[00:00:39] hey ejegg - would you mind sending the LYBUNT deployment email to major gifts? i think they'll be super stoked and it'll help them a lot starting like, right now [00:00:58] Sure, will do! Sorry, got sidetracked [00:01:11] no worries, just wanted to make sure it didn't fall off the radar :) [00:01:20] is there an mg list? [00:01:47] i think so? but fr-all is probably ok/better anyay [00:02:39] Oh no! I've been bifurcated. [00:03:25] ejegg: Were you able to figure out what happened with the wp form in the last deploy? [00:03:59] yep, sent you an email with a workaround patch [00:04:34] we should fix form chooser at some point, but a tiny settings change will get us back up for now [00:04:39] ejegg: In that case, can you review the legal language changes? [00:04:43] Oh good. I can... [00:04:59] ok, will review your patch [00:05:36] ejegg: Sweet! If you're going to deploy this other thing, we should get the english changes in there too. [00:05:52] OK. I actually have to take off in a few. [00:06:09] I'll get that CR in first though [00:06:18] happy trapeze-ing ejegg!! [00:06:25] thanks! [00:06:58] hey atgo Megan and I are trying to plan email stuff [00:07:11] We were just almost talking about that. [00:07:14] hi ccogdill! yeah that's what K4-713 and ejegg are working on :P [00:07:14] can we send tomorrow morning? what’s the timeline for gc? [00:07:26] Probably not morning. [00:07:45] ccogdill: is it possible to plan on the afternoon? [00:08:00] 300,000 emails on a friday afternoon isn’t great [00:08:15] yeah, we have a lot of ground to make up from losing these past 2 days [00:09:25] (PS3) Ssmith: Show actual SQL instead of query string in db modal [wikimedia/fundraising/dash] - https://gerrit.wikimedia.org/r/173632 [00:13:10] (CR) Ssmith: [C: 2] Add debug command line switch [wikimedia/fundraising/dash] - https://gerrit.wikimedia.org/r/173425 (owner: Ejegg) [00:13:17] Yeah, I think we all know this isn't ideal. [00:13:21] (Merged) jenkins-bot: Add debug command line switch [wikimedia/fundraising/dash] - https://gerrit.wikimedia.org/r/173425 (owner: Ejegg) [00:15:06] (CR) Ejegg: [C: 2] Changing legal text in en [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/174837 (owner: Katie Horn) [00:15:37] ejegg: I'm sorry, I appear to have lost my brain again. What patch did we need to get in to fix the form issues for WP? [00:15:49] Is it in the formsettings? [00:15:53] https://gerrit.wikimedia.org/r/174845 [00:15:55] yep [00:15:57] AHA. [00:16:03] That's why it's not in DonationInterface. [00:16:10] I feel less bad now. [00:16:15] just need to set a selection_weight on something or formChooser goes mad with indecision [00:16:24] I know the feeling. [00:16:44] ok, I gotta run, but I'll keep an eye on email [00:16:44] But... that's still weird, isn't it? [00:16:52] yeah, FormChooser needs fixing [00:17:03] but I figured settings workaround for now [00:17:06] I thought if you specified a ffname, it would select even if the selection weight is zero. [00:17:14] In fact, that's what selection weight = 0 is for. [00:17:47] oh phooey, i really need to run [00:17:57] Yeah, shoot. I'll look at it. [00:18:01] i'll be checking email though [00:18:08] All the "test" forms should be set to zero so they don't get chosen. [00:18:08] see ya later [00:18:13] late [00:18:21] yeah, they still should be in that patch [00:18:32] bc of line 503 [00:18:45] Oh, yeah. Okay. [00:20:35] so K4-713 are you recommending that ccogdill and meganhernandez not send email tomorrow? [00:20:40] or ccogdill and meganhernandez when do you need to know? [00:21:00] Well, the bug we need to solve to not break WP when we update the language, is still outstanding. [00:21:24] Still looking at that. [00:21:25] ahhh [00:21:50] If I can get this all deployed tonight, there won't be a problem. [00:22:00] What language are the emails going to be in? [00:22:06] english, just US [00:22:09] okay [00:22:12] so the legal text for GC is all we need [00:22:16] Can I let you know in about three hours? [00:23:28] let me know if you guys need anything - i've got to run [00:23:31] i'll be on email though [00:23:53] sure… I guess what we don’t understand is is this worldpay work related to the legal text? [00:24:03] I *think* getting the US back up is the most important thing [00:24:18] no... but if you want to be up in france WP also needs to get fixed [00:24:43] and K4-713 correc tme if i'm wrong, but thye're in the same world so deploying the 2 together is nic [00:24:49] you mean the world pay error from earlier today or the world pay legal text? [00:24:52] typing. [00:24:53] (Merged) jenkins-bot: Changing legal text in en [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/174837 (owner: Katie Horn) [00:25:05] It's actually going to be a huge PITA not to do both at this point. [00:25:36] lmk if you guys need anything. i'm out of IRC [00:40:39] ccogdill: The issue is that ejegg had to un-deploy something earlier. [00:40:55] for worldpay, right? [00:41:02] Yep. [00:41:48] So, fixing that and going forward is a ton easier than trying to shave all that off, go backwards, go forwards again with the new thing we want, and then redo all the stuff we skipped for the WP issue. [00:41:53] Does that make more or less sense? [00:42:45] more or less… do you mind if Megan and I come find you? :) [00:42:51] Sure, go ahead. [00:42:57] k we’ll be right down [00:43:00] But, I'm eyeball deep in trying to get this done for tomorrow... [00:46:30] OK, just let us know if you think it’s something that would happen today or tomorrow / if you know what time tomorrow so we can plan the send and banner tests. We have the English and French legal text. I’ll reply to the thread with that. [00:46:52] Yeah, I'm getting that you're very interested in these changes. :) [00:47:53] can someone share the URL for the first staging site for civi? I seem to only have a link to the 2nd one [00:48:30] I don’t know about first and second, but here’s the one I have Caitlin_Virtue https://civicrm.frdev.wikimedia.org/?q=user [00:49:27] aha yes, thanks!! [00:52:03] (PS1) Katie Horn: Merge branch 'master' into deploy [extensions/DonationInterface] (deployment) - https://gerrit.wikimedia.org/r/174882 [00:52:41] (CR) Katie Horn: [C: 2] "+2 for deploy" [extensions/DonationInterface] (deployment) - https://gerrit.wikimedia.org/r/174882 (owner: Katie Horn) [00:52:50] (CR) Katie Horn: [V: 2] "+2 for deploy" [extensions/DonationInterface] (deployment) - https://gerrit.wikimedia.org/r/174882 (owner: Katie Horn) [01:02:53] Oh fer... [01:03:00] >:[ [01:14:57] ccogdill: I'm trying the deploy again now. [01:15:04] ooh! [01:16:46] so this is the WP thing so we can then focus on the legal text? [01:16:50] or both? [01:16:54] or neither? :p [01:19:13] !log updated payments to 4d6afa865b5e8 [01:19:19] Logged the message, Master [01:19:26] ccogdill: Both. And it appears to be done, without breaking anything. So far. [01:19:39] Now I get to stare at it for an hour, though. [01:19:49] whoa! that’s awesome k4-713 [01:20:13] wow go K4-713 [01:20:19] :) [01:20:36] My team left some cool deploy traps for me, too. [01:20:41] Remind me to thank them in the morning. [01:20:50] But, it should be okay now. Want to help verify? [01:21:07] sure, how can we help test? [01:21:19] Well, first of all, making sure the legal text in English is okay. [01:22:17] sorry, i just realized what i was asking [01:22:19] it is! [01:22:25] And that the worldpay forms you want to use, are loadable. :) [01:22:34] https://payments.wikimedia.org/index.php?title=Special:GlobalCollectGateway&appeal=JimmyQuote&ffname=cc-vmad&recurring=&payment_method=cc&utm_source=.default~default~default~default~control.cc&utm_medium=&utm_campaign=&utm_key=&referrer=&language=en&country=US&returnto=Thank_You%2Fen&amountGiven=¤cy_code=USD&frequency=onetime&amount=3&uselang=en - legal text is right [01:23:25] k4-713 it’s not updated in French yet [01:23:27] we tried WP and GC [01:23:40] but I think US was the top prioritiy for this... [01:24:00] Yeah, that would be sort of impossible at this point. I don't think anybody sent me the french text, so I thought it was going to come through translate wiki. [01:24:05] ohhhhh [01:24:20] Megan sent it like 20 minutes ago so that makes sense [01:24:26] Aw, nuts. [01:24:44] yeah, that was after my last team member able to review those things took off, too. [01:24:49] yeah [01:25:01] it’s okay for now [01:25:16] thank you! [01:25:35] Sure. Do you know if it would be helpful for me to send the label names to Jessica? [01:26:14] Oh I’m not positive how that works to be honest. I thought just emailing you the text would do it. [01:26:23] heh [01:26:42] So, there are a couple ways this can work. [01:27:17] But, we sort of took off down one road a few hours ago. [01:34:18] Is there something we should get jessica to do? would that help? [01:42:18] meganhernandez: Sorry; conversation hijacked. [01:42:34] I'll send an email to Jessica about what label we should push on. [01:49:28] meganhernandez: email update sent. [01:58:31] ejegg|away: back! :) [01:59:47] * K4-713 blinks [02:00:31] thanks K4-713 [02:00:49] so are we good to put english banners back up? and blast off a bunch of emails tomorrow? [02:02:13] Yep! If it's English. [02:02:33] Maybe more when we get more translations through. [02:02:56] I'd certainly like to update more before the weekend if they get in there. [02:09:35] K4-713: the email tests tomorrow are gonna be so interesting [02:09:39] 15 tests! [02:09:43] Woah. [02:10:06] When are you going to fire those off? [02:14:56] 7am K4-713 [02:18:26] meganhernandez: Cool. I'll let you know immediately if alarms start going off in the middle of the night. [02:18:27] ...again. [02:18:40] Actually, you know what? [02:19:00] what’s that? [02:19:04] I'm marking the week before thanksgiving on the calendar for next year. This happens every time. [02:19:29] which happens? payment processor weirdness? or legal? or... [02:19:49] It's the payment processors doing *exactly what we're doing* in preparation for OMG Thanksgiving Shopping Weekend!!!1one1!1. [02:23:20] I'm putting a giant alarm on my calendar with reminders a week before. [02:27:13] meganhernandez: Okay, I've stared at the logs for the requisite hour, and everything seems solid. Going home now. :) [02:27:24] Please do let me know if anything else comes up before the email frenzy. [02:27:36] ok, thank you for making that happen! [02:27:49] i’ll get the english banners back up in just a bit [02:28:00] Cool. Good night! [02:28:03] please get some rest, night! [10:23:21] (PS5) AndyRussG: Use per-campaign buckets, with smooth transition [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/174581 [11:26:22] (PS6) AndyRussG: Use per-campaign buckets, with smooth transition [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/174581 [12:05:08] (CR) AndyRussG: "Smoke tested on desktop and mobile, seems to work as advertised! :)" [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/174581 (owner: AndyRussG) [14:13:25] Electrician is gonna turn off my electricity to fix some wires, shouldn't take longer than 15 minutes, he says [15:56:08] (PS1) AndyRussG: WIP Schema change for bucket reshuffle date [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/174958 [15:56:59] hey ejegg :) [15:58:35] (CR) Ejegg: "code's looking very good so far, just concerned with adding tons of possibly irrelevant campaigns to cookie" (2 comments) [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/174581 (owner: AndyRussG) [15:59:30] AndyRussG: any chance we can filter a bit before storing buckets in the cookie? [16:00:02] Hi ejegg, thanks! Uh, hmmm [16:00:10] I'd hate to keep details for campaigns with only ipad banners in desktop user cookies around the world [16:00:25] Ah I see [16:00:46] Good point! Hadn't thought of that [16:00:50] Yeah we can definitely do that [16:01:03] wonder if maybe the filter function could be called in two passes, one for geo and device, one for bucket [16:01:17] or maybe a geo/device filter and a bucket/flatten filter [16:01:49] either that, or we only add new campaigns to the cookie when a banner is selected, but that seems like a bigger code change [16:02:05] actually I like the additive way better [16:02:17] that works for users whose ip hops between countries [16:04:37] so maybe add a persist flag on the bucket in bucketsByCampaign [16:04:48] set the persist flag on anything read out of the cookie [16:05:23] and also on the bucket used for the selected banner [16:05:34] then don't call storeBuckets() until the banner is selected [16:05:46] Wouldn't be too bad a change, I think. [16:06:15] just one more step in storeBuckets to filter on the persist flag [16:08:33] hmmm not sure what you mean about the persist flag? [16:08:57] So forget what I said about doing filtering earlier [16:09:16] (BTW in 5 minutes my electricity is gonna get cut again for about 15 minutes. There's an electrician here fixing some stuff that was almost starting to smolder...) [16:09:35] oh shoot. I can email this idea too [16:10:00] sure! If we get cut off and ur in the middle of typing, that'd be great :) [16:10:14] the flag is set on any campaign we want to put in the cookie [16:10:35] these include campaigns we've hit before (and are not expired) [16:10:46] and the campaign for the banner we get now [16:11:30] Ah so just a flag in the in-memory data, and then use that at the end to decide what to store? [16:11:30] Hmmm [16:11:33] So in retrieveBuckets, we add a 'persist' property to each campagin we get from the cookie [16:11:35] Maaaaybe.... [16:11:55] My concern is about campaigns that are not consistent [16:11:56] and once we get the banner, we set the persist property on the campaign we just selected [16:12:12] So if you have a campaign where some buckets are getting desktop, and others are getting iphone [16:12:35] Obviously doesn't make sense from the user's perspective, but it's easy to do because of our topsy-turvy "domain" model [16:13:43] Huh, tricky [16:13:48] The two-pass filter option wouldn't have that issue... we'd only filter out campaigns where none of the banners could possibly apply to the user [16:14:11] The persist-flag option is clever! But I'm worried about introducing unpredictability [16:14:36] As in, let's say you in-memory choose a bucket for a campaign [16:14:48] And happen to get the bucket that goes to a banner that's not for your device [16:15:02] So the bucket isn't stored, and next time you go back, you get a random bucket again [16:15:09] right, i get the scenario [16:15:27] seems like a terrible use of buckets! [16:15:28] As opposed to someone on exactly the same device, who randomly did get a banner and thus a stored bucket [16:15:45] Yeah! It would be indeed, comletely nonsensical [16:16:18] But I actually did just that by accident last night at about 3:49 am and 26 seconds [16:17:02] So yeah in that scenario you've got some people with correctly chosen and sticky buckets and others with buckets that are getting unassigned an randomly re-assigned until they get a banner that's good for them [16:17:06] With the initial geofilter, users who cross borders with their device would have their old campaigns wiped out though, right? [16:17:31] I'm just worried about targeting European commuters [16:17:36] Only if we purge buckets from other countries, which we wouldn't need to do [16:17:59] We could continue to limit the purging to when a bucket is past expiry? [16:18:11] Definitely do that [16:19:35] So you're saying a combination of persisting those already in the cookie and filtering current choiceData by device/geo [16:20:56] oh, I see, if you just filter the choiceData, we don't need any persist magic [16:21:49] cause the cookie stuff comes in with retrieveBuckets, then you look at new choiceData to add to bucketsByCampaign, then you purge expired, then you store them [16:22:31] Yeah, so an inital filter pass before processBuckets would be all we'd need to keep the cookies small [16:26:51] Hey ejegg, back now :) [16:27:33] So should we go with the two-pass filtering? [16:27:42] Hi, was just realizing that the initial geofilter on choiceData wouldn't kill stuff that was already in the cookie [16:27:51] so yeah, two-pass filtering would be great [16:27:59] OK doing that :) thanks [16:38:17] (CR) AndyRussG: "Thanks!!! :)" (2 comments) [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/174581 (owner: AndyRussG) [17:39:00] Blurg... I keep forgetting that my laptop battery lasts 2 hours, not 2 years.. [17:58:59] (PS7) AndyRussG: Use per-campaign buckets, with smooth transition [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/174581 [18:01:28] (CR) AndyRussG: "Divides filtering of banner choices into two stages, to remove inappropriate campaigns before buckets are chosen. :)" [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/174581 (owner: AndyRussG) [18:01:48] ejegg: ^ :) [18:01:58] Yep, reviewing it now! [18:02:01] thanks! [18:05:04] likewise! :) [18:05:56] ejegg... I'm scheming... Are you in or out? [18:06:08] for deploy? [18:06:26] Sure, at least to wmf9 [18:06:48] heheh for deploy "plus ultra", as they say in Mex [18:07:58] Scheming and plotting... about getting in small but big features that built on this, too... [18:09:52] what, deploy more than just this today? [18:10:06] Mwahahahahaha [18:10:21] Feel free to say no, though [18:10:42] I'd be very wary. Was hoping we could deploy this fairly soon [18:11:10] Yeah! like in a couple hours, no? [18:11:32] Didja see my (/me cackles furiously) schema change? [18:12:02] I did... [18:13:43] Not much to make it work, it'd allow reshuffling buckets within a single campaign [18:14:23] The other change that was supposed to have come first, and is ummm also technically not in the sprint was setting per-campaign bits of JS to run ahead of banners being fetched [18:14:54] Hey, should we use toISOString instead of toUTCString? [18:15:12] Hmmm [18:15:17] toISOString's format can differ between platforms / languages [18:16:17] derp, i meant toUTCstring can differ [18:16:33] oh, but toISOString isn't in IE till 9. Curses [18:17:03] We really need it human-readable, and can't use a timestamp? [18:17:22] ejegg: we can certainly use a timestamp... [18:17:37] more compact that way too [18:20:08] ejegg: what sort of timestamp do u suggest? [18:20:16] unix timestamp? [18:20:25] Ah [18:20:32] But how do we do those in JS? [18:20:42] We want to be able to get it in and out of a JS Date object... [18:20:52] Date.getTime() [18:21:08] and Date.setTime() [18:22:15] ooh, old IE will also barf on making dates from the ISO string in choiceData [18:23:01] Hmm... I guess whatever is most reliable... [18:23:14] IE 8 is our earliest required IE target [18:23:19] for JS functionality [18:23:19] I'd say the unix timestamp [18:23:48] yeah, IE 8 doesn't do that according to http://stackoverflow.com/questions/5802461/javascript-which-browsers-support-parsing-of-iso-8601-date-string-with-date-par [18:26:04] Blarg! OK, well that'll do it... thanks... :) [18:26:31] I was gonnay say I'd choose readability over a few bytes, but then compatibility of IE8 trumps anything else [18:27:23] the bytes /are/ in a cookie, too... [18:27:32] True... [18:27:43] https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Date just said it's compatible throughout, but I now that it's more complex [18:27:49] http://dygraphs.com/date-formats.html [18:28:37] http://blog.dygraphs.com/2012/03/javascript-and-dates-what-mess.html [18:30:01] ejegg: K I'll do the timestamps... [18:30:39] phew, that situation /is/ a mess [18:33:57] (CR) Ejegg: "As per IRC discussion, need to serialize all dates as unix timestamps for browser compatibility and cookie compaction." [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/174581 (owner: AndyRussG) [18:36:49] pizzzacat1: hey [18:48:12] ejegg: something like this? campaignStartDate = new Date(); campaignStartDate.setTime( campaign.start * 1000 ) [18:48:39] mmm, i think so [18:49:34] OK, TS_UNIX is also guaranteed to be in UTC on the PHP side [18:49:45] yep! [19:05:59] (PS4) Ssmith: Show actual SQL instead of query string in db modal [wikimedia/fundraising/dash] - https://gerrit.wikimedia.org/r/173632 [19:09:37] (PS8) AndyRussG: Use per-campaign buckets, with smooth transition [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/174581 [19:17:57] ugh, am i nuts or is there a nasty parens problem on bannercontroller.js line 78ish, in the window.Geo = ... apply argument? [19:18:50] checking... [19:18:56] We should close the parens after the regex, not after the || [] [19:19:57] unrelated to your changes now, but throwing errors when I try to test IE8 with no geoip cookie primed [19:20:19] AndyRussG, ejegg: Hola. I don't want to interrupt, but could I get a brief update when you're not trying to power through the work I want to talk about? :) [19:20:39] K4-713: hola! [19:21:55] Hi K4-713: code is solid and ready to go except for one change: using unix timestamps for campaign dates instead of more verbose strings [19:22:02] brief update is: Deploy slots are all free today last time I checked, this is the patch under review that would make use of what we've done https://gerrit.wikimedia.org/r/#/c/174581/ [19:22:04] mostly for browser compatibility [19:22:10] ejegg: Ah that's done :) [19:22:30] oops, missed it! Will pull the latest and keep testing in old IE [19:22:40] Ah K! thanks much! [19:22:58] ejegg: Unix timestamps are always the right answer. :) [19:23:13] K4-713: I think i found an unrelated bug though - see backscroll at 14:17 [19:23:23] Even if the question is, for example: "What is the capitol of Earth?" [19:23:32] heh [19:23:44] ejegg: looking now... FWIW, that code has been there for a while... [19:23:48] ejegg: Unrelated, huh? How unrelated? [19:23:58] I guess Hitchhiker's got it wrong then :) [19:24:07] Ha, I was about to ask if it was a pre-existing bug or not. [19:24:08] K4-713: well, it's been there forever [19:24:29] but it throws JS errors the first time you hit CN and have no geoIP cookie [19:26:58] 1427246580! [19:27:38] AndyRussG: :) [19:27:54] I'm just getting ready in case you have more questions [19:28:18] About that bug... dang. [19:28:30] Easy fix? [19:28:36] yep [19:28:53] ejegg: yeah that's definitely a mistake [19:29:09] Okay. If it wasn't easy, I'd strongly suggest making a new card for it. [19:29:50] You don't want to be sending an empty array into a JS match() call. Though of course ( /regex/ || [] ) should always give you the regex... [19:30:12] nice catch!! BTW [19:30:23] * K4-713 agrees [19:30:24] thank IE8 developer tools! [19:30:31] no! really? [19:30:33] Do I have to? :) [19:30:43] 1410135780! [19:30:43] hey, just trying to cover our testing bases [19:30:53] Indeed. [19:32:18] This morning chrismcmahon sent up a change for improving our browser tests [19:32:48] (PS1) Ejegg: Fix parens when setting geo data [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/175006 [19:32:48] I was thinking of leaving 'em for now since we don't need to deploy code to use them, I think... [19:33:40] AndyRussG: Yeah, agreed. And anyway, this is not the time for scope creep, no matter how cool it is. [19:34:51] K4-713: K... [19:35:32] AndyRussG: Anyway, something like that, we will probably be able to do somewhat sneakily when nobody is looking directly at... [19:35:39] >_> [19:35:40] <_< [19:35:55] Ahhh OK iiinteresting... [19:36:00] * K4-713 whistles innocently [19:36:14] Mmm did you want to talk about the work of which you want to talk? Any particular medium, also? [19:36:49] Nah, just interested in the final throes of the CN deploy you wanted to do today. [19:36:57] Also, re: being sneaky? Ask me later about the time I deployed the original orphan slayer. [19:37:26] Will you answer with a Unix timestamp? [19:37:47] I wonder if NaN is ever allowed as an answer... [19:38:06] I guess only if one's brain is on Javascript [19:38:31] Well, I can't find any other issues in that code. [19:39:18] :) [19:39:20] I've smoke tested and also it passes the allcation fixture tests, but if you feel like smoke testing all the expected functionality too (if you haven't yet) that might be fun... [19:40:14] AndyRussG if you and K4 are agreed on the parens fix, want to roll that in? [19:40:37] ejegg: I'm cool, I'd just test it a few times tho... [19:40:52] works in IE8 :) [19:41:05] lemme test in something less archaic [19:41:33] Nice! (/me streches the limits of the possibile positive statements following a sentence that mentions IE8) [19:42:05] bahaha [19:42:41] K4-713: ejegg: WRT sneakiness, I was thinking of plotting to do at least one more of the changes that FR wanted, that our banner choosing change was meant to allow [19:42:59] The simplest one involves a wee schema change [19:43:06] I realized [19:43:09] AndyRussG: I really don't want to deploy much later than pretty soon [19:43:14] That... might not be sneakable. [19:43:27] K... [19:44:07] Actually, I should draw a diagram of relative sneakability. [19:44:32] The other also unsneakable would have been the adding an editable snippet of JS associated with each campaign, to run before a banner is loaded, to avoid missed impressions with their new fullscreen-then-top-banner one-two punch [19:44:41] But I got stuck on that [19:44:43] I'll call it "Don't EVER do any of these things!" and then explain how likely it is you'll get away with different things anyway. [19:44:52] Heheheh [19:45:33] Mmm I got stuck thinking about what's the right way to do it, either a Wiki page like banners or an in-DB string [19:46:00] Hum. [19:46:04] NEway, I think I'm doing again what I apologized for doing before, bbgglll [19:47:46] so... targeted doesn't have a double t [19:47:47] (Yeah the Wiki page route would be best for having an easy handle on history and changes of the JS snipped, but the DB field keeps access and permissions stricter and simpler.) [19:47:58] ehh, forget it [19:48:09] Wat? no that's important! [19:48:45] i'm sure there are no other misspelled variables in Mediawiki! [19:50:29] yeah, that's an easy search & replace. Let's fix it [19:50:37] Oh it wasn't my fault... Mmm should be geotargot [19:50:42] :P [19:58:02] OK, tested parens fix in chromium & ff. Definitely an improvement [19:59:21] ejegg: agreed! Just out of curiosity, where are you seeing errors on that? I was just trying to reproduce... [19:59:36] AndyRussG: you need to clear your geoIP cookie [19:59:37] the errors [20:00:09] oh, maybe some browsers are fine with a null 2nd arg to apply [20:01:11] I'm going to sign us up for the 13:00 PST slot [20:01:23] Ah OK sounds fantastic, yeah thanks! [20:04:17] yea https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function/apply#Browser_compatibility [20:05:34] ah, so it's supposed to allow null [20:06:17] but not fully compatible apparently [20:07:23] did you want to s/geotargetted/geotargeted/g ? [20:08:58] Mmmm maaaaybe, don't feel especially strongly about it either I guess [20:09:33] (CR) AndyRussG: [C: 2] "Excellent catch!!" [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/175006 (owner: Ejegg) [20:09:37] it's bugging me now that I noticed it, and we spell it correctly on the admin page [20:09:44] ty for the CR! [20:09:53] np! [20:10:21] (Merged) jenkins-bot: Fix parens when setting geo data [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/175006 (owner: Ejegg) [20:10:41] I guess I'd be happier trying to get at least a rough version of my moar bucket feature sneakiness, even assuming that we wouldn't deploy it today... [20:11:15] It's just if we deploy with the misspelling, it will be very difficult to fix later [20:11:44] hmmm OK good point! [20:12:30] could make it another PS if you don't want to mess with the perfectly working one [20:12:51] yeah, sorry. I'll +2 yours and make a new PS for the spellcheck [20:13:09] Ahh OK [20:13:13] Thanks :) [20:13:27] Yeah the misspelling does go back further [20:13:34] (CR) Ejegg: [C: 2] "Nicely done, sir!" [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/174581 (owner: AndyRussG) [20:20:27] (PS1) Ejegg: Spelling fix: geotargetted -> geotargeted [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/175015 [20:20:51] OK, now to make sure spellcheck didn't bust your functionality! [20:24:25] Appears to be in order. [20:25:04]