [03:57:32] I may start harping on Conway's Law again. I was just passed a URL that I'm planning to read later: https://wingolog.org/archives/2015/11/09/embracing-conways-law [05:04:58] [15:34:47] It wouldn't be "difficult", it would just be a giant amount of work <-- it's extremely difficult to find reviewers for it https://gerrit.wikimedia.org/r/#/c/181521/ [05:23:50] i get a lot of flack for playing fast and loose with +2s, but holy shit are these anaemic +1s pitiful [05:25:58] a +1 can have actual signal value when a patch is genuinely controversial [05:26:17] what does +1 mean anyway? [05:26:17] Reviewer has a working mouse. [05:28:32] it means http://media4.giphy.com/media/ASd0Ukj0y3qMM/giphy.gif [06:53:45] * robla +1s the Ralph image [07:53:51] ori: I am confused (not sure if offended) by above discussions. First long discussion whether TitleValue is useful (this seems controversial), and then ranting against +1 on a patch to use TitleValue (no longer controversial?). I am not sure what I should do if I just want chime in and point a thing that might be missed by other reviewers. [18:17:39] ori: I think this -- https://gerrit.wikimedia.org/r/#/c/252359 -- will bandaid over our json encoding exception problems until I can get a better fix done with upstream. As a bonus it will bandaid over other logging encoding/transport problems in the future as well. [19:29:06] tgr|away: if you are still feeling crappy we can skip the 1:1 today. Up to you [20:29:51] bd808: no, back to work [20:30:23] ok. talk to you in 30m then [21:14:29] tgr: is my side broken or yours? I can't tell [21:18:48] hangouts are not our friend today I guess [21:18:58] * bd808 shuts off video [21:23:04] legoktm: ping? [21:26:49] urandom: pong [21:27:46] legoktm: re: https://phabricator.wikimedia.org/T116786, i think rcfeed will still work for this, but if you think we're able to make breaking changes here, i'd just as soon do it right [21:28:37] mostly because it seems like there are way too many different approaches to this problem [21:29:45] legoktm: any ideas here? [21:30:41] I've been thinking about the separation between the formatters and engines over the past day, and right now it appears that you can use them all interchangably, but in reality, you can't (e.g. the IRC colorful formatter one won't work with redis). [21:31:50] I'll post something on the bug sometime today, I need to think about it a bit more [21:32:31] legoktm: what did you think about using EventRelayer here? [21:33:12] it seemed like another abstraction layer that might not really be necessary [21:33:39] just to play around, i added a $wgRCRelayer to configure EventRelayer instances, and just iterated over them in addition to wgRCFeeds [21:33:53] seemed like it would work [21:34:36] legoktm: the reason i was thinking of ER here, was that it seemed pointless to maintain a formatter and engine distinction, which i think is what you were just saying [21:35:02] that a single interface that accepted a map would do it... [21:35:09] and that we already had EventRelayer [21:39:10] yeah. [23:07:29] TimStarling: I'd like to get https://gerrit.wikimedia.org/r/#/c/247178/4 in and do any improvements (if I think of them) later if that's OK [23:08:25] ok [23:50:07] Hey bd808, what would bawolff need to do to get access to logstash? He's signed an NDA with us, just need someone to update his ldap group? [23:53:34] needs an ldap admin or ops to add either the wmf or the nda ldap group, yeah [23:57:53] csteipp: yeah, what Krenair said [23:59:00] tl;dr ping ostriches on irc [23:59:47] * ostriches feels pinged