[00:02:50] i am thinking maybe logger->info() doesn't make it to the log server? [00:03:15] looking [00:07:01] XenoRyet: and cwd we don't need this right this minute. If you want to look for this tomorrow that is cool too [00:08:53] cwd: Yea, it looks possible that wouldn't end up on the log server. [00:09:49] dstrine: Good to know, we won't kill ourselves over it tonight. [00:10:21] cool and thanks for looking [00:10:56] i might be able to find them in the web logs [00:11:00] i'll take a quick look [00:15:57] I'm running a debug session locally and can catch the XML, but again it's not certain that's what's happening on live. [00:16:21] Funny, because again the really authoritative thing is whatever ends up landing on their server. But alas, that's not to be. [00:18:30] yeah i think our best approach is asking them to look for it, since changing the log would require a code change and a deploy [00:19:10] XenoRyet: http://i.imgur.com/GnMP2zR.png [00:19:15] and dstrine ^ [00:19:19] looks like worldpay was the exception [00:19:57] Makes sense that they were the exception, since they could be trusted not to do their own jobs. [00:20:05] indeed [00:20:17] But yea, we don't want to do a deploy just for this, so let's see if we can push this back on them. [00:20:46] sounds sensible to me [00:21:33] so what do we say to them? [00:22:02] They really should have it, somebody probably just gets flack for bothering their engineers. [00:22:39] XenoRyet: lol like right now? :) [00:22:49] ;-) [00:23:40] ok so we don't log exactly what we send them. that's the issue right? [00:24:48] yep [00:25:06] Probably say something along the lines of: We're not currently logging that particular bit of XML in our production systems, so they'll need to find it on their end or possibly make do with what we're getting on dev machines. [00:25:27] Leave that last bit out if we're really trying to make them do the legwork. [00:25:55] hehe ok [00:25:56] we can change it but we'd have to deploy. low likelihood of failure just an unfortunate week for it [00:26:29] Would violate our wikimania mini-freeze [00:27:42] also we have a new server in the pool. i assume we could just specify it with the rsync tool but it'd be prudent to confirm with jeff [00:27:42] it's cool. This is a side request from a conversation. Nothing critical. [00:29:14] cwd: nah I'm going to shoot off this email and hold this for a few days [00:29:22] Thanks ! [00:33:03] sounds good. i'm gonna bounce for the evening, gnight! [00:33:14] Catch ya later. [00:34:16] seeya [16:25:09] Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Easy: Clone button for CN campaigns - https://phabricator.wikimedia.org/T91078#2421735 (Pcoombe) @DStrine This is the task I mentioned yesterday. Any chance someone could take a look at it? Seems like it ought to be fairly easy, and would have... [16:29:13] Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Easy: Clone button for CN campaigns - https://phabricator.wikimedia.org/T91078#1073823 (DStrine) Thanks for the note! I have moved this back into our analysis column. I'll try to get an estimate. I can't determine priority until then. [20:27:43] dstrine: hi! sorry I missed standup yesterday and today... We just (almost) finished moving everything.... [20:28:11] Also I have no Internet, yet, just phone tethered.... [20:29:03] I'll be semi-here for the next while... [20:44:09] AndyRussG|sortof: it's cool. Good luck on the move. Check out my spec on CN sequence feature when you get a chance [20:57:52] Fundraising-Backlog, FR-PayPal-ExpressCheckout, Epic: Epic: Add semantic prefix to order_id - https://phabricator.wikimedia.org/T138013#2422669 (Danny_B) [20:58:09] Fundraising-Backlog, MediaWiki-extensions-DonationInterface, Epic: EPIC: mediawikiless-donationinterface - https://phabricator.wikimedia.org/T137586#2422675 (Danny_B) [20:58:22] Fundraising Sprint House of Pain, Fundraising Sprint Indigo Girls, Fundraising Tech Backlog, MediaWiki-extensions-DonationInterface, and 2 others: [Epic] Make the forms come from one template - https://phabricator.wikimedia.org/T86248#2422681 (Danny_B) [21:12:21] dstrine: cool! Where is it? [21:12:31] in email [21:12:38] I'll add it to the task [21:13:37] Fundraising Sprint Killing Time, Fundraising Sprint Licking Cookies, Fundraising-Backlog, MediaWiki-extensions-CentralNotice: Spike: how to implement banner sequence feature - https://phabricator.wikimedia.org/T135398#2422698 (DStrine) Spec: https://docs.google.com/document/d/1YlNwpFZ4gPyO1BYxYRE... [22:24:51] awight: hiya! yes i think we noticed that with 1005 in the pool its orphans were orphaned, but the slayer would clear the local redis [22:25:18] it is now depooled and running with staging config [22:28:18] which is to say clearing out 1-3 and not scheduling any itself. it didn't occur to me why copying that section from staging wouldn't work for running the front end and the cron job [22:35:08] cwd: hey [22:35:38] so 1005 is a frontend + the rectifier? [22:35:48] not at this point, just rectifier [22:36:10] http://www.metal-archives.com/images/2/1/2/212.jpg?5958 [22:36:27] ok that sounds good then [22:36:37] i thought that was going to be a mesa boogie dual rectifier [22:36:46] might as well not screw with making it work any better [22:36:52] cos we're rewriting so soon [22:37:03] hehe. mesa boogie is for princes [22:37:17] does make a real nice sound though [22:37:32] awight: it did look to me like you could combine the queues sections from config and have it do front end + rectifier [22:37:47] because when 1005 was in the pool, it was both writing to its local redis and clearing it out with the cron job [22:37:53] (ignoring 1-3_ [22:38:02] s/_/)/ [22:38:49] this is the only kid I know who had a MB--I see he's moved on quite a bit since our high school whiny metal band. http://www.alekseybogdanov.com/ [22:39:35] thanks for keeping rain out of the cellar! [22:40:31] wtF. "He has performed at the United States Supreme Court and CIA Headquarters" [22:41:02] If only they knew... [22:41:17] blocked 35 scripts [22:41:21] now I must drink one hundred fifty milliliters of vodka. [22:41:35] really? with that janky regex tool? [22:41:53] "splendidly slimy" [22:41:56] The Wall Street Journal [22:41:58] nice [22:42:13] awight: noscript [22:42:20] i'm fighting the good fight [22:42:26] manually approving every script [22:43:24] O_o [22:43:35] Bass Soloist in Handel's Messiah at Carnegie Hall [22:43:37] dang [22:43:46] Thanks for pointing that out, horrifyingly I just discovered that this browser allows "non-intrusive advertising" [22:44:09] non-intrusive until my wife gets on the spamlist for viagra [22:44:20] yeah the ad gaskets are blowing all over the place [22:44:53] people say "ublock origin" is the new thing [22:45:08] or a curated list of dns blocks for your /etc/hosts [22:45:11] cool, thx [22:45:23] hehe yeah that's how I did it in the dark ages [22:45:30] a whole lotta localhost [22:45:49] "?" [22:45:59] sometimes simpler is better [22:46:18] i found this today and got happy: http://lcwireless.net/ [22:46:54] available in the twigs for miles around santa fe [22:47:01] loners [22:47:14] with friends [22:47:36] reminds me of when the guy moved into my apartment building and stuck a giant ham mast on the roof. [22:47:56] I was getting German over my electric keyboard the rest of my tenure there. [22:48:49] hehe [22:48:53] gotta go wash my one pair of underwear in the sink again... we're still recovering from massive multitheft [22:49:02] oh shit [22:49:14] your stuff got stolen? [22:49:15] unrelated incidents! [22:49:17] yah [22:49:21] like everywhere we went [22:49:28] it's sort of hilarious how light we're traveling now [22:49:38] oh man i'm sorry to hear that [22:49:48] nah it's a nice family bonding experience [22:49:56] heh i can see it [22:49:59] ... and we've become hell of street. [22:50:27] 50€ notes tucked up in all the private places [22:50:36] cardboard umbrellas [22:51:21] the kid *loves* it [22:51:33] like we're living out Oliver Twist or something [22:51:56] heh totally idyllic childrens book fare [22:52:43] Even the european elevators are starting to be funny. this hotel has one so small I can only get a side-thumb onto the buttons once we all pile in [22:53:04] cannot rotate the arm into position to beep properly [22:53:12] stairs ftw [22:53:31] (PS1) XenoRyet: Fix Handling of Curriencies Without Minor Units. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/297106 (https://phabricator.wikimedia.org/T137607) [22:53:44] i like when the stairs are properly circular so you can just roll down [22:54:12] k. feel free to email weird questions, if anything else comes up. it's nice to pretend i have important work things waiting for me ;) [22:54:34] thanks for stopping in! makes me feel better about letting this sit for the weekend [22:54:38] * awight waves at XenoRyet [22:54:43] Hey there [22:55:05] that x100 standard is pretty awesome [22:55:14] Right? [22:55:19] (CR) jenkins-bot: [V: -1] Fix Handling of Curriencies Without Minor Units. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/297106 (https://phabricator.wikimedia.org/T137607) (owner: XenoRyet) [22:55:27] I'm always happy to find that PSPs have as hard a time with the edge cases as they ought to [22:56:03] Be nice if they all decided to handle it the same way, but that's asking too much I suppose. [22:56:25] XenoRyet: hope you're reunited with Jr., it must have been both great and horrible to temporarily separate [22:56:37] We'd be out of a job :p [22:56:53] 125 test fails, 126 test passes, wat [22:57:05] Yea, what's that about? [22:57:20] supersymmetry. plus one [22:57:25] awight: Yea, nice to have some time off from parenting, but good to be back too. [22:57:42] considering that 125 is EOL i wouldn't be shocked if CI for it was broke [22:58:03] 22:54:22 Error: your composer.lock file is not up to date, run "composer update" to install newer dependencies [22:58:09] pinche > 22:54:22 wikimedia/composer-merge-plugin: 1.0.0 installed, 1.3.1 required. [22:58:21] I saw that the other day. what oh [22:58:31] man. [22:58:59] And another thing that makes me angry that my career involves computers: I tried to do the simplest freaking thing with a single Symfony component the other day [22:59:45] of course, I was forced to immediate suck the entire "framework-bundle" into scope, create a configuration file, fire up a "httprequest" application and so on. [22:59:51] mortifying [23:00:07] talk about illegal and undocumented [23:00:17] heh [23:00:40] ugh [23:00:47] this is why we end up with n+1 everything [23:01:04] It was funny cos it had been long enough that I had forgotten the pain of staring at a thing that should be incredibly simple, and realizing that it could take an indeterminate amount of suffering to solve [23:01:16] awight: cwd; XenoRyet boo! [23:01:28] there could be n^3 bad beneath the surface [23:01:36] awight: really sorry to hear about your stuff losses [23:01:39] AndyRussG|sortof: bonjour! [23:01:53] like the other day when i was like "surely there's a stupid simple static image gallery generator with no dependencies out there" [23:01:58] a few hours later i was like https://github.com/caseydentinger/galsh/blob/master/gal.sh [23:01:58] * awight knocks on remaining personal item [23:02:07] heh [23:02:21] wow [23:03:00] the thing that happened to me was, I wanted a CI job that linted yaml. https://gerrit.wikimedia.org/r/#/c/296766/ [23:03:03] *not* [23:03:30] awight: if you have prod ssh keys with you and a yankering for running quick SQL prod query, lmk :) [23:04:06] I had to fashion a cape out of sewage instead: https://gerrit.wikimedia.org/r/#/c/296775/ [23:04:20] AndyRussG|sortof: mmm. sure, can do [23:04:36] awight: cool!! Lemme think about the query now... ;p [23:05:27] Do you have access / yubi issues I can help with, though? [23:06:01] Mmmm administrativia-allergy issues, I haven't even made the task to get new prod keys [23:06:16] This is for prod prod, not FR prod [23:06:22] cwd: you had me at "files=(*jpg)" [23:06:44] Basically it's just to check how long ago, if ever, anyone used the "banner weights" function of CN [23:06:53] :D [23:07:00] or how often it's used [23:08:02] If u recall, it's the thing where, if you have more than one banner in the same bucket, targeting the same slice of users (wrt logged-in status and device) [23:08:36] AndyRussG|sortof: it used to be used quite heavily [23:08:39] then you can give them different weights, to control the likelyhood of spawning one banner or the other [23:08:56] Seddon: ah hmmmm.... And these days? [23:09:11] AndyRussG|sortof: sorry to rat you out, but: https://phabricator.wikimedia.org/T139213 [23:09:41] AndyRussG|sortof: cool, I can probably figure out that query [23:09:55] awight: ahhhhh I'm getting hives... I mean, access to Hive... [23:10:01] hehehe [23:10:03] awight: I mean, thx a million! :) [23:10:11] AndyRussG|sortof: probably less so. It was used to do what the traffic limiting ability we have now [23:10:14] is hive the prod you are talking about? [23:10:45] cwd: no... It's the prod db for CN on meta [23:11:08] gotcha [23:11:25] cwd: But awight just requested access 4 me again, that'll give me back access to Hive [23:11:52] i have hive keys but i don't know what else they are good for [23:12:01] this was one thing https://people.wikimedia.org/~cwdent/ [23:12:02] AndyRussG|sortof: something like this? [23:12:02] select * from cn_notice_log where notlog_end_weight != 25 order by notlog_timestamp desc; [23:12:52] P.S. My internet is verrry dicey, if I don't answer it's 'cause my phone battery died again (though weechatmux should capture backscrill) [23:13:12] dang, that came up empty. [23:13:22] Huh! [23:13:42] awight: maybe find all unique values of that field? [23:13:50] I guess that should just confirm, 25 [23:13:59] Seddon: do u know how long ago that feature was used? [23:14:17] looking empty--maybe i'm in the wrong table, that must be campaigns [23:14:52] aww, it's json now. [23:14:52] [notlog_end_banners] => {"B1516_0519_enmlWW_mob_lw_sm_ind":{"weight":25,"bucket":0},"B1516_0519_enmlWW_mob_lw_sm_cnt":{"weight":25,"bucket":1}} [23:14:53] Seddon: the idea is to maybe remove the weighting feature, to make it simpler to add banner sequencing to the UI [23:16:29] awight: how about a join on cn_assignments and cn_notices, ordering by not_end (campaign end date)? [23:16:52] Seddon: what do you think of killing the feature? [23:17:23] AndyRussG|sortof: I'd need to look into what it's usage has been over the last twelve months [23:18:03] Seddon: yeah exactly.... :) [23:18:49] gotta go, have a good weekend everyone [23:18:59] AndyRussG|sortof: aha, the assignments table has the weights, thank you. [23:21:05] AndyRussG|sortof: To be honest I would rather not see a feature be dropped for the sake of UI [23:21:27] Seddon: even if it were a sinful feature? :p [23:23:14] Seddon: it's a good principle if it's a useful feature... The flip side is, if no one is using it, it may be sucking effort out of people by making the UI harder to use.... [23:23:57] If there are legit use cases that we can't cover otherwise, we should keep.... [23:24:20] It'll also be possible to clean up the UI for banner sequences even if we keep banner weights... [23:25:35] Seddon: looking for the mockup dstrine-away made, one sec... [23:29:13] Awight: Then we should fix the feature. Getting rid of something that has at least in the past proven it has it's use in place of something that at the moment is untested seems a tad silly [23:30:07] Seddon: the only use I'm aware of was to balance banners against "blank" banners, which was crazy enough that it drove me to implement the "throttle" feature to deprecate it [23:30:27] Seddon: I fully agree that we shouldn't surprise anyone by taking away a feature they use, though [23:31:02] If we decide it's unused, we should still announce a deprecation period or something, etc [23:31:26] * awight suspects Seddon is also on CEST [23:31:30] what are you doing awake... [23:31:31] awight: If noone has used it in the last 18-24 months then I am happy to give way [23:31:51] awight: I am awake at all times of the day [23:32:05] awight: don't tell PPena though :P [23:32:07] yah there's no need to do anything sudden here, I think AndyRussG|sortof's timeline is probably on the months scale [23:32:10] hehehe [23:33:56] Seddon: awight: https://phabricator.wikimedia.org/T135398 [23:34:54] AndyRussG|sortof: I think I have the answer, and the last usage was 2014 [23:35:41] awight: fantasmic! thx much! [23:36:38] I emailed the two queries and results, happy to iterate tomorrow if you see any obvious problems with the methodologemonon [23:36:51] awight: thx!!!! [23:37:15] it's the least I could do! [23:37:17] awight: Seddon: we could at least make sequences and banner weights not together-enableable [23:37:28] +1 [23:37:45] now, to warsh the last unstolen clothing ;)