[00:01:30] looks to be working on meta.beta, at least [00:03:51] what is meta.beta? [00:04:13] I don't see the checkbox on meta special:BannerAllocations [00:05:17] AndyRussG: http://meta.wikimedia.beta.wmflabs.org/wiki/Special:CentralNotice [00:05:18] I do see it on mediawiki [00:06:15] Woohoo!! [00:06:22] * awight pops open a very small champagne [00:06:23] http://meta.wikimedia.beta.wmflabs.org/wiki/Special:BannerAllocation?project=meta&language=en&country=US [00:06:48] Fiddle with the checkbox to say goodbye to allocation quantization [00:07:08] awight: I actually have one of those in the fridge. [00:07:08] awight: wow [00:07:15] a checkbox? [00:07:19] AndyRussG: mediawiki has the client choice stuff now, I mean [00:07:25] K4-713: overprepared! [00:07:39] I don't know where it came from! [00:07:40] AndyRussG: but I don't see the banners. Maybe div#sitenotice is missing! [00:07:48] K4-713: the future, clearly [00:07:53] or no. that would be an empty bottle. [00:07:54] Oh. Right. [00:08:04] Inbetween future. [00:08:49] awight: U don't see banners on mediawiki... should you? [00:08:56] I see lots of choices. [00:08:59] And I see a siteNotice [00:10:31] wait... on mediawiki it's live and enabled [00:10:36] yep [00:10:41] awight: wait, I don't see it in the list of projects to target from meta [00:10:41] sorry ;) [00:10:55] ejegg: it's listed as 'wikimedia', I think [00:11:00] ohhh [00:11:08] just noticed that today. [00:12:26] AndyRussG: geotargetting was working correctly. [00:12:27] http://www.mediawiki.org/wiki/Team_Practices_Group/Health_check_survey?country=AC [00:12:29] ejegg: awight: K4-713: banners work! [00:12:30] https://www.mediawiki.org/wiki/MediaWiki?debug=true&country=FR [00:12:32] you should see WLA [00:12:33] yes! [00:12:42] jinx again [00:12:53] I know. we've been working a bit too closely ;) [00:13:07] nice [00:13:15] OK, I'll quit messing with esperanto [00:13:34] OK. Statutory high-fives, and I think I'll be off, now. [00:13:38] you mean you've been working and I've been mostly watching and giving Krinkle material for his next review [00:13:43] congrats! [00:13:44] awight: congrats [00:13:52] ejegg: congrats [00:14:14] Thank you and sorry for any intentional shoveling of your mostly-baked code into the spotlight [00:14:42] Heh sometimes it's the only way [00:14:50] the big question is whether this data refreshes when the campaigns change... [00:15:42] K4-713: thank you for letting the show go on... :) also please deposit empty popcorn containers in the bin [00:15:52] AndyRussG: You just annihilated two years of cruft [00:15:54] !! [00:16:15] ah! lemme try mobile! [00:16:22] good call [00:16:22] -_- [00:16:54] I'll announce in their channel [00:16:58] https://www.mediawiki.org/wiki/MediaWiki?debug=true&country=FR&useformat=mobile [00:17:14] looks like it's all good [00:17:53] * AndyRussG hopes that the "two years of cruft" that awight mentioned isn't the mobile web site [00:18:06] 8D [00:18:59] bahaha [00:19:46] Really, this enthusiasm is inspiring. [00:20:10] K4-713: that slots thing has been rotting around my neck... [00:20:19] It was fun while it lasted. [00:20:20] I was so proud of it at first... [00:20:22] yeah [00:20:22] ...kind of. [00:20:32] well not really. Explaining it was the most embarrassing part. [00:20:42] Then, trying to work around the caching implications was tha arse. [00:20:51] For some reason, explaining it in terms of shoe boxes seemed to work okay. [00:21:10] "so, imagine your house is this shoe box. You're evicted, sucka." [00:21:14] * K4-713 shrugs, fishes metaphorical junior mint out of the couch [00:21:28] "Also, I have all your shoes." [00:21:42] WHAT NOW, FOOL. [00:21:45] :) [00:21:59] That... should go in some official documentation. [00:22:02] My parents fished a few of my stuffed animals out of a dumpster in Somerville, once... [00:22:18] and look at your immune system now! [00:22:24] ha [00:22:25] Ewwww. [00:22:30] Noooooooes. [00:22:58] awesome. see you in a week! [00:23:02] Wow OK, way to go folks... I'm very, very impressed [00:24:28] K I'm also satisfied that things are stable enough.... gonna spend some time doing other stuff for a bit, will check back, plot the next steps and send some e-mail summaries later! [00:24:40] Thanks once again, all!! Cya soon :) [00:25:13] I just enabled the esperanto campaign on mediawiki.org to make sure our version stamp updates when there's new info [00:25:23] or within 5 minutes, anyway [00:26:09] AndyRussG|away: ejegg could one of you update megan, etc on the thread about where the deployment is at? [00:26:25] OK, will do [00:27:20] is the new plan reliable enough that they should postpone starting the israel campaign? [00:27:55] It's looking good so far [00:28:32] I'm just looking at one last thing to make sure updated data is sent along to clients without too much delay [00:29:28] aaaah yes thanks ejegg [00:31:52] ejegg: how are u monitoring that? I guess looking for changing version numbers on the request to the RL module? [00:32:07] that, and different choices data [00:33:00] ejegg: that'd only be relevant if a campaign ends or starts or if someone changes something, no? [00:33:21] so ejegg AndyRussG - awight gave me a 30 second synopsis a few minutes ago. but am i understanding correctly that this should go out wednesday? [00:33:30] yep, i'm enabling a very-throttled campaign for esperanto users [00:34:10] atgo: yes, it's in the 1.25wmf8 version tree, which will roll out to the world Wednesday [00:35:00] sweet [00:36:06] AndyRussG: looks good! Got the new choiceData JSON and a version bump for ext.centralNotice.bannerChoiceData [00:36:16] thanks guys! [00:36:53] ejegg: thanks! [00:37:42] atgo-away: ejegg: this is the big change in the way CentralNotice works. However it is not yet the actual smallish features that we needed this framework for [00:38:11] Oh, was the Israel campaign waiting on one of those smaller features? [00:38:18] Not sure [00:38:39] Not quite, I think it's also so we can test that banners continue to operate as expected [00:38:46] via actual monitoring of impressions [00:38:50] the israel campaign is waiting to test out the new buckets [00:39:06] get a feel for it before December [00:39:41] atgo-away: OK, the new buckets are what's not there. What is partly rolled out and will automatically go out on Wednesday is the new infrastructure to support all that [00:40:18] Oh, and that rtl stuff would be nice to have too for Israel. [00:40:45] ok, so AndyRussG what's the hold up on the buckets? [00:40:51] when's that supposed to happen? [00:41:07] the RTL would be nice... sounds like that's done and just needs to be in the next DI deploy, right? ejegg [00:41:16] it would be great to push out some of the PD cards anyway :) [00:41:53] Are we not waiting for the rtl patch too? [00:41:57] atgo-away: not quite done, needs some tweaking per Amir's email advice [00:42:14] ok.. but my understanding was that Amir was going to do the tweaking when he approached me about this at all [00:42:15] ...it would help if I had ever learned to read. [00:42:23] hehe [00:42:26] and the fields in the iframe from GC need some production work (a css file on foundation wiki and an account settings change) [00:42:34] ok so hold on [00:42:52] there are 2 pieces to the RTL issues. 1 is our part. there is still work to be done there? [00:42:58] atgo-away: the buckets are not on hold. It's a small change but it's not been imlemented yet [00:42:59] It's not merged. [00:43:04] because Amir told me that he would be doing all of this and just need some review and deployment. [00:43:29] it was dependent on all the big changes that are done, as are a large number of tests [00:43:55] atgo-away: The only thing I saw from him were comments on a small patch that ejegg wrote, that would take care of our side... and some suggested changes in the mingle card for what we should put up on the GC form css. [00:44:15] ok so, let's focus on our side. ejegg you ended up writing the fixes and now he's commenting them? [00:44:35] then we can figure out GC [00:44:39] There's one outstanding issue, and that's the email input. [00:44:43] Yeah, I can update that change with a new PS [00:44:53] That is apparently an exception to the rtl rules. [00:44:56] grumble.. he said he was going to do this. [00:45:24] email = always ltr. Which I guess makes a certain kind of sense. [00:45:36] or we can ask him to take over the change - I think there's just some reluctance to steal a gerrit change that someone else initially uploaded [00:45:54] ahh ok [00:46:03] That's kind of normal, I think. [00:46:16] well, whatever works. i just know you guys are sort of... swamped with all this other stuff and when he approached me about it i said "please feel free" [00:46:41] Like, you have one person to write it, and if the main reviewer alters it directly, you have to pull in another reviewer or get sucked in to something that resembles a self-merge a little too much. [00:46:46] word. ok [00:46:47] If it looks like it'll take me more than 10 min, I'll ask him to take it over [00:46:55] cool [00:47:01] so we have a clear path forward on that [00:47:17] this is not a blocking issue for Israel, but definitely if we cean push it out before they start that would be cool [00:47:24] Are we blocking on getting this out too? ...is what I just typed. [00:47:49] so for GC - it was my understanding that we couldn't have different CSS in the form. can we do it differently for different languages? [00:47:53] K4-713: :P [00:47:56] I can't figure out if you answering questions as I'm typing them is cool, or creepy. [00:48:16] atgo-away: Yeah, I don't know that. [00:48:22] the-wub might. [00:48:32] I think he looked at those docs most recently. [00:48:34] atgo-away: there was a mention in the hosted merchant link PDF of different css for different languages [00:48:40] Or, ejegg did. [00:48:40] ahh ok [00:48:48] that's cool [00:49:03] Woah. It's almost like they have had... other clients ever. [00:49:05] so i've got a card about new CSS for GC anyway that should go into the next sprint or so, assuming we get to it [00:49:06] but it didn't say exactly how [00:49:11] bahaha, or not. [00:49:33] ok.. we can ask them. i'd really rather this not become a giant time suck going into December, ya know? :/ [00:49:45] heh [00:49:49] that said, we've now been in IRC for like 10 min anyway haha [00:50:07] so yeah, neither of the RTL issues are blockers for Israel [00:50:26] but the buckets could be... so AndyRussG what's the status there? [00:50:41] the structural/dangerous changes go out Wednesday? [00:50:43] and the rest, when maybe? [00:51:11] atgo-away: well, it's a question of implementing testing and deloying [00:51:37] I don't really know is the best I can do I'm afraid [00:52:00] and you can't do that before the Wednesday thing goes out? [00:53:02] atgo-away: it's pretty small but I don't think it would be right to promise that it can be done, or that it can't [00:53:16] I did have an idea tho [00:53:43] right, fair enough [00:53:50] Which is how a smooth transition from the old buckets to the new ones could work [00:54:02] i'm trying to outline a 'best case' scenario. right now i think that the thing to do is recommend that they put up tests in israel anyway [00:54:09] what do you think of that? [00:54:49] I think they could put up their tests and make the new bucket transition smooth. That way if the new buckets are there before a week is out, it would be basically the same [00:55:01] I mean, we can always take things down if there is a problem, right? [00:55:43] My usual impulse is to leave things up unless we're doing something that will very specifically benefit from no banner/donor traffic. [00:55:53] word [00:56:03] they had wanted to start the campaign with the new buckets for the sake of practice [00:56:04] Like, rebooting the queue serer or something. [00:56:36] Ah, I see. Well... Can't they just pretend to start over when/if we get it? :) [00:57:19] that's my suggestion [00:57:23] i'm going to send an update :) [00:57:27] I think they should start and we can also use this to monitor that banners are being shown on the new syste [00:57:29] system [00:58:00] cool :) [00:58:02] The first issue they have I think with buckets is that they should last. If we make a smooth transition to the new system before a week it out, then that will indeed be the case, even if they start now [00:59:46] a week is a tight deadline :P [01:02:16] AndyRussG: whay the week? [01:02:18] why [01:02:34] a week because that's how long the old-style buckets last [01:03:08] Stupid shoe boxes. [01:03:12] So if their interest in this case is that the buckets last, we have a week to make a transition [01:03:34] heh, yah now we'll be transitioning to golden shoe bags [01:04:12] mmm shoe bags [01:04:31] Neoprene booties? [01:04:48] I mean, I heard it was clean now. [01:05:23] today we had our first snowstorm, so boots it is [01:05:35] I miss snow. [01:05:45] ...until I have to try to drive in it. [01:06:23] hehe biking in it is tough too [01:07:21] I had one of these for a while: http://www.autosportracetech.com/RX-7/carpic.bmp [01:07:29] Try driving *that* in snow. [01:07:59] It's pretty much made out of torque. [01:08:27] It's fantastic if you want to make a lot of funny noises and not really go very far. [01:09:00] ha ejegg|away our emails just crossed midair [01:09:33] atgo-away: Aha. You keep reading people's minds. [01:09:37] I thought it was me. [01:09:41] But no. [01:09:44] MUA HA HA [01:10:03] sorry for jumping the gun on your email (that i asked you to write) ejegg|away :( [01:10:34] heh, no worries! [01:10:42] :) [01:10:49] alright.. i've got to run [01:10:55] K4-713 you think you [01:10:58] (also if anyone would like to cc it on over to me, too, that'd be great... thanks!) [01:10:59] will be wfh tomorrow? [01:11:07] AndyRussG: i had you cc'd on mine :) [01:11:11] atgo-away: I don't know. 50/50. [01:11:18] Mostly I'm worried about the bart. [01:11:21] yea atgo-away thanks!! [01:11:34] And, the ice issue. :) [01:11:35] ejegg|away: is the standoffish one here [01:11:39] (jk) [01:12:06] ahh K4-713 makes sense [01:12:08] well we miss you [01:12:26] I want to come in! But it might get weird. [01:12:35] yeah fair enough [01:12:41] don't take this as an obligation at all [01:12:56] 'course not. :) [01:13:03] I am sick of this COUCH. [01:13:08] ...is all. [01:13:54] Though, I do love having meetings with my "NO LOAFING" banner in the background. [01:14:03] yeah tha'ts hilarious [01:14:06] ok... see you cats soon [01:14:11] later! [01:14:48] Bye! [01:17:54] ejegg|away: have you detected changes in the banner choice data module version param? I'm seeing mine stay stable :( [01:19:31] ...okay, I'm off too. Good night, y'all. [01:19:38] Bye! [04:41:46] (PS2) Ejegg: Fix input direction for rtl langs when sitedir=ltr [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/173435 [18:20:03] So quiet in here today. [18:24:41] * ejegg drops pin [18:26:50] I'm gonna do a civi deploy, unless there are any objections. Just a new custom field and a Adam's custom LYBUNT report, so low impact. [18:32:10] (PS1) Ejegg: Merge branch 'master' into deployment [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/174179 [18:32:30] ejegg: Yeah, no objections [18:32:50] (CR) Ejegg: [C: 2] Merge branch 'master' into deployment [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/174179 (owner: Ejegg) [18:34:14] ok, here goes! [18:34:40] !log update crm from e9e81a828d50e8bddf98eae699c925e09b25927b to 71ec68e8da1de289c4e7adca090c0fdbccbd8b8a [18:34:41] Logged the message, Master [18:41:21] aah, needs a submodule bump for adam's cross-db civicrm core fix [18:43:49] (PS1) Ejegg: CiviCRM submodule bump for cross-db reporting aliases [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/174183 [18:44:19] (CR) Ejegg: [C: 2] "non-code self-merge" [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/174183 (owner: Ejegg) [18:45:25] hi ejegg, K4-713 [18:45:32] Hi AndyRussG [18:45:33] Heeeey. [18:45:42] just heard a pin drop [18:45:49] It was loud. [18:45:54] (PS1) Ejegg: Merge branch 'master' into deployment [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/174185 [18:46:05] bowling pin [18:46:12] (CR) Ejegg: [C: 2] Merge branch 'master' into deployment [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/174185 (owner: Ejegg) [18:47:14] In 15 minues the new CN stuff ride the 1.25wmf8 train to sites other than WP (Wiktionary, wikitravel, etc) [18:47:34] !log updated crm [18:47:37] Logged the message, Master [18:47:47] !log updated crm from 71ec68e8da1de289c4e7adca090c0fdbccbd8b8a to ff89895638a0dd0600b2e4c0b6adfd1b8e402df5 [18:47:47] The feature switch is only enabled for mediawiki (already deployed there) and aa.wikibooks, tho [18:47:49] Logged the message, Master [18:49:18] AndyRussG: Exciting! [18:49:26] yeah! [18:49:32] Need any backup or anything? [18:50:19] no that I know of... backup like data backup or backup like help? [18:50:35] Er. The second one. [18:51:14] I think we're OK. I'm not deploying, it's already merged into 1.25wmf8 so whoever's doing that train deploy will be putting it out there [18:51:54] I think it's pretty clear that with the feature de-activated nothing terrible happens, though we should still ask ops to keep their eyes open [18:51:57] I think [18:52:07] This morning I did ask on the ops channel and everyone ignored me [18:52:19] Which is at least a sign that nothing horrible and obvious has happened [18:53:30] What we would need to do is watch for any drops or strangeness in impressions once the disable feature goes out to all sites, tomorrow [18:53:35] K4-713: ^ [18:53:55] And then after that decide whether and when to actually turn it on [18:54:02] Okay, cool. [18:54:34] Sorry you got ignored. :/ [18:54:58] Meh... thanks, it's OK [18:55:05] Kind of. [18:55:23] Maybe I didn't ping the right folks... do you know who would be most on top of any caching issues in ops? [18:55:46] Not really. Thank you, siloing. [18:55:58] Usually I just ask Jeff_Green who he thinks we should be talking to. [18:56:07] whut [18:56:11] Hi. :) [18:56:13] caching...probably Brandon [18:56:13] Ah hi! [18:56:15] hey [18:56:31] caching rl modules [18:56:33] OK [18:56:36] thanks! [18:56:52] rl modules? [18:57:00] I should probably know what that is... [18:57:08] ResourceLoader [18:57:11] oic [18:57:19] he'd be a good place to start anyway [18:57:22] relaoded leaves [18:57:35] wait a sec actually, [18:57:40] :) [18:57:48] Rehabilitated Lizards. [18:57:54] heh [18:57:59] ...I was slightly late to that party. [18:58:18] well! I didn't even know if was a party [18:58:23] It is now! [18:58:27] heh [18:58:45] bblack on irc, not the other Brandon [18:58:53] ah yes [18:59:06] the one without the iphone app [18:59:14] o_O [18:59:17] ha [18:59:47] K4-713: ejegg: on my plate today I have, reading the ResourceLoader code that deals with the new change we made yesterday at the last minute (I know you and awight are on top of it, but I feel the need to be one with more of the nitty-gritty of it) [18:59:59] And writing the first actual bucket change [19:00:08] And the other thing that I can think of to do related to all this [19:00:32] Is read through the tests to get an idea of how close we are to full coverage of code paths and complex interacting criteria [19:01:49] So K4-713 to finally answer ur question, those might be things that I could use a hand with, though I'm also fine with them if other priorities have now come to the fore :) [19:07:38] I really want to ask Anne about this, too... [19:08:11] I get the very strong sense that all of our priorities got knocked on their heads while I was gone. [19:13:15] (CR) Ejegg: [C: 2] make gauge range selection more clear [wikimedia/fundraising/dash] - https://gerrit.wikimedia.org/r/171982 (owner: Ssmith) [19:15:54] (CR) Ejegg: [C: 2] Reset form when modal is reopened [wikimedia/fundraising/dash] - https://gerrit.wikimedia.org/r/172938 (owner: Ssmith) [19:16:10] (Merged) jenkins-bot: Reset form when modal is reopened [wikimedia/fundraising/dash] - https://gerrit.wikimedia.org/r/172938 (owner: Ssmith) [19:20:54] K4-713: Hmmm... so, summary, when you were gone we did a lot but not what we were supposed to? :) [19:21:09] Er, sort of? [19:21:23] I mean, we wanted to have CN frozen already. [19:21:34] Yes [19:21:40] I think this is the first time it went in to November at all, actually. :/ [19:21:57] So, presumably, #1 priority is setting that down. [19:22:05] OK [19:22:34] Makes sense [19:22:41] But, I don't actually know what's going on at this exact moment. I mean, there was a ton of unrelated stuff I was racing to get done too. [19:23:32] Hmm [19:28:43] (CR) Ejegg: [C: -1] "Stray remaining yellowRange() causing JS errors" (1 comment) [wikimedia/fundraising/dash] - https://gerrit.wikimedia.org/r/173064 (owner: Ssmith) [19:29:56] K4-713: do you need more details about what has happened w/ CentralNotice? I don't know much about stuff unrelated to CN, but am happy to work on whatever [19:30:38] K4-713: ejegg: got pinged in wikimedia-visualeditor about CN weirdness on beta labs [19:30:47] ut-oh [19:30:54] how so? [19:31:35] AndyRussG: I don't know that more CN details will help me right now. Mostly, it's going to come down to what Megan and Lisa need, I think. Anne should be all up on those conversations. [19:32:31] Presumably, they are interested enough in the CN changes to entertain the idea that we're going to get that done, or we would have been stopped by now. [19:33:33] Jeff_Green: regarding the long-running silverpop export queries, are we only concerned with splitting up those that select from the civi database, or also with the silverpop-db-only dedupe and stats stuff? [19:33:59] I'm not entirely sure what the latter ones are [19:34:15] K4-713: OK...! sounds good [19:34:33] AndyRussG: What's the weirdness all about? [19:35:04] Some funny thing from the banner when you enable Visual Editor, probably related to banner contents and not to our change [19:35:06] Jeff_Green: We pull in all the contacts with selects from the civi db, then massage that table a bit, including de-duplicating email records, with queries that don't touch civi [19:36:37] Tho it's not my department, I'd say the bucket changes have the potential to improve A/B tests quite a lot [19:36:46] ejegg: there are two main things I'm concerned about--one is anything that blocks replication, and anything that bogs down civicrm [19:37:48] we might have to look case by case [19:37:57] OK. [19:38:16] AndyRussG: That was the sense I had gotten, too. [19:38:21] replication might be fine if it works in temp tables [19:38:28] I'll look at batching the selects in from Civi first [19:38:35] cool! [19:38:39] ejegg: cool [19:39:30] things that beat on temp tables might not be a big deal in terms of replication, but I'm not 100% sure [19:39:47] ...ah. Good thing I was concentrating on my inbox. Apparently I have to do some anti-harassment training. [19:39:53] cheapest batching should be order by and where > on the pk with some limit, right? [19:40:04] * K4-713 harasses everyone one last time [19:40:36] ejegg: that sounds right, springle would know better though [19:40:45] ok, thanks [19:40:55] next time we can expect to be anti-harassed? [19:41:06] might be worse [19:41:17] That sounds like it hurts. [19:41:20] ejegg: I know innodb has some oddball behavior around primary key [19:42:00] ok, i'll look into it [19:42:44] it does something insane re. row counts without a where clause [19:44:47] 8p maybe anti-harassment is when one gets excessively complimented and made to feel great about one's work, and so one's ego goes through the roof... in which case, likely painful for others... [19:46:06] Anti-harassment = conspicuously ignored? Hmm. [19:46:26] I'm going to lose sleep over this. [19:51:30] howdy folks [19:51:42] Hi atgo! [19:52:35] :) [19:58:04] hey atgo [19:58:19] quick question: are banners supposed to show up when you edit a page, too? [19:58:41] for some reason i have such a hard time staying connected in IRC [20:00:24] was there any weirdness with SSL turning off for GC yesterday? K4-713 ejegg? [20:00:37] I didn't see anything. [20:00:43] cool [20:10:30] K4-713: ejegg: atgo: the CN wierdness the VisualEditor folks were seeing on beta labs are just expected behaviour of the the test campaign enabled there... :) [20:10:43] oh awesome! [20:10:52] Good news. [20:11:42] phew [20:12:18] yeah! [20:13:26] ejegg: what I am seeing is a page bump on banner display, maybe more than before? not sure. what do you think of the idea of adding a URL param switch to use the old system and write some profiling data on the browser console? [20:14:31] AndyRussG: you mean more delay between initial load and banner appearance? [20:15:56] yeah. Well, that's my feeling about what I'm seeing on beta labs. But it could be my imgination. Or my wifi today. Or.. more worser infrastructure used by labs... [20:17:17] Can't imagine the client-side calculation would be the bottleneck. And loading the single requested banner should be faster than going through bannerRandom, if anything. Huh [20:18:36] ejegg: maybe the timing of when the call to bannerRandom is made and when the banner is inserted? There's a lot of JS in there, must be some startup time, not sure where we are and where we were in all that [20:19:36] If the call to get the banner were somehow happening a lot later due to RL modules' startup code juggling, that might make a difference? [20:20:03] I'd kinda like to profile somehow... [20:20:54] Might be possible with Chrome's dev tools [20:21:42] though that's not gonna help you switch algorithms [20:27:41] atgo wanna grab lunch? [20:27:56] hi pizzzacat! i actually grabbed a sandwich on my way in and inhaled it [20:28:02] but i'll join you for a bit when you come back [20:28:04] that sounds…painful [20:28:09] ok :) [20:28:12] i've been training [20:28:19] hehe [20:29:20] ejegg: yeah we'd need a param for that.. though now that some sites are on and some are off, we could maybe... [20:34:07] Jeff_Green: batching some of those queries is going to be a real pain. Just learned that updates with joins can't use limit and order by [20:40:01]