[00:00:10] K4-713: that's code you wrote? [00:00:22] Probably true [00:01:11] AndyRussG: So... kind of. I had to take the extension and make it work with multiple payment processors. [00:01:18] When I got here, there was just one, really. [00:01:29] Ah interesting [00:01:40] And the first new processor, was globalcollect. [00:01:56] After I worked for several months on the EducationProgram extension, I felt any exchange I had with jeroen (the author) was completely one-sided [00:02:16] Since I had some deep knowledge about him but not vice versa [00:02:24] ...heh [00:03:25] Yeah... it's weird in DonationInterface. [00:04:02] ...or so I've heard. [00:04:08] It usually makes sense to me. [00:04:20] Hmmm [00:04:56] Well, for one thing, if you don't enjoy working with object inheritance, it's going to be just... awful. [00:05:06] AndyRussG: I'm not sure if u were kidding earlier, but allocations should have been 100% regardless of device... [00:05:24] It should be a distinction like bucket A vs B [00:05:32] awight|mtg: Ah hmm [00:07:59] apologies... [00:08:21] (it's been a long day...) [00:09:37] I guess first I should probably ask what that data is that indicates there's a problem? [00:12:15] Did u see the card? I gave my customary bad explanation there :) [00:12:34] I'm not sure there's a problem yet :) [00:12:39] awight|mtg: thanks, no not yet, lemme see [00:13:19] the issue is either * allocations in GlobalAllocation don't take device into account, and do not match reality (very likely) or * actual allocation algorithm is not accounting for device correctly (also likely) [00:13:28] We need to figure out which. [00:14:13] OK gotcha [00:14:17] or both ;) [00:14:33] I'm afraid it's the latter, though, because we're using BannerChooser directly in GlobalAllocations, and not normalizing the probabilities. [00:16:24] Right... yeah I have seen there is BannerChooser in there... What do you mean by normalizing probabilities? [00:17:27] I mean, AFAIR, nothing in GA takes the probabilities coming from the chooser and tries to force them to add up to 100% [00:17:39] so those should be the real probabilities of delivering banners [00:18:16] Ah right [00:18:35] That's a **load of campaigns (looking at link on card) [00:19:12] Hmm they all start with C14 and banners are all B14 [00:19:16] all enUS [00:19:24] and all _FR [00:19:50] And the variant part is the device... [00:20:23] mob is mobile... ipd is ipad, dsk is desktop, what does "lw" stand for? [00:20:28] no clue [00:20:43] OK I'll just try to imagine something ridiculous [00:20:55] last waterbuffalo [00:21:33] lewd wrestler [00:22:01] linguistics wombat [00:23:08] All the lw's are getting no allocations [00:25:38] * awight|mtg blinks back tears [00:25:45] those _lw's were not there a minute ago [00:25:48] All the lw's are priority high and the non-lw's are priority emergeny, that's why [00:27:03] we can just pretend the lw's are still not there [00:29:49] emergency? ugh, so rude [00:30:27] ok maybe they were there before, and I'm going crazy but not blind [00:30:38] awight|mtg: Maybe it's a foodmergency. [00:31:55] man all day every day is a foodmergency [00:32:42] I know. [00:32:47] That's life, I guess. [00:33:37] yeah... indeed, no foodmergencies in death [00:36:37] awight|mtg: neither I think http://meta.wikimedia.org/wiki/Special:CentralNoticeLogs [00:41:43] Okay... I'm running away. [00:42:00] Also, I'm not taking my computer home tonight, so for the love of Pete, don't break anything. [00:42:10] K promise :) [00:42:11] ciao [00:55:15] awight|mtg: it's a problem with how GlobalAllocation shows results, not with actual banner display [00:56:30] I ran 3 campaigns locally, each with a different banner, all on the same priority. First I activated every banner for all devices and randomly got one of the three served to my subscribing wiki. [00:57:01] Then I made 2 of the 3 banners for other devices (iphone and android) [00:57:22] And on my local desktop subscribing wiki, I got banner 1, i.e. campaign 1, 100% of the time [00:57:53] hrmph [00:58:06] can u dump the BannerChooser array for that? [00:58:15] So it doesn't look like users are getting short-changed on their banner provisions [00:58:35] Getting all the banners as expected [00:58:40] Dump? [00:58:48] Ah hmm OK [00:58:53] Yes [16:36:45] * K4-713 waves [16:38:29] hii [16:38:58] Hey ho [17:59:49] Hey guys... give me a second to see if I can grab the closet. [18:10:46] ejegg: Hey, would you mind moving our 1:1 up to... nowish? I don't want to lose the closet. [18:21:05] K4-713: oh shoot, missed your message! [18:21:11] 'sokay. [18:21:11] any time is good though [18:21:22] closet still free? [18:21:26] I'm still in there, yes. [18:21:30] *whew* [18:21:34] be back on in one sec [20:48:33] PPena: Would you do me a huge favor, and ping me on IRC if we get any new information from WP before 3pm? [20:48:40] I am likely to miss email activity. [20:48:50] K4-713 ABSOLUTELY [20:48:56] Thank you! :) [20:48:57] :) [21:02:33] (PS1) Awight: Matching gifts added to the English letter again [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/165873 [21:02:35] (PS1) Awight: less panic [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/165874 [21:04:53] (CR) Awight: "@the-wub: do you want 'class="plainlinks"' around this? Please send yrself a test thank-you to see the external link icon... Sorry I forg" [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/165873 (owner: Awight) [21:11:59] ejegg: Oh hey: Incoming. [21:12:01] (PS1) Katie Horn: Worldpay cleanup: Antifraud filters firing too much [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/165886 [21:13:39] Wow. [21:13:45] Thanks for the whitespace noise, netbeans. [21:17:41] taking a look! [21:18:58] I would suggest Eclipse, but I'm not sure it's any better [21:19:05] Eclipse blows. [21:19:10] ... [21:19:26] I mean: "I'm sure Eclipse is lovely." [21:19:30] >_> [21:19:36] Actually... [21:19:42] I used to use it a lot. [21:20:10] Until they got to Europa, and it just kept crashing on me during what are clearly going to be the most inopportune moments to have a crash. [21:20:50] My boss at the time (previous job), said: "Oh, well, see: Your problem is that you've somehow downloaded the European version." [21:21:02] bahaha! [21:21:10] He wasn't kidding, either. [21:21:46] It saddens me, on many levels, that he is probably gainfully employed somewhere. Right now. [21:22:21] Also... I really want to know what he thought would be, you know, different. [21:22:27] ...in Europe. [21:41:33] (CR) Ejegg: "looks good once you get past the whitespace, trivial comments inline" (3 comments) [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/165886 (owner: Katie Horn) [21:46:08] ejegg: ...for the nth time, I wish there were unit tests for my comments making sense. [21:46:23] :P [21:46:40] I realize both were copypasta [21:47:17] Guilty as charged. [21:47:30] Actually... [21:47:38] ...why not unit test for identical comments? [21:47:40] ha. [21:48:13] err, i guess you could do that! [21:48:39] That would catch my bonehead maneuvers 99% of the time, I think. [21:51:27] (CR) Katie Horn: "Yep; Thanks. Round 2 forthcoming." (2 comments) [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/165886 (owner: Katie Horn) [21:55:14] (PS2) Katie Horn: Worldpay cleanup: Antifraud filters firing too much [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/165886 [21:55:22] ejegg: ^^ [21:56:29] (CR) Ejegg: [C: 2] Worldpay cleanup: Antifraud filters firing too much [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/165886 (owner: Katie Horn) [21:56:44] good stuff! [21:56:47] (Merged) jenkins-bot: Worldpay cleanup: Antifraud filters firing too much [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/165886 (owner: Katie Horn) [21:56:53] Yaaaaay. [21:57:04] I think... yep, it's only 3pm here. I'm going to try to deploy this. [21:57:19] ejegg: Unless you want to do another deploy. :) [21:57:35] But, maybe I should push the button today. Jeff did just mess with the deploy scripts. [21:57:41] oh, I'll let you have this one! [21:57:48] hehe [21:57:54] Okay, here I go. [21:58:20] ejegg: Thanks for the quick review, there. [21:58:33] yw [21:59:56] it's almost alarming to look in the error log and see actual errors instead of amazon form junk [22:01:51] "Your branch is behind 'DI/deploy-payments_1.22' by 1140 commits, and can be fast-forwarded." [22:01:52] wtf [22:02:04] You guys have been busy. [22:02:06] ehh? [22:02:09] :p [22:02:18] ohh! [22:02:29] That was awight merging in master [22:02:47] it's actually mergable now! [22:02:55] Hurm. [22:03:20] and gerrit doesn't make you re-review each change if you just merge (or it worked for me like that once) [22:04:39] Hurm again. [22:04:45] Let's see... [22:05:05] It does give you the scary list when you say git review [22:05:28] but when you type 'yes' it only reviews the merge commit, I think [22:10:15] (PS1) Katie Horn: Composer packaging to make us available as a library [extensions/DonationInterface] (deploy-payments_1.22) - https://gerrit.wikimedia.org/r/165900 [22:10:17] (PS1) Katie Horn: fixups for unfork [extensions/DonationInterface] (deploy-payments_1.22) - https://gerrit.wikimedia.org/r/165901 [22:10:19] (PS1) Katie Horn: Worldpay cleanup: Antifraud filters firing too much [extensions/DonationInterface] (deploy-payments_1.22) - https://gerrit.wikimedia.org/r/165902 [22:10:34] ejegg: Hum. [22:10:43] Is that weird? [22:10:45] :p [22:11:15] oh, huh [22:11:19] I mean, it's... fine. [22:11:24] well, least it's just a few. [22:11:33] It's just if you expected it to do something different... [22:11:37] * K4-713 shrugs [22:11:51] (CR) Katie Horn: [C: 2] "deploy" [extensions/DonationInterface] (deploy-payments_1.22) - https://gerrit.wikimedia.org/r/165902 (owner: Katie Horn) [22:11:52] you did a git merge master? [22:12:03] (CR) Katie Horn: [C: 2] "deploy" [extensions/DonationInterface] (deploy-payments_1.22) - https://gerrit.wikimedia.org/r/165901 (owner: Katie Horn) [22:12:09] ejegg: Yep. [22:12:17] ok, i have no idea [22:12:20] (CR) Katie Horn: [C: 2] "deploy" [extensions/DonationInterface] (deploy-payments_1.22) - https://gerrit.wikimedia.org/r/165900 (owner: Katie Horn) [22:13:46] it would be awesome to be able to restrict command history search to commands run in the current directory [22:14:04] ejegg: ...yes. Yes it would. [22:14:38] .bash_history would need some annotation [22:24:21] Well, that was uneventful. I think. [22:24:36] !log updated payments wiki to 17f822a64742bd13e [22:24:41] Logged the message, Master [22:25:33] K4-713 summary: They are still waiting for the new set up to be ready. Will escalate it and see if that gets done tomorrow. Think due to the time of the year (close to code freeze) that team is overloaded... [22:25:42] K4-713 hopefully will have an update tomorrow [22:25:47] PPena: Got it. :) [22:25:51] So, guess what? [22:26:06] K4-713 you can forget that F1 question on card type- they were dreaming [22:26:10] K4-713 what?? [22:26:18] bahaha [22:26:38] PPena: The thing I was having you guess? I just pushed out the only changes we know we need to make on this end. [22:26:53] K4-713 thats cool-ahead of the game! [22:26:55] So... it's *possible* that when they get the accounts fixed, we'll be good to go right away. [22:27:12] What's more likely is that they'll unblock this one logjam, and we'll find another one behind it. :/ [22:27:23] ejegg: We're blowing up. [22:27:28] K4-713 ha [22:27:30] K4-713 gotch [22:27:34] K4-713 thanks anyway [22:27:44] ejegg: ...not to alarm anyone, but... exchange rates failmail avalanche? [22:27:49] PPena: Sure. :) [22:27:51] K4-713: sorry, that was my fault [22:28:04] Ran the backfill script on dev without OANDA credentials [22:28:13] nothing to do with your deploy [22:28:20] Excellent timing, though. [22:28:25] heh :P [22:28:31] You should do that on purpose next time. [22:28:46] * K4-713 files that under "Hazing rituals" [22:29:00] when it's on purpose, i'll choose something more plausibly related [22:29:35] Yeah, I'd totally know that was unrelated. [22:29:46] But if, say, amazon seemed to explode... [22:30:29] teehee [22:32:35] ok, let me try that for just one day [22:32:57] I'm sure there will be a *really good* time to pull that one. [22:33:26] My goodness, these logs are quiet. [22:33:28] cool, no new failmail. [22:33:29] :) [22:51:08] ok, going to meet some folks for dinner, but i'll check email in case of delayed kaboom [22:51:41] ejegg|dinnr: Have a good one! [22:51:51] thanks!