[00:00:22] 1.28-wmf.3 is at e72c61d8bbf9d2dda546a5e03a3988961191df9c [00:00:31] (these are CentralNotice commits btw) [00:03:42] I'm asking Greg for a deployment window to push the revert patch again... [00:04:44] awight: thx! Yeah I see flying of organic matter on -operations [00:04:53] (related to some other thing) [00:05:06] Someone left the fan set to high... [00:05:46] And ate from the wrong taco stand [00:07:07] awight: yeah the secret keys to my deployment rights were eaten by the SSD mold [00:08:57] or more likely they went to the wrong poutine food truck [00:09:26] AndyRussG: looks like a window may be open right now, I'll know in a minute. [00:09:47] Confirming that I should deploy 50b7a37df47a1ef798efdfc1186d66c9d75c1d51, current wmf_deploy? [00:09:57] awight: checking.... [00:13:00] awight: yes that's right [00:13:24] k.. the submodule pointer is very wrong! I'll be fixing that. [00:13:41] I think it's cos people don't know about our wmf_deploy branch. [00:16:20] awight: I think the last deploy they moved the submodule pointer back [00:16:23] to revert [00:16:28] I mean, when stuff was reverted [00:21:48] it didn't work, though... maybe failed second revert was caused by not running submodule update actually. [00:22:01] cos I found that the tree said "new commits" for extensions/CentralNotice [00:22:08] the deployment tree. [00:22:31] wat? [00:23:23] awight: would it help to Hangout screenshare? (don't bother if u don't think it really would of course) [00:25:51] https://gerrit.wikimedia.org/r/#/c/292304/ looks fine [00:27:42] AndyRussG: would you mind testing the registerTest bug? [00:27:48] I believe I've reverted correctly now. [00:28:24] awight: yep! dunno how long RL takes to roll over these days [00:28:31] Ah I think there's a way to bypass the cache... [00:28:41] * AndyRussG mourns lost notes [00:31:26] gurp [00:31:53] I think there used to be a timestamp on RL URLs [00:32:11] awight: works gut! [00:32:35] whew. Whadda mess [00:33:08] I don't feel the need to belabor this, personally, but we could do a postmortem if you think it's helpful? [00:34:58] hmm [00:35:00] Only two things jump out as fixable: * we should reconsider the wmf_deploy branch, cos it causes trouble with other WMF staff supporting us, and * we need a clone of you, someone who really understands recent CN work and can do smart things [00:38:02] awight: hmm [00:38:16] * AndyRussG runs from undeserved compliments [00:38:19] lol [00:38:25] sorry to pigeonhole you [00:38:33] I can take it back :p [00:39:50] Heh thx it's OK I may still find a way to become more jaded, NEway [00:40:32] I think a the CI issues addressed while I was melting in NY were a great start [00:40:48] We should continue along those lines with moar better browser tests [00:41:33] +1 [00:44:07] Maybe also have a formal checklist for smoke and automated tests for patches [00:44:36] Sounds like similar issues are being discussed on -operations for the whole cluster [00:51:33] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM: Undo WmfException::isNoEmail hack - https://phabricator.wikimedia.org/T136784#2347810 (awight) [00:52:31] (CR) Awight: "This is great! My only questions are about the choice of "fatal" error as our threshold, and using the batch size knob as a disable toggl" (2 comments) [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/281842 (https://phabricator.wikimedia.org/T131200) (owner: Ejegg) [00:55:09] awight: just verified the CN code on another wmf.4 wiki and it's good [00:55:29] thank you! [00:56:20] (CR) Awight: "This looks perfect. I'll smoke test it tomorrow..." [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/287147 (https://phabricator.wikimedia.org/T133197) (owner: Ejegg) [00:56:53] That's it. I'm through. See you tomorrow! [00:57:48] (PS3) Awight: Import errors are logged and recoverable [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/283112 (https://phabricator.wikimedia.org/T88460) [00:58:41] Fundraising Sprint Ghostbusting , Fundraising Sprint Hermit Crab Husbandry, Fundraising Sprint Internet Exploring, Fundraising Sprint Jabberwock Slaying, and 4 others: Can the CIVI import function show what rows in a file aren't imported? - https://phabricator.wikimedia.org/T88460#2347846 (awight... [00:59:08] awight: cya! [00:59:13] Thx! [00:59:56] (CR) Awight: [C: 2] Fix time and message limits [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/288449 (https://phabricator.wikimedia.org/T133965) (owner: Ejegg) [01:01:21] (Merged) jenkins-bot: Fix time and message limits [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/288449 (https://phabricator.wikimedia.org/T133965) (owner: Ejegg) [09:35:46] Fundraising Sprint Killing Time, Fundraising-Backlog, Patch-For-Review, WMF-deploy-2016-06-07_(1.28.0-wmf.5): Ingenico iframe styling has changed - https://phabricator.wikimedia.org/T135388#2348505 (Pcoombe) Looks good now, thanks @cwdent! [14:23:53] ejegg|away: morning! I'm juggling git to try to get a patch ready for SWAT this morning that only fixes the CN test API thing, pls LMK if u have the chance to give some input... thx! [14:58:05] (PS1) AndyRussG: ext.centralNotice.display: API for registering tests [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292367 (https://phabricator.wikimedia.org/T134286) [14:58:07] (PS1) AndyRussG: jshint, gruntfile, and compatibility fixes (partial, wmf_deploy branch only) [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292368 [14:58:24] cwd: ejegg|away: ^ [14:59:12] cwd: hi! lmk if you'd have some thoughts on a strategy I'm using to fix up CentralNotice's wmf_deploy branch while bringing in as little as possibel from master [14:59:39] ...and messing up the relationship between wmf_deploy and master as little as possible, too [15:01:04] (CR) jenkins-bot: [V: -1] jshint, gruntfile, and compatibility fixes (partial, wmf_deploy branch only) [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292368 (owner: AndyRussG) [15:11:19] AndyRussG: i will take a look! [15:12:07] cwd: cool beans thx! [15:12:39] AndyRussG: can you explain what's broken? is it described somewhere? [15:12:51] cwd: basically it's the two gerrit changes in scrollback here. The first re-applies the patch that caused an outage for IE11 and earlier [15:13:22] The second should have been a cherry-pick for the fix that was merged to master, which fixes several compatibility issues, not just the one that broke stuff [15:13:49] However, there's a lot of stuff in master, and I don't want to deploy it all just yet, especially considering the stuff that's been flying about production recently (not just from us) [15:14:22] So to play it safe, the second is a "partial cherry-pick" (not a real thing) of only the parts of the fix that don't conflict [15:14:43] aah ok, i think i understand [15:14:55] thanks! i'll check it out now [15:15:04] There are two potential problems with this, but I think they're both lesser than what we'd get with merging all of master to production [15:15:05] i have to run to an appointment in 15m [15:15:12] ah K np of course! [15:15:12] what are those? [15:15:52] 1) there will be a commit on wmf_deploy that isn't on master (the partial cherry-pick). We normally don't do that [15:16:30] 2) For some reason (I can't see quite why) after this master won't merge cleanly into wmf_deploy. It borks on something or other [15:16:37] Also something we try to avoid [15:17:22] merge conflict? [15:17:29] Yea [15:17:40] A simpler solution might just be to make the re-application of the feature that broke stuff a fixed version [15:17:54] i'm confused about the diff here: https://gerrit.wikimedia.org/r/#/c/292368/1/resources/infrastructure/ext.centralNotice.adminUi.campaignPager.js [15:18:00] That would definitely have issue 1), but might avoid issue 2 [15:18:05] why does line 150 show changes? [15:20:11] is there breakage right now? and why don't you want to deploy master? [15:20:21] a ton of new code? [15:20:23] Maybe newline at the end? dunno.... [15:20:47] that's the kind of strange diff behavior that causes merge conflicts i don't understand [15:21:10] yeah I don't see any diff in meld [15:21:23] on that line [15:21:44] weird [15:21:46] anyway [15:21:51] in master there's new stuff that I want to test more on the beta cluster and kick hard on various platforms [15:22:18] does https://gerrit.wikimedia.org/r/#/c/292367/ fix live breakage? [15:26:57] cwd: no, it re-instates it, but is an exact cherry-pick from master. So the follow on is as near as possible to a cherry-pick of the fix on master [15:27:25] I guess here I'm prioritizing keeping the commits in wmf_deploy as similar as possible to master, rather than just simplicity [15:28:00] Yeah the other strategy would be that, making https://gerrit.wikimedia.org/r/#/c/292367/ a fixed version i.e. w/out the breakage, and forget the follow-on partial cherry-pick [15:28:26] Maybe that's better [15:28:34] Clearer for anyone reading the commit log [15:28:35] AndyRussG: hmm, what about a commit to master that undoes the parts of https://gerrit.wikimedia.org/r/#/c/292368/1 you don't want? [15:28:56] you could keep fast forward history that way [15:29:29] Hmmm [15:29:32] interesting! [15:29:53] then reapply them later if desired [15:30:00] anyway i have to run! good luck! [15:30:01] but then how would I cherry-pick from master? [15:30:04] thanks!!! [15:30:30] back in a bit and would love to try and help [15:32:56] (CR) AndyRussG: [C: -2] "Still working out what's the right strategy here..." [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292368 (owner: AndyRussG) [15:33:29] (CR) AndyRussG: [C: -2] "Still working out what's the right strategy here..." [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292367 (https://phabricator.wikimedia.org/T134286) (owner: AndyRussG) [15:46:33] AndyRussG: good moring. Are you still working on the CN deploy? [16:07:38] dstrine: hi! [16:08:04] I've been working on it, but it's not going out this morning, I think [16:08:20] ok when is the next deploy? [16:08:45] There are complications due to a divergence between our deployment version and development version ("master") [16:08:58] I don't want to deploy everything in the development version now, because I want to test it some extra [16:09:31] It can be deployed as soon as we decide how to solve the divergence. cwd|afk gave me a good idea for a strategy, but I want to check it with others, first [16:09:59] There is a deploy slot this evening [16:10:32] Sorry for the delay.... [16:10:59] ok it's pretty important for reading and the-wub to know when this will go out. Would Adam be able to help? [16:12:39] if we can get this out this evening that would be great. If not, we're really blocking a lot of people. [16:13:00] dstrine: basically anyone else on the team who wants to give a thumbs-up to the strategy for fixing the divergence between versions [16:13:37] There's a proposal for that on Gerrit, but I'm trying first another strategy that cwd|afk recommended [16:15:34] dstrine AndyRussG : it's not that big a deal if it gets held up a bit. I'd prefer extra testing to avoid any more problems with the site [16:17:27] dstrine: there were a bunch of issues on the cluster with a few deploys yesterday, so I don't want to stir the soup too much, that's why I don't want to deploy a lot of things at once [16:17:53] the-wub: thx! and ^ [16:18:45] ok thanks all [16:18:52] the-wub: yeah exactly [16:20:50] AndyRussG: early next week would probably be fine. we don't plan to launch any banners until 14th, and I can do some prep even before the test registration is deployed [16:21:48] AndyRussG: one thing: do you have any documentation of how we get the information on tests out? will it be stored with the impression numbers in the pgehres database? [16:30:08] the-wub: aaaarg gooood question ;p [16:31:15] No, for now, the reading test impression numbers will only be available in Hive [16:35:04] I found awight :) [16:35:28] d'oh [16:35:33] there goes my cover [16:36:35] aaannnd I'm dstrine again thanks irc [16:37:15] AndyRussG: I hear the morning deploy ran afoul? [16:38:32] the-wub: When is that test running? I can get Adyen in IL working in a few minutes if it's still useful. [16:40:21] awight: "afoul" would be a bit too strong.... I don't feel OK about merging all of master, but it's complicated to only deploy a fixed test API and maintaint our good wmf_deploy branch practices [16:40:41] dstrine: https://calendar.google.com/calendar/embed?src=8bhpbpjtu1k7agbck05j59lmv4%40group.calendar.google.com&ctz=America/Los_Angeles [16:41:25] 8bhpbpjtu1k7agbck05j59lmv4@group.calendar.google.com [16:41:32] There are a couple patches in the backscrill that sort of OK, but the subsequent merge wouldn't work perfectly, though almost so. https://gerrit.wikimedia.org/r/#/c/292367/1 https://gerrit.wikimedia.org/r/#/c/292368/ [16:42:22] Other two options are (1) just fix it on wmf_deploy only (simplest) [16:43:31] Good question. We almost never make patches on the deployment branch, that part is true. Could we cherry-pick the original broken patch then a fixup for it? [16:44:21] or (2) revert the patch with all the compatibility fixes, then re-apply in two parts, one of which can be cherry picked to wmf_deploy without merge conflicts, then cherry pick that one to wmf_deploy [16:44:28] ejegg: ^ [16:44:43] the second option suggested by cwd|afk btw [16:45:49] I'm tending towards the second one. I can send up the patches so everyone can look at them [16:45:53] I'm okay with side-stepping our deployment branch hygiene, personally. [16:45:55] dstrine: did I miss a meeting? [16:45:58] So either 1 or 2. [16:46:44] If (1), then a good additional step to take is to try merging all of master into wmf_deploy after we make the patch and before merging to gerrit, to evaluate how bad the conflict will be. I don't think it'll amount to much, since we're only talking about one function. [16:46:58] awight: K hmmm... Well, 1 has the advantage of making history easier to understand [16:47:35] You already said you don't want to, but just to keep it on the table there is (3) smoke-test everything on master and if it looks good, merge it all. [16:47:37] Also (1) easier to re-revert on production if necessary, probably a more important consideration than feel-good branch hygiene [16:47:42] +1 [16:48:55] K I'll go with option 1 and test the merge first [16:48:56] Yeah I did test the merge with the option currently on Gerrit, it has an unfathomable merge conflict over an invisible (non-existent?) difference [16:49:07] (I guess that'd be option (0)) [16:49:08] wheee! [16:49:49] awight: sorry, had stepped away. if you could activate Adyen for IL that would be great. test probably won't be until at least 1800 UTC, since we have our meeting with Kathleen today [16:51:22] OK perfect. It wasn't the simple fix I was hoping for, but I'll figure it out and deploy within the hour. [16:56:16] (PS1) Awight: Correct typo in Adyen form chooser settings [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292386 [16:56:17] Semi-urgent CR ^ [16:57:36] ejegg: XenoRyet: If you have a moment, here's an easy one ^ [17:00:07] on it [17:00:23] thanks! [17:00:43] (CR) Ejegg: [C: 2] "d'oh! Good catch." [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292386 (owner: Awight) [17:01:06] oof, so maybe it wasn't that selection_weight silliness after all [17:01:10] hahaha [17:02:54] (Merged) jenkins-bot: Correct typo in Adyen form chooser settings [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292386 (owner: Awight) [17:03:45] (PS1) Awight: Merge master into deployment [extensions/DonationInterface] (deployment) - https://gerrit.wikimedia.org/r/292388 [17:03:55] (CR) Awight: [C: 2] Merge master into deployment [extensions/DonationInterface] (deployment) - https://gerrit.wikimedia.org/r/292388 (owner: Awight) [17:04:27] Fundraising Sprint Killing Time, Fundraising-Backlog, FR-Adyen, Unplanned-Sprint-Work: Failed charges from adyen form in japan - https://phabricator.wikimedia.org/T136845#2349733 (DStrine) [17:05:41] awight: ejegg: the real drawback to option (1) is that we don't get a lot of good compatibility fixes that ejegg added in the fix-it-up patch. Most of them _can_ be put into wmf_deploy without merge cornflakes [17:05:55] K so now I'm back to option (2) [17:06:05] AndyRussG: those conflicts were pretty minor [17:06:19] (Merged) jenkins-bot: Merge master into deployment [extensions/DonationInterface] (deployment) - https://gerrit.wikimedia.org/r/292388 (owner: Awight) [17:06:26] and a diff with master after resolving them is totally clean, just the .gitreview upstream [17:09:41] ejegg: which conflicts exactly? You mean on the patch that's now on Gerrit? [17:10:43] AndyRussG: oh, I meant the merge to wmf_deploy I did yesterday [17:10:46] and abandoned [17:11:14] (PS1) Awight: Update DonationInterface submodule [core] (fundraising/REL1_25) - https://gerrit.wikimedia.org/r/292389 [17:11:58] ejegg: ah right... yeah the issue I have there is that I want to test those changes more on the beta cluster [17:13:03] ah, ok [17:13:36] !log update paymentswiki from d26426c4225080c95f0bd5a6a31c54e4826287b1 to de86eadcd98922ee4207a0c46112585f3ba5c48d [17:13:41] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [17:21:01] (PS2) Awight: Support certificate authentication [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292296 (https://phabricator.wikimedia.org/T136743) [17:22:49] (CR) jenkins-bot: [V: -1] Support certificate authentication [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292296 (https://phabricator.wikimedia.org/T136743) (owner: Awight) [17:24:00] GAH [17:24:11] I can never remember how to fix the "No tests executed" bug [17:30:25] (PS2) Awight: Wrap gateway URL in an accessor function [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292272 (https://phabricator.wikimedia.org/T136743) [17:31:47] (CR) jenkins-bot: [V: -1] Wrap gateway URL in an accessor function [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292272 (https://phabricator.wikimedia.org/T136743) (owner: Awight) [17:33:32] (PS2) AndyRussG: ext.centralNotice.display: API for registering tests [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292367 (https://phabricator.wikimedia.org/T134286) [17:42:34] awight: where's that new integration checklist? [17:43:07] i should add 'fr-tech provides test links' - eliminate one source of confusion [17:44:06] ejegg: I jammed it in here somewhere, https://www.mediawiki.org/wiki/Fundraising_tech/New_integration_manual [17:44:26] thanks! [17:44:26] ejegg: did I mess up some urls? :( [17:44:31] np [17:45:09] dstrine: nah, not an issue now, i just think we should give em out up front so nobody has to guess at it [17:45:25] ok cool [17:51:34] general heads up: adyen test in France & Israel in 10 minutes [17:52:19] Thanks! [17:52:35] Of course, there's some meeting that's supposed to embroil the entire Product team. [17:53:18] awight: on vagrant? [17:53:44] happens when i'm on the deployment branch [17:57:28] hey! [17:57:32] hey [17:57:33] hrm [17:57:42] maybe not the issue, but thank for the idea! [17:58:03] awight: https://www.mediawiki.org/wiki/Fundraising_tech#Running_PHPUnit_tests_under_vagrant [17:58:24] it says FIXME: not working. [17:58:28] which i think i might have wrote [17:58:32] but it does work [17:58:55] * awight kicks things [17:59:20] see you in the "Product Team Meeting" /me crosses eyes [17:59:25] non-team. [18:00:35] Adyen test is up! [18:01:12] thanks, the-wub [18:10:33] (PS1) Ejegg: Make isValidForm self-contained [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292398 (https://phabricator.wikimedia.org/T136254) [18:11:58] a lot of clicks coming in from these tests. don't think we'll need to go over an hour [18:12:58] not seeing any adyen donations in civi yet though... [18:13:34] they all come through the listener so it could take a bit i think [18:14:25] okay, thanks cwd [18:16:18] plenty at portal [18:30:44] Fundraising Tech Backlog, Fundraising-Backlog, fundraising-tech-ops: Adjust modsecurity rules to work for us - https://phabricator.wikimedia.org/T110227#2350343 (Jgreen) [18:33:02] Fundraising Tech Backlog, Fundraising-Backlog, Epic: [epic] PCI gap analysis and Improvements - https://phabricator.wikimedia.org/T77912#2350366 (Jgreen) [18:33:32] halfway through. MBeat: everything still looking okay to you? [18:33:57] yep, most of them are processing fine @ portal [18:34:23] the rest look like bank rejections [18:34:43] anything in civi yet? it looks like the job runner is firing [18:34:55] Fundraising Tech Backlog, Fundraising-Backlog, fundraising-tech-ops: Adjust modsecurity rules to work for us - https://phabricator.wikimedia.org/T110227#2350368 (Jgreen) Open>stalled Stalling this ticket because we've reigned in false positives with sane rules wherever possible, but the outst... [18:35:16] Fundraising Tech Backlog, Fundraising-Backlog, fundraising-tech-ops: Adjust modsecurity rules to work for us - https://phabricator.wikimedia.org/T110227#2350370 (Jgreen) p:High>Low [18:35:20] (PS1) AndyRussG: Revert "jshint, gruntfile, and compatibility fixes" [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/292401 [18:35:22] (PS1) AndyRussG: Re-apply "jshint, gruntfile, and compatibility fixes" (part 1) [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/292402 [18:35:24] (PS1) AndyRussG: Re-apply "jshint, gruntfile, and compatibility fixes" (part 2) [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/292403 [18:35:46] Fundraising-Backlog, FR-Smashpig: Add 1 day banner history queue to expiry job - https://phabricator.wikimedia.org/T136646#2350377 (Jgreen) Open>Resolved deployed 6/1/2016 [18:35:51] Fundraising-Backlog, MediaWiki-extensions-DonationInterface, Patch-For-Review: Refactor form chooser - https://phabricator.wikimedia.org/T136254#2328603 (cwdent) If we are doing a refactor I'd love to get down to the fundamentals. Can we try using Flow to make a page that describes this particular p... [18:35:55] one interesting thing, we are seeing some Ingenico donations from France coming in. nothing in Civi for Adyen yet, cwd [18:36:13] hmmm [18:37:24] (PS1) AndyRussG: Re-apply "jshint, gruntfile, and compatibility fixes" (part 1) [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292404 [18:37:43] Fundraising Sprint Killing Time, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Browser-Support-Internet-Explorer, and 6 others: CentralNotice failing in older browsers due use of ECMAScript 6 syntax - https://phabricator.wikimedia.org/T136387#2332929 (thcipriani) 50b7a37df47a1ef798efdf... [18:38:00] (CR) jenkins-bot: [V: -1] Re-apply "jshint, gruntfile, and compatibility fixes" (part 1) [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/292402 (owner: AndyRussG) [18:39:28] (CR) jenkins-bot: [V: -1] Re-apply "jshint, gruntfile, and compatibility fixes" (part 1) [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292404 (owner: AndyRussG) [18:41:32] Fundraising Sprint Killing Time, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Browser-Support-Internet-Explorer, and 6 others: CentralNotice failing in older browsers due use of ECMAScript 6 syntax - https://phabricator.wikimedia.org/T136387#2350409 (AndyRussG) Hi @thcipriani! Yes, th... [18:44:32] MBeat: that's expected, we're running a banner with Ingenico there to compare. sorry if that wasn't clear [18:44:44] ah, thanks [18:51:02] i don't really know how to read these logs. ejegg - do you know how long it should take for an adyen donation to hit civi? [18:51:52] cwd maybe 10-15 min? [18:51:57] lemme see [18:52:13] the job runner looks like it's working [18:52:15] adyen job runner logs would be good to look at [18:52:24] ok, let's see... [18:52:28] yeah, i have been reading them [18:52:34] seems like the right stuff... [18:53:33] Fundraising Sprint Killing Time, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Browser-Support-Internet-Explorer, and 6 others: CentralNotice failing in older browsers due use of ECMAScript 6 syntax - https://phabricator.wikimedia.org/T136387#2350494 (thcipriani) [18:55:06] still not seeing any adyen donations in civi [18:57:23] !log disabled adyen job runner [18:57:28] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [19:06:13] (PS2) AndyRussG: Re-apply "jshint, gruntfile, and compatibility fixes" (part 1) [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/292402 [19:06:15] (PS2) AndyRussG: Re-apply "jshint, gruntfile, and compatibility fixes" (part 2) [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/292403 [19:12:22] (PS2) AndyRussG: Re-apply "jshint, gruntfile, and compatibility fixes" (part 1) [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292404 [19:19:16] (PS1) Ejegg: Slightly less horrible Adyen AVS defaults [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/292406 [19:19:34] cwd ^^ [19:22:28] (CR) Cdentinger: Slightly less horrible Adyen AVS defaults (1 comment) [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/292406 (owner: Ejegg) [19:25:44] (CR) Ejegg: "That's why 'slightly less horrible'... Not sure how much config we want to give away - let's revisit that if we notice any 3rd party inte" [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/292406 (owner: Ejegg) [19:27:16] ejegg: cwd: awight|sequester: Here's my proposal for a git way to deploy only fixes and test API, while keeping wmf_deploy pretty spiffy: on master (1) revert and re-apply as two separate patches the compatibility fixes: https://gerrit.wikimedia.org/r/#/c/292401/1 https://gerrit.wikimedia.org/r/#/c/292402/ https://gerrit.wikimedia.org/r/#/c/292403/ [19:27:46] If you're on the last patch of those 3 patches, git diff HEAD~3 should show no changes [19:29:35] AndyRussG: seems reasonable [19:29:38] on wmf_deploy (2) cherry-pick the bad test API patch and the first half of the fixes patch (split so as to be cherry-pickable without merge conflicts): https://gerrit.wikimedia.org/r/#/c/292367/2 https://gerrit.wikimedia.org/r/#/c/292404/ [19:29:53] yep yep [19:30:22] ejegg: thx!!! Yeah I feel like I'm really in nitland here, but I didn't want to not deploy as many of those ES3 fixes as I could [19:31:13] (Abandoned) AndyRussG: jshint, gruntfile, and compatibility fixes (partial, wmf_deploy branch only) [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292368 (owner: AndyRussG) [19:32:16] This approach is really cwd's brainchild [19:32:25] :) [19:33:41] k, i'll re-enable the job runner for 1 job, see how that goes [19:34:44] capture successful [19:34:58] ramping up... [19:37:39] !log re-enabled adyen job runner [19:37:43] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [19:38:05] nice [21:00:06] fundraising-tech-ops, Operations, ops-eqiad: Rack and setup paylvs1005-8 - https://phabricator.wikimedia.org/T136881#2350904 (Cmjohnson) [21:00:13] ejegg: So, question about the bad override. Did "Freak out when merging config nodes of different types" fail, or were the nodes somehow of the same type? [21:00:22] fundraising-tech-ops, Operations, ops-eqiad: Rack and setup paylvs1005-8 - https://phabricator.wikimedia.org/T136881#2350918 (Cmjohnson) [21:01:39] awight: the fraud-filters node was under /adyen/payment-provider/adyen, instead of just under /adyen [21:02:03] so it wasn't trying to merge with anything, just adding more leaves [21:02:21] fundraising-tech-ops, Operations, ops-eqiad: Rack and setup new fundraising queue servers - https://phabricator.wikimedia.org/T136882#2350932 (Cmjohnson) [21:02:50] fundraising-tech-ops, Operations, ops-eqiad: Rack and setup new fundraising queue servers - https://phabricator.wikimedia.org/T136882#2350957 (Cmjohnson) [21:03:31] oof [21:09:00] (PS2) Ejegg: Slightly less horrible Adyen AVS defaults [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/292406 [21:23:08] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM: Turn on CiviEvent on staging - https://phabricator.wikimedia.org/T136888#2351077 (DStrine) [21:29:30] (CR) XenoRyet: [C: 2] Slightly less horrible Adyen AVS defaults [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/292406 (owner: Ejegg) [21:31:29] (Merged) jenkins-bot: Slightly less horrible Adyen AVS defaults [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/292406 (owner: Ejegg) [21:39:04] AndyRussG: I don't think the revert/reapply patches are necessary on master [21:40:28] Starting from fixed wmf_deploy (https://gerrit.wikimedia.org/r/292404), I tried merging master as it is now, and master + revert/reapply [21:41:00] both had the exact same tiny conflicts in display.js [21:41:37] ejegg: ah hmmm [21:41:45] Yes, I mean, that makes sense [21:42:07] Considering that the master patches leave it in exactly the same state as before [21:42:36] The only purposes of the patches on master is so that on wmf_deploy we can point to a patch from master that we can say was cherry-picked [21:42:58] yeah... it seems a bit fundamentalist to insist on that, given it's applied to a different state [21:42:59] I only saw one conflict in display.js, and I didn't understand it [21:43:11] ejegg: I guess you're right... [21:49:06] (PS1) Ejegg: Use DefaultAppeal config setting [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292486 [21:49:14] gotta relocate, back soon! [21:49:56] ejegg: do you understand what's causing the merge conflict? [21:57:55] cwd: Thanks for setting me on the right path with that test failure--I solved it by using mw-core#master rather than the fundraising/REL1_25 branch. [21:58:04] I still don't understand. but it works. [21:58:12] wat [21:58:43] i guess extensions aren't submodules on there, just gitignored? [21:58:50] that makes sense [21:59:13] garrgh [21:59:20] It was something about Hooks [22:00:11] When I looked at the intermediate result of ExtensionsTestSuite UnitTestsList, none of the extensions had injected their tests. [22:00:15] every time that happens i just flail wildly until it works then forget what i did [22:00:41] that actually describes like half the things i do [22:00:42] +1 [22:00:54] it's pretty infuriating to live only in the present [22:02:21] dstrine: Here's a fun one that we should really jam into Q1... I'd forgotten until this test stuff jogged my trauma. We need to upgrade everything to php5.5 or better, HHVM. [22:02:42] Otherwise we're going to get badly out of sync with mw-core and no longer be able to run tests or backport security fixes. [22:03:03] T130130 [22:03:03] T130130: Upgrade payments to MediaWiki 1.26 + PHP 5.5 - https://phabricator.wikimedia.org/T130130 [22:05:16] yes [22:05:28] once we do that we should really try to get on the train too [22:05:44] or at least within a reasonable distance of it [22:06:22] yah I set that as a blocking task for now [22:06:32] hobos of the world, unite! [22:06:39] :D [22:07:06] !feel free to annotate that as stuff occurs to you [22:07:52] The train is also a scary place to be--I guess we'd have to lobby the whole org if we need to revert a breaking change, for example [22:08:04] and then we get a lot of "No. my way is more correct" [22:08:39] that's true [22:08:58] my fear of any other approach is we'll just stop doing it [22:09:11] stop doing updates? [22:09:18] if we don't have a compelling reason to use the new code we'll just fork it and end up here again [22:09:20] yeah [22:09:35] we've been sort of good about that, but really freaking slow [22:09:48] maybe that's better [22:10:06] I guess I don't have to worry so much about the train, cos we have the ultimate ace in the hole: everyone's paychecks. [22:10:15] it's just one of those things that's hard to prioritize [22:10:23] no features [22:10:25] That is a motivation to agree with reverts etc. [22:10:31] yep [22:10:32] down with features!! [22:11:45] why does that ticket say 1.26? aren't they rolling 1.28? [22:13:06] mmm hehe This ticket is old [22:13:36] I think I was going with REL_ branch availability. But if we opt for the train, obviously we can jump ahead [22:14:25] the day that we [array] donation interface is going to be fun [22:15:37] I think we already kicked most of the fun into YAMLland [22:16:33] certainly a lot of it [22:16:45] but doesn't CI enforce short array...everywhere? [22:17:58] It does, but ejegg|afk hacked us up a zend5.3/mw-core1.25 CI job for DonationInterface [22:18:14] oh right on [22:18:33] ...our core branch uses short array? [22:18:36] it's a towel in the dike [22:18:41] no, not yet [22:19:19] we have a modified job for that too? [22:20:49] Fundraising-Backlog, fundraising-tech-ops: Upgrade payments to MediaWiki 1.26 + PHP 5.5 - https://phabricator.wikimedia.org/T130130#2351308 (DStrine) a:Jgreen [22:21:32] cwd: I'm not sure that we do any CI on that branch :( [22:21:51] awight: yeah, there's definitely some! [22:22:08] cwd and I were fighting it after the last round of security patches [22:22:40] got it wrangled to use the same basic test suite as normal rel1_25 [22:23:01] that's impressive! [22:23:14] We need to let that funeral ship sail tho [22:24:00] yeah, i'm not attached [22:24:12] ooh, we'll want to revive your extension.json patch [22:24:14] hehe [22:24:17] d'oh [22:24:32] sorry we let that one rot... [22:24:47] oh it was just an experiment anyway [22:25:20] the CN one will be fun, though--it requires a hook to conditionally set globals depending on is_infrastructure [22:25:58] k yeah I see the php5.3 CI jobs https://gerrit.wikimedia.org/r/#/c/292389/ [22:26:03] last time we hopped from 1.23 to 1.25. Any reason to go to 1.26 next instead of 1.27? [22:26:14] alls I care about is that we have a REL_ branch [22:26:34] but (see backscroll)--the very enticing alternative is to jump on the train [22:27:13] I really like the idea of cutting maintenance burden loose from fr-tech [22:27:16] same deploy timing as enwiki? or call our own whistle stop further out? [22:27:29] interesting... yeah probably the 3rd stage [22:27:35] (with the wikipedias) [22:27:50] yeah seems like we could just deploy whatever stabilized last week [22:28:01] also, what does that mean for quick DI fix deploys? [22:28:12] good question! [22:28:20] we could cut our own branch every monday [22:28:26] that is last week's core branch [22:28:35] We wouldn't want DI to be in the wmf.N release branch cos it's not on wikipedias [22:28:41] dang. [22:29:24] could just disable that in wmf-config though [22:29:31] or we choose a new back end for donation interface :D [22:29:43] Fundraising-Backlog, fundraising-tech-ops, Release-Engineering-Team: Spike: Talk about moving payments to WMF production MediaWiki - https://phabricator.wikimedia.org/T130658#2351327 (awight) [22:29:45] hahaha [22:30:17] just route a few more things through WmfFramework... [22:32:37] Fundraising Sprint Killing Time, Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM: Running the Donor Summary report it blocks us with an incorrect validation error - https://phabricator.wikimedia.org/T136893#2351334 (Eileenmcnaughton) [22:34:29] Fundraising-Backlog, fundraising-tech-ops, Release-Engineering-Team: Spike: Talk about moving payments to WMF production MediaWiki - https://phabricator.wikimedia.org/T130658#2351353 (awight) @Jgreen We're talking about trying to get ourselves on the train of WMF-standardized deployments for MediaWi... [22:36:55] Fundraising-Backlog, fundraising-tech-ops, Release-Engineering-Team: Spike: Talk about moving payments to WMF production MediaWiki - https://phabricator.wikimedia.org/T130658#2351359 (awight) @mmodell Sure thing, thank you for noticing this task. We're trying to identify ways in which our process m... [22:37:06] I noted a few of the things we were talking about ^ [22:38:30] thanks! [22:38:58] (CR) Krinkle: "Why? What broke?" [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/292401 (owner: AndyRussG) [22:39:22] AndyRussG: awight: I'm trying to find out why the revert. Looked everything (I think?) but can't find it. [22:39:28] 'twould be thrilling. The fresh air of the boxcar life [22:39:30] everywhere* [22:39:57] Krinkle: sorry--this was just bookkeeping-- [22:40:16] We wanted to cherry-pick clean patches to the deployment branch, but not take the entirety of what's on master. [22:40:33] Krinkle: it was re-applied in two follow on patches [22:40:36] ^ that [22:40:53] But I think we'll abandon those, sorry for clogging your code review inbox [22:40:55] The revert and two follows-ups are not merged. [22:40:59] and wasitng ur time [22:41:09] OK, so the commit didn't break anything? [22:41:18] Yeah abandoning [22:41:32] ok, no worries :) Thanks [22:41:47] (Abandoned) AndyRussG: Revert "jshint, gruntfile, and compatibility fixes" [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/292401 (owner: AndyRussG) [22:42:08] (Abandoned) AndyRussG: Re-apply "jshint, gruntfile, and compatibility fixes" (part 1) [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/292402 (owner: AndyRussG) [22:42:28] awight: isn't donation interface included in other wikis for some message translations or something? [22:42:47] (Abandoned) AndyRussG: Re-apply "jshint, gruntfile, and compatibility fixes" (part 2) [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/292403 (owner: AndyRussG) [22:45:33] cwd: good call. It's included on donatewiki just for i18n strings, but we usually don't care about the deployment schedule so let the train run on master [22:45:38] DI#master [22:46:21] DI#saster? [22:47:09] (CR) XenoRyet: [C: 2] "Makes sense." [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292486 (owner: Ejegg) [22:49:23] lol [22:49:37] (Merged) jenkins-bot: Use DefaultAppeal config setting [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292486 (owner: Ejegg) [22:49:58] awight: ejegg: I'm unable to smoke test the patches for deploy due to local install bork :( :( :( [22:50:17] "There seems to be a problem with your login session; this action has been canceled as a precaution against session hijacking. Go back to the previous page, reload that page and then try again. " [22:50:24] (on my local wiki) [22:52:03] If anyone can smoke test this https://gerrit.wikimedia.org/r/#/c/292404/ and +2 it and its predecessor, I'll request and test the deploy [22:52:08] (swat window in 10 minutes) [22:52:12] Sorrrrrrrry!!!!!!!!!! [22:52:37] It seems very harmless.... but so did the last one that killed stuff, of course [22:52:48] If u prefer we can wait 'till next week, as the-wub suggested [22:53:23] ok smoking now... [22:53:42] awight: thx so much!!!!!! [22:56:08] (CR) Awight: [C: 2] ext.centralNotice.display: API for registering tests [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292367 (https://phabricator.wikimedia.org/T134286) (owner: AndyRussG) [22:56:19] AndyRussG: Can you release the CR-2s... [22:56:31] AndyRussG: lemme take a peek [22:56:40] (CR) Awight: [C: 2 V: 1] "Runs locally with no console errors!" [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292404 (owner: AndyRussG) [22:57:02] awight: wat CR-2s? [22:57:13] AndyRussG: https://gerrit.wikimedia.org/r/#/c/292367/ [22:57:15] Ahh right [22:57:33] (CR) Awight: ext.centralNotice.display: API for registering tests [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292367 (https://phabricator.wikimedia.org/T134286) (owner: AndyRussG) [22:57:42] (CR) Awight: [C: 2] ext.centralNotice.display: API for registering tests [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292367 (https://phabricator.wikimedia.org/T134286) (owner: AndyRussG) [22:57:50] tis ok, I removed you as a reviewer :) [22:57:51] (CR) AndyRussG: [C: 1] ext.centralNotice.display: API for registering tests [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292367 (https://phabricator.wikimedia.org/T134286) (owner: AndyRussG) [22:58:01] AndyRussG: want to reserve the SWAT slot? [22:58:06] yep [23:00:28] (Merged) jenkins-bot: ext.centralNotice.display: API for registering tests [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292367 (https://phabricator.wikimedia.org/T134286) (owner: AndyRussG) [23:00:30] (Merged) jenkins-bot: Re-apply "jshint, gruntfile, and compatibility fixes" (part 1) [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/292404 (owner: AndyRussG) [23:02:23] i hope you guys are listening to some exciting music right now [23:02:31] like the bourne identity soundtrack [23:02:48] awight: slot deploy reserved... [23:02:53] cwd: KDVS [23:02:56] it's been a good show [23:03:23] awight: I see ur deploying? [23:03:33] yah [23:03:36] K [23:03:48] i'm volunteering to do SWAT once every ten Thursdays or something. Too much stress :) [23:04:04] Extra stress to put FR stress in perspective? [23:04:12] I'll bother u about details like the core patches on wikimedia-operations, then :) [23:09:42] AndyRussG: Do you know how to access mw1017 to smoke test? Cos i don't... [23:11:11] awight: https://wikitech.wikimedia.org/wiki/Debugging_in_production and https://wikitech.wikimedia.org/wiki/X-Wikimedia-Debug [23:11:18] I know how to smoke test from here, yes [23:11:18] awight: links coming in one sec [23:11:51] ^ [23:11:53] also https://wikitech.wikimedia.org/wiki/Debugging_in_production#Pushing_code_to_mw1017 [23:15:14] (PS1) Awight: [WIP] Clean up how redirects are performed [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292501 [23:15:27] Thank you! [23:17:20] (CR) jenkins-bot: [V: -1] [WIP] Clean up how redirects are performed [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/292501 (owner: Awight) [23:20:29] (PS6) Ejegg: Send failmail and set no_thank_you on TY errors [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/281842 (https://phabricator.wikimedia.org/T131200) [23:21:08] (CR) Ejegg: Send failmail and set no_thank_you on TY errors (2 comments) [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/281842 (https://phabricator.wikimedia.org/T131200) (owner: Ejegg) [23:33:49] back in a bit! [23:54:27] Fundraising-Backlog, fundraising-tech-ops, Release-Engineering-Team: Spike: Talk about moving payments to WMF production MediaWiki - https://phabricator.wikimedia.org/T130658#2351641 (mmodell) @awight: scap3 deployment would be the best bet I think. Regular mediawiki train deployments will eventuall...