[00:50:56] Fundraising Sprint Hermit Crab Husbandry, Fundraising Sprint Internet Exploring, Fundraising Sprint Jabberwock Slaying, Fundraising Sprint Killing Time, and 10 others: [Epic] Consolidate queue abstractions - https://phabricator.wikimedia.org/T131271#2464824 (awight) [00:50:58] Fundraising Sprint Internet Exploring, Fundraising Sprint Killing Time, Fundraising Sprint Licking Cookies, Fundraising Sprint Muggle Baiting, and 5 others: Implement AtomicReadBuffer for choice of PHP-Queue backend - https://phabricator.wikimedia.org/T133964#2464823 (awight) Open>Resolved [00:51:46] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, FR-ActiveMQ, FR-Smashpig, and 3 others: [Epic] Deprecate old pending queues - https://phabricator.wikimedia.org/T133433#2464825 (awight) [00:53:07] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, fundraising-tech-ops, FR-ActiveMQ, and 3 others: [Epic] SPOF: Replace ActiveMQ donation queues with a more robust software stack - https://phabricator.wikimedia.org/T108229#2464829 (awight) [00:53:08] Fundraising-Backlog, FR-ActiveMQ, Patch-For-Review: Write PHP-Queue module for new backend - https://phabricator.wikimedia.org/T131269#2464827 (awight) Open>Invalid These are already implemented for Predis + Redis. [00:53:31] Fundraising Sprint Hermit Crab Husbandry, Fundraising Sprint Internet Exploring, Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, and 4 others: [Epic] Formalize schema for all message types - https://phabricator.wikimedia.org/T131272#2464830 (awight) [01:01:15] Fundraising-Backlog, FR-ActiveMQ, Epic: Spike: Investigate potential for banner impressions rewrite - https://phabricator.wikimedia.org/T131278#2464835 (awight) [01:01:57] Fundraising-Backlog, Epic: [epic] Make WorldPay more robust - https://phabricator.wikimedia.org/T77909#2464839 (awight) @DStrine Please tell me I can take the WP tasks off of life support? [01:04:05] andrewbogott: got it, thx! [01:04:07] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, FR-ActiveMQ, FR-Smashpig, and 3 others: [Epic] Consolidate "pending" queue usages - https://phabricator.wikimedia.org/T130897#2464843 (awight) [01:06:06] Fundraising-Backlog, FR-ActiveMQ: Upstream whatever we can to PHP-Queue - https://phabricator.wikimedia.org/T133754#2464850 (awight) [01:06:08] Fundraising-Backlog, FR-ActiveMQ: Remove KeyValueStore from PHP-Queue - https://phabricator.wikimedia.org/T133190#2464851 (awight) [01:06:33] AndyRussG: hi! [01:07:00] I left a note on the Purge wiki... those machines look like fair culling to me [01:08:07] dstrine: loving the silly new gantt view, btw! [01:11:29] dstrine: hold on; no way. But when more than 100 tasks are involved, it just shuts off rather than falling back to simply listing parents + children [01:11:33] T108229 [01:11:34] T108229: [Epic] SPOF: Replace ActiveMQ donation queues with a more robust software stack - https://phabricator.wikimedia.org/T108229 [01:11:41] garbage. [01:12:03] * awight sulks off to "edit parent tasks" [01:12:31] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, fundraising-tech-ops, FR-ActiveMQ, and 3 others: [Epic] SPOF: Replace ActiveMQ donation queues with a more robust software stack - https://phabricator.wikimedia.org/T108229#2464854 (awight) [01:40:23] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, FR-ActiveMQ, Epic: [Epic] Deploy new queue to a pilot topic - https://phabricator.wikimedia.org/T130306#2464951 (awight) [01:42:10] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, fundraising-tech-ops, MediaWiki-extensions-ContributionTracking, and 2 others: [epic] SPOF: Use Redis as backend store for contribution_tracking - https://phabricator.wikimedia.org/T119556#2464957 (awight) [01:42:12] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, fundraising-tech-ops, MediaWiki-extensions-ContributionTracking, MediaWiki-extensions-DonationInterface: Deploy Redis 3 to frack - https://phabricator.wikimedia.org/T120464#2464956 (awight) stalled>declined [01:42:13] awight: hey [01:42:53] cool, thx... yeah I didn't think we had much in the way of currently used instances, or at least not for CN... Didn't know if there was a risk of losing more than instances [01:50:01] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, fundraising-tech-ops, FR-ActiveMQ, and 3 others: [Epic] SPOF: Replace ActiveMQ donation queues with a more robust software stack - https://phabricator.wikimedia.org/T108229#2464988 (awight) [01:56:50] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, FR-ActiveMQ, FR-Smashpig, and 2 others: [Epic] Payments and other components should send a copy of outbound queue messages a Redis queue - https://phabricator.wikimedia.org/T130287#2464998 (awight) [01:57:02] (CR) AndyRussG: "I think it'd make sense to squash this commit with the follow-on patch (I179e9ddab2). There's stuff that's changed here, then gets moved a" (2 comments) [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/186899 (https://phabricator.wikimedia.org/T108721) (owner: Awight) [01:58:10] AndyRussG: True, there might be a risk of blowing away home directories and unsaved work. I think homedirs are kept on a central server, though? [01:59:15] awight: no idea at all... For me, I have nothing worth saving there. [01:59:32] Fundraising-Backlog, FR-ActiveMQ, FR-PayPal-ExpressCheckout, FR-Paypal, FR-Smashpig: Rewrite PayPal listener under SmashPig - https://phabricator.wikimedia.org/T130851#2465015 (awight) > DStrine moved this task from Sprint +3 to Sprint +1 on the Fundraising-Backlog board. Great idea! Knocki... [01:59:37] What I meant tho is that at least we keep a project/privileges to create instances, no? [02:00:32] Did something change with our varnish caching or RL module adding and it no longer takes a month to roll over? I can't find any cached HTML with bannerControllerz [02:00:44] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, fundraising-tech-ops, FR-ActiveMQ, and 3 others: [Epic] SPOF: Replace ActiveMQ donation queues with a more robust software stack - https://phabricator.wikimedia.org/T108229#2465016 (awight) [02:01:03] awight: howdy! way to take the day off [02:05:34] Fundraising-Backlog, FR-ActiveMQ: fundraising-tools queue module should be abstracted and support Redis - https://phabricator.wikimedia.org/T130308#2465020 (awight) [02:06:51] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, FR-ActiveMQ, MediaWiki-extensions-DonationInterface, Epic: [Epic] Migrate contribution tracking to new queue - https://phabricator.wikimedia.org/T131279#2465022 (awight) [02:08:36] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, FR-ActiveMQ, MediaWiki-extensions-DonationInterface, Epic: [Epic] Migrate contribution tracking to new queue - https://phabricator.wikimedia.org/T131279#2161951 (awight) [02:08:38] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, fundraising-tech-ops, MediaWiki-extensions-ContributionTracking, and 2 others: [epic] SPOF: Use Redis as backend store for contribution_tracking - https://phabricator.wikimedia.org/T119556#2465044 (awight) [02:10:41] seeya! [02:11:01] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, FR-ActiveMQ, FR-Smashpig, Epic: [Epic] Rewrite all queue clients to use a single shim library, improve library - https://phabricator.wikimedia.org/T133108#2465045 (awight) [02:13:47] Fundraising Sprint Muggle Baiting, Fundraising-Backlog, Patch-For-Review: Make RapidHTML work with mediawiki 1.26+ - https://phabricator.wikimedia.org/T139551#2465046 (awight) [02:18:56] cwd: Anything I can do to help push us onto 1.27 turf? [02:19:32] If it's just a straight backport, I can do that... [02:19:48] awight: ejegg cut the new branch today [02:21:01] i do not know of any necessary backports [02:22:14] awight: here was the main FR overrides https://gerrit.wikimedia.org/r/#/c/299035/ [02:22:20] sick! [02:22:48] ooh, I like that style. [02:22:58] Might as well keep it concise... [02:23:21] i think this was diff if FR 25 vs tarball 25 applied to 27? [02:23:29] *of [02:23:40] We can discuss a naughty Friday deployment, as long as there's a way to rollback to 1.25 without ops handholding [02:24:07] awight: wellllll i think there is going to be a CI mess [02:24:33] (CR) Awight: "I like this squashed style!" [core] (fundraising/REL1_27) - https://gerrit.wikimedia.org/r/299035 (https://phabricator.wikimedia.org/T138016) (owner: Ejegg) [02:24:56] cwd: right. Thanks for patiently re-reminding me :) [02:25:31] On that note, do you think it'd be greedy to just ffwd to trunk? [02:25:49] While we're applying putty and paint to the CI anyway... [02:26:01] instead of deploying 27 you mean? [02:26:06] yah [02:26:26] my gut feeling is it wouldn't hurt anything [02:26:33] ... this upgrade is starting to become a psych thriller rather than technical challenge... [02:26:43] hehe you got that right [02:26:58] 1/2 the CI headache appeals to me big-time... [02:27:31] if the goal is still getting on the train [02:27:32] Friday's not a goot day to get Euro help, unfortunately [02:27:37] yessir [02:27:39] but we could also just pin to 27 and hang out for awhile [02:28:00] , or as easily hang out on 28 [02:28:19] yeah that is true [02:28:29] i don't see a good reason not to [02:29:47] where is that short array cutover? that's compelling in both directions... [02:30:08] i think it's 27? [02:30:21] https://github.com/thomasbachem/php-short-array-syntax-converter [02:30:32] Would be good to get to the promised land quickly, vs. our migration will be hairy [02:32:08] do we cut all the submodule branches (besides DI) at the same spot as core? [02:32:46] Short array is 1.27. [02:33:01] cwd: that part is disturbing me. Oh--hold on [02:33:03] it's only the tests that care right? [02:33:08] it's fine now that we have php5.5 [02:33:20] we can migrate the extensions before mw-core [02:33:22] and CI [02:33:26] nice. [02:33:35] I can prepare that patch [02:33:57] might there be regressions in new extensions vs. old core? [02:34:12] cos of short array sugar? [02:34:21] that or...whatever [02:34:25] depending on new behavior [02:34:51] interesting. I think that would only be possible if it were doing things at the PHP parser level [02:35:31] i mean if core extension branches depended on new mw behavior [02:35:40] since 1.25 [02:35:51] I bet they do [02:36:02] ah I see the confusion, sorry [02:36:22] by "migration" I meant, refactor using short array syntax. [02:36:28] that was the wrong word ;) [02:36:50] ooh, yeah defintiely [02:37:10] I don't even remember... do we have to do all that, now? is that enforced by the default CI jobs? [02:37:15] in fact...isn't DI deployed on the wikis? is it weird that it passes CI with old arrays? [02:37:21] wat. [02:37:32] maybe it's totally optional. [02:38:04] i thought it was enforced but i don't know why i think that [02:39:11] Something traumatic must have happened to us. [02:41:03] I don't see it. I can't watch. [02:41:20] Maybe I'll get the "new" out of my system by irrevocably committing to traits somewhere. [02:41:42] * awight circles prey [02:41:43] https://gerrit.wikimedia.org/r/#/c/133508/ [02:42:20] cwd: I still donno what we should try to migrate to, now. Jeff foiled our conservative non-plans. [02:43:29] fwiw https://etherpad.wikimedia.org/p/Fundraising_MW_version_embarrassment_of_riches [02:47:50] gotta run in a second, my kid is charging towards me with Candy Land [02:49:03] hi5 mari [02:50:34] awight: i think 27 would get us a big enough window to roll our own app [02:51:00] and without the constant breakage of trying to keep up with the train [11:30:53] Fundraising Tech Backlog, Mail, Operations: Add granularity limiter (g=) to wikimedia.org DKIM record(s) - https://phabricator.wikimedia.org/T140316#2465591 (fgiunchedi) p:Triage>Normal [14:20:45] hi awight ! Early for ya [14:21:50] ejegg: I don't know why I'm computing. [14:22:03] it's more fun to compute? [14:22:07] compute [14:22:09] Got excited about deleting some code, I'm afraid to say [14:22:10] compute [14:22:14] hehe, yeah [14:22:26] (PS1) Awight: [WIP] Use traits to consolidate TestingAdapters [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/299150 [14:22:44] oho, taking advantage of 5.5 already! [14:23:14] ejegg: cwd and I were chatting for a moment about where to go now that Jeff slayed all the dragons [14:23:23] https://etherpad.wikimedia.org/p/Fundraising_MW_version_embarrassment_of_riches [14:23:49] (CR) jenkins-bot: [V: -1] [WIP] Use traits to consolidate TestingAdapters [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/299150 (owner: Awight) [14:24:01] awight: I'm pretty solidly for 1.27 now, then talk about keeping up with train or totally decoupling [14:24:16] 1.27 should be a pretty quick deploy and buy us a lot of time [14:24:23] ejegg: u know that 1.27 is short array syntax, though? [14:24:42] sure, there's a converter script out there that worked pretty well locally [14:24:44] The thing I can't tell is whether that actually affects us at all--can't see offhand where CI enforces it. [14:24:47] yah [14:25:14] What's the argument for 1.27? Just that it has a REL branch? [14:25:14] i figure wait till the rapidHTML stuff is merged, then run that [14:25:22] 1.27 is LTS [14:25:47] so if all sorts of crazy happens and we can't do either of the other good things, we are still sitting pretty [14:26:10] it's LTS, but that doesn't guarantee anything CI-wise [14:26:51] yah, but fixing busted CI is less of an emergency than 'no more security patches' [14:27:09] a tangent: I realized we can experiment with deploying any version, without needing Jeff to tweak the f_c_u config. We can version lock to commits on any branch. [14:27:24] oho, nice find! [14:27:35] right, x=abc123 [14:27:37] ejegg: that's not quite what we're weighing, though. [14:27:56] oh, sorry it totally is. [14:28:11] but both of those things count against 1.27 [14:28:33] eh? LTS means security patches, if not immediate [14:28:52] err, that we'll still be stuck in the security patch grey zone rather than being on the version that WMF ops cares about [14:29:44] AFAICT, the only advantage to stopping at 1.27 is that we'll be able to continue doing our REL backporting--but I can't convince myself that's a positive thing [14:29:55] Did we hash out exactly what the process is for staying up to date with trunk? [14:31:03] We'd definitely want a script to convert a wmfx branch to our fr style, 'cos it's not quite trivial [14:31:22] Not exactly but pretty close [14:31:24] though it didn't take me more than an hr or two to make 1_27 [14:31:35] goot point. btw I really like the squashed style you used [14:31:50] yeah, nice to see all the changes in 1 commit [14:33:07] what are u guise doing up so early [14:33:19] scheming [14:33:34] So CI-wise, if we are on trunk, what, we get to delete ALL of the DonationInterface exceptions and it Just Works? [14:33:46] whoa [14:33:59] I've been wanting to try that [14:34:11] I was gonna put the normal tests under experimental: [14:34:17] uhuh [14:35:58] so, if we can get up to master by tuesday or wednesday and it slays our CI headaches, I'd say sure, skip 1_27. Otherwise I'd like to deploy 1_27 monday [14:36:43] Right. Since you've already written the patch for 1.27, it shouldn't hurt to deploy it, and then we're on solid ice at least. [14:37:41] ejegg: was that squashed commit basically the changes from FR 25 vs tarball 25? [14:37:45] but applied to 27? [14:38:07] cwd yep, I just tried to mirror the 1_25 diffs [14:38:15] nice, good thinking [14:38:54] ejegg: It's possible that we could do a normal merge to hop between .wmf release versions [14:39:07] There shouldn't be conflicts w/o the REL branch mechanism [14:39:41] oh yeah, that oughtta work! We can leave the READMEs in place to avoid conflicts there [14:40:08] though don't the wmf branches have extension submodules? [14:40:18] ooh aargh [14:40:31] ok i needa ponder life on Bart [14:40:36] k [14:40:42] yeah there is submodule hell...one sec [14:41:14] https://github.com/wikimedia/mediawiki-tools-release/blob/master/make-wmf-branch/MakeWmfBranch.php so gnarly [14:41:35] we could make our own version of that thing [14:41:58] does frig have the scaffolding to do some of it already? [14:42:01] submodule hell is almost enough to sic me off of master for now [14:42:34] also keeping up with lib updates :S [14:43:04] some of the code in frig could certainly be used to do this [14:43:09] see ya in an hour or two [14:43:35] turning the wmf branch into a FR branch is probably a lot uglier than making our own from master [14:43:50] removing all the submodules and re adding some vs. just adding them [14:44:07] hmm, lightncandy and monolog are at the same version in trunk and 1_27 [14:44:36] hmm, would be even easier if we didn't have to include ParserFunctions and decide which commit to take [14:45:41] wasn't parser functions pretty light on responsibility? [14:45:44] I'm still leaning to deploying 1_27 monday and taking our time making tools for keeping up with master [14:46:06] partly so we can all start looking at queue stuff asap [14:46:12] yeah [14:46:26] my only coherent argument against that would be it's twice the CI drudgery in the long run [14:46:41] but it's the quickest way out of the hole we're in [14:49:27] ejegg: i will have the rtbt stuff reviewed some time this morning. by the time i got the vpn working last night my brain was shutting down. [14:49:36] thanks! [14:49:45] back in a bit... [14:49:58] me too [15:12:30] Fundraising-Backlog, fundraising-tech-ops, Operations, Patch-For-Review: Allow Fundraising to A/B test wikipedia.org as send domain - https://phabricator.wikimedia.org/T135410#2466161 (Jgreen) I think all the necessary DNS config changes are in https://gerrit.wikimedia.org/r/#/c/298500/ and that... [15:54:51] cwd: going to take a crack at a 1_27 / php55 CI job [15:56:38] ejegg: there should be one already no? just need to point it at our release [15:59:16] suddenly, inexplicably: The page you are trying to view cannot be shown because it uses an invalid or unsupported form of compression. [15:59:23] gonna be one of those days i guess [16:00:12] eh?? this is on local mediawiki? [16:00:24] oop, K4 time [16:01:18] oh vagrant [16:01:37] hehe [16:01:43] awight: does bart from oakland go under water? [16:01:55] did it poop on the carpet? [16:02:00] yeah [16:02:05] lots of very shallow water. [16:02:12] that's neat. can you see out through the water?? [16:02:20] hehe [16:02:30] in the future perhaps, when *only* rich people take the train [16:02:35] hehe, dang [16:02:37] wait. that's not happening. [16:02:42] that woudl be so cool though [16:02:53] slash terrifying [16:03:19] The bay area tech ideal seems to be, amusingly, a regression to taking the private bus to work. [16:03:37] s/bay area// [16:03:39] cwd: things jsut get dark and echo-y as you go under the bay. and it feels like a scene from Esacpe from New York usually :) [16:03:48] There's a really cool train in NYC that goes over the East River tho, and the view is unobstructed by vehicular traffic. [16:03:53] hehe [16:03:57] do your ears pop? [16:04:09] sometimes a little. It's not very deep though [16:04:31] probably the pressure from how fast the train is displacing air causes more effect than the depth [16:04:55] shallow so that it will be easier to recover the lost train after an earthquake traps it in the tube [16:05:02] bd808: btw, thanks again for the Scrutinizer recommendation. *way* more informative than codeclimate, it seems [16:05:04] D: [16:05:20] awight: awesome [16:05:23] yeah those parts are worth serious money--only military contractor Bechtel can do anything for Bart [16:05:37] cos they accidentally used a different gauge track than... well anything else ever built. [16:05:49] "accidentally" [16:05:49] nice [16:05:52] contract lock in [16:06:26] Bart has this hilarious "failsafe" background context... 7 computers talking to each other. How could that evar fail? [16:06:45] ugh [16:07:01] at some point xp service pack 1 will get installed [16:07:03] bart gridlock [16:07:23] yah you see windoze boot screens on Bart equipment all the time. [16:07:35] cwd: so why do you ask? You jelly that we have public transit at all? [16:08:08] heh yeah [16:08:13] We're also HQ to this historic wonder of the world: http://idlewords.com/2007/04/the_alameda_weehawken_burrito_tunnel.htm [16:08:19] i saw this at malpensa airport: http://i.imgur.com/409PI8Z.jpg [16:08:37] hey that's not so ancient [16:08:55] nah not too bad [16:09:23] ah i love everything this guy writes [16:09:29] i would love to see him talk [16:10:03] I seriously need to plug a few of these genioses into my RSS reader. I get so teary subsisting on just the real-world news. [16:11:07] i think rss is dead [16:11:10] i think google killed it [16:11:18] atom, whatev [16:11:26] long live the feed! [16:12:00] jesus, are you suggesting that pubsubhubbub is the future? [16:13:31] * awight quickly sticks head in hot CI sand to cauterize the wound [16:14:45] i think google reader was a ploy to centralize rss and kill it [16:14:56] the feed workflow in general [16:15:01] oh sick. https://pthree.org/2008/03/04/static-windows-in-irssi/ [16:16:11] i can no longer hang with console irc [16:16:48] no terminal i know can divide the screen vertically [16:17:04] i love to have scrolly nick/chan lists so much [16:17:06] isn't that what your WM is for? [16:17:15] ah [16:18:10] i mean you can have vertical nick list in console [16:18:17] but then there's a link that spans lines [16:18:25] you can't select it [16:18:31] without getting some names too [16:19:14] there's some way to test CI stuff without it getting merged, right? generate XML with JJB & do...something? [16:20:23] trying to see if https://gerrit.wikimedia.org/r/299170 is what we'd want for 1.27 DI tests [16:20:34] ejegg: yeah, you can push directly to the CI server using your API key [16:20:49] ah, lemme set that up... [16:20:50] but that does make it live [16:21:06] ejegg: yeah, stick your new job under the "experimental" trigger. [16:21:24] ok, lemme see [16:26:44] ... I wanted to try the generic jobs too, would you mind sticking those after your custom job? [16:27:17] just "extension-unittests-generic" [16:27:53] or we can do that in a separate test, I'm fine either way [16:29:52] brilliant https://ourincrediblejourney.tumblr.com/ [16:31:04] nicely consolidated blog! [16:31:14] clearly they're gunning for selloff as well [16:31:21] "buy us and we shut up" [16:32:04] (CR) Ejegg: "check experimental" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/299083 (owner: Ejegg) [16:32:52] hehe, devious [16:32:59] huh, that's running against hhvm [16:33:25] i got the job deployed to jenkins, but I think I'm missing a step to tell zuul it's part of the experimental trigger [16:34:34] ejegg: it goes in https://gerrit.wikimedia.org/r/#/c/299170/1/zuul/layout.yaml,unified [16:34:39] at line 2962 [16:34:40] * cwd shudders [16:34:47] "experimental:"... [16:35:22] awight: ah, I just push that up for review and it looks for it? [16:35:36] ymm [16:35:37] umm [16:35:40] hold up [16:36:01] I'm confused now. [16:36:04] well, here goes [16:36:16] ejegg: check line 2420 [16:36:20] it's like that. [16:36:22] (CR) Ejegg: "check experimental" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/299083 (owner: Ejegg) [16:36:30] Then you compile and push to the server with jjb [16:38:24] ah, I see you were able to push [16:38:55] huh, jjb doesn't look at the patch in review, does it? [16:38:58] seems to be missing the "experimental" trigger though [16:39:11] ejegg: no. but it looks like you pushed successfully? [16:39:15] k, I'll re-push now that I amended the patch [16:41:01] didn't change anything in jenkins [16:41:10] dang. Can I see the path? [16:41:13] patch [16:42:01] ejegg: that experimental: section looks right to me [16:42:05] yah, something definitely needs to be deployed to zuul [16:42:39] ejegg: so you're doing this, right? jenkins-jobs --conf etc/jenkins_jobs.ini update config/ '*-example' [16:42:47] s/example/real/... [16:43:07] yarp [16:43:37] ejegg: ok, maybe there's a zuul step also. [16:43:38] https://www.mediawiki.org/wiki/Continuous_integration/Zuul#Deploy_configuration [16:43:42] I've never done that one [16:44:20] huh, so merge is actually necessary [16:44:39] oh good, I'm not allowed to log into gallium. Probably for the best. [16:44:43] err, wait, i should be able to trigger thru jenkins, right? [16:44:55] awight: it might be darmstadtium now [16:44:58] I think so, but the ZUUL_ params might be jacked [16:45:24] ah dang [16:45:41] yeah, guess I shouldn't go fabricating uuids etc [16:48:04] asking in -releng [16:48:21] oops, I already dd [16:48:24] *did [16:50:04] thanks! Guess I'll drop this till monday [16:51:17] awight: I rebased that AstroPay signature thing - it is an improvement, and gets rid of some getTransactionSpecificValue abuse [16:51:23] https://gerrit.wikimedia.org/r/293465 [16:51:42] also, I'm not sure I understand your objection to https://gerrit.wikimedia.org/r/298549 [16:53:48] nice. okay I'll take looks [16:59:49] anyone want to get on the call? [17:00:12] agenda? [17:00:30] sure! [17:00:38] i'll be on in 5 [17:00:41] lemme see if I can book a conf room [17:00:56] Yea, give me a couple of mins and I'll get on too. [17:00:56] if not I'll join and just lurk/type [17:02:58] cool, got a room, be on in a sec [17:04:28] (PS2) Awight: [WIP] Use traits to consolidate TestingAdapters [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/299150 [17:05:33] (Abandoned) Awight: DO NOT MERGE: (requies PHP 5.4.0) Use traits to consolidate TestingAdapters [extensions/DonationInterface] (php54_test_adapter_collapse) - https://gerrit.wikimedia.org/r/133508 (owner: Awight) [17:06:04] (CR) jenkins-bot: [V: -1] [WIP] Use traits to consolidate TestingAdapters [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/299150 (owner: Awight) [17:20:58] Fundraising Sprint Muggle Baiting, Fundraising-Backlog, MediaWiki-extensions-DonationInterface, Patch-For-Review, and 3 others: Convert GlobalCollect credit card forms from RapidHTML to Mustache - https://phabricator.wikimedia.org/T123002#1918994 (Ejegg) [17:30:19] (CR) Awight: [C: 1] "PS 14..17 looking great!" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298633 (https://phabricator.wikimedia.org/T123002) (owner: Awight) [17:32:57] Fundraising-Backlog, MediaWiki-extensions-DonationInterface: All form submission should be handled consistently through donation.api - https://phabricator.wikimedia.org/T140482#2466470 (Ejegg) [17:35:12] (PS2) Ejegg: Composer validation fixup [wikimedia/fundraising/php-queue] - https://gerrit.wikimedia.org/r/295948 (owner: Awight) [17:35:31] (CR) Ejegg: [C: 2] Composer validation fixup [wikimedia/fundraising/php-queue] - https://gerrit.wikimedia.org/r/295948 (owner: Awight) [17:36:20] (Merged) jenkins-bot: Composer validation fixup [wikimedia/fundraising/php-queue] - https://gerrit.wikimedia.org/r/295948 (owner: Awight) [17:40:29] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, FR-ActiveMQ, FR-Smashpig, and 3 others: [Epic] Consolidate "pending" queue usages - https://phabricator.wikimedia.org/T130897#2466517 (Ejegg) [17:40:31] Fundraising Sprint Hermit Crab Husbandry, Fundraising Sprint Internet Exploring, Fundraising Sprint Jabberwock Slaying, Fundraising Sprint Killing Time, and 7 others: Write pending queue consumer and schema - https://phabricator.wikimedia.org/T133197#2466516 (Ejegg) Open>Resolved [17:44:10] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, FR-ActiveMQ: CiviCRM should read 'completion messages' from pending DB, not activeMQ - https://phabricator.wikimedia.org/T140484#2466525 (Ejegg) [17:44:55] cwd: lemme help with the vagrant hell... [17:45:02] ugh sorry i missed the last few minutes of that [17:45:16] this freaking connection [17:45:18] it should be reflected in the agenda [17:45:29] one thing I forgot I really wanted to do today was > Create columns on the overhaul workboard https://phabricator.wikimedia.org/project/board/1895/ [17:45:46] My thought was to tease out isolated threads of work... [17:45:47] great [17:45:51] yeah [17:46:42] Could not find gem 'mediawiki-vagrant (= 0.14.0) ruby' in any of the gem sources listed in your Gemfile or available on this machine. [17:46:58] this is a fresh copy [17:47:08] did you run ./setup.sh ? [17:47:45] cwd: my "gem sources" just says, https://rubygems.org/ [17:47:47] bd808: that's what i get from ./setup.sh [17:47:58] arrgh [17:48:09] that bug has been biting anomie for months [17:48:18] I don't know how to debug it [17:48:22] cwd: you on a chromebox or something? [17:48:24] glad it's not just me :) [17:48:32] heh no this is a regular computer [17:48:57] you could try gem sources -u [17:49:00] I don't know anything about it tho [17:49:05] i'll start pulling on some strings [17:49:21] i don't know nothing about ruby [17:49:24] but i'm good at google [17:49:28] https://rubygems.org/search?utf8=%E2%9C%93&query=mediawiki-vagrant [17:49:38] bd808: ^ you know anything about that? [17:49:42] where *does* the gem come from? [17:49:53] the gem is local in the lib folder [17:50:16] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, FR-ActiveMQ, FR-Smashpig, and 3 others: [Epic] Consolidate "pending" queue usages - https://phabricator.wikimedia.org/T130897#2466549 (Ejegg) [17:50:18] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, FR-ActiveMQ: CiviCRM should read 'completion messages' from pending DB, not activeMQ - https://phabricator.wikimedia.org/T140484#2466548 (Ejegg) [17:51:26] https://phabricator.wikimedia.org/T132449 [17:51:41] I'm crept out that mediawiki-vagrant isn't mentioned in the Gemfile. [17:52:12] cwd: just for fun, ruby -v -> ruby 2.1.5p273 (2014-11-13) [x86_64-linux-gnu] [17:52:22] gem -v -> 2.2.2 [17:52:41] yep [17:53:11] interesting [17:53:17] (PS2) Ejegg: Switch to caret semantic versioning [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/298418 (owner: Awight) [17:53:22] I wonder if it not being in the gemfile is somehow the problem. it's in gemfile.lock [17:53:49] but this only happens on some systems? [17:54:17] group :plugins do gemspec end -- that's the bit in the gemfile [17:54:19] ruby is voodoo to me :/ [17:54:36] me too [17:54:55] must be how the search path is getting built [17:55:05] maybe something in the env affects it [17:55:49] I wonder if we can change back to not actually installing as a plugin [17:55:53] looks like we could be using git submodules for the local gem? https://github.com/bundler/bundler/issues/1935 [17:56:32] or bundler... I can't watch. [17:56:51] vagrant uses bundler internally [17:56:57] k thx [17:57:03] but like I said that part is all voodoo to me [17:57:09] me too D: [17:57:38] (CR) Ejegg: [C: 2] Switch to caret semantic versioning [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/298418 (owner: Awight) [17:58:01] dduvall is the ruby guru for mw-vagrant :/ [17:58:46] (Merged) jenkins-bot: Switch to caret semantic versioning [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/298418 (owner: Awight) [17:58:50] He would have a lot of answers [17:59:08] (CR) Ejegg: [C: -1] "squashed into I69b741e939bd4 - abandon?" [wikimedia/fundraising/crm/drupal] - https://gerrit.wikimedia.org/r/297522 (owner: Awight) [17:59:30] (Abandoned) Awight: (FR #1070) Drupal registry supports namespaces [wikimedia/fundraising/crm/drupal] - https://gerrit.wikimedia.org/r/297522 (owner: Awight) [18:01:32] hmmm... maybe the mediawiki-vagrant.gemspec is messed up? [18:01:47] $LOAD_PATH.push File.expand_path('../lib', __FILE__) [18:01:56] but the lib should be in the same dir as lib [18:02:19] cwd: can you try editing that in the gemspec file? [18:02:48] or does the .. just drop the filename and get to the dir? [18:03:02] fwiw, I can remove ~/.gem and run setup.sh successfully [18:03:07] - Installing the 'mediawiki-vagrant-0.13.2.gem' plugin. This can take a few minutes... [18:03:10] - Installed the plugin 'mediawiki-vagrant (0.13.2)'! [18:03:12] i feel like you guys will appreciate this: http://www.theverge.com/2016/7/12/12161792/ferret-candy-vaccine-drones-montana (and especially K4, though she's not here) [18:03:15] lemme see if I can make it verbose [18:03:54] atgomez: oh hi! /me jumps on next plane to find laced M+Ms [18:04:01] atgomez: That's definitely the best headline I've read in a looooong time. [18:04:47] "The USFWS has constructed a device that can shoot peanut-butter covered M&M's in three different directions" -- awesome! [18:04:57] (CR) Ejegg: [C: -1] "Looks cool! Don't understand why adding 'scripts' makes composer.lock out of date, probably just a crappy filetime test. Needs one more " (2 comments) [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296775 (owner: Awight) [18:06:10] (PS2) Awight: Cheap yaml lint job. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296775 [18:06:19] (PS2) Ejegg: Deprecate generate-version; this is done by our deployment scripts. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296751 (owner: Awight) [18:06:33] (CR) jenkins-bot: [V: -1] Cheap yaml lint job. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296775 (owner: Awight) [18:06:55] hilarious [18:07:07] awight XenoRyet bd808 glad you appreciate as much as i do :) [18:07:58] * ejegg sulks, wonders why nobody shoots peanut butter M+Ms his direction [18:09:02] * awight prepares to send a candy: http://arstechnica.co.uk/science/2016/03/pigeon-london-air-quality-crisis/ [18:09:23] bd808: yeah this mediawiki-vagrant.gemspec looks like the business [18:09:31] either this needs to change or something needs to load it [18:10:06] group :plugins do gemspec end -- that's the bit in the gemfile that should load it [18:10:49] none of this looks like it should be non-portable [18:13:07] cwd: maybe try this -- https://gerrit.wikimedia.org/r/299184 [18:13:56] I don't see how it's really different, but that's what the head of a file generated by `bundle gem foo` looks like [18:14:21] (CR) Ejegg: [C: 2] "deprecate with extreme prejudice..." [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296751 (owner: Awight) [18:14:49] bd808: same thing :-\ i tried plugging several paths in there [18:15:12] can't get it to find the "gem" [18:15:21] is the gem the ruby file, or the dir of ruby files? [18:15:33] (PS3) Awight: Cheap yaml lint job. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296775 [18:15:43] (CR) Ejegg: "bummer, needs manual rebase" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296775 (owner: Awight) [18:16:18] it's the dir of files and the mediawiki-vagrant.gemspec file [18:16:19] (Merged) jenkins-bot: Deprecate generate-version; this is done by our deployment scripts. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296751 (owner: Awight) [18:16:39] (PS4) Awight: Cheap yaml lint job. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296775 [18:17:15] gotcha [18:18:48] another thing to try... change "gemspec" in group :plugins to `gem "mediawiki-vagrant", path: "."` [18:18:54] (CR) Awight: "TODO: add composer-test to integration-config once this is merged" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296775 (owner: Awight) [18:19:26] (CR) Ejegg: [C: -1] "license string fix is good! I merged your other version patch, though (93a380134e1b31) so that part needs to go" [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/295928 (owner: Awight) [18:19:40] bd808: with the back ticks? [18:19:51] no. see https://github.com/mitchellh/vagrant-aws/blob/master/Gemfile [18:20:31] no sauce [18:20:47] blerg [18:21:02] i don't mean to drag you into this, i can dig in this afternoon [18:21:09] what happens if you delete Gemfile.lock completly? [18:21:20] and the try setup.sh again [18:21:40] (CR) Ejegg: [C: 2] "Awesome, thanks!" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296775 (owner: Awight) [18:21:46] same deal [18:22:01] i know a couple super ruby nerds, i could see if they can reproduce [18:22:22] that would be awesome. Brad has been pissed about this for a long time [18:23:14] i will do that [18:23:21] (Merged) jenkins-bot: Cheap yaml lint job. [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/296775 (owner: Awight) [18:26:15] ejegg: hashar's around if you want to ask about zuul experimental... [18:27:13] cwd ugh, I just got the latest vagrant and am experiencing same borkenness [18:27:19] oh wow [18:27:36] i just pinged a few people about it, i bet one of them will know what's up [18:27:37] * awight rushes to upgrade [18:27:59] not borken :( [18:28:00] I'm running 1.8.4 and things work for me but ... [18:28:14] /usr/lib/ruby/2.3.0/rubygems/specification.rb:946:in `all=': undefined method `group_by' for nil:NilClass (NoMethodError) [18:28:47] during vagrant plugin install mediawiki-vagrant-0.14.0.gem [18:30:00] (PS4) Ejegg: Don't assume configured directories end in a slash [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/241109 (owner: Awight) [18:30:52] (PS2) Awight: Fix license string [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/295928 [18:31:04] ejegg: wat, that's a different error from me [18:31:10] but same file [18:31:47] my funness is that I can "install" succesfully--but locate tells me that it doesn't install to anywhere [18:32:15] ori might have some insight? [18:34:37] (CR) Ejegg: [C: 2] Don't assume configured directories end in a slash [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/241109 (owner: Awight) [18:35:20] (CR) Ejegg: [C: 2] Fix license string [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/295928 (owner: Awight) [18:36:10] cwd: terrible thing to try: https://gerrit.wikimedia.org/r/299186 [18:37:09] oh nvm, it's already including the ./*gemspec from the last group in the gemfile, i guess [18:37:26] (Merged) jenkins-bot: Don't assume configured directories end in a slash [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/241109 (owner: Awight) [18:37:28] (Merged) jenkins-bot: Fix license string [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/295928 (owner: Awight) [18:37:47] ejegg: Thanks for taking out my trash!! [18:40:20] cwd: If you're stuck on rtbt review, want me to pick up the obt task? [18:40:52] awight: be my guest [18:41:07] * awight sharpens knives for the slaughter [18:41:40] awight: did you get a chance to look at https://gerrit.wikimedia.org/r/298549 and https://gerrit.wikimedia.org/r/293465 again? [18:41:48] cwd: you could get a bare-metal paymentswiki set up perhaps... [18:41:54] ejegg: ah thx for the reminder [18:42:32] awight: yeah i've got one, just thought it might be nice to fix this bug [18:42:40] but if you want me to hurry up i will switch over [18:43:05] (CR) Awight: [C: 2] Always capture cURL verbose output (1 comment) [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298549 (https://phabricator.wikimedia.org/T139698) (owner: Ejegg) [18:43:11] (PS5) Awight: Always capture cURL verbose output [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298549 (https://phabricator.wikimedia.org/T139698) (owner: Ejegg) [18:43:22] cwd: not at all, I'm only thinking in terms of unblocking [18:43:37] I don't personally care if u want to geek out on Ruby :p [18:44:11] i'm bad at not doing this [18:44:15] i am easily blocked [18:44:32] that's pride f**ing with you [18:44:48] things worked out ok for Butch [18:44:53] in the 3rd round, your ass goes down [18:45:00] :D [18:45:03] a bit rough the next afternoon but he made it out [18:45:11] yeah he's pretty much the only person who makes out ok in that movie [18:48:17] (CR) Awight: "It looks great, but I'm getting an invalid request error from Adyen." [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/293465 (https://phabricator.wikimedia.org/T137164) (owner: Ejegg) [18:51:14] awight: try VPN? I just got a successful redirect with that patch (using MXN CC link to avoid fiscal_number bs) [18:51:15] (CR) XenoRyet: [C: 2] Trim fields before sending to server [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/291154 (owner: Ejegg) [18:51:18] (PS5) Awight: AstroPay signature as staging helper [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/293465 (https://phabricator.wikimedia.org/T137164) (owner: Ejegg) [18:51:26] (CR) Awight: [C: 2] "Works great! I had bad config." [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/293465 (https://phabricator.wikimedia.org/T137164) (owner: Ejegg) [18:51:55] ejegg: I'm in the office, which is pretty luxurious cos of my outbound IP [18:52:05] It was just stale confi [18:52:07] g [18:53:32] ah, cool [18:54:13] (PS2) Awight: Add bpay logo and css class [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298889 (owner: Cdentinger) [18:54:15] (PS5) Awight: [WIP] Mustache for obt_bpay [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298572 (owner: Cdentinger) [18:57:00] (CR) Ejegg: "Oh yeah, we want this! Want to rebase and add a caret?" [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/293249 (owner: Awight) [18:57:32] (CR) jenkins-bot: [V: -1] [WIP] Mustache for obt_bpay [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298572 (owner: Cdentinger) [18:59:17] (PS6) Ejegg: Always capture cURL verbose output [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298549 (https://phabricator.wikimedia.org/T139698) [19:00:26] awight: Thanks for all the CR! that last-second rebase always trips up gerrit though - https://gerrit.wikimedia.org/r/298549 needs +2 removed/re-added [19:01:09] (CR) Awight: [C: 2] Always capture cURL verbose output [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298549 (https://phabricator.wikimedia.org/T139698) (owner: Ejegg) [19:01:13] likewise! [19:03:06] I'm going to take off. Have a great weekend fr-tech! [19:03:21] See ya later [19:03:55] see ya! [19:04:22] have fun! [19:07:15] (Merged) jenkins-bot: Always capture cURL verbose output [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298549 (https://phabricator.wikimedia.org/T139698) (owner: Ejegg) [19:07:16] cwd: your patch works perfectly locally, brb after lunch [19:14:46] i also must relocate to somewhere with food [19:41:56] (CR) Awight: "check experimental" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/299083 (owner: Ejegg) [19:46:54] (PS5) Awight: Add the predis library [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/293249 [19:49:28] (PS6) Awight: Mustache for obt_bpay [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298572 (owner: Cdentinger) [19:49:57] cwd: ^ I think that does it. I didn't have to make any changes, just removed some old files to feel good about meself [19:50:33] (CR) Awight: [C: 2] "Elegant!" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298572 (owner: Cdentinger) [19:50:50] argh. rebase conflict [19:51:45] (CR) jenkins-bot: [V: -1] Mustache for obt_bpay [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298572 (owner: Cdentinger) [19:52:36] (PS3) Awight: Add bpay logo and css class [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298889 (owner: Cdentinger) [19:52:38] (PS7) Awight: Mustache for obt_bpay [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298572 (owner: Cdentinger) [19:53:04] (CR) Awight: [C: 2] Add bpay logo and css class [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/298889 (owner: Cdentinger) [19:53:42] (PS2) Awight: Update libs to match REL1_27 [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/299083 (owner: Ejegg) [19:53:48] (CR) Awight: "check experimental" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/299083 (owner: Ejegg) [20:02:02] ejegg|away: ^ tests performed as advertised, and passed to boot! [20:03:59] Doorbell, going to be a min late to standup [20:32:33] !log disable donations queue consumer [20:32:37] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [20:38:52] Fundraising Sprint Muggle Baiting, Fundraising-Backlog, MediaWiki-extensions-DonationInterface, Unplanned-Sprint-Work: Donations message with empty body! - https://phabricator.wikimedia.org/T140515#2467166 (awight) [20:40:06] Fundraising Sprint Muggle Baiting, Fundraising-Backlog, MediaWiki-extensions-DonationInterface, Unplanned-Sprint-Work: Donations message with empty body! - https://phabricator.wikimedia.org/T140515#2467166 (awight) We should test for this condition and throw a reasonable error. [20:42:31] ������ [20:42:39] makes me want to play mario [20:43:02] Fundraising Sprint Muggle Baiting, Fundraising-Backlog, MediaWiki-extensions-DonationInterface, Unplanned-Sprint-Work: Donations message with empty body! - https://phabricator.wikimedia.org/T140515#2467186 (awight) Another one, this was not redelivered: ``` 20:41:47 [headers] => Array 20:41:... [20:43:10] unless you guys have those glyphs [20:46:27] (PS1) Awight: Fail if the message body is missing [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/299261 (https://phabricator.wikimedia.org/T140515) [20:46:42] ^ terrible solution but might get us through the weekend? [20:46:50] s/solution/hackfest/ [20:49:35] Fundraising Sprint Muggle Baiting, Fundraising-Backlog, MediaWiki-extensions-DonationInterface, Patch-For-Review, Unplanned-Sprint-Work: Donations message with empty body! - https://phabricator.wikimedia.org/T140515#2467196 (awight) p:Triage>Unbreak! [20:52:12] fr-tech: could use a quick CR: https://gerrit.wikimedia.org/r/299261 [20:57:38] On it [20:58:29] thanks! [21:00:30] (CR) XenoRyet: [C: 2] Fail if the message body is missing [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/299261 (https://phabricator.wikimedia.org/T140515) (owner: Awight) [21:00:58] No worries [21:02:31] (PS1) Awight: [WIP] Remove RapidHTML [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/299263 [21:02:56] (Merged) jenkins-bot: Fail if the message body is missing [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/299261 (https://phabricator.wikimedia.org/T140515) (owner: Awight) [21:06:25] (CR) jenkins-bot: [V: -1] [WIP] Remove RapidHTML [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/299263 (owner: Awight) [21:07:37] (PS1) Awight: Merge master into deployment [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/299264 [21:07:46] (CR) Awight: [C: 2] Merge master into deployment [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/299264 (owner: Awight) [21:08:26] (Merged) jenkins-bot: Merge master into deployment [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/299264 (owner: Awight) [21:11:17] !log update civicrm from cea316cc57c511c645a92a003028c95e19cac877 to a386eb5a76ec97b3b01c46a49309dfa39bbc58b0 [21:11:22] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [21:12:14] !log reenable donation queue [21:12:18] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [21:18:24] Fundraising Sprint Muggle Baiting, Fundraising-Backlog, MediaWiki-extensions-DonationInterface, Patch-For-Review, Unplanned-Sprint-Work: Donations message with empty body! - https://phabricator.wikimedia.org/T140515#2467219 (awight) The robust consumer patch is deployed now, so future empty b... [21:19:18] Fundraising Sprint Muggle Baiting, Fundraising-Backlog, MediaWiki-extensions-DonationInterface, Patch-For-Review, Unplanned-Sprint-Work: Donations message with empty body! - https://phabricator.wikimedia.org/T140515#2467220 (awight) p:Unbreak!>High [21:31:34] be present again at 3:30 Pacific... [22:23:37] Error: /usr/bin/git clone --recurse-submodules https://gerrit.wikimedia.org/r/p/mediawiki/core.git /vagrant/mediawiki returned 128 instead of one of [0] [22:23:50] vagrant up fails [22:23:52] sigh [22:24:25] awight|distract: I may be away-ish still until 3:35 PDT or so [22:41:38] awight|distract: back! [22:52:34] cwd: why did it do that? [22:53:02] ¯\_(ツ)_/¯ [22:53:11] i logged into the vm and ran the same command and it was successful [22:53:50] grrp [22:53:55] well that'll work for now, at least [22:53:57] reconfiguring vms is the #1 thing where i forget how much it sucks every time [22:54:09] what reconf do you do? [22:54:46] whatever thrashing is required that particular time [22:54:59] heh. [22:55:04] plus the hour wait for git to work [22:55:48] try not to rm-rf that dir I'd say [22:55:56] yeah [22:56:04] but how else to really reset? [22:56:25] lots of gitignored stuff [22:57:22] Heh for CN I've always fallen back to local wikis [22:57:55] awight: cwd: so the idea was that queue reform would happen first in vagrant? [22:58:20] How up-to-date is this beautious doc? https://www.mediawiki.org/wiki/Fundraising_tech/Message_queues [22:58:50] AndyRussG: yeah i think that was the plan until we have a queue server [22:59:03] cwd: u can make a deep clone of your local mw-core, is one shortcut [22:59:30] AndyRussG: yeah I think any piece we migrate should be able to work under vagrant before we deploy it [22:59:39] just cos there are multiple moving parts for each. [22:59:52] Plus, having all this stuff work under vagrant will be a huge win in its own right [23:00:08] AndyRussG: that doc should be current. [23:00:17] sounds like a good plan to me [23:00:18] cool beans! [23:00:26] AndyRussG: err--actually, with the exception of the "proposed migration" column [23:00:40] K gotcha [23:00:56] lemme update that... [23:01:39] So would this include a feature vagrant branch w/ needed changes to the fundraising vagrant role? [23:03:38] seems wise [23:05:21] AndyRussG: yeah, usually it'll be quite minor, e.g. https://gerrit.wikimedia.org/r/#/c/208856/ [23:05:29] i.e. mostly configuration [23:05:34] hmmmm [23:05:46] cool [23:06:02] yeah to keep whatever setup we're all working with consistent [23:06:16] Maybe also to foreshadow prod puppettry [23:11:14] Did I also understand correctly that the transition would involve running both old and new systems in parallel? [23:12:10] I think so... [23:12:16] sorry, distracted for 20 [23:25:59] vagrant provision hangs [23:26:02] i seem to remember this [23:26:58] that sounds like git clone? [23:27:11] turn that puppet deug on... [23:28:07] *debug [23:30:19] cwd: vagrant version all good? install vagrant guest something plugin, update guest inside vagrant? [23:30:55] AndyRussG: yep i just updated vagrant today cause there was a bug [23:30:59] https://github.com/mitchellh/vagrant/issues/74933 [23:31:21] cwd: PUPPET_DEBUG=1 vagrant provision [23:31:36] that shows you what step is actually happening [23:32:10] https://github.com/mitchellh/vagrant/issues/7493 fwiw [23:32:28] oops! sorry [23:32:42] omg that was the whole issue? or that's an additional turtle? [23:32:44] i guess it didn't really hang it just took a lot time doing something not networky [23:33:02] awight: that was the whole issue. bug in a lot of distributed builds [23:33:04] "builds" [23:33:14] nice [23:33:29] glad you were able to get some closure before the weekend! [23:34:43] sigh [23:34:48] well it's just downloading stuff agian [23:35:47] clone that mw-core yourself! [23:36:20] glad i came to the coffee shop for this anyway [23:36:22] internet is better her [23:36:24] e [23:36:49] ruthless [23:37:13] aww. mediawiki-vagrant/ included a big directory of cached packages. sorry bra [23:37:34] 1.3G cache/ [23:37:54] oh well. sorry to be such a slug lately [23:39:01] i have been using vnstat to map my network usage and see what i actually need. happy to find that most days i don't use more than 400MB. [23:39:26] but today i vagrant upped: 1.66 GiB [23:39:48] nice tool for finding leaks! [23:40:11] yeah it's pretty dang cool [23:42:16] pip install -r /vagrant/srv/fundraising-tools/silverpop_export/requirements.txt returned 1 instead of one of [0] [23:42:22] hmm [23:42:40] cwd: you have an old version of the tools repo [23:43:21] cwd: oh hey, to address your earlier thing about implementing the mirrored producers, the easiest place to start will be DonationInterface, cos DonationQueue already exists there. Then you can decide what to do with that module and how to generalize... [23:43:51] ok, i will do that [23:44:31] I may have really screwed us, I'm realizing [23:44:35] ? [23:44:47] I asked hashar to deploy the REL1_27 stuff as voting [23:44:51] should have gone with Kafka? [23:45:01] but 1.25 patches don't have compatible libraries [23:45:03] AndyRussG: hehehe [23:45:20] If you actually think so, we should chat! [23:45:22] regex adding or removing trailing zeros? [23:45:25] no jk! [23:46:23] What is it that Glowing Bear has against word wrap? [23:46:40] Maybe it's passé [23:46:49] that sounds bad [23:47:50] weird, CI seems to still work. Must have been something devious cwd did with vendor... [23:48:30] uhm [23:48:36] 27 CI passes agianst 25 code? [23:49:04] it was the vendor libraries I was most worried about... [23:49:29] My samples are https://gerrit.wikimedia.org/r/#/c/298572/ https://gerrit.wikimedia.org/r/#/c/299083/ [23:50:06] seems to work fine, spooky! [23:50:23] takes 2.5x as long, weird [23:50:31] and i wouldn't think it would pass! but hey that's nice [23:51:09] I did dig into the zuul-cloner lines, and it really is running mw 1.27 [23:51:19] depending on the zuul config it will try to check out the same branch for every submodule [23:52:21] AndyRussG: just catching up with your question from earlier--yeah I was thinking that it would make sense to migrate by first rewriting producers to * send the old message to old queue and * send a new message to new queue [23:52:31] I can't say that strategy has worked well for me in the past, though... [23:52:49] The last time I tried it, with the orphan rectifier, I ended up breaking both the old and new streams for a month. [23:53:02] I spent a lot of time rewriting the code to "do the old thing" using new code :( [23:53:44] awight: --project-branch mediawiki/vendor=fundraising/{branch} [23:54:01] i think that will do the right thing [23:54:10] i guess we werent' able to use vanilla vendor? [23:56:57] We *should* be able to, probably just make a FR branch out of habit. Please keep an eye on that!