[17:41:39] mwalker: You're going to throw knockout in DI, aren't you? [17:43:54] K4-713, yep [17:44:16] ...Thought so. [17:44:18] but... like I was saying in the other channel; you don't actually need the KO runtime on the client if you don't actually need it [17:44:31] Yeah, I lurked that whole thing. [17:44:45] I mean, we might. [17:44:57] But probably not. [17:45:39] not really... there's some stuff we could do with dynamic forms based on card types [17:45:41] Oh dear. Driving volunteers to DI will be... entertaining. [17:45:43] but it's probably overkill [17:45:55] I was going to do unsubscribe actually [17:45:59] or centralnotice [17:46:14] Unsub is probably a better starting point... [17:46:37] Less insane going on there, probably easier for people to set up locally. [17:47:42] Actually, if you're trying to make a point that people can poke at, *anything* is easier to set up locally than DI. [17:48:33] unsub has three templates; it's a great starting point for demonstraing basic client side functionality [17:48:39] sorry, server side functionality [17:49:34] Grand. I'm good with that. [17:57:32] Jeff_Green, you had a comment yesterday that we missed each other one -- about all the mysql queries causing a pager storm if the DB went down [17:57:52] I was really hoping icinga somewhere had a method for declaring test dependencies [17:58:21] it seems to, but we don't use it [17:59:07] looks like it has to be hand configured in conf files, not via web UI, and I don't see any evidence we've ever tried to use it--nothing puppetized, nothing shows up in the UI for dependencies [18:02:03] hmmm... so basically I'm in a world of pain if I suggest we try it out [18:04:29] imo it would be simpler to make a test that has its own simple logic [18:05:04] is there any reason it would not be good enough to have a script that runs a series of queries for all the civi queue stuff and warns/criticals if any of them are out of bounds? [18:18:51] Jeff_Green, it's just reinventing the wheel [18:18:59] there's nothing wrong with it [18:19:12] it'll be a wheel [18:19:22] it'll be a wheel either way :-) [18:20:03] * awight checks if we are using BSD now... [18:20:15] or if there is a major cheese order en route... [18:23:03] #1383: (AW) O:AW|TS:B|P:MH|TT:CT|IU:N|T:TT Description changed -- https://wikimedia.mingle.thoughtworks.com/projects/online_fundraiser/cards/1383 [18:25:20] mwalker: I made something for you :p ^^ [18:29:02] heh [18:32:14] any time a program tells you that "simplification failed", we're going to get an angry phone call :p [18:32:24] Computer, simplify dammit! [18:41:39] K4-713: hehe we need to templatize "caused {some} recurringdonations to pause for {suck-period} ...." "it will process {normally_or_not} going forward" [18:42:15] "we will {fudge-probability} probably not charge you for the skipped {suck-period} months" [18:42:35] Bah [18:42:57] I mean, the first line seems to be accurate. [18:43:07] oh and... "if you would like to recommend a friend for a new fr-tech opening, please call this 800 number" [18:43:10] jesus [18:43:37] The latter, though, might be better off "We certainly won't charge you for missed months, and you might want to check your billing statements to make sure we actually got you this time." [18:43:48] lol [18:44:15] If you think we are failing to take your money, *do not start another subscription* because we will almost certainly boff that as well [18:44:58] "If you would like to cancel your subscription, please call your credit card company to have your entire account cut off." [18:45:52] There are definitely reasons I don't get to write copy. [18:46:13] (if there aren't any, there should be) [18:46:16] <_< [18:46:20] >_> [18:47:01] oh man that was a good one. i chuckled long enuf that the baby started laughing [18:47:33] ... I love when she gets really tired and just makes a bunch of beavis n butthead sounds... [18:47:40] bahaha [18:48:47] If I look to see how many things we have undeployed in DI, what reaction am I likely to have? [18:49:46] I ask, because this mega patch I'm working on in master is behaving very oddly in several ways, and I can't see how it could have been anything I did. [18:49:56] "It appears you have begun a recurring subscription to the wikimedia foundation, a group of overpaid bozos who are doing wonderful things with your money, while chronically underfunding the team actually responsible for processing your contribution. If you think this charge was in error, please panic and reply-to sue gardner. [18:49:56] But it's clearly not happening in prod. [18:50:04] K4-713: you are scaring me now [18:50:29] * K4-713 chokes [18:50:43] See, there's reasons *none* of us get to write copy! [18:51:05] hmm, DI is suspiciously up-to-date [18:51:12] Which as we know is not possible. [18:51:16] Hm. [18:51:23] WTF? [18:51:25] Maybe I need to... local... [18:51:31] Wait. [18:51:42] When you reverted all of the things, did you do it in the branch, or just on boron? [18:51:55] no I'm thinking something really sinister is going on with the merge counters [18:51:59] Last weekend. [18:52:01] I didn't actually revert--what happened [18:52:09] is that I accidentally deployed DI-master [18:52:19] * awight loosens collar [18:52:25] I'm still really... confused re: how that happened in the first place. [18:52:49] I think cherry -v is lying, and we have to actually diff [18:52:49] I didn't think you could even do that on purpose. [18:52:56] I always diff. [18:53:12] um, I put the payments DI submodule pointer on DI-master [18:53:17] pretty easy to do [18:53:24] OH. [18:53:33] yeah including tests and everything :/ [18:53:33] Yeah, that would do it. [18:54:12] oops [18:55:32] well... diff says the only outstanding change is the appendLanguageAndMakeURL http_build_url stuff you and mwalker found. [18:55:45] (not including tests/ of course) [18:55:54] Huh. [18:55:57] Okay, cool. [18:56:22] IMO we should revert that on master if we can't deploy it [18:57:16] 3d1d12f5 [18:57:55] okay cherry -v does admit that is not merged [18:57:57] Yeah, I think we probably should do that. [18:58:03] but it's waaaay up the list of commits [18:58:09] Heh. You'll probably like some of the things that I had to do. [18:58:20] card_type, for instance, now only exists on payment forms. [18:58:20] omg, for GC OID? [18:58:35] Why that was related to oid, is a longass story. [18:58:48] hehe it sounds like I might actually like, yes [19:00:03] Seems that this is going to contain the logline refactor, too. [19:00:11] ha ha [19:00:23] It's like pulling a loose end on a sweater. [19:00:26] u want to make small commits for my sanity? [19:00:32] Sadly, that's... [19:00:38] and for rollbackability... [19:00:47] ...there is no cutoff that wouldn't break the whole thing so far. [19:01:02] that's crazy talk. [19:01:10] Rolling back isn't going to be easy/pretty, whatever we do. [19:01:31] I know the feeling... but making logline refactor a predecessor cannot actually be coupled to other work? [19:01:34] oh god [19:01:35] But, honestly, most of my time I'm spending trying to figure out how *not* to include things in the initial commit. [19:01:37] I can imagine [19:02:11] Happily, I am in cleanup Round 1. [19:02:16] There is a switch. [19:02:27] Isn't there some Oddysey parallel to this? Hehe, circumspect Penelope unraveling the shawl [19:02:29] no that's not it [19:02:39] In each gateway. When it's on, we generate oids the old way. When it's off, we get the new way. [19:02:46] nice [19:02:56] We... might have to go by payment type, though. [19:03:07] Which should be trivial after the first mess goes out. [19:03:17] Did I tell you we really did run out of numbers? [19:03:31] yeah I saw [19:03:32] unreal [19:03:56] Seriously. [19:04:00] Sascha is great, I have high hopes [19:04:09] She can be great. [19:04:34] She can also demand a phone call in which she reads my day-earlier email back to me, and says: "Huh. I don't know. I'll have to look in to that." [19:04:38] hehe. Well apparently she'll be their most experienced tech person when you're done with the grilling. [19:04:50] crap. I was hoping like she was not that same person [19:05:06] It's got to be some kind of bullshit internal SLA metric. [19:05:11] I've seen this before. [19:05:23] no c'mon [19:05:30] Some high-up says "You must get them on the PHONE within 24 hours, email is NOT ACCEPTABLE." [19:05:42] I was in one of those. [19:05:52] The moves are... easy to spot from the outside. [19:06:26] Tell her you moved to a swiss banking island and it's illegal to make phone calls [19:06:45] Particularly when multiple people from the same org have the same moves... [19:06:52] mein gott [19:07:10] * awight suddenly finds work to do [19:07:15] bahaha [19:07:16] okay [19:07:20] I see how it is. [19:07:23] :p [19:07:24] this is making me uncomfortable, I'm going to have to ask you to stop [19:08:05] hehe on that note, I wrote a two-page parenting tract on NERF last night, narrowly held myself back from actually sending [19:08:42] Oh yeah. Hold that one for a... good moment. [19:08:52] Email smokebomb. [19:09:38] yeah however I think not only do I have a creepily airtight thesis, but the main intervention is going to have wild success [19:09:52] will discuss irl :p [19:09:58] Interesting. [19:10:27] but long story short, I'm going to host a gun trade-in, and bring a small bandsaw to the office. [19:10:35] we'll fucking make desk art suckas [19:10:55] Sorry... I'm stuck on "portable bandsaw". [19:11:01] no shit [19:11:09] I've taken weirder things on BART [19:11:15] I believe you. [19:11:40] 15 flats of strawberries and a 8'x3' bike trailer... [19:12:06] I'm always deeply in awe and respect of people who move via multiple garbage bag or even garbage cans [19:12:42] I have also been studying ferry plaza drumkit guy. Nesting buckets! who would have thunk. [19:13:37] * awight fleetingly wonders if can be more of a chatty cathy occupational hazard in IRC or RL? [19:14:28] Logged forever? Probably depends on the topic. [19:14:30] :D [19:14:54] woot [19:15:36] mwalker: you want that TY UI thing, or shall I drag it into my lair? [19:16:41] * awight dullens fillet knives [19:16:50] I dont presently have time I dont think [19:16:56] so it's yours if you want it [19:17:01] #1381: (AW) TS:DR -- https://wikimedia.mingle.thoughtworks.com/projects/online_fundraiser/cards/1381 [19:17:01] #1381: (AW) AT:AW|TS:ID -- https://wikimedia.mingle.thoughtworks.com/projects/online_fundraiser/cards/1381 [19:17:53] AGH. That change to appendLanguageAndMakeURL() got me again. [19:18:01] mwalker: if sounds fun, it can wait [19:18:07] for your availability [19:18:18] if a chore, I don't mind at all [19:22:02] Hehe baby is shouting into her broccoli-talkie [19:24:02] #1186: (AW) O:MW -- https://wikimedia.mingle.thoughtworks.com/projects/online_fundraiser/cards/1186 [19:24:02] #1381: (AW) O:MW -- https://wikimedia.mingle.thoughtworks.com/projects/online_fundraiser/cards/1381 [19:24:28] (PS1) Katie Horn: Revert "FR #1113 Rather more intelligent thank you link generation" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/114769 [19:24:55] awight, mwalker: CR please. This is killing me. [19:25:22] (CR) Adamw: [C: 2] Revert "FR #1113 Rather more intelligent thank you link generation" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/114769 (owner: Katie Horn) [19:25:27] (Merged) jenkins-bot: Revert "FR #1113 Rather more intelligent thank you link generation" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/114769 (owner: Katie Horn) [19:26:41] well... that was one way of doing it [19:26:43] the other was https://gerrit.wikimedia.org/r/#/c/114053/ [19:27:23] we can still do that? [19:28:25] does your followon patch fix the thing katie was reporting wrt: TY language-basedURL? [19:29:25] it fixes the crash caused by a missing function; I'm not aware of other issues [19:29:38] it's in the revert commit message [19:30:34] it was working for me on my local? [19:30:47] Were you going to https://wikimediafoundation.org/wiki/Thank_You ? [19:31:05] K4-713: is it possible your TY url global variable is borken? [19:31:12] No way. [19:31:26] $wgDonationInterfaceThankYouPage = 'https://wikimediafoundation.org/wiki/Thank_You'; [19:31:42] Took me to https://wikimediafoundation.org/wiki/En instead. [19:32:02] Also logged the hell out of it and got back to that function. [19:32:05] So... [19:32:43] I guess I can test with the new suggestion and see if it gets me to the right TY language variant. [19:32:56] Why did we go down this road in the first place? [19:33:03] just... keep going on OID :p [19:33:15] I CANT ITS BROKEN BECAUSE THANK YOU. [19:33:24] * awight looks into combat pay raises [19:33:27] ^^ worst sentence ever [19:34:09] So, yeah: I'm blocked on this. [19:34:34] nvm, I was trying to say: revert is committed, run like hell! [19:34:38] har [19:34:44] okay [19:34:58] * K4-713 does branch voodoo to retest [19:35:12] ancient mysteries... [19:36:51] Oh good. Merge conflicts. [19:38:58] you're correct -- the original patch didn't work... what I meant to say was that when I tested the new patch it worked as expected [19:39:49] oic [19:41:39] ...hm. Why am I getting a security alert on the english TY page? [19:42:02] the-wub: ^^ [19:42:22] so frustrating... anyone here have experience with MariaDB 10.x causing about 2000 transfers/sec trying to do basic innodb stuff? I did change the fsync mode to O_DSYNC and turned off binary logging. Not happy! [19:42:45] K4-713: I don't. double-checking... [19:42:50] Uh. Weird. [19:43:03] oh i do! [19:43:23] hehe it's mixing https and http, first of all [19:43:30] yep. Probably not ideal. [19:43:39] K4-713: Special:HideBanners is loaded insecurely [19:43:57] Is this normal and I just never noticed? [19:44:00] Oh the wildcard is for *.wikipedia.org [19:44:11] That's... the weak sauce [19:44:17] I asked for loud sauce [19:44:39] yes, it has been like this forever afaik [19:44:46] oic [19:44:50] * K4-713 heads back for the cave [19:45:01] thank you for your donation! you are now entering shark-infested waters ;) [19:45:01] I think peterg tried to fix it at some point, but had to revert his changes. not sure why. [19:45:03] Eh, we should probably throw a card in for this, though. [19:45:33] what lovely shadows we cast today [19:45:41] I wonder if anybody ever complains about it. [19:46:51] https://wikimediafoundation.org/wiki/Template:Hide_banners is the scene of the crime [19:47:00] :) [19:48:38] I wonder if the reason we have it explicitly http is for shared cookies... [19:52:53] And, of course there aren't any comments in the template. Ah well. [19:53:21] That would have been too easy. [20:00:50] so... k4; the plan is to translate the I'm sorry message, send that out as a one off job; and then start recurring people when they next would recur? [20:03:32] ah, K4-713 ^ (pinging actually helps) [20:03:53] hurf? [20:04:22] Ah. Actually I was thinking that would go slightly differently. [20:04:33] Unless I'm misunderstanding what you just said. [20:05:35] I think we should take the apology sections and conditionally add them to the TY when we detect we have blown it. Then set that group to recur as normal, and get the modified TY. [20:06:08] But, not as a special job. The only intervention I was picturing was resetting their next recurring date to this month. [20:06:20] ...and letting the regular stuff pick it up. [20:07:09] Also: We should probably failmail ourselves when we apologize > 0 times in the recurring job. [20:08:00] 'course, the problem-detecting code is probably more complicated than it seems like it should be. [20:08:23] ok; so I'm going to add this text to the thank you letter on metawiki; I don't know how many non english people get it; but it means we probably wont be able to re-enable recurring till late next week [20:08:54] Why do we need to turn it off? [20:10:42] we won't be turning it off [20:10:52] re-enable... the people we lost? [20:10:55] I just can't reset the broken jobs until we have the translations [20:11:14] Aha. Yeah, that makes sense. [20:11:27] And, I mean, we shouldn't get them until their selected calendar date anyway. [20:11:45] 26th, give or take... [20:11:52] Yeah, that works for me. [21:42:35] (CR) Pcoombe: [C: 2 V: 2] New thank you messages (1 comment) [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/114515 (owner: Mwalker) [22:12:28] !log updated civicrm from eb3536eb32cbc7400e4e5884d56fbf104e38fc2b to 41dce289bc15ea1ca638c37b29ff2e3e709a2251 for thank you templates [22:12:35] Logged the message, Master [22:30:19] aah crap. https://wikimediafoundation.org/wiki/Special:LandingCheck?landing_page=Cancel_or_change_recurring_payments&basic=true&country={{ country }}&language={{ language }} [22:30:44] Is that actually sensitive to country? [23:10:21] (PS1) Katie Horn: Big Order ID Refactor * This thing does a lot of things. Srsly, look out. * GC seems to be okay with either OID generation method now. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/114894 [23:10:54] (Abandoned) Katie Horn: Big Order ID Refactor * This thing does a lot of things. Srsly, look out. * GC seems to be okay with either OID generation method now. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/114894 (owner: Katie Horn) [23:29:20] (PS1) Katie Horn: Revert "Big Order ID Refactor" This was headed for a sandbox and ended up in Master. Oops. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/114895 [23:32:29] (CR) Adamw: [C: 2 V: 2] "You are not crazy." [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/114895 (owner: Katie Horn) [23:42:16] (PS1) Mwalker: let's not allow us to push things directly to master shall we? only allow pushing to the deploy branch and sandboxen. [extensions/DonationInterface] (refs/meta/config) - https://gerrit.wikimedia.org/r/114898 [23:44:21] what the... ^^ [23:44:31] (CR) jenkins-bot: [V: -1] let's not allow us to push things directly to master shall we? only allow pushing to the deploy branch and sandboxen. [extensions/DonationInterface] (refs/meta/config) - https://gerrit.wikimedia.org/r/114898 (owner: Mwalker) [23:46:27] mwalker: ^^ isn't there a gerrit UI knob for this? [23:46:42] yes; I'm using it [23:46:56] how. what. [23:47:14] configuration is stored in git [23:47:19] you can either have it commit directly [23:47:25] or have it go through a review process [23:47:29] https://gerrit.wikimedia.org/r/#/admin/projects/mediawiki/extensions/DonationInterface,access [23:47:34] I chose to review it to see if I knew what I was doing [23:47:43] oh cool [23:47:45] and since it turns out that what I did has side effects [23:47:51] I've not yet applied it :) [23:47:57] until I understand what they are... [23:51:03] yeah it looks like those changes would still allow us to push to master? very hard to say