[00:01:51] milimetric: right, iirc you guys were thinking to do the hql and refinery wireup. bearND|afk, when you get back tomorrow could you quick hit us up to make sure we have a good hand off here? gotta run... [00:18:52] rmoen: is last modified / max width stuff ready for review- i just saw it but it's column hasn't updated so i didn't know to review :/ [00:24:37] jdlrobson: the last-modified-bar is in and can be reviewed. The max-width is in progress [00:25:27] okay. Just checking cos they'll need to be done together (I guess we can squash the commits before merging) [00:25:37] I'll take a look tomorrow [00:25:40] as i need to head out soon [00:26:15] kristenlans: you in office still? [00:36:08] Hey I'm nearby. Coffe/beer? [00:36:18] kristenlans: where can i find you? [00:37:11] I'll meet you outside the lobby in 5? [15:21:08] niedzielski: howdy! could you remind me what issue you were running into when running the docker tests in the content service deployment meeting? i'm setting up the vm on my main laptop and may be running into the same issue [15:22:54] mdholloway: sure! i just sent you an email i recorded when it happened [15:23:13] mdholloway: i haven't tried it since so maybe it works better now [15:24:01] got it, thanks! looks like my issue is different, but good to have a reminder of this in case it comes up, too [15:24:11] mdholloway: what issue are you seeing? [15:25:48] niedzielski: the test suite is hanging in the jshint portion and then crashing, saying it might be because I'm using the legacy node package [15:25:57] niedzielski: (which i just uninstalled and tried again; no dice) [15:26:03] niedzielski: (brb) [15:29:18] mdholloway: i always get the legacy node warning, even just running the service locally (no deploy), which i assumed to be pretty typical. i don't know what the issue is, but if you like i'm happy to try to repro on my native linux system [15:33:14] niedzielski: sure, if you wouldn't mind -- it's happened both times i've tried running the tests with ./server-js docker-test [15:33:21] niedzielski: (the service itself works fine) [15:33:51] niedzielski: i didn't run into this on the loaner [15:33:55] mdholloway: ok will do. give me a couple minutes as it's been a while since i tried to do a deploy [15:34:03] mdholloway: that's a good note [15:37:57] mdholloway: hm, so i run "./server.js build --deploy-repo --verbose" and get the same error as in that email, although the exit status is good [15:40:08] mdholloway: does it print out each step as it goes along? maybe we could try the last command it fails on [15:43:19] niedzielski: it hangs for ~30s during the test suite, right after the 'jshint' heading, then , then a tests-failed error, with the warning: npm ERR This failure might be due to the use of legacy binary "node" [15:43:46] mdholloway: can you access the internet from your vm? [15:44:36] niedzielski: with a suggestion to read /usr/share/doc/nodejs/README.Debian for further info, which it's probably time to go back and do [15:44:41] niedzielski: yep [15:45:39] mdholloway: i mean, the setup instructions for RB _say_ to install nodejs-legacy so i think that warning is expected [15:46:33] niedzielski: hmm, true, might be a nonissue. let me try building the deploy repo. [15:47:23] mdholloway: ./server.js build --verbose has about the same output as with "--deploy" [15:47:58] mdholloway: i don't think my system gets to the test phase. it quits pretty quick [15:49:26] niedzielski: no trouble when doing ./server.js build --deploy-repo; just says it's up to date (which makes sense, i just pulled all the code down onto the vm this morning) [15:50:24] niedzielski: so as long as the tests are passing maybe i won't stress the docker test issue for the moment [15:50:34] stress *about [17:08:16] jdlrobson: so the dev branch is a bit behind, should I just update it ? Or any objections to putting the max-width works in a new branch based of master? [17:16:34] niedzielski: bearND|afk dbrant_ coreyfloyd mhurd_afk mdholloway bgerstle sorry connection issues [17:17:42] hey rmoen no problems whatsoever [17:17:44] whatever's easiest [17:17:58] it was just a suggestion [17:18:17] i have no real issues with working on master either i just thought it might be a better way to move this along since there are lots of moving parts [17:18:29] jdlrobson: my only question is howdoes that make sign off easier in MF ? Beta labs is running off master [17:19:05] otherwise i'm totally fine to put it in another brnahc [17:19:12] branch. sorry not quite awake ye [17:19:14] yet [17:20:26] :) [17:25:39] rmoen: can i get a merge for https://gerrit.wikimedia.org/r/236037 if your browser tests are working? should be straightforward [17:25:51] keen to keep our browser test builds looking nice :) [17:26:19] yeah [17:37:36] mbinder: is the app story prio meeting happening right now android only? [17:37:46] no but we can do them first half [17:37:57] i only saw android people when i poked my head in a second ago [17:40:50] pop in at the top of the hour [17:41:04] coreyfloyd: you too ^ :) [17:43:52] jdlrobson: Tests are running. Login features passes. Though a bunch of others are failing for me. Likely on my end. [18:00:19] coreyfloyd: bgerstle we're ready for ya [18:00:34] * bgerstle prepares to make grand entrance [18:06:17] rmoen: thanks. Gotta keep looking good on those browser tests :) [18:33:27] coreyfloyd: mhurd bgerstle http://hatjitsu.wmflabs.org/78399 [19:14:07] niedzielski: good morning :D [19:14:31] hashar: hello! [19:16:49] niedzielski: I commented on https://gerrit.wikimedia.org/r/#/c/231624/3/zuul/layout.yaml,unified :D [19:17:02] niedzielski: the lint job is non-voting, so there is no point to have it in the gate-and-submit pipeline [19:17:25] niedzielski: so it is better to comment it out and leave a note next to the voting: false flag [19:17:34] beside that, kudos :-} [19:18:51] hashar: ah, i see. i will make the change momentarily. thanks for reviewing! [19:19:09] dr0ptp4kt: Stephen is acing the CI for the Android app :-} [19:19:27] niedzielski: and the jobs looks legit enough :) go ahead and deploy them [19:19:39] once Zuul layout is amended I am going to deploy it [19:23:42] :) [19:25:11] hashar: gr8! updated. i think everything should be in order [19:25:31] niedzielski: yeah [19:25:39] have you deployed the job or should I do it ?:-} [19:26:27] hashar: i'm happy to if you like. i'm not sure if you do anything extra to make the deployment more official or not :) [19:26:37] nothing more than running jjb [19:27:09] hashar: i know! good job niedzielski [19:27:22] deploying [19:27:41] \o/ [19:29:32] Job apps-android-wikipedia-.+ not defined [19:29:33] grbmblb [19:31:36] niedzielski: even review doesn't catch everything https://gerrit.wikimedia.org/r/#/c/231624/4/zuul/layout.yaml,unified :d [19:31:47] niedzielski: gotta bring my kid to bed, will be back in a few [19:32:10] hashar: nuts. i'll take a peek, thanks [19:32:24] yeah that issue is mind numbing [19:32:29] confusing / whatever :D [19:36:40] hasharAw: ok patch is revised [19:40:17] niedzielski: that made it [19:41:37] hasharAw: awesome! thanks for all the help! [19:42:32] hasharAw: it _should_ be the last android ci patch for a little while. we'd like to eventually get a release server up (there's a long email thread) which will likely involve releng and it would be nice to also have the gerrit promotions working for our restbase service [19:43:39] hasharAw: anyway, i will bother you again one day for sure but it will probably be a little ways out. thanks again! [19:44:28] niedzielski: yeah releasing to stores is a similar story [19:44:51] niedzielski: the problem is holding the credentials safely. But potentially you can come up with an internal Jenkins that is restricted [19:45:57] hasharAw: yeah, i'm not quite sure how that will all work yet. the privacy issue seems to really complicate matters. does our infrastructure support deploying to internal and external boxes? [19:47:52] niedzielski: Zuul / CI doesn't do any deploy beside a few git pull here and there or the auto update of the beta cluster [19:48:17] we could well set up a private Jenkins that would listen for Gerrit event [19:48:24] or automatize the publishing process [19:50:44] hasharAw: hm, at some point i will pick your brain on it to understand what's practical [19:52:10] niedzielski: I have it to read and reply to the mail you send on the QA list "Keeping secrets safe on Jenkins" [19:53:16] hasharAw: thanks, me too. dan left a mail there i need to follow up on [19:53:17] we had the same issue with the mediawiki tarballs or building jar and hosting them on a central repo [19:56:58] time for me to vanish [19:59:00] hasharAw: have a good one! thanks so much for all the help [20:15:01] mdholloway: hey! i'm looking at your database saved pages patch. i wanted do a quick test but it looks good to me. the only thing was that there's now an unnecessary menu member variable. if you can remove that, i think we can merge this today! [20:15:24] niedzielski: awesome! [20:15:39] mdholloway: https://gerrit.wikimedia.org/r/#/c/229959/13/app/src/main/java/org/wikipedia/page/PageFragment.java [20:15:54] mdholloway: we don't need that optionsMenu now that we call the invalidate menu, right? [20:17:05] niedzielski: that sounds right, but gimme one sec, just finishing up a commit msg on a content service wip thing [20:17:23] mdholloway: ok thanks! [20:19:24] niedzielski: yep, you're right, i'll update to remove it! [20:20:17] mdholloway: yay! the rest of the patch and testing looked good. as soon as you update the fragment, we can merge [20:21:56] niedzielski: awesome! [20:22:10] niedzielski: while we wait on that, i was just going to talk to you about content service stuff [20:22:37] mdholloway: sure! what's up? [20:22:52] or do you want to hang out? [20:23:46] niedzielski: i've been looking at a task to disable client-side JS transforms when loading from RESTbase and started a patch for that, but I think first we need to actually wire up the transforms on the server side, and I remembered that being your main department-related quarterly goal [20:25:03] niedzielski: i mean, i could still write the patch and just comment out the relevant lines, but just wanted to touch base with you on what you're thinking on that -- still planning on making that a focus? [20:27:10] mdholloway: for the transforms, bernd's already copied all the code over to the restbase service and we've been wiring them up into the routes. his most recent app patches allow us to actually point at the restbase service. in terms of my specific goals, i'm most interested in removing js code from the app and moving it to the service, or anything that helps that effort. i think you had very similar quarterly goals and have no [20:27:18] qualms about any overlap [20:28:07] mdholloway: if there's something that progresses us further, by all means do it! [20:28:29] niedzielski: cool, thanks! [20:29:17] mdholloway: np thank you! [20:30:57] niedzielski: hmm, i see what you mean about that dark mode patch in the nat'l assembly of pakistan article. back to the chrome device inspector... [20:31:23] mdholloway: i think it's just because that's in a table [20:32:09] mdholloway: there was that preexisting isInTable logic and i wasn't quite sure if that was right [20:36:04] bgerstle: what part of FL are you in? [20:36:12] delray beach [20:36:23] mbinder: why? [20:36:54] ah, the eastern side. I am going to be near sarasota at the end of oct [20:37:14] but it's basically as far as NY to Boston [20:37:58] right [20:39:41] hey dr0ptp4kt do you know if we have event logging on table of contents for either app? [20:39:52] kaity: i think we have some TOC logs for iOS [20:40:03] i think it's just whether we open/close, maybe if we also tap an item in TOC [20:40:22] bgerstle: sweet, can we compare to other article actions or search/W menu? [20:40:27] not sure [20:40:44] maybe dr0ptp4kt could you look into that? [20:47:07] niedzielski: yeah, i have to admit the business about stripping colors from elements in tables never made a whole lot of sense to me. [20:59:49] kaity: apologies for delay. one moment [21:00:00] cool thanks! [21:03:26] kaity: the list of app event logging schemas is available at https://meta.wikimedia.org/w/index.php?title=Special%3APrefixIndex&prefix=MobileWikiApp&namespace=470 . it looks like MobileWikiAppToCInteraction captures open/close/click actions, along with the app install id. [21:07:00] niedzielski: hmm, the post-merge builds are failing. [21:07:42] mdholloway: i posted a follow up on the phab. since this is a new job, i don't think this should impact anything [21:07:54] niedzielski: ok [21:08:03] mdholloway: it's weird issue with proguard missing in the otherwise populated sdk. i was wondering if it's anything to do with nfs [21:09:27] dr0ptp4kt: is it possible to see "open" ToCInteraction compared to "start" Search and "open" NavMenu [21:11:43] jdlrobson with the max with work the skins.minerva.tablet.beta.styles module is basically empty, am I correct in thinking we should remove? Reason I'm asking is because skins.minerva.tablet.beta.styles>ui.less is already empty [21:31:54] rmoen: i suspect so yeh [21:32:36] sounds good [21:32:44] be sure to move the styles in there to more relevant places [21:33:07] eg pointer overlay stuff should go with pointer overlay code [21:33:16] right [21:33:19] got that [21:33:22] although i'm not sure how feasible that is :) [21:33:45] it should probably be using a more generic class anyhow [21:34:21] brb meeting [21:35:05] kaity: in principle, yes, i believe that's possible with some sql-fu. a person would have to take into account the sampling rates for the different tables and isolate the analysis to versions of the app that use all three of those logging schemas [21:49:51] niedzielski: You had some issues merging code at first, right? As in, you'd give something +2 but Jenkins didn't merge it? [21:50:32] Deskana: when i first joined you mean? [21:53:16] Deskana: i had a _lot_ of account issues but i'm not sure if i had that issue. i think the +2 issues i had were 1) can't get +2 permissions on volunteer account 2) some gerrit jobs require manual vote +2, verify +2, and submit [21:55:29] Hmm, perhaps I'm thinking of someone else. [21:55:41] Someone on my team was having issues, but we fixed it now anyway. [21:56:00] He could give CR+2, V+2 but couldn't "Publish and submit" so nothing actually got merged [22:04:59] can't get +2 permissions on volunteer account [22:05:00] um [22:05:05] that can't be right [22:06:59] Deskana|Away: plenty of repos don't grant the "submit" permission to prevent people from overriding jenkins. [22:07:06] at least two volunteers can +2 anything anywhere in gerrit [22:07:09] Krenair: er, well i haven't been able to yet. still trying to get my volunteer nda in T106064 [22:07:20] a group of others can +2 anything in a mediawiki-related repository [22:07:38] +2 has nothing to do with NDAs [22:08:22] Krenair: maybe i misunderstood. i'd love to get +2 access on my volunteer account [22:08:31] then request it? [22:09:06] what repositories do you want +2 on? [22:10:12] Krenair: i have in the past but couldn't really get anywhere with it. i will re-request it [22:11:52] There's some stupid line of gerrit config somewhere which gives WMF engineers automatic +2 access to some repositories. [22:11:57] Krenair: apps/android/wikipedia, mediawiki/services/mobileapps, and apps/android/java-mwapi would cover my needs currently i think [22:12:01] It really needs to die [22:12:39] If you have split accounts in gerrit and already own those on one account, you should be able to include your other account on them [22:14:58] Krenair: i can +2 on my wmf account. i'm not sure how to include the volunteer account [22:15:06] Hmm [22:15:10] You can +2 on apps/android/wikipedia because all mediawiki devs are - that same group includes all of ldap/wmf which you are in [22:15:57] the other group allowed is apps-android-wikipedia, containing YuviPanda and BearND [22:16:02] can get added there [22:16:06] Krenair: weird, i thought that took some setting up but maybe it was just getting the right ldap group [22:16:11] no [22:16:47] I think they should be able to add you to the group [22:18:04] Oh wow: https://gerrit.wikimedia.org/r/#/admin/projects/apps,access [22:18:12] Who set that up.... [22:18:28] Demon [22:18:29] dammit [22:18:42] yeah, so good luck with apps/android/java-mwapi [22:18:51] would probably require a gerrit admin to fix that [22:20:14] Krenair: hm, i think we're going to deprecate that repo long term so maybe i can just bug yuvi for apps-android-wikipedia [22:20:28] what're you going to do with it? [22:21:40] Krenair: oh i try to keep my activity on one account as much as possible. right now i have to login to +2 things or push tags [22:22:00] no, I mean the apps/android/java-mwapi repo [22:25:06] Krenair: oh! it doesn't have a whole lot in it right now. we recently made some changes to the way we're performing network requests so we'll probably either rewrite the lib or scrap it. i'd really like to see us provide 3rd parties with more compelling componentized libs. whether that will be reworking something old or writing something new, i'm not sure [22:25:44] I'd really like to see a generic mediawiki API with the wikipedia app just being a layer on top with WP-specific things [22:25:58] sorry, generic mediawiki app [22:27:42] Krenair: if our restbase service is a success, i don't think we'll be working on MW as much in the short term. either way, i want hello world wikipedia / mediawiki apps to be as easy as _any_ hello world apps [22:28:29] Krenair: hopefully we can make a really great restful api that people seek out