[01:02:33] drdee: i hear there was some dutch mail? what's the latest? [01:02:58] i am still plowing through my email, haven't' responded yet [01:03:05] * jeremyb_ took the bus with some people from noord holland today. [01:03:18] ok, i would have asked maarten but he's not on now [01:05:13] nice! [13:03:00] mooooorning [13:08:04] morning ottomata [13:09:31] morning [13:10:56] btw, RobH is now looking for a eqiad stat1 replacement :p [13:12:49] nice [13:50:12] morning [13:50:49] merrrnin [13:51:40] mnng [14:01:17] ng [14:01:36] g [14:06:56] - [14:11:28] [14:14:04] milimetric: how hard is the limn bug ? [14:14:14] average_drifter: got a new report? [14:14:18] brb [14:14:27] hard for me, dschoon says it's easy though [14:14:54] it's because I'm not sure what got added to the object that now has a cyclic reference, so I'd have to look through git history or all the properties one by one [14:15:16] dschoon just probably knows since he added it [14:24:38] back in 15ish [14:43:26] brb, restarting [15:03:52] average_drifter: ping [15:06:53] pong, not yet [15:47:05] New patchset: JGonera; "Add Rendering time graph (in new Performance tab)" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/58305 [15:47:29] gm drdee! [15:48:06] Could we find some time to chat about which cards we'll plan for next sprint later today? [15:48:59] yessir! [15:51:00] At ease. ;) [15:51:58] Also made some comments on your Leader Standard Work from yesterday. If you have a moment to talk now before I chat w milimetric, that's great, otherwise we'll chat later. [15:52:38] let's do it now [15:53:24] ok [15:53:52] drdee: https://plus.google.com/hangouts/_/2da993a9acec7936399e9d78d13bf7ec0c0afdbc [16:02:25] kraigparkinson: meeting? [16:02:29] yep omw [16:02:35] New review: Yuvipanda; "Folders should be independent and self contained. This makes it easy to have dependencies of config ..." [analytics/limn-mobile-data] (master) C: -2; - https://gerrit.wikimedia.org/r/58246 [16:05:49] New review: Yuvipanda; "LGTM but the SQL doesn't account for non-consecutive days. But that's okay, I guess - I suppose ther..." [analytics/limn-mobile-data] (master) C: 1; - https://gerrit.wikimedia.org/r/58305 [16:16:01] New review: Ori.livneh; "Note that this will grab both desktop and mobile numbers. If you want only mobile numbers, you want ..." [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/58305 [16:20:40] New review: Jdlrobson; "probably also want to limit to stable only to be useful?" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/58305 [16:21:21] New review: Ori.livneh; "> probably also want to limit to stable only to be useful?" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/58305 [17:00:53] poke average_drifter: scrum [18:30:25] ping erosen [18:30:30] yo yo [18:30:32] ready? [18:30:35] yep [18:30:37] hangout! [18:30:39] woo! [18:31:05] https://plus.google.com/hangouts/_/a838646099396288ed3c328afe579a90585d1aab [18:31:35] erosen ^^ [18:31:40] working on it [19:45:42] milimetric, erosen: thx for putting that list together! [19:46:01] no prob [19:46:18] we went through each item, so the lack of items on the list is a good sign [19:47:31] question is: what's the next step :) [19:49:14] drdee: milimetric and I are gonna start working on a few mingle cards on Friday [19:49:42] yeah, mingle cards are the priorities [19:49:42] the two for april 15: 430 and 429 [19:49:51] and the list just identifies stuff that we want to keep an eye [19:49:56] ack [19:49:57] on [19:50:03] and I might start working on 388 [19:50:07] drdee: ? [19:50:43] looking [19:51:11] yeah totally! [19:51:27] drdee: you mean you don't have all the mingle card #s in memory? [19:51:58] i was looking in my brain [19:52:16] but my random read access time is a couple of seconds [19:52:20] i see [19:52:21] because my brain is vst [19:52:32] vst = vast [19:52:49] i was like Visual Studio? [19:53:10] with autocomplete :) [19:54:54] off to lunch, fyi [20:17:40] New review: JGonera; "OK, I see your concern, I'll change it." [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/58246 [20:34:08] ping average_drifter [20:37:47] pong [20:37:55] drdee: ^^ [20:38:44] what's the status of adding support for the final two edge cases? [20:39:07] it's in progress [20:40:08] do you need help? [20:40:13] brain bounce? [20:40:43] well I still have some code to write uhm [20:40:52] here is a small representation of the regex used [20:40:53] http://i.imgur.com/tIuTi9P.png [20:41:02] it will be used inside the documentation [20:41:31] the groups #1 up to #8 are the captures [20:42:28] it was generated with regexper.com so I didn't spend much time on it [20:42:35] milimetric: can you have a look at the url that aveage_drifter pasted? [20:43:16] looking [20:43:25] holy $%^&*( [20:43:34] that is awesome! [20:43:42] :) [20:44:18] so you should have page=|titles= or something like that for the title capture right? [20:44:33] those are present too, but not in the big regex [20:44:51] but they are present in the code [20:48:54] I'm not 100% regexes are the solution but for what we need at this moment .. they are [20:49:33] if it turns out that we need more complex logic, we'll drop regexes and we'll use a module which parses URLs into [20:49:49] domain, path fragments, url key/value pairs [20:49:54] protocol [20:52:36] i've got two questions about this diagram, WHICH DOES ROCK! [20:52:57] go ahead [20:53:00] is group #3 optional? [20:53:09] eys [20:53:12] *yes [20:53:12] ok [20:53:37] then the ".m." shouldn't that be ".m. || .zero. " [20:53:43] erosen ^^ [20:53:54] can you also look at that diagram [20:54:09] I'm not sure exactly what the right solution is either. Diederik went the way you're describing with Java (module parsing the URL into components) [20:54:09] it's super verbose and hard to follow. I think it's worth it to do a spike and talk about possible solutions [20:55:13] open a card milimetric :D [21:08:14] erosen: http://i.imgur.com/tIuTi9P.png [21:08:16] erosen: welcome back [21:08:25] erosen: 23:53 < drdee> then the ".m." shouldn't that be ".m. || .zero. " [21:08:35] erosen: 23:53 < drdee> can you also look at that diagram [21:08:56] * erosen is looking [21:10:52] average_drifter: what is the context for this exactly? [21:11:18] average_drifter: that is, what do you do with things that match this regex (I assume this is a regex diagram?) [21:11:25] ok, outttyyyy latas [21:11:49] erosen: the captures have different meanings [21:12:06] erosen: one of the major things in the pageview definition is how the url field relates to the referer field [21:12:28] erosen: we found that out more clearly yesterday from feedback from the mobile team [21:12:51] erosen: so after the regex is applied to both url and referer we can compare the captures of both and find out if it's a pageview or not [21:13:19] i see [21:13:53] erosen: just forwarded you an e-mail with some concrete examples [21:16:29] average_drifter: i'm not sure how to comment on this yet [21:17:07] average_drifter: is it safe to say that the zero site is identical to the mobile site from a page tracking persepective? [21:17:13] pageview* [21:18:03] erosen: the only thing that differs is .zero. instead of .m. [21:18:05] so imho, yes [21:18:29] cool, then it seems reasonable to count them together [21:19:19] ok [21:19:31] average_drifter: also, it might be worth noting that this regex won't perfectly match some '"projects", like meta, species, and a few others [21:19:45] they look like http://species.m.wikimedia.org/ [21:20:01] not sure if that matters [21:21:03] oh, wasn't counting those [21:21:06] they can be easily added [21:22:24] erosen: http://stats.wikimedia.org/EN/TablesPageViewsMonthlyMobile.htm [21:22:35] erosen: Erik Zachte doesn't count them [21:23:06] k [21:23:06] erosen , drdee should we count "meta" and "species" ?