[00:00:16] I *think* so...like adding members to a project under your username namespace? [00:00:19] legoktm: it sure looks like it. [00:00:19] yeah [00:00:46] cool then, I'll move some stuff over from gitlab.com :) [00:01:16] yeah, should be possible :) [00:01:55] I don't think we're quite to the point where we're encouraging folks to move stuff, but it should be possible [00:03:16] yeah, we'd like to get using it for some releng stuff and then there's the list of people who've asked about early adoption (T282842) [00:03:17] T282842: Early adoption signup for WMF GitLab - https://phabricator.wikimedia.org/T282842 [00:03:28] I'm just trying to jump my spot on that list :p [00:03:32] but then also i'm kinda to the point where i'm like: let's just get to using this and bash out the details. [00:03:46] as far as namespacing goes, I hope we can keep the shared mediawiki/{extensions,skins,php,libs} namespace [00:03:54] yeah, we're planning to. [00:04:24] we wanted top-level namespaces like that to be "functional areas" but it also seems in practice like a lot of stuff outside of mediawiki/* will probably boil down to team names. [00:05:55] I think that's reasonable given that GitLab allows to rename projects or move things around when teams do change, so we're not stuck with mediawiki-api-team/ forever like we would in Gerrit like operations/ [00:06:21] but mediawiki/* is shared ownership for better or worse [00:06:22] yeah, it feels a lot more flexible in that regard. [00:08:18] I would just like to not see MediaWiki extensions split all over the place depending on what team is trying to maintain them at the time [00:09:15] are there already CI runners on gitlab.wm? and will I cause issues if I push a repo that already has a .gitlab-ci.yml? [00:09:56] * bd808 wishes his gitlab name was bd808 and not BryanDavis [00:10:42] is that not adjustable? [00:10:52] soooo CAS [00:10:56] there are not CI runners just yet [00:12:03] hm, it picked "Legoktm" for me [00:12:09] uid=bd808, CN=BryanDavis. CAS knows this [00:12:27] thcipriani: ok, I just want to not cause problems by pushing a repo with a .gitlab-ci.yml already [00:14:06] so it's going with CN. it might be that we could tweak that to use uid instead, but the process of integrating with CAS has been one of those cans of worms i haven't wanted to reopen. if we were going to do so i imagine it'd have to be... soonish. [00:15:01] How happy is gitlab with usernames that contain whitespace and unicode? [00:15:31] uid is ascii and no whitespace. cn is full unicode and whitespace possible [00:16:11] honestly for most folks in the movement, CN is the more correct name for them to be known by [00:16:24] I'm just a weirdo [00:17:08] is changing the CN difficult/impossible? or was that the uid? [00:17:16] the uid [00:17:30] changing the CN is "easy" except for gerrit [00:17:54] just want to make sure that if people want to change their display name it's straightforward to do so because of ...recent events. [00:18:51] uid is easy too until you become a Cloud VPS/Toolforge member and then it gets more complex [00:19:17] i log in with "Brennen Bearnes"; gitlab treats that as "BrennenBearnes". i think it collapses spaces; i'm guessing unicode is... actually, i'm not going to guess until i find docs or test it. [00:19:51] and yeah, we do, i think, need a procedure for changing this when someone changes CN otherwise. [00:21:06] If I ever get to convert wikitech to a SUL wiki and detach it from LDAP then there is a possibility that a whole lot of CNs could change [00:21:43] it would make much more sense for the Developer account CN to become the SUL account name then [00:27:33] hrm, gerrit uses cn as well [00:27:49] and I wish it used uid [00:27:50] boo. wikitech doesn't let me create a new account with CN=🦄 [00:28:12] if I find missing things in GitLab (e.g. no privacy policy or CoC links), will it be useful if I file bugs for them or would that just be disruptive? [00:28:19] thcipriani: I think netbox uses uid via CAS [00:28:51] legoktm: it would be be helpful, but please give the workboard a quick scan first - if i remember right i've got a ticket for that [00:28:56] * brennen digs up workboard link [00:29:09] https://phabricator.wikimedia.org/project/view/5212/ [00:29:19] https://phabricator.wikimedia.org/T285354 [00:29:49] in general: very happy to get bugs as people start to use it [00:30:06] will do [00:30:28] thanks! [00:37:02] i'm off for the day, but thanks bd808 and legoktm for helping us think about this stuff. [00:37:23] ^ I think I'm going to do likewise. My thinking is not going to get better from here on :) [00:37:44] y'all are smrt and I will do the same [00:38:09] o/ [00:41:28] 10Release-Engineering-Team, 10Cloud-VPS (Project-requests), 10GitLab (Initialization), 10User-brennen: Request creation of gitlab-integration VPS project - https://phabricator.wikimedia.org/T285913 (10brennen) [00:41:38] (03CR) 10Thcipriani: [C: 03+2] "best brackets of seen today!" [tools/release] - 10https://gerrit.wikimedia.org/r/702468 (https://phabricator.wikimedia.org/T282824) (owner: 10Ahmon Dancy) [00:42:25] 10Release-Engineering-Team (Radar), 10Cloud-VPS (Project-requests), 10GitLab (Initialization), 10User-brennen: Request creation of gitlab-integration VPS project - https://phabricator.wikimedia.org/T285913 (10brennen) [00:42:34] (03Merged) 10jenkins-bot: Initialize state/train-versions.json file [tools/release] - 10https://gerrit.wikimedia.org/r/702468 (https://phabricator.wikimedia.org/T282824) (owner: 10Ahmon Dancy) [00:43:17] (03CR) 10Thcipriani: "> Patch Set 1: Code-Review+2" [tools/release] - 10https://gerrit.wikimedia.org/r/702468 (https://phabricator.wikimedia.org/T282824) (owner: 10Ahmon Dancy) [05:19:24] 10Gerrit: Can't log in into Gerrit — "Cannot assign user name" - https://phabricator.wikimedia.org/T49385 (10Hattieowler) Cannot assign user name, Also cannot know what you are going to see from [[ https://www.superiorpapers.com/ | this source ]]. I would like to handle those things which are giving me different... [05:23:06] ^ removed spam comment and disabled user [07:08:17] 10Release-Engineering-Team (Radar), 10Cloud-VPS (Project-requests), 10GitLab (Initialization), 10User-brennen: Request creation of gitlab-integration VPS project - https://phabricator.wikimedia.org/T285913 (10Legoktm) Is there a reason to keep this separate from the existing integration project? [07:47:03] kostajh: I have finally arrived [07:47:36] bonjour! [07:47:52] google meet time ? ;D [07:49:25] hashar: yes we could try, but it would have to be short as I need to get to a bunch of other things today [07:56:33] 10Phabricator, 10Tools: Publicly log account bans made using the phab-ban tool - https://phabricator.wikimedia.org/T200856 (10DannyS712) Disabled https://phabricator.wikimedia.org/p/Jace0411/ for nonsense creations and per global lock of SUL account [08:28:49] hashar: What's your take on Qualcom's depends-on plugin ? https://groups.google.com/g/repo-discuss/c/TweKZknojmc/m/V9cjTAywAAAJ?pli=1 I'd totally love if they developed a more generic plugin and we can drop Gerrit's `zuul` plugin. Less code to maintain :) What do you think? [08:29:08] Good morning btw :-) [09:46:39] 10Gerrit: Can't log in into Gerrit — "Cannot assign user name" - https://phabricator.wikimedia.org/T49385 (10Krystavans) Cannot assign user name, Also cannot know what you are going to see from [[ https://www.superiorpapers.com/ | this source ]]. I would like to handle those things which are giving me different... [09:49:08] again? ^ phab admin please delete [10:11:34] qchris: sorry completely missed your ping. I was in video calls all this morning [10:12:41] majavah: done [10:15:31] 10Gerrit: Can't log in into Gerrit — "Cannot assign user name" - https://phabricator.wikimedia.org/T49385 (10hashar) [10:16:03] and restricted edits [10:19:06] o/ does wmf currently have a browser testing future stratergy? [10:19:44] I see lots of things relating to "cypress" and we are considering moving our legacy ruby browser tests to cypress [10:20:33] I guess https://phabricator.wikimedia.org/T230729 would be the task im questioning about, but it hasn't seen all that much movement in the past 6 months [10:20:44] addshore: My understanding is that zeljkof has looked at some of this, but he's not in this channel. [10:21:19] i guess I can just ping on phab! [10:21:26] and dvpita [10:21:59] addshore: so yeah browser testing tooling is owned by "Quality and Test Engineering" [10:22:04] the sister team of release engineering [10:22:13] is there another channel? or :D [10:22:37] we have a private IRC channel for both teams + performance [10:23:01] addshore: IRC is dead, didn't you hear? [10:23:08] but they are not connected [10:23:09] >.> [10:23:14] there is Slack [10:23:16] but well [10:23:19] *cough* [10:23:31] Clearly we should point people to a random Discord server instead. [10:23:38] SMS! [10:23:48] oh yeah ther eis discord as well [10:23:58] and a telegram group [10:23:59] Discord is SMS or "So Much Shouting", yes. [10:24:04] * James_F sobs. [10:24:33] so yeah hmm best is to ping whatever phabricator task has cypress mentionned [10:24:40] done! [10:24:42] but I have no ide awhether cypress was just being evaluated [10:24:46] or being a poc [10:24:51] * James_F nods. [10:24:55] or considered a serious candidate to migrate out of wdio [10:25:04] I don't know enough to have an opinion. [10:25:04] we already use it in some of our codebases [10:25:18] so we have some level of opinion :P [10:25:27] James_F: why reinvent the wheel? why just not use wiki talk pages? [10:25:29] addshore: Ideally you'd have had a standards-setting conversation first, hmm? ;-P [10:25:42] addshore: Join our IRC channel: #wikimedia-qte connect. [10:25:44] majavah: Which of the seven "talk page" technologies? [10:26:19] majavah: UseMod? MW vanilla? MW LQT? MW SD? MW vanilla+DT? All of them terrible. [10:28:56] qchris: I guess I will eventually reply on that mail thread, I haven't processed that email inbox yet :] [11:00:27] hashar: Thanks! [11:55:41] 10Phabricator, 10Developer-Advocacy (Jul-Sep 2021): Evaluate initial Phab form and subtype usage in 09/2021 after changes made in T253744 - https://phabricator.wikimedia.org/T285939 (10Aklapper) 05Open→03Stalled p:05Triage→03Lowest [11:59:20] 10Phabricator, 10Developer-Advocacy, 10PM: Decrease issues created many years ago with no recent activity (aka stale tickets) - https://phabricator.wikimedia.org/T248034 (10Aklapper) [12:03:34] (03PS1) 10Arturo Borrero Gonzalez: zuul: add debian-glue to cloud/toolforge/jobs-framework-cli [integration/config] - 10https://gerrit.wikimedia.org/r/702640 [12:29:25] 10Continuous-Integration-Infrastructure: CI failures with "No space left on device" - https://phabricator.wikimedia.org/T285942 (10Tgr) [13:32:44] 10Continuous-Integration-Config, 10Patch-For-Review: Introduce non-voting jobs with quibble+apache - https://phabricator.wikimedia.org/T276428 (10kostajh) >>! In T276428#7183437, @Addshore wrote: > I'd be interested to look at the VM running the tests and just see if it is even more stuff like {T281122} and VM... [13:37:18] 10Continuous-Integration-Infrastructure: CI failures with "No space left on device" - https://phabricator.wikimedia.org/T285942 (10hashar) 05Open→03Resolved a:03hashar Indeed, there was a rogue workspace on integration-agent-docker-1004 and I have deleted it (`sudo rm -fR /srv/jenkins/workspace/workspace/... [13:37:53] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team: CI failures with "No space left on device" - https://phabricator.wikimedia.org/T285942 (10hashar) [13:39:35] (03CR) 10Hashar: [C: 03+2] zuul: add debian-glue to cloud/toolforge/jobs-framework-cli [integration/config] - 10https://gerrit.wikimedia.org/r/702640 (owner: 10Arturo Borrero Gonzalez) [13:40:38] (03Merged) 10jenkins-bot: zuul: add debian-glue to cloud/toolforge/jobs-framework-cli [integration/config] - 10https://gerrit.wikimedia.org/r/702640 (owner: 10Arturo Borrero Gonzalez) [13:40:45] 10Gerrit, 10Upstream: Editing via the web interface, "Drag and drop a file here" does not accept files in "Add a new file or open an existing file" - https://phabricator.wikimedia.org/T274699 (10Paladox) Change merged, back ported it to https://gerrit-review.googlesource.com/c/gerrit/+/311043 [13:42:09] (03CR) 10Hashar: "Deployed!" [integration/config] - 10https://gerrit.wikimedia.org/r/702640 (owner: 10Arturo Borrero Gonzalez) [13:46:36] (03CR) 10Hashar: "Awesome, thank you James" [integration/config] - 10https://gerrit.wikimedia.org/r/702375 (owner: 10Jforrester) [13:48:29] 10Continuous-Integration-Config, 10Wikidata, 10wdwb-tech: Investigate moving some Wikibase tests to @group Standalone - https://phabricator.wikimedia.org/T285950 (10Lucas_Werkmeister_WMDE) [14:03:46] 10Release-Engineering-Team (Doing), 10Release, 10Train Deployments, 10User-brennen: 1.37.0-wmf.12 deployment blockers - https://phabricator.wikimedia.org/T281153 (10Lucas_Werkmeister_WMDE) It looks like {T285951} is a regression in wmf.12. I don’t know if it’s significant enough to block the train. [14:56:16] (03PS1) 10Ahmon Dancy: update-train-versions: Operate on state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702676 [14:56:32] (03CR) 10Ahmon Dancy: [C: 03+2] update-train-versions: Operate on state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702676 (owner: 10Ahmon Dancy) [14:58:16] (03Merged) 10jenkins-bot: update-train-versions: Operate on state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702676 (owner: 10Ahmon Dancy) [15:03:44] (03PS1) 10Ahmon Dancy: Review access change [tools/release] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/702405 [15:05:04] 10Release-Engineering-Team (Doing), 10Release, 10Train Deployments, 10User-brennen: 1.37.0-wmf.12 deployment blockers - https://phabricator.wikimedia.org/T281153 (10brennen) > It looks like T285951: Some section links in search results are redlinks is a regression in wmf.12. I don’t know if it’s significan... [15:05:32] 10Release-Engineering-Team (Doing), 10Release, 10Train Deployments, 10User-brennen: 1.37.0-wmf.12 deployment blockers - https://phabricator.wikimedia.org/T281153 (10brennen) [15:05:42] (03PS2) 10Ahmon Dancy: Add wmf-deployment group to forge committer identity [tools/release] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/702405 [15:06:01] (03CR) 10Ahmon Dancy: [C: 03+2] Add wmf-deployment group to forge committer identity [tools/release] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/702405 (owner: 10Ahmon Dancy) [15:06:10] (03CR) 10Ahmon Dancy: [V: 03+2 C: 03+2] Add wmf-deployment group to forge committer identity [tools/release] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/702405 (owner: 10Ahmon Dancy) [15:09:05] (03PS1) 10TrainBranchBot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702680 [15:09:07] (03CR) 10TrainBranchBot: [C: 03+2] Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702680 (owner: 10TrainBranchBot) [15:11:38] (03Merged) 10jenkins-bot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702680 (owner: 10TrainBranchBot) [15:31:46] Project mediawiki-core-doxygen-docker build #26066: 04FAILURE in 20 min: https://integration.wikimedia.org/ci/job/mediawiki-core-doxygen-docker/26066/ [15:50:23] Hello, I search to know "Howto test my extension in jenkins with two docker instances of Mediawiki via docker compose ". There is already an example with an extension that uses docker compose in jenkins in order to execute tests ? [15:55:38] 10Release-Engineering-Team (Doing), 10Release, 10Train Deployments, 10User-brennen: 1.37.0-wmf.12 deployment blockers - https://phabricator.wikimedia.org/T281153 (10Reedy) [15:55:40] 10Continuous-Integration-Config, 10Patch-For-Review: Introduce non-voting jobs with quibble+apache - https://phabricator.wikimedia.org/T276428 (10kostajh) Comparing index.php?title=Item:Q2 on Apache and PHP built-in server for the tainted-ref.js test failure. I see that the JS editing UI isn't loading in the A... [16:24:02] 10Release-Engineering-Team (Doing), 10Release, 10Train Deployments, 10User-brennen: 1.37.0-wmf.12 deployment blockers - https://phabricator.wikimedia.org/T281153 (10Reedy) [16:25:08] Yippee, build fixed! [16:25:09] Project mediawiki-core-doxygen-docker build #26067: 09FIXED in 20 min: https://integration.wikimedia.org/ci/job/mediawiki-core-doxygen-docker/26067/ [16:26:45] (03PS1) 10TrainBranchBot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702692 [16:26:47] (03CR) 10TrainBranchBot: [C: 03+2] Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702692 (owner: 10TrainBranchBot) [16:27:56] (03Merged) 10jenkins-bot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702692 (owner: 10TrainBranchBot) [16:28:31] 10Release-Engineering-Team (Doing), 10Release, 10Train Deployments, 10User-brennen: 1.37.0-wmf.12 deployment blockers - https://phabricator.wikimedia.org/T281153 (10brennen) Rolling back to group0 for T285951, per IRC discussion. [17:23:45] 10Continuous-Integration-Config, 10Patch-For-Review: Introduce non-voting jobs with quibble+apache - https://phabricator.wikimedia.org/T276428 (10Addshore) > Comparing the network requests, the only difference I see is that PHP built-in server has a final call to: /load.php?lang=en&modules=wikibase.mediainfo.g... [17:26:43] Krinkle / addshore we could chat here [17:26:45] o/ [17:27:22] so it loads on php built in, not on apache? [17:27:26] addshore: we're trying to figure out why the wikibase.mediainfo.getDeserializer isn't loading [17:27:29] addshore: yeah [17:27:53] also, if I understand the code right, wikibase.mediainfo.getDeserializer shouldn't be needed for the editing UI to show up, but I guess that's the only clue we have to go on at the moment [17:28:08] yeah, only clue [17:28:33] o/ [17:28:47] I haven't compared our apache config in quibble (in integration/config) with what we use in production, but that's probably another place to compare [17:29:05] I was looking at various .done() and using(,fn) calls, which can sometimes fire too early (synchronous, before returning) instead of async. [17:29:23] I mean, I see this in both of those pages https://usercontent.irccloud-cdn.com/file/GTewLeVz/image.png [17:29:27] which could play a role in the serial vs concurrent behaviour difference between the two servers [17:30:18] yeah I added a console.log statement for that to see if it was the same [17:30:29] gotcha [17:30:33] just removed it. if you need other console.log or extra logging let me know [17:30:35] I dumped the live state machine at https://gist.github.com/Krinkle/0f42fad6b08342d887153c8a7ee05302 [17:30:43] which is the registry + all currently failed/loading/ready modules [17:31:00] and it looks like there aren't any other modules that e.g. got into an error state or something like that [17:31:08] addshore: does Wikibase make some internal HTTP calls, and perhaps one of them is happening in a way that is not processed right by apache? I assume lots of other things would break if that's the case [17:31:18] nah [17:32:01] if they share a cache it migh be that wikibase doesn't recognise itself given the servername differences [17:32:19] but that would be an issue specific to the dual setup, not the root cause [17:32:28] yeah, and this happened before I added the built-in server [17:32:35] whereas as I understand it, the apache server had this issue even befor ethe php one came alongside it [17:32:38] so we can ignore that for now [17:32:39] I see that a VisualEditor edit fails with "Error contacting the Parsoid/RESTBase server (HTTP 404)" [17:32:42] Krinkle: yeah [17:33:21] setting wgServer to use a public IP for both might help increase confidence that it isn't causing some indirect issue though [17:33:43] but re: Parsoid, I think that's because I'm not using this patch https://gerrit.wikimedia.org/r/c/mediawiki/core/+/701898/1/includes/DefaultSettings.php [17:33:51] and then restart either so as to clear any php-apcu cache and run update.php to clear objectcache table [17:34:15] ok [17:34:24] https://www.irccloud.com/pastebin/vR2czjz6/ [17:34:43] but I'll just hardcode the address for now [17:35:41] dashing to dinner, bu "i'll be back" [17:36:24] ok, it's hardcoded to 9413 port, and I ran update.php plus restarted php-fpm [17:37:11] and I shutdown the built-in server [17:41:22] > "config.json":{"entityTypes":{"types":["item","property","mediainfo"],"deserializer-factory-functions":{"mediainfo":"wikibase.mediainfo.getDeserializer"}}}} [17:41:56] so this and the loader code are getting loaded as part of wikibase.EntityInitializer [17:42:08] perhaps you can put a debugger in a couple of places there to see where it goes [17:42:18] wikibase.EntityInitializer.js [17:43:01] http://142.93.162.97:9413/load.php?lang=en&modules=wikibase.EntityInitializer&debug=2&version=1kmlj [17:43:05] it's a fairly short file [17:44:41] I'd want to know: does EntityInitializer get constructed, does getEntity get called. If both yes, we'd be close to the figuring out why it won't load. [17:44:56] If both no, then we'd have to find where it normally get called and debug from there. [17:49:30] Krinkle: added console.trace & console.log statements to there [17:49:33] so yes, it gets constructed [17:50:04] getEntity gets called [17:53:05] Krinkle: oh but it looks like _getDeserializer doesn't get called [17:53:28] try a `debugger;` in getEntity() [17:53:34] I think the promise doesn't resolve [17:53:39] whatever that arg is, it doesn't get resolved [17:55:04] Krinkle: added [17:55:38] (03PS1) 10Dduvall: Allow list comprehensions to destructure into multiple variables [integration/pipelinelib] - 10https://gerrit.wikimedia.org/r/702705 [17:56:35] kostajh: ok, so the call is coming from "repo/resources/wikibase.ui.entityViewInit.js" calling wb.EntityInitializer.newFromEntityLoadedHook().getEntity() [17:56:55] (03CR) 10Ahmon Dancy: [C: 03+2] Allow list comprehensions to destructure into multiple variables [integration/pipelinelib] - 10https://gerrit.wikimedia.org/r/702705 (owner: 10Dduvall) [17:57:12] that's nice because it means the construction is also local to the wb.EntityInitializer.js file [17:57:20] keeps it a bit simpler [17:57:33] var entityPromise = $.Deferred(function(deferred) { [17:57:33] mw.hook('wikibase.entityPage.entityLoaded').add(function(entity) { [17:57:33] deferred.resolve(entity); [17:57:33] }); [17:57:33] }).promise(); [17:57:34] return new EntityInitializer(entityPromise); [17:57:37] So I guess that hook never fires [17:58:08] (03Merged) 10jenkins-bot: Allow list comprehensions to destructure into multiple variables [integration/pipelinelib] - 10https://gerrit.wikimedia.org/r/702705 (owner: 10Dduvall) [17:58:22] indeed, if I ignore breakpoints and run `mw.hook('wikibase.entityPage.entityLoaded').add(console.log)` there is no memory fire and also no later fire during the page load [17:58:27] * addshore back [17:58:49] sounds like you are getting somewhere [17:59:04] https://codesearch.wmcloud.org/deployed/?q=wikibase.entityPage.entityLoaded.%2Bfire&i=nope&files=js%24&excludeFiles=&repos= [17:59:11] I need to step away in a minute [17:59:11] https://gerrit.wikimedia.org/g/mediawiki/extensions/Wikibase/+/f13b4df1f0c5096d6dd338c80685b7e1db97affb/repo/resources/wikibase.entityPage.entityLoaded.js [17:59:14] this is where it is normally fired [17:59:20] addshore: so why would this not get run? [17:59:33] https://gerrit.wikimedia.org/g/mediawiki/extensions/Wikibase/+/f13b4df1f0c5096d6dd338c80685b7e1db97affb/repo/resources/wikibase.entityPage.entityLoaded.js#74 [17:59:34] oh bingo [17:59:38] `if ( wbEntity ) {` [17:59:48] specialEntityDataPath = mwConfig.get( 'wgArticlePath' ).replace( [17:59:52] that looks fishy for sure [17:59:54] oooo [18:00:01] yeah [18:00:02] we should see a bad url in the network request logs then [18:00:16] http://142.93.162.97:9413/index.php?title=Special:EntityData/Q2.json&revision=13 [18:00:17] 200 OK [18:00:25] but garbled response [18:00:28] `"/index.php/$1"` [18:00:29] https://usercontent.irccloud-cdn.com/file/eHZ2IAOs/image.png [18:00:31] yup [18:00:32] is wgArticlePath [18:00:50] Transfer-Encoding: chunked [18:01:01] Accept-Encoding: gzip, deflate [18:01:08] wait, is this this same bug [18:01:09] It's not saying that there is a compressed response [18:01:11] *finds the thing* [18:01:12] but compresses it anyway [18:02:00] https://phabricator.wikimedia.org/T282154 [18:02:25] addshore: depending on how normal it is expected to be for this JS to run but no enitity to be in the response, may be worth tacking on a .catch() handler and else statement that both call mw.log.warn() [18:03:22] yeah, that seems like it [18:04:17] crosslinked [18:04:41] 10Continuous-Integration-Config, 10Patch-For-Review: Introduce non-voting jobs with quibble+apache - https://phabricator.wikimedia.org/T276428 (10Addshore) After debugging the Wikibase issues @Krinkle and @kostajh found this is related to T282154 [18:07:35] dashing out again but back in a short while [18:10:26] nice work, addshore & Krinkle [18:10:49] I did nothing, only like to a ticket !!! [18:19:11] (03PS1) 10TrainBranchBot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702727 [18:19:13] (03CR) 10TrainBranchBot: [C: 03+2] Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702727 (owner: 10TrainBranchBot) [18:20:04] (03Merged) 10jenkins-bot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702727 (owner: 10TrainBranchBot) [18:26:31] (03PS1) 10TrainBranchBot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702728 [18:26:33] (03CR) 10TrainBranchBot: [C: 03+2] Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702728 (owner: 10TrainBranchBot) [18:28:19] (03Merged) 10jenkins-bot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702728 (owner: 10TrainBranchBot) [18:34:10] addshore: linking the ticket is half the battle :) [18:43:02] so did we decide it was something with article path? no? just something with encoding / compression? [19:05:00] OMG i see more gitlab repos [19:07:26] could someone make a "wmde" group or whatever? and add me? [19:07:54] we were discussing groups yesterday [19:08:10] yeah, the groups stuff was all 404 when i looked at it the other day [19:08:48] I wonder where the "mwcli" repository should end up [19:09:26] is there a reason to change the current heirarchy? [19:10:04] good point! *looks at where it is* [19:10:46] mediawiki/tools/cli [19:11:03] i guess that needs to change a wee bit because of all of the /s? [19:11:15] GitLab allows for subgroups in groups [19:11:20] oh nice! [19:11:29] so a mediawiki/tools group with a cli repo? [19:11:35] e.g. https://salsa.debian.org/php-team/pecl/php-redis [19:11:39] ty [19:13:05] i could quite get my head around if I was opening up an attack vector or not with the way i played around with workers etc on my test repo [19:13:09] so it would be a MediaWiki group, with Tools subgroup, and then "cli" or "mwcli" [19:13:14] +1 [19:14:05] and it's not like Gerrit where you need to pick the perfect name on the first try, since renaming or transferring repos is trivial [19:14:51] are there any estimates when people not in cn=nda/wmf can access gitlab? [19:15:36] (03PS1) 10TrainBranchBot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702737 [19:15:38] (03CR) 10TrainBranchBot: [C: 03+2] Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702737 (owner: 10TrainBranchBot) [19:17:23] (03Merged) 10jenkins-bot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702737 (owner: 10TrainBranchBot) [19:25:51] (03PS1) 10TrainBranchBot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702743 [19:25:53] (03CR) 10TrainBranchBot: [C: 03+2] Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702743 (owner: 10TrainBranchBot) [19:27:37] (03Merged) 10jenkins-bot: Update state/train-versions.json [tools/release] - 10https://gerrit.wikimedia.org/r/702743 (owner: 10TrainBranchBot) [19:57:23] 10Release-Engineering-Team (Doing), 10Release, 10Train Deployments, 10User-brennen: 1.37.0-wmf.12 deployment blockers - https://phabricator.wikimedia.org/T281153 (10brennen) [20:01:45] majavah: unless i am mistaken (which is possible), people wikitech accounts generally ought to be able to log in to gitlab. [20:01:50] people WITH [20:01:52] that is [20:02:56] ...i'll find out whether i'm mistaken. [20:14:52] brennen: nope, "Service access denied due to missing privileges." for me [20:15:23] majavah: ah, k. will dig into that. [20:15:41] (probably at a time when i'm not staring at production error logs.) [20:18:10] brennen: it comes from https://github.com/wikimedia/puppet/blob/4a06eb771dc968c5ce768fa8003520279681f56e/hieradata/role/common/idp.yaml#L169 [20:23:36] majavah: right, thanks. on reflection, i've got to make sure we're actually ready to throw open the floodgates before we change that (a few things with logging, monitoring, and backups need tightened up yet), but Soon™ [21:19:40] 10Release-Engineering-Team (Radar), 10Cloud-VPS (Project-requests), 10GitLab (Initialization), 10User-brennen: Request creation of gitlab-integration VPS project - https://phabricator.wikimedia.org/T285913 (10thcipriani) >>! In T285913#7189159, @Legoktm wrote: > Is there a reason to keep this separate from... [22:04:21] 10Release-Engineering-Team (Radar), 10FR-Docker, 10Fundraising-Backlog, 10Gerrit-Privilege-Requests, and 3 others: dev-images +2 rights and Docker registry credentials for FR-Tech - https://phabricator.wikimedia.org/T274303 (10thcipriani) 05Open→03Resolved a:03thcipriani Apologies for the delay here.... [22:33:02] 10Release-Engineering-Team (Radar), 10Cloud-VPS (Project-requests), 10GitLab (Initialization), 10User-brennen: Request creation of gitlab-integration VPS project - https://phabricator.wikimedia.org/T285913 (10bd808) A separate project from the Jenkins job runners seems ok to me. Soon enough™ the Jenkins th... [22:35:04] addshore: I think we found it. https://gerrit.wikimedia.org/r/c/mediawiki/extensions/Wikibase/+/701899 [22:35:32] =o [22:36:38] hehe, thats been there since June 2013 :D [22:37:39] actually, it was there before that patch too! https://gerrit.wikimedia.org/r/c/mediawiki/extensions/Wikibase/+/66122/6/repo/includes/specials/SpecialEntityData.php#b308 [22:38:26] addshore: https://integration.wikimedia.org/ci/job/wmf-quibble-apache-selenium-php72-docker/1982/console [22:38:46] such green? [22:40:11] Wikibase tests passed \o/ [22:40:41] (03PS1) 10Ahmon Dancy: train-dev clean: Add message about deleting files [tools/train-dev] - 10https://gerrit.wikimedia.org/r/702769 [22:41:10] (03CR) 10Ahmon Dancy: [V: 03+2 C: 03+2] train-dev clean: Add message about deleting files [tools/train-dev] - 10https://gerrit.wikimedia.org/r/702769 (owner: 10Ahmon Dancy) [22:41:33] (03PS1) 10Ahmon Dancy: Only use bind mounts on Linux hosts [tools/train-dev] - 10https://gerrit.wikimedia.org/r/702770 [22:43:57] kostajh: thats awesome! [22:44:08] fixed at least 2 bugs but just remove 1 line :D [22:44:59] (03PS2) 10Ahmon Dancy: Only use bind mounts on Linux hosts [tools/train-dev] - 10https://gerrit.wikimedia.org/r/702770 [22:59:34] 10Release-Engineering-Team (Radar), 10Cloud-VPS (Project-requests), 10GitLab (Initialization), 10User-brennen: Request creation of gitlab-integration VPS project - https://phabricator.wikimedia.org/T285913 (10thcipriani) >>! In T285913#7191447, @bd808 wrote: > A separate project from the Jenkins job runner... [23:01:03] 10Release-Engineering-Team (Radar), 10Cloud-VPS (Project-requests), 10GitLab (Initialization), 10User-brennen: Request creation of gitlab-workers VPS project - https://phabricator.wikimedia.org/T285913 (10bd808) [23:01:45] 10Release-Engineering-Team (Radar), 10Cloud-VPS (Project-requests), 10GitLab (Initialization), 10User-brennen: Request creation of gitlab-workers VPS project - https://phabricator.wikimedia.org/T285913 (10brennen) If we're bikeshedding the naming, I don't have strong feelings here, but "gitlab-runners" doe... [23:09:50] brennen: I likely jumped the gun on renaming your task. sorry. No complaints from me about -runners, -workers, or pretty much anything else that makes it reasonably clear that the project is about the exec nodes used by the gitlab ci system. [23:23:06] 10Project-Admins: Please create User-MediaJS - https://phabricator.wikimedia.org/T285994 (10MediaJS) [23:23:50] 10Project-Admins: Please create User-MediaJS - https://phabricator.wikimedia.org/T285994 (10MediaJS) p:05Triage→03Lowest No rush. [23:31:52] bd808: no worries! thx for the check on that. i concur that "integration" isn't super obvious. [23:32:51] phab really needs a bikeshed token. [23:33:09] https://bikeshed.toolforge.org/ [23:33:21] a perennial favorite. [23:34:32] (03PS1) 10Dduvall: Enforce pipefail on all run step commands [integration/pipelinelib] - 10https://gerrit.wikimedia.org/r/702778 [23:35:41] (03CR) 10Dduvall: "Note I did not implement a functional test as our current systemtests cannot test stage failure. :(" [integration/pipelinelib] - 10https://gerrit.wikimedia.org/r/702778 (owner: 10Dduvall) [23:37:26] 10Release-Engineering-Team (Deployment Training Requests): Deployment training request for **cjming** - https://phabricator.wikimedia.org/T285898 (10thcipriani) 05Open→03Resolved ## {icon check-circle-o} Training completed! Hey @cjming thanks for attending deployment training I have a list of the next step...