[01:23:54] from what I saw having one long running bug didn't seem to get much traction it was more of a tourist attraction than a work magnet :) [01:33:27] oh yea, nobody would ever want to take one of those [01:33:31] because they are unresolvable [01:33:44] that's the point of a tracking bug... [01:34:16] no, a tracking bug can have an end [01:34:21] when all subtickets are closed you can close it [01:34:28] but "missing doc" will never ever be resolved [01:59:40] tracking bugs are no longer useful [01:59:45] they should be tag style projects [01:59:46] imo [01:59:51] that is the translation of the idiom [02:00:28] that is a broken idea in bugzilla from what I can tell in a lot of ways [02:42:56] you can create a parent task with a bunch of blocking tasks to track the different pieces of a larger task, that certainly works well enough. it gives you some functionality that a 'tag project' doesn't provide like notifying subscribers of the parent task when the blocking tasks get closed. [02:43:28] what's the advantage of tags over blocked-tasks? [02:43:42] (I could be missing something, obviously) [02:43:51] none really I just think it's teh closest translation to how people use 'tracking' issues in bz [02:44:12] they really should be called "stalking" issues [02:44:47] I think that's mostly what they are used for. Cross project theme tracking [02:45:10] i.e. tagging :) [02:45:15] yup [02:45:31] but I will call tagging cross project theme tracking from now on [02:45:58] It's got better marketing buzz :) [02:46:07] synergy!!!!!!!! [02:46:19] I know how to talk to VCs and VPs ;) [02:47:00] * bd808 notes this drives some of his team members crazy [02:47:59] I've spent so much of my career dealing with non-tech jagon loving execs I default to that mode when writing emails and proposals [02:48:29] the weird part is, I used to do suit and tie presentations [02:48:30] *jargon [02:48:35] and we would fix something huge, massive [02:48:40] earth shattering [02:48:47] and then some widget that sparkled would blow their mind [02:48:53] and one guy did it last weekend in a day [02:48:59] heh. yup [02:49:01] and that justified us for 6 months [02:49:42] I added gravatars to a workflow tool and got a trophy for it [02:49:43] never underestimate a sparkly widget [02:49:49] truth [02:50:05] we could either meet all project deadlines [02:50:14] or fix thsi one thing the ceo complained about last week that doesn't matter [02:50:17] you choose #2 [02:50:26] unfortunately for life [02:50:32] my friend damon did really well with dluxwebcandy.com ... the content was almost entirely sparkly widgets [02:51:49] customizable sparkly widgets even. that you could post on your myspace or even facebook (though facebook disabled the sparkles, obviously. because facebook hates kittens) [02:52:04] they do I've seen it [02:54:36] http://yoursmiles.org/gsmile/unicorn/g3103.gif [02:54:44] your hired! [02:54:48] one of my favorites [02:55:52] * bd808 goes back to debugging some weird parsoid thing [03:23:05] bd808: #wikimedia-kawaii ! [03:23:35] I've heard of such a place but fear to tread there [03:23:54] It's a very friendly place [03:29:02] too many kittens I bet. [04:01:27] definitely friendlier than /dev/full [17:03:32] qgil: meeting? [19:04:06] chasemp, twentyafterfour: around? meeting. [19:05:33] yep [19:15:48] hey there, i just merged that patch about the default timezone [19:15:52] UTC for the win [19:16:12] also, T497 , RT #8628 updated [19:16:32] imaginaryIRCbot: would be nice if you were here to report that [19:35:17] Ah, :) [19:35:42] chasemp, any objections on Close,Resolved https://phabricator.wikimedia.org/T572 ? [19:35:55] I was about to press the button, but then I thought I'd better ask [19:36:00] should we wait for any of the ppl to reply? [19:36:07] robla or faidon [19:36:17] I was going to give them at least a day [19:36:50] ok, then you close it -- I'm in an exceeding Friday-Night mode, and should restrain myself. :) [19:37:58] heh ok man [19:38:47] greg-g, legoktm YuviPanda what do you know about the IRC bot related tasks? Is there someone that is working on them? I'm trying to keep the tasks up to date because there seems to be movement, but it's unclear [19:39:59] valhallasw was working on it last, there's a phawikibugs outputting into -feed right now, but it needs work [19:41:05] (also I haven't seen him online in the past few days, so I assume he's busy IRL) [20:17:19] what am I the bottleneck for? [20:18:48] * greg-g reads [20:23:12] https://phabricator.wikimedia.org/T572 [20:24:11] robla: chasemp wanted you/faidon to respond on https://phabricator.wikimedia.org/T572 before closing [20:24:23] andre__, I just realized that "Needs Info" is "Open,Needs Info", is this ok? i.e. https://phabricator.wikimedia.org/T498 [20:25:38] robla, "wanted to respond", or acknowledge, etc. This came when I asked chasemp whether we could close the task; he was more polite. :) [20:27:17] so...I think anything short of a full user migration is going to stir up a conversation on wikitech-l. the question then is this: [20:27:52] do you want to frame this conversation, and have it deliberately beforehand, or reactively if/when people learn for the first time that a compromise was made here? [20:28:26] that's an admittedly biased way of asking it...so let me unbias it a shade: [20:29:36] it may be that everyone generally agrees with the tradeoff that chasemp laid out [20:29:49] I suspect, both from my conversations with people and hunches based on experience, that this will be at least a little controversial [20:29:59] ^d had the same reaction as faidon [20:30:06] (for example) [20:31:36] so...that's my longwinded way of suggesting that discussing this on wikitech-l may be useful, and that even if you don't think it will be, you may end up sucked into a conversation anyway [20:31:59] I also don't want to speak for faidon and ^d [20:33:01] it would be good to put that all in the ticket, but I'm goign to leave the how of greater feedback solicitation to qgil or andre__ [20:33:21] there is a scheduled period of feedback solicitation during which this is all clear [20:33:37] so it's not a hunt for red october type mission by any means [20:33:53] but I generally don't know the best approach for wider criticism [20:35:06] that also doesn't state your opinion just says to ask for the opinions of others :) [20:37:33] +1 to a wider discussion than that one ticket with those few people [20:37:53] (this is the first I've seen it) [20:38:22] it kind of feels like when HR sent out the "vote" on loomio for how we would do FSA in the future (we had a choice) and the poll was only open for 4 days [20:38:35] I'm not entirely sure how it works, but this isn't really a vote scenario? [20:39:29] not saying it's a vote, but, like rob-la said, you might have to talk about it anyways, so better to frame it and discussion it before hand rather than reactively when someone comes at you with pitchforks [20:39:38] discuss* [20:39:54] greg-g: +1 [20:40:02] my sense of things is this. My gut tells me that Faidon is correct, and that even if it takes more time, it's potentially worth doing. [20:40:03] just plus one'ing rob-la's comment, really [20:40:16] however, it's not a hill I'm going to die on [20:40:18] what is faidon correct about? [20:40:42] "IMHO, I don't think asking thousands of users to migrate their account before the migration is a sane process, when it's really us that should be making this step for them." [20:40:53] we aren't? [20:41:13] did you read https://phabricator.wikimedia.org/T572#10081 [20:41:18] or the updated proposal in description [20:41:40] I read T572#10081, but I may have missed something [20:43:25] chasemp: does Phabricator has a feature to merge two accounts? [20:43:29] no [20:43:32] :-( [20:43:43] and renaming has weird consequences for any account w/ any history [20:43:49] is merge on the Phabricator roadmap? [20:44:02] no [20:44:03] I thought about importing all bugzilla accounts with a name of Bugzilla- being associated with the email for that account [20:44:14] so people login would reclaim their old comments [20:44:24] they can via the proposal now [20:45:50] accounts are always a thought issue :-/ [20:46:50] tough? [20:46:58] yeah that one [20:47:18] :) [20:47:24] although it is thought to come through all those similar terms [20:47:30] I'll let the community guys noodle on it, I'm really not up to snuff on knowing best plan [20:47:34] chasemp: to be clear, I think the solution your proposing is an ok compromise. it's a little hacky, but we have plenty of things that are (more than) a little hacky [20:47:38] damn I made the mistake again :/ [20:48:03] if there is one thing I know for sure it's that there isn't a way to do this that isn't a little hacky [20:48:31] hashar: Your english is much much better than my French ;) [20:49:01] bd808: you should work for Wikimedia France, you will greatly improve your french ! [20:49:21] * bd808 doesn't even have enough French to start a bar fight [20:49:22] ok thanks robla, if you could drop notes and I imagine qgil will respond [20:49:25] telecharges mediawiki [20:49:35] writing now [20:49:37] they are the barometer for this to my thinking [20:54:04] bd808: gilles once did a talk called "how to be rude in french" when we were there [20:54:06] fun stuff [20:54:30] just ask us whether we have color TVs and know about internet [20:59:15] I don't think one person was rude to me in france, other than maybe a person who looked at me like I was evil was reaching for my own croissant at this stand place [21:01:41] well hashar, do you? [21:03:50] robla: I do sir :] [21:04:01] btw, "Act Three" of this seems appropriate: http://www.thisamericanlife.org/radio-archives/episode/173/transcript [21:04:04] we have a lot of cliché regarding americans [21:05:38] chasemp: since you are around, does Phabricator has any stream of events that can be consumed ? [21:05:58] yes [21:06:10] chasemp: context is continuous integration. We would need to get a way to trigger something when a patch is proposed via Arc or a patch is approved [21:06:23] though I have exactly ZERO clue how Arc and Phab handles patches :( [21:06:39] https://secure.phabricator.com/T5462 [21:08:06] http://www.guywarner.com/2014/05/integrating-jenkins-and-phabricator.html [21:08:09] https://secure.phabricator.com/D7368 [21:08:11] maybe [21:08:24] yeah though our CI is slightly more compilcated [21:08:49] we have a daemon between Gerrit and Jenkins which listen for events (i.e. new patch uploaded), takes decision as to which job to run and trigger them in Jenkins :D [21:09:05] that's not so bad [21:09:12] depending on if you want to keep that [21:09:22] yeah that is going to be the next big question [21:09:36] what the hell do we do with our current CI system :D [21:09:42] burn it [21:09:45] hehe [21:09:56] as long as I don't burn with it, I am fine with that plan bd808 ! :D [21:10:02] in the generic sense we could still trigger an event for you [21:10:06] and you could still pull the diff [21:10:09] and do your thing [21:10:15] that's all abstractable [21:10:35] the daemon between Gerrit and Jenkins (named Zuul) is tightly coupled with Gerrit though [21:10:37] hashar: I'll give you the matches. You can take any safety precautions you'd like :) [21:10:40] so might not be easy to migrate [21:11:21] qgil told me there is a Phabricator brick to handle CI but I am not sure it is robust enough for now nor whether it will match our needs [21:11:57] we have some kind of epic meeting on Oct 20th, I guess we will start talking about it [21:12:35] bd808: meanwhile, did you get the mediawiki/vendor.git patches merged in? CI is ready for them :-] [21:13:11] I need to update them one more time. The @since 1.24 tags are out of date now :( [21:13:27] I already had to change them from @since 1.23 to 1.24 previously [21:13:49] but I should update them today or tomorrow [21:17:17] s/tomorrow/tuesday/ [21:17:51] heh. No tomorrow [21:18:04] Ihope they will land just fine and open the gate to get more 3rd party libraries [21:18:12] and phase out a bunch of our legacy code [21:18:59] we'll just have to find the parts we don't like that someone else did better [21:19:22] I'm more excited about make some of our stuff usable outside of MW [21:19:28] *making [21:28:56] bd808: yeah that is going to be a good thing [21:29:04] we would need to market those bricks somehow though [21:29:24] That's part of the deliverables for my Q2 project [21:29:45] https://www.mediawiki.org/wiki/Library_infrastructure_for_MediaWiki [21:30:59] bd808: I can imagine a dev / library portal of Wikimedia software that people can pick from [21:31:07] there is a dev hub being prepared by some people [21:32:08] I'd loce to end up with something like http://symfony.com/doc/current/components/index.html eventually [21:32:14] ahaha [21:32:20] *love [21:32:26] I had to delete my current sentence to say "ahaha" [21:33:03] a friend that cowork at the same place as me, has been the release manager for Symfony 1 [21:33:09] and know the author "Fabien Potencier" [21:33:37] Neat. They have made some nice stuff [21:33:51] my friend is using Silex, a micro framework that let you compose your application yourself [21:33:55] Some is a bit over engineered in places but solid code over all [21:34:01] you "just" have to pick some Symfony bundles [21:34:14] I like silex, but I like slim more [21:34:46] so in the end, I don't think we can compete with Symfony / Zend framework [21:34:52] but can surely start reusing their code [21:35:08] and extract some of our code along the same model they are using [21:35:13] Zend is mostly crap in my experience [21:35:33] isn't it popular in US? [21:35:38] (regardless of crapiness) [21:35:40] too many knobs to turn and not enough perfomance [21:36:04] I think php is mostly a roll your own language unfortunately [21:36:30] Very few large projects are built on another project's framework [21:36:41] probably because PHP by itself offers very little [21:36:59] Yeah and so many projects are grown rather than designed [21:37:09] our case :D [21:37:27] I think Drupal (or maybe it is Wordpress) is migrating to use Symfony components [21:37:27] "I'll just make a quick hack to X" and then 3 years later it gets an IPO :) [21:37:45] I think I heard Drupal was, yeah [21:37:45] Bugzilla is telling people to put their Bugzilla emails to Phabricator. [21:38:00] But https://www.mediawiki.org/wiki/Phabricator/Help#Creating_your_account says, "This section only applies after RT and Bugzilla tickets have been imported into Phabricator. This import has not happened yet." [21:38:03] Which is it? [21:38:42] If you want attribution for past messages, add the email(s) you used in bugzilla [21:38:49] superm401: I would trust the banner [21:38:57] Thanks, bd808. I'll fix MediaWiki.org. [21:39:20] superm401: you might want to poke qgil / andre__ by email about it [21:39:31] bd808: [21:39:40] hashar, why, do you think the Bugzilla banner might be wrong? [21:39:41] The links won't happen until there is content imported obviously [21:39:46] hmmm, I should create my personal phab account... [21:39:50] bd808: maybe I will reimplement python argparse in PHP I like it too much :) [21:39:57] ( to claim my greg@grossmeier.net bug stuffs) [21:39:59] bd808: and get rid of our Maintenance options [21:40:16] hashar: Heh. Or look at the symphony cli classes [21:40:33] greg-g, not certain what I'm going to do about that. Currently I only have one Bugzilla, and it includes stuff from before I worked here. [21:40:33] superm401: na i think the bugzilla banner is correct and that the wiki doc is lagging out or might be have a different purpose [21:40:33] hashar: http://symfony.com/doc/current/components/console/introduction.html [21:40:38] bd808: yeah :-] [21:40:50] hashar, alright, I'll email them just to let them know. [21:40:50] bd808: but then we would need to add that component to /vendor.git :] [21:41:22] bd808: that could be a good use case since maintenance script do not impact end users that much [21:41:27] superm401: \O/ [21:42:05] superm401: /me nods, I separate pretty religiously [21:42:19] hashar: Yeah. It would be a reasonable place to experiment. Or my your lib, but don't do it without looking for some decent existing php argparse code [21:42:21] superm401: before the whole "(WMF)" everywhere discussion :) [21:42:25] *Or make [21:42:51] greg-g, I was one of the holdouts back when that was still allowed. And I changing my existing Bugzilla email to Wikimedia. [21:43:08] (I had a long-time single account on the main SUL). [21:43:24] I think legoktm is going to go single-account on Phabricator, though, IIRC. [21:43:40] bd808: yeah I probably don't want to create a 15th standard [21:44:18] yes [21:44:27] creating two accounts would be a PITA [21:44:32] greg-g: do I need to change my bugzilla account to Antoine Musso (not really WMF) ? [21:44:52] I guess I can just tweak my username [21:45:22] Antoine "hashar" Musso (WMF) [21:45:25] \O/ [21:45:55] https://phabricator.wikimedia.org/p/bd808/ [21:46:04] yep, free for "name" field in Phab is awesome [21:46:22] free form* [21:46:30] free jorm! [21:46:43] https://phabricator.wikimedia.org/p/hashar/ \O/ [21:47:11] phabricator has a field for "IRC nick" [21:47:29] yup just noticed that [21:47:29] That's new [21:47:39] though if logged with LDAP we can't set our mw user page :-) [21:47:42] Probably so bots can poke you [21:48:13] hashar: You can link both. https://phabricator.wikimedia.org/settings/panel/external/ [21:48:34] And you can add 2factor auth -- https://phabricator.wikimedia.org/settings/panel/multifactor/ [21:48:52] ohhhh [21:49:04] https://phabricator.wikimedia.org/p/greg/ vs https://phabricator.wikimedia.org/p/greg.grossmeier/ [21:49:19] if only our LDAP sul account had some oauth portal with two factor [21:49:37] so first you use labs user, [21:49:46] then you realize you might also connect mw sul user [21:50:01] then you realize you gotta decide if it's the "WMF" user or not :p [21:50:17] so you ask to support login via google wikimedia.org [21:50:24] :p [21:50:26] schizophrenia ! [21:50:27] Yeah. It should really let you connect more than one SUL account [21:50:47] I guess folks will ask to log with their Facebook identityy as well [21:51:13] But it's really corporate software in a FOSS disguise so many of our use cases are new [21:51:14] rejected: wikipedia ain't quora /me hides [21:51:17] but I can imagine everyone already has a mediawiki.org account anyway [21:51:29] They will soon. SUL! [21:51:55] That OAuth endpoint really needs to be moved to meta [21:52:12] Having it on mw.o is confusing to some people [21:52:15] it is [21:52:59] w00t, the task list I wrote down at midnight last night as I was falling asleep is now done [21:54:25] greg-g: qgil proposed a meeting on Oct 20th to write some epics stories [21:54:29] for gerrit migration [21:55:02] yeah, I rsvp'd yes [21:55:07] bah forwarded [21:55:09] ok :-] [21:55:11] :) [21:55:15] thanks :) [21:55:15] not sure whether I will be able to attend [21:55:51] so in the end sleeping time [21:55:56] yeah, bad time for you [21:55:57] enjoy your week end [21:56:07] you too! [22:01:31] * qgil reads back... [22:02:19] robla, chasemp there are discussions that can't be avoided and can change things, and there are discussions that can't be avoided that can't change things [22:02:31] and the truth is that, at this point, we have little to no choice [22:02:52] besides, I don't think it is going to be *that* controversial [22:03:32] (but I shouldn't write this anywhere, you know) :) [22:03:48] yeah, I'll be quoting that :-) [22:04:06] * hashar giggles [22:04:33] I think it won't be that controversial if it's communicated in a thoughtful way [22:05:15] I think there's a real danger that, if you try to minimize attention to this tradeoff, people will wonder what other tradeoffs you're trying to minimize [22:05:18] it's a trust thing [22:12:42] We are not trying to minimize attention at all (I believe @chasemp was the first one using pink Unbreak-Now! in a real task in phab), but I don't think it is worth promoting any big wide wikitech-l debate on the void either [22:13:20] I'd rather discuss this topic in the context of all the rest, in front of an actual Phabricator test instance with a bunch of authentic tasks automatically migrated from Bugzilla [22:14:54] Alright, it's official: Saturday! [22:15:14] good night!