[04:57:01] anybody hanging around? [04:57:12] I'm looking for advice on avatar storage [05:13:17] /dev/null [17:38:13] robla, about? [17:38:19] heya, yup [17:38:24] Reedy: what's up? [17:38:35] Hey, any idea who is the main contact with the calcey lot? [17:38:40] moi [17:38:51] Awesome [17:39:40] robla, can you get them to fix their svn autoprops. Other people shouldn't have to be cleaning up after them http://www.mediawiki.org/wiki/Special:Code/MediaWiki/77657 http://www.mediawiki.org/wiki/Special:Code/MediaWiki/77592 [17:40:19] http://www.mediawiki.org/wiki/Subversion/auto-props [17:40:19] yup, can do. [17:40:26] Cheers :) [17:41:06] I'm explicitly pointing them at CodeReview, so any comments there should get to them [17:41:24] I did comment on their first one [17:44:36] one thing I'll be evaluating them on is how well they integrate into the community; a big part of that, obviously, is how well they follow our standards and how reachable they are [17:45:01] Yup, definately [17:45:10] IIRC there was issues with similar stuff before [17:45:23] It's alright as a mistake, and most people usually pay attention when someone points it out :) [17:45:30] What're you guys talking about? [17:45:35] Calcey [17:45:39] RoanKattouw, Calcey not setting auto props [17:46:34] Ah [17:46:43] It's not a major issue, it's just annoying [17:46:44] Well it's hardly the first time Calcey's annoying us [17:47:33] I think I'll ask them to make sure they're on IRC every time they make a commit, and for an hour or two afterward, at least for a while [17:47:50] But to be fair, we haven't done a great job of pointing out what they're doing "wrong", and it's usually a mix of different people doing stuff under one name [17:48:06] Are they sharing accounts? [17:48:12] On BZ they used to [17:48:12] If so, that should be rectified [17:48:15] I don't think so [17:48:15] Not in SVN I think [17:48:44] When we worked with Calcey on the usability project there was one BZ account for all of Calcey IIRC [17:48:49] lol [17:48:52] I really could do with 3 monitors now [17:48:55] They have personal accounts on SVN, yes [17:48:55] *Reedy groans [17:48:56] Nadeesha is the one actually making commits; I don't know if she's writing all of the code, or if she's taking code from others [17:51:11] other than the whitespace issues, anyone have an opinion about the code itself (and most importantly, what its doing?) [17:51:41] I haven't looked at the code [17:51:50] My major annoyances with Calcey in the past were on BZ [17:52:17] Bad bug reports, filing dupes, filing dupes of their own bugs (!!), filing things that weren't bugs, that kinda thing [17:52:38] Which is why I believe an in-house QA person is much preferable over an external firm [17:53:00] Especially if that external firm's timezone is UTC+5:30 or whatever it is [17:53:34] robla, other than the commiting "test files" for no real reason :P [17:54:46] and then not immediately deleting them [17:56:29] Yeah they deleted one, I deleted the other [18:03:52] Blah, what would be the simplest way to test langlinks on a dev wiki? [18:04:37] interwiki link to self? [18:04:55] Add interwiki table entry 'de' => 'localhost/wiki/$1' , add [[de:foo]], done [18:05:12] Ahh, that'll do [18:06:58] Cheers roan :) [18:56:09] do we know who kingturtle is? admin/beauracrat on en? [18:57:14] Reminiscent of the turtle-y man from the "Six Flags" commercials. [19:03:20] hey, it's shirley. you've been quiet lately. [19:04:01] I guess. I've been poking my head into #mediawiki occasionally. [19:04:03] And boggling. [19:10:05] playing boggle, or being surprised at stuff? [19:10:57] The latter. [19:11:25] I don't BOGGLE® much. [19:11:48] what's surprising you, then? [19:16:02] Same things as always, I think. Attempted increases in bureaucracy, general slowness, lack of progress, lack of transparency, etc. [19:16:05] Those chestnuts. [19:16:18] ah. [19:16:39] it's the holiday and travel season. everything is going to be slow. [19:16:59] though i'm not sure where the increases in bureaucracy comes from. [19:17:24] Too many "managers," I think. [19:17:56] hrn. not seeing it. my command chain is me -> danese -> sue [19:18:26] You don't have an EPM or whatever? [19:18:31] Most people do. [19:18:58] And then there are the people like Howie, who .. also manage things, I guess. It's all a bit fuzzy. [19:19:18] Shirley: I think the current status of CR is a lack of proper management, rather than an excess of it [19:19:26] for example [19:19:28] I don't really understand the EPMs' positions... it seems like they officially aren't in the chain of command but in practice I feel I take my marching orders more from Alolita than anyone else [19:20:13] Bryan: Well, I think your keyword there might be "proper." :-) [19:20:16] i have an epm (alolita) but she's not my manager. i don't report to her. [19:20:40] so, wait. on one hand, we don't have enough management, and on the other, too much? [19:20:51] Bryan: I don't have anything against managers. I have things against unnecessary, ineffective, or bureaucratic ones, though. [19:20:52] Code review is hardly managed AFAICT [19:21:10] Danese and Alolita are anything but those things, IMO [19:21:27] The suggestion to add more layers to the code deployment process yesterday is still haunting me. [19:21:38] Realistically, that won't happen [19:21:50] I think all main developers would oppose that [19:22:06] (I'm assuming you meant robla's extra BZ state suggestion) [19:22:09] Yes. [19:22:09] i didn't see that discussion. [19:22:32] RoanKattouw: You were doing stats recently. What is the code review backlog status? [19:22:33] Basically robla wanted to draw the distinction between fixed on trunk and fixed on deployment with two different BZ resolution states [19:22:53] Shirley: Mind if I run those after I finish my review queue? :) For yesterday's numbers, see [[mw:User:Catrope/CR stats]] [19:23:01] Shirley: http://toolserver.org/~robla/crstats/crstats.html [19:23:05] *RoanKattouw goes back to work, 162 revs to review [19:23:20] RoanKattouw, isn't that a bug? ;) [19:23:34] Isn't what a bug? [19:23:41] Review queues [19:23:59] brion (and/or ^demon|away) was there a time in mediawiki history where mediawiki would write something other than an anon user's IP into the rev_user_text field of the revision table? [19:24:00] Shirley: calm down about the BZ suggestion [19:24:06] robla: Graphs aren't helpful. I want to know when the next actual deployment is going to happen. [19:24:32] nimish_g: for stuff imported from old UseModWiki, yes [19:24:40] *robla is writing up a wikitech-l mail about that very thing now [19:24:40] Like a real deployment. Not just "special case this code has to go live today!!!" deployment. [19:24:44] it would i think do a reverse DNS and store the hostname sometimes [19:24:46] Okay. :-) [19:24:53] Reedy: Yes, review queues in software is a bug. GMail kinda works for a personal review queue when your queue is "anything that's not semantic, in chronological order" [19:25:00] heh [19:25:13] Shirley: It's been said for a long time we intend to do a 1.17 release+deploy in January [19:25:20] Because there's some pressure to get ResourceLoader live too [19:25:32] I think that's partly for the wrong reasons, but whatever [19:25:33] Last I heard, RL was supposed to be tested in November. [19:25:44] RoanKattouw: I'm actually thinking March is more realistic [19:25:44] <^demon|away> brion: It would? [19:25:47] We're also building the new version of ArticleFeedback to depend on RL [19:25:50] March, bah [19:25:50] <^demon|away> Must've been before my time. [19:25:58] January is sort of like March if you squint really hard. [19:26:03] haha [19:26:05] (but I'll get to that in my mail) [19:26:15] ArticleFeedback, woo! That's what the users want. [19:26:15] Srsly though, I need to set up an RL prototype with WMF extensions and such [19:26:30] ^demon|away: long before your time :) [19:26:35] 2001 [19:26:47] <^demon|away> Speaking of other old stuff... [19:26:52] AF is not necessarily what the users want but it's what PPI wants for sure, and that's being built to depend on RL too now [19:27:00] <^demon|away> I'm tempted to buy an Apple II case, going for about $10 on eBay [19:27:07] <^demon|away> Kinda wanna put my PC inside it :) [19:27:07] PPI is stupid. [19:27:15] so, brion, for REALLY old edits, if the rev table says it was from user 0, irrespective of what the user_text says, it should mean it's an anonymous edit right? [19:27:16] *RoanKattouw wants such grant obligations/restrictions to go away already [19:27:23] i would totally do that, chad. [19:27:34] except that i'm trying to get RID of computers these days. [19:27:41] hmm [19:27:52] Computers are evil. [19:27:55] Shirley: You saw what we did with UW, right? Deploy an "experimental" version on the very last day of a month? Three guesses why that was [19:27:57] nimish_g: well also user edits from those days didn't all get user_ids added i think [19:28:03] so there may be some logged-in edits from then too [19:28:15] they'll have the username (not guaranteed to be unique) as rev_user_text [19:28:17] RoanKattouw: I knew about the November deadline. [19:28:25] hmm...ok, thanks brion =) [19:28:29] Deadlines + Wikimedia --> :-( [19:28:34] Fairly consistently. [19:28:43] *^demon|away goes back to playing SNES games, only popped in because of the ping [19:30:22] So... maybe by March there might be a software deployment of some kind? [19:30:36] That'll be nearly a year since the last real update, I think. [19:30:46] Nearly, yes [19:31:14] March sounds like a looong time away though [19:31:28] I realize this month is mostly shot, but still [19:31:30] <^demon|away> We should've branched a month ago. [19:31:34] And yet I'd put money on April. [19:31:37] <^demon|away> [19:31:45] Given historical trends. [19:31:46] I think we should branch soon [19:31:52] And if we do, it shouldn't take 3 months to stabilize [19:31:54] I think we should stop branching. [19:32:22] <^demon|away> Shirley: We branch before every release, and have been for years now. [19:32:46] There's 1.17 and there's what Wikimedia is running. [19:32:50] I was talking about Wikimedia. [19:33:21] From an operations perspective, I think the 1.NwmfM branches are good [19:33:26] RoanKattouw: Reedy: is it ok if we move our meeting up 30min? [19:33:35] But they shouldn't be allowed to get that far out of sync [19:33:38] TrevorParscal: You mean 30 mins later? [19:33:42] TrevorParscal, I'd completely forgotten [19:33:45] earlier [19:33:49] so, now? [19:33:49] jorm: you ok with that? [19:33:50] So that's like 6 mins ago [19:33:52] as in now... [19:34:01] there is indeed no problem with branching if we could branch at least every month [19:34:01] yeah, i'm fine with it. lemme pee real quick. [19:34:03] I'd have to grab my headset and go to a quieter room [19:34:15] I know jorm wants to start getting on his way, and It would help me clear something else [19:34:19] k [19:34:21] me too [19:34:27] It's fine, I'll move upstairs [19:36:01] ok [19:36:06] call into 2001 when you can [19:36:36] RoanKattouw: Reedy - dail in [19:38:38] Coming [20:31:36] hi RoanKattouw ... what's the status on the per-developer review tweak to the CodeReview system that you started at Hack-a-ton? [20:32:38] It's in trunk [20:32:40] robla, it needs prettyfying and reviewing :) [20:32:42] Called sign-offs [20:32:52] And yes, it needs code review (ironically) and Trevor love [20:33:05] lol [20:33:07] It seems to work fine now (I found the initial minor oversights) [20:33:13] i.e. someone needs to think about how people would actually interact with this [20:33:21] And it needs to look a bit nicer maybe [20:34:59] RoanKattouw: how easy would it be to deploy that to prototype? I'm not asking you to do it; I'm just wondering if there's anything surprisingly complicated [20:35:15] Easy [20:35:20] I'll make a note to that soon [20:35:30] Right now my marching orders are to fully focus on resourceloader [20:35:39] So I'm gonna set up a resourceloader prototype but it wouldn't be hard to enable CR there too [20:35:48] cool [20:35:49] And import some MW history into that instance [20:36:40] I may end up even doing it myself....just wanted to know if there was something ugly about it [20:37:15] No, not really [20:37:24] You just need a trunk install of CR with a dataset [20:37:55] screen php svnImport.php MediaWiki [20:37:56] Wait :P [20:38:18] That'll be fine, just run it overnight [20:38:25] Provided prototype isn't short on disk space or anything [20:38:46] I'll set it up once I finish... reviewing code :) [20:38:58] It'd, err, be much easier to use a smaller repo if you just need a dataset. [20:39:03] Like pywikipedia or something. [20:40:00] Yeah well [20:40:09] You can also just Ctrl+C the MW import after a few thousand revs [20:40:18] At least we (MW) have a reliable history of using rNNNN references [20:41:21] RoanKattouw, it doesn't take long from here [20:41:31] or didn't from me... [20:41:35] *for [20:41:53] Took some time for me [20:42:04] Hopefully faster from prototype, as it's right near the SVN server [20:42:18] So svn log in batches of 400 revs shouldn't take that long [20:42:44] For me the imports themselves were fast, but it hung on svn log each time [20:43:27] Mine does like 60 revs/s when doing updates [20:43:43] RoanKattouw, or just dump and reimport the CR tables from the cluster [20:43:55] then update.php against them [20:44:02] hah [20:44:05] Yeah that could work [20:44:07] I'd have comments too [20:44:11] Actually that's a very good idea [20:44:11] :) [20:44:21] *RoanKattouw sets out to do just taht [20:45:53] robla, that do you? :P [20:46:05] that's fantastic [20:46:45] https://bugzilla.wikimedia.org/show_bug.cgi?id=26220 There are too many files in the root of phase3/includes [20:46:52] If anyone has any sane ideas [20:47:03] delete files? :) [20:47:22] Haha [20:52:12] Reedy needs bug-writing lessons. [20:52:23] Pfft [20:52:35] My bugs are fome [20:52:36] *fine [20:52:41] Not that I can seemingly type today [20:52:44] Fscking cold hands [20:52:49] When #c0 doesn't include the summary text and the summary changes, it confuses people. [20:53:00] Fucking * [21:10:51] okay. off. [21:43:26] hi RoanKattouw [21:43:43] ok, sorry, it was pretty Tolly in there [21:43:47] No worries [21:43:52] Yeah, Salsman... [21:44:12] anyways [21:44:29] so, as per my most recent email [21:44:49] I also intend to set up an RL prototype very soon, including the CR signoff stuff because it's little extra work to set up and robla wants to play with it [21:45:06] we need to push to get 1.17 out in January, I'm not really feeling comfortable with this whole "push 1.17 into the future" trend... [21:45:34] Me neither [21:45:40] See my reply in 1.17 [21:45:41] RoanKattouw: so, when reviewing code, we need to limit how much code you are reviewing so you can return to focusing on resource loader deployment [21:45:46] I think we should just focus my effort [21:45:48] Yeah [21:45:55] yes, we agree [21:45:58] Well my limit currently is to just 'catch up' [21:46:07] so, what projects are you reviewing for right now? [21:46:10] I've got 130 candidate revs left, many of which have already been reviewed probably [21:46:15] Pretty much everything I understand [21:46:30] Mostly core [21:46:31] we need to at least temporarily narrow that [21:46:40] Well there's rigth now and right now [21:46:53] Right now, like today, I'm just catching up on backlog and reviewing without prejudice [21:47:00] Normally, I'm more picky in what I review [21:47:01] cool [21:47:07] ok - fair enough [21:47:20] *Reedy wonders how much more can be deferred in extensions [21:47:22] So I'm now reviewing like simple installer revs but that's a one-off [21:47:28] we need to assess what needs to be resolved for 1.17 to go into production [21:47:37] Reedy: I say we don't waste our time on that and get stats for phase3 + WMF exts instead [21:47:52] TrevorParscal: Exactly. Current numbers are either for the entire repo or for phase3 only [21:47:57] Although the phase3 backlog is still about 800 revs [21:48:06] where is chad in all of this? [21:48:08] RoanKattouw, I've just been doing it when i notice something new, and clearing whole paths [21:48:12] I think WMF should really, really lead the push on this [21:48:14] where they all are irelevant [21:48:27] That means freeing people like Chad from some of their responsibilities so they can focus on 1.17 more [21:48:37] yes [21:48:51] RoanKattouw: this is what Chad is supposed to be focused on at this point [21:48:54] Total number of results: 244 in /branches alone... RoanKattouw what do we do with "reviewing" 1.16wmf4? [21:48:59] Can I just mark all as deferred? [21:49:00] OK that's cool [21:49:08] Not really [21:49:16] According to the review guide, merges of OK things are OK [21:49:22] But let's not even care about that right nwo [21:50:06] I have 3 things that are critical - 1. 1.17 deployment in January (blocks ArticleFeedback for sure), 2. ArticleFeedback (scheduled release date needs to be met) 3. UploadWizard 0.2 ideally being released in January as well [21:50:17] Yes [21:50:34] I personally want WMF to put as many resources into 1.17 as they reasonably can [21:50:46] robla: If we have almost all of Chad's time for that, that'd be great [21:50:49] That said, the changes to ArticleFeedback's UI are pretty simple, I will be done with the new UI code next week [21:51:49] so I have about 1.5 weeks before I go on vacation where I can be multi-tasking between ArticleFeedback (start poking at new survey system, make design tweaks, fix browser bugs) [21:51:59] and 2 weeks when I return to also multi-task [21:52:32] I am going to start mapping out this stuff (like right now) [21:52:40] robla: are you around? [21:52:45] yup [21:53:04] what's chad's status? [21:53:07] we need his help [21:53:31] Code review is his top priority [21:54:03] he has some minor Pending Changes stuff, but that cna mostly be punted [21:54:24] We need to focus on 1.17 [21:54:38] I don't think we should be pushing it back at all [21:54:45] TrevorParscal: sure, but... [21:54:50] ? [21:55:31] is this a matter of focus or do we have some deeper problems we need to fix [21:55:54] such as? [21:56:01] oh.. sorry [21:56:03] read that wrong [21:56:14] i.e. is it just a matter of doubling down on code review, or do we need to rethink our process? [21:56:29] There's a lot of old-ish backlog to clean up [21:56:49] Chad can coordinate that I guess, and I'd be willing to work on it at well although RL is supposed to be my top priority [21:57:20] what's left to do on RL? [21:57:29] fix bugs! [21:57:31] Well it breaks Monobook in IE6 [21:57:34] There are some trivial bugs I think [21:57:42] More testing, test WMF-deployed extensions with it [21:58:04] I think the right way to go - as Roan has suggested - is to setup an RL prototype [21:58:18] then we can test against something more production-like [21:58:27] Yes [21:58:37] It could be the generic deployment prototype in fact [21:58:41] Once we branch 1.17wmf1 [21:58:42] isn't http://prototype.wikimedia.org/en.wikipedia.org/Main_Page that? [21:58:45] And I think that should be done soon [21:59:01] robla: Well we'd create prototype.wm.o/1.17/Main_Page or something [21:59:09] why? [21:59:26] well...I suppose if we're branching right away [21:59:31] Well once we branch that is [21:59:48] ok...so for now, trunk prototypes will work just fine? [22:00:09] Yeah for now it'd be a trunk prototype [22:00:12] Used mostly to test RL [22:00:21] Once we branch 1.17 it'd run the 1.17 branch [22:00:31] robla: how is the new installer going? [22:00:36] that's the other part here right? [22:00:57] yup, but I don't think that's a blocker really [22:01:14] Well it has to work right in order to be released [22:01:21] Although you're right it doesn't block WMF deployment [22:01:47] I'm assuming 1.17wmf1 will be our first focus, then we can pivot to 1.17 tarball release [22:01:48] yes, we need to deploy in January, but the release can follow later [22:02:16] so, what is keeping us from branching? some ResourceLoader bugs? [22:03:06] Bugs won't keep us from branching [22:03:13] Branching is so we can stabilize [22:03:27] right [22:03:37] Maybe, once we branch, we can assign Chad to be merger-in-chief or something? [22:03:37] so, as Bryan just said - can we just branch right now? [22:03:48] i.e. to merge bugfixes to the branch [22:03:49] I like that title [22:04:02] RoanKattouw: that's normally called release manager :p [22:04:09] Yeah and technically that's Tim [22:04:28] But Tim's availability isn't at the level where he can coordinate a release branch right now, AFIAK [22:04:46] What's Tim's status? [22:05:11] he appears to be away in the secure channel? [22:05:11] But Tim's availability isn't at the level where he can coordinate a release branch right now, AFIAK [22:05:15] Tim can probably do this, but Chad doing merge work would help [22:05:24] right [22:05:24] I'm not sure what Tim's status is [22:05:32] He works when I sleep [22:05:46] Well, Tim needs to be involved in this of course, but let's not just wait until he calls the shot [22:05:55] He needs to be involved for sure [22:06:06] We can deploy before we release, and we need his help with deployment [22:06:08] I'm just wondering whether he's available enough to be able to lead it [22:06:18] personally I think Tim is much more useful as coder and engineer than as manager [22:06:30] Bryan: he would probably agree with you [22:06:44] and we need him in an engineering capacity in this [22:06:51] primarily [22:06:56] Not manage people, manage code in this case :) [22:07:46] well, Tim should have the final say, but Roan and Chad should do the dirty work for him :) [22:08:06] I just realized the reason Tim's away is probably because it's Saturday morning for him [22:08:14] yeah, there's that problem [22:08:38] where on the wiki (is|should) the planning for deploying 1.17? [22:08:59] *robla looks [22:09:00] mediawiki.org for sure [22:09:14] for sure [22:09:16] could branch on Tuesday, then everybody has had their first office day of the week to do their say [22:09:33] Bryan: sounds sensible [22:09:38] Sounds good [22:10:13] Then we should announce this intention on wikitech-l like right now [22:10:19] yes [22:10:25] And roughly define the time on Tuesday [22:10:58] http://eiximenis.wikimedia.org/MediaWiki1-17Branch - draft time [22:11:18] it should be more imporrant to define the person who is going to do it than the time [22:11:44] http://www.mediawiki.org/wiki/MediaWiki_roadmap is where it should be parked [22:11:53] (at least, linked to from) [22:12:40] How about Chad and I do it together [22:12:48] so...btw, Chad's next two weeks are going to be pretty busy with finals [22:12:53] Hmpg [22:12:57] OK so how about I do it [22:13:00] thanks robla [22:13:32] Mid-afternoon UTC on Tuesday I guess [22:15:32] haha <3 Etherpad [22:19:41] btw....I'll note that Tim doesn't usually park on this channel, but he's on #mediawiki [22:21:07] hexmode: you around? [22:21:58] New in Extensions: 1314 [22:22:02] does this channel get logged the same way that #mediawiki does? [22:22:14] http://prototype.wikimedia.org/logs/%23wikimedia-dev/ [22:22:16] Topic says so [22:22:22] duh...thanks [22:23:25] oh....that's my excuse. my client is truncating the topic [22:24:07] all new.. Total number of results: 2326 [22:24:19] Tim should be here... when he's around... [22:24:29] Why wouldn't he idle here? [22:26:24] you'd have to ask him [22:28:14] well, maybe he's never been invited? [22:28:34] *TrevorParscal hopes we didn't exclude anyone here... [22:28:39] there are a few other presumably interested people not here [22:28:42] ialex, hashar [22:29:22] does anybody have a problem broadcasting the etherpad link in #mediawiki? [22:29:32] nope [22:29:41] Go ahead [22:33:38] sounds like a good idea to me [22:42:02] RoanKattouw: Reedy: jorm2_ http://www.mediawiki.org/wiki/Article_feedback/Public_Policy_Pilot/Tasks [22:43:23] TrevorParscal: Nice; but that's just your tasks :) [22:44:18] I'm still adding to it man! [22:44:44] :) [22:45:47] RoanKattouw, we could just add "Everything else" and assign that to TrevorParscal too... ;) [22:46:05] haha [22:46:41] Take that! [22:46:44] (see wiki page) [22:48:19] Total number of results: 1923 [22:48:25] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/status/new :) [22:48:45] ouch [22:50:00] Bryan, new 2,828 [22:50:07] at 20:35 [22:50:51]