[06:30:19] is there a brion (it's not an emergency whatsoever)? [17:17:02] RoanKattouw: howdy [17:17:10] Ohai [17:17:25] So I guess I'll be in the meeting after all [17:17:32] Grabbing dinner now-ish [17:17:41] good stuff [17:18:06] I will be calling in too, I'm at home [17:18:26] (Been wrestling with a change Jan Paul made to our raytracer that broke the default camera model because the defaults weren't sane and (0, 0, 0).normalize() == (NaN, NaN, NaN) ) [17:18:40] Fun stuff to do when you're on a train with no wifi anywya [17:18:44] ha ha [17:18:57] are you dividing be zero again Roan? [17:19:15] Dividing a float by zero is fine [17:19:25] a/0 == +/- Infinity as long as a != 0 [17:19:30] 0/0 == NaN however [17:19:41] so what is 0.0/0 [17:19:43] indeed [17:20:02] And yeah normalize() was implemented as return (*this) / length(); without regard for the zero case [17:20:27] sounds like trouble waiting to happen [17:22:56] Alright, 30 mins till meeting, dinner [17:59:17] RoanKattouw_away, back from dinner? [17:59:39] we're on x2003 [18:01:50] Coming [18:13:30] m??f [20:08:37] TrevorParscal: Ping. Deploying your AFT patches to test now [20:08:53] cool [20:08:56] looking [20:09:06] Hold on, not there et [20:09:08] *yet [20:09:11] :) [20:09:12] ok [20:09:18] Do you realize you're marked as away? [20:10:18] OK deployed [20:10:43] I guess I'll have to clear the message blob tables, hrm [20:15:37] oddly, I'm not seeing cookies being set for pitches, nor am I getting pitches.. [20:16:19] hmm [20:16:25] chrome's dev tools suck sometimes [20:17:02] good old firefox, you never let me down.. unless you are rendering to canvas... then you suck horribly [20:17:21] Firebug is one of the few reasons I've stuck with Firefox [20:17:39] "Did you know that you can edit this page?" [20:17:42] Getting a pitch just fine [20:17:52] (I manually put in a "hide survey" cookie thuogh) [20:18:48] yeah, chrome hides cookies under certain unknown circumstances [20:19:07] Use $.cookie in the console then [20:23:08] hmm [20:25:00] yeah [20:25:09] checked them, deleted them, saw the tool reset them [20:25:17] but chrome doesn't show them [20:25:20] so, boo on chrome [20:27:54] OK [20:28:01] Are the changes OK? [20:28:43] double checking the pitches [20:29:15] looks good [20:29:31] not that it was that big of a code change or anything :) [20:29:58] heh no [20:30:01] OK I'll push it [20:32:11] TrevorParscal: BTW how far did you get with Narayam? [20:32:25] As in, is it deployable? Do you have local changes that would make it so? [20:32:34] well, I was messing with the CSS and caught myself before I got too carried away [20:32:49] OK [20:32:59] it's stable, but needs some UI help [20:33:02] (AFT changes pushed now, they'll take the usual 5 mins to propagate) [20:33:04] if we aren't "allowed" to work on this [20:33:14] then I can at least provide some input to whoever is [20:33:18] It's something that can be improved later, right? [20:33:25] afaict, yes [20:33:28] Didn't you also have local changes for renaming stuff? [20:33:40] no [20:33:42] i tossed them [20:33:52] *aude wishes there was an easy option to clone pages on wikipedia [20:33:53] are any of the dialects rtl? [20:34:11] i didn't test all languages [20:34:19] I tested a few [20:34:22] but none were RTL [20:34:27] so I didn't do any RTL testing [20:35:10] We have no keymaps for RTL langs currently, so it's not an acute issue [20:35:28] Other than maybe for an RTL contlang and an LTR userlang [20:35:54] But given the extension's use case that doesn't really make sense [20:36:26] right [20:36:32] it's all about contlang [20:36:41] Alright, so I'll schedule it for deployment tomorrow then [20:36:41] I actually think it should be triggered on contlang, not userlang [20:36:45] cool [20:37:00] It triggers on either [20:37:04] I need to get the patch for showing expiry in the UI for AFT done today so we can do that [20:37:09] ah - ok [20:37:23] Basically it's intersect( languages with keymaps, array( contlang, userlang ) ) [20:37:33] nice [20:37:38] I am going to bed nowish [20:37:38] ok, I missed that when looking it over [20:37:42] I was mostly looking at JS [20:37:44] ok [20:37:48] Yeah that part is in PHP [20:37:56] get some sleep - I will be working at 7am again [20:38:01] So I will schedule deployment of Narayam for tomorrow [20:38:02] we can deploy then [20:38:17] is there anyone else who has to buy in on that? [20:38:17] OK but I'll have to leave before 9 [20:38:22] understood [20:38:26] both should be simple to do [20:38:38] On deploying Narayam? Not really. Ext is good per you and me, community wants it, so go [20:38:49] Just on the Malayam wikis (IIRC) as a start [20:39:08] *RoanKattouw fires up Gmaps to estimate when he has to leave [20:39:50] OK I guess I'll aim to leave at 8:30am PDT [20:39:58] ok [20:39:59] It's a 1-hour drive apparently [20:40:14] well, in the case that we don't get it all done in the am [20:40:27] will you be around pm at all? [20:40:31] No [20:40:40] I might pop in around 2 [20:40:45] k [20:40:49] good to know [20:40:50] But not around for practical purposes [20:41:05] I guess I'll deploy Narayam earlier then [20:41:12] While you're asleep [20:41:23] And deploy your AFT patch around 7-7:30, whenever you show up [20:41:27] k [20:41:31] good stuff [20:41:50] GerardM-: Per the above, I will deploy Narayam some time tomorrow afternoon (CET); I'll look at the calendar now and schedule a time [20:45:30] Scheduled for 1pm CET / 5am PDT (4:30pm IST?) [20:51:23] Cool [21:10:08] Ryan_Lane, tomaszf: http://rt.wikimedia.org/Ticket/Display.html?id=674 [21:31:09] Reedy: how's it going [21:31:14] i have an AFT question [21:31:47] TrevorParscal, shoot [21:31:52] it looks like we never actually made ratings expire in the totals [21:32:02] like, there's a stale flag in the query API [21:32:07] based on number of revisions [21:32:09] Indeed [21:32:17] but the totals always count everything forever [21:32:20] The stale was based on rev count since last count? [21:32:39] number of revs since when you rated last [21:32:48] ya [21:32:59] the design no longer needs a stale state [21:33:03] it needs an expired one though [21:33:16] hmm [21:33:24] and the expired thing seems to have never been implemented in the old tool either [21:33:33] so we'll do sum ratings where date > now - arbitary value [21:33:57] hmm [21:34:21] well it's supposed to be that we only include ratings which have a rev number within 30 of the current one [21:34:35] I just wanted to check that I'm not taking crazy pills [21:35:01] because brandon seemed to think it was included in the old version - or at least I think that's what he was saying [21:35:08] and while I have seen designs that call for it [21:35:17] I see not code that actually implements it [21:35:25] I was hoping you might confirm this [21:35:39] before I go and reinvent the rating expiration wheel [21:36:08] Errm [21:36:10] Let's have a loook [21:36:19] (baring in mind it's probably 6 months since i touched the code) [21:38:28] if ( $res && $dbr->numRows( $res ) > $wgArticleFeedbackStaleCount ) { [21:38:28] //it's stale! [21:38:47] That's set at 5 actually [21:38:58] yeah, that was the "stale" feature [21:39:02] we don't need that anymore [21:39:25] *Reedy mumbles something about customers sucking [21:39:27] we are getting rid of that state, it was making it too complex [21:39:31] ha ha [21:39:33] seriously [21:39:37] Seriously [21:39:40] 24 week project [21:39:42] 19 weeks in [21:39:46] No signed specification [21:39:47] Fu----- [21:39:57] ha ha [21:40:13] And they have the goalposts on some moving walkway or something [21:40:14] Anyway [21:40:59] AFAIK that's the closest we got to "expired" [21:41:58] ok [21:42:04] Surely you could just move it around with some sort of date related filter.. [21:42:15] Well, the spec is 30 revisions [21:42:23] Oh [21:42:24] well, the spec also calls for all sorts of other calculations [21:42:34] but, the minimum is 30 revisions afaict [21:42:35] Change "stale" -> "expired"? :D [21:42:42] yes [21:42:44] that's 1 thing [21:42:53] then I have to also exclude the expired ratings from the totals [21:43:01] which is far from simple [21:43:15] because the aggregates are calculated on write [21:43:20] i think... [21:44:43] Yeah [21:48:37] hi folks... GerardM- and I are discussing the genderized user namespace issue. the code is apparently already checked into trunk, but I'm wondering if anyone knows off the top of their head which rev it is [21:48:57] *robla pokes around himself [21:50:31] hexmode: bug 28052, based on the title, looks like it might be the right place to keep track of it, but that bug looks like more of a general bug about all gender localization issues rather than being specific to this one issue [21:50:41] siebrand told me that it is in trunk.. ithe code is likely by Nikerabbit [21:50:49] r82029 btw (if you havn't found it yet) [21:51:15] bawolff: awesome, thanks [21:52:24] ah...17160 was really the bug number I was after [21:54:01] GerardM-: this looks like it's on track to go into 1.18 already [21:55:51] When the software goes into 1.18 the localisations are needed to be deployed. [21:56:11] When I am to enthuse people it does not work when it goes online in 6 months [21:56:31] so it is best when there are a few moments when they can go live [21:56:50] and the objective is to be complete at the 1.18 tarball release [21:58:13] GerardM-: what are you asking for? [21:58:49] three releases of implementing namespaces for specific languages [21:59:30] one as early as possible after the 1.17 tarball release and one after two months and one prior to the 1.18 tarball release [22:01:00] we will have to branch 1.18 from trunk just after the 1.17 release [22:01:20] will there be enough sizeable changes though? [22:01:23] then review the changes in 1.18 and once finished makes it 1.18wmf1 for live deployment [22:01:36] that would be the first release [22:02:16] if the site run 1.18 we could then deploy i18n fastly :b [22:02:43] Reedy: what do you mean by "sizeable changes" ? [22:02:53] Enough to warrant a release [22:02:54] Reedy: do you mean enough new features? [22:02:57] when people see that things go live, we can REALLY get them going [22:02:58] sort of, ish [22:03:07] I mean, I'm not suggesting we wait 6 months or anything [22:03:17] Reedy: I don't think that should be a factor [22:03:24] but if there's nothing really changed, releasing a 1.18 seems daft [22:03:41] I thought we had a consensus to release often (3 or 4 releases per year) regardless of the # of new bugs^H^H^H^Hfeatures released [22:03:51] we have a feature...gender-correct namespaces :) [22:03:57] Sure, but we need to tie down those specifics [22:04:00] I know the RELEASE-NOTES for trunk is already getting large [22:04:14] But it'd be crappy to go, oh, 1.18 is out, upgrade if you want [22:04:22] As people will ask "what's in it for me?" [22:04:39] "too much stuff to tell" [22:04:40] :) [22:04:59] Reedy: I think I would love to have that problem at this point [22:05:02] We want women to be more involved [22:05:33] GerardM-: my wife does not want to fill the gender field because she does not want to get annoyed online [22:05:33] the theory is that this will make them notice that we take them seriously by addressing them properly according to the rules of their language [22:05:40] GerardM-: ;b [22:05:45] it is a choice [22:05:46] Do people actually not edit because of gender issues in messages [22:05:51] yes [22:05:59] I would still edit if the software called me a she [22:06:08] I mean, sure, we shouldn't be getting as far as 1.17 was in such a huge leap [22:06:39] well it is too late for me to talk about an 1.18 release. But we should surely have it branched once 1.17 tarball is released [22:06:52] *Reedy shrugs [22:06:56] Maybe not quite so [22:07:04] But certainly looking towards doing so [22:07:15] Depends how long it takes to stabilise 1.17 too.... [22:08:31] given the 8months code took 3 or 4 months to review, 1 month to deploy and will probably take one month to stabilize on live ... [22:08:39] Hmmm :P [22:08:39] we should branch asap [22:09:01] (and we havent released 1.17 yet, which will come with more bugs!!) :p [22:09:07] heh [22:09:37] thus 1.18 will probably be easier to deploy/stablize/release [22:09:42] and 1.19 will be rocket mw [22:10:06] then we can take a 2 years lag to rewrite the parser under Brion guidance and release ... [22:10:12] MEDIAWIKI TWO DOT ZERO !!!!!!! [22:11:23] well, check out http://toolserver.org/~robla/crstats/crstats.trunkphase3.html [22:12:04] prior to branching, we were down to 500 unreviewed in trunk/phase3 (from a peak of almost 1500) [22:12:13] we're already back up to 750 [22:12:29] neil raised that issue on the list [22:12:39] we do post commit review when we should probably do pre commit review [22:12:52] let me find it [22:13:04] I know what you're talking about [22:13:06] (at least it is not raising) [22:14:19] it's drifting up [22:15:01] MEDIAWIKI TWO THOUSAND [22:15:24] :))) [22:15:40] robla: I am planning to take a week off in april to review stuff [22:15:53] I might be able to get 250 or 300 less :b [22:16:19] heading bed, sorry for the rant :( [22:16:41] that was a rant? [22:17:04] lol [22:17:18] The pre-commit/post-commit review isn't something we can do really on svn [22:17:27] we'll just end up with a load of patches on bz etc [22:17:31] so the git model works a lot better there [22:17:54] yup [22:18:18] but it may play havoc with translatewiki.net [22:18:25] so that's a "tough until we get there" [22:18:51] anyway ... my bottom line question, when can the first genderised namespaces be created. [22:22:30] because that decides when I can start enthusing people to work on the support for gender in the localisation of their language. [22:30:44] brion: Who plays Michael Might, riding the MITT ? [22:30:58] MediaWiki Industries Two Thousand [22:31:14] well heck [22:31:19] now i totally want to hire david hasselhoff [22:31:43] need to write a job description [22:32:15] Atleast we've got a role for FLAG/Foundation already :D (Foundation for law and Government) [22:32:33] Actually Wikimedia is much like the Knight Rider franchise :P [22:32:40] Who's Bonny though ;-) [22:33:41] Edward Mulhare as Jimbo xD [22:36:26] lol, MediaWiki Industries Two Thousand [22:36:37] [[ MITT ]] [22:38:12] A shadowy flight into the dangerous world of a man... who does not exist.