[02:37:32] (03CR) 10DannyS712: [C: 03+1] "LGTM" [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596868 (owner: 10Daimona Eaytoy) [02:38:14] (03CR) 10DannyS712: [C: 03+1] "LGTM" [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596829 (owner: 10Umherirrender) [02:39:16] (03CR) 10DannyS712: "Can tests for the proper spacing not triggering warnings also be added?" [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596822 (owner: 10Umherirrender) [02:40:31] (03CR) 10DannyS712: [C: 03+1] "LGTM" [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596825 (owner: 10Umherirrender) [05:30:49] (03PS8) 10DannyS712: Move FunctionCommentSniff annotation checks to FunctionAnnotationSniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/593746 (https://phabricator.wikimedia.org/T250958) [05:30:52] (03PS12) 10DannyS712: Move FunctionCommentSniff EmptySees check to new EmptyTagSniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/593980 (https://phabricator.wikimedia.org/T250958) [05:30:56] (03PS8) 10DannyS712: FunctionCommentSniff: Move some sniffs to new separate sniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/595737 (https://phabricator.wikimedia.org/T250958) [06:36:15] (03PS1) 10Thiemo Kreuz (WMDE): Fix UnsortedUseStatements sniff failing with leading backslashes [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596997 [07:00:40] (03PS1) 10Thiemo Kreuz (WMDE): More resource friendly UnsortedUseStatements implementation [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/597000 [07:09:06] (03PS1) 10Thiemo Kreuz (WMDE): Add leading empty token test cases for unused "use" sniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/597001 [07:12:38] 10MediaWiki-Codesniffer, 10Patch-For-Review, 10User-DannyS712: Add sniff to disallow specific tags with no text after them - https://phabricator.wikimedia.org/T252914 (10thiemowmde) > `@param`, `@return`, and `@throws` should already be covered by the FunctionCommentSniff Sure. I think it's critical to have... [07:27:30] 10MediaWiki-Codesniffer, 10PHP 7.4 support, 10PHP 8.0 support: Create a sniff for: "PHP Deprecated: The behavior of unparenthesized expressions … will change in PHP 8" - https://phabricator.wikimedia.org/T248110 (10thiemowmde) > […] or it is better to require the parenthesis for all arithmetic operators […... [07:36:46] 10Phabricator (Upstream), 10Upstream: On small screens, add a button to remove an action from a list of actions - https://phabricator.wikimedia.org/T224276 (10Aklapper) p:05Medium→03Lowest @Jonteemil: Do you [plan to work on fixing this task](https://www.mediawiki.org/wiki/Phabricator/Project_management#Se... [07:45:51] 10MediaWiki-Codesniffer, 10User-DannyS712: Add a sniff to warn when public methods lack `@since` - https://phabricator.wikimedia.org/T251785 (10thiemowmde) > warn when public methods lack `@since` Oh no, please don't do this. I feel such a sniff would be heavily counter-productive in pretty much **all** codeb... [07:51:57] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (CI & Testing services), 10Release-Engineering-Team-TODO (2020-04 to 2020-06 (Q4)): Cleanup CI puppet manifests - https://phabricator.wikimedia.org/T252942 (10Dzahn) I am really interested in removing `class { '::apache::mod::rewrite':` fro... [07:55:37] (03CR) 10Thiemo Kreuz (WMDE): [C: 03+1] "Fine for me." (034 comments) [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596771 (https://phabricator.wikimedia.org/T252948) (owner: 10Umherirrender) [07:56:05] (03CR) 10Thiemo Kreuz (WMDE): [C: 03+2] Fix unused use statement sniff for complex @phan-var [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596741 (owner: 10Umherirrender) [07:56:46] (03Merged) 10jenkins-bot: Fix unused use statement sniff for complex @phan-var [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596741 (owner: 10Umherirrender) [07:56:48] (03CR) 10Thiemo Kreuz (WMDE): [C: 03+2] Add test case for !is_null and parentheses must be kept [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596829 (owner: 10Umherirrender) [07:57:27] (03Merged) 10jenkins-bot: Add test case for !is_null and parentheses must be kept [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596829 (owner: 10Umherirrender) [07:57:48] (03CR) 10Thiemo Kreuz (WMDE): [C: 03+2] "> Can tests for the proper spacing not triggering warnings also be added?" [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596822 (owner: 10Umherirrender) [07:58:20] 10Project-Admins, 10User-DannyS712: Create #mediawiki-global-scope - https://phabricator.wikimedia.org/T249995 (10Aklapper) 05Open→03Declined Boldly declining per last comment. [07:58:25] (03Merged) 10jenkins-bot: Add test for SpaceAfterControlStructureSniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596822 (owner: 10Umherirrender) [08:03:31] (03PS9) 10DannyS712: FunctionCommentSniff: Move some sniffs to new separate sniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/595737 (https://phabricator.wikimedia.org/T250958) [08:03:42] (03CR) 10Thiemo Kreuz (WMDE): [C: 03+2] Include += in PlusStringConcatSniff (031 comment) [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596830 (owner: 10Umherirrender) [08:04:20] (03Merged) 10jenkins-bot: Include += in PlusStringConcatSniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596830 (owner: 10Umherirrender) [08:04:41] (03PS10) 10DannyS712: FunctionCommentSniff: Move some sniffs to new separate sniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/595737 (https://phabricator.wikimedia.org/T250958) [08:05:41] (03PS11) 10DannyS712: FunctionCommentSniff: Move some sniffs to new separate sniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/595737 (https://phabricator.wikimedia.org/T250958) [08:06:38] (03PS12) 10DannyS712: FunctionCommentSniff: Move some sniffs to new separate sniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/595737 (https://phabricator.wikimedia.org/T250958) [08:09:07] 10Phabricator (Upstream), 10Upstream: On small screens, add a button to remove an action from a list of actions - https://phabricator.wikimedia.org/T224276 (10Jonteemil) @Aklapper I wouldn't know how to fix this task unfortunately. I didn't know that "the Priority field summarizes and reflects reality and does... [08:20:55] (03CR) 10Thiemo Kreuz (WMDE): [C: 04-1] Simplify IfElseStructureSniff (036 comments) [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596825 (owner: 10Umherirrender) [08:22:13] (03CR) 10Thiemo Kreuz (WMDE): [C: 03+2] UnusedUseStatementSniff: Add @property-read and @property-write [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596868 (owner: 10Daimona Eaytoy) [08:22:20] (03CR) 10jerkins-bot: [V: 04-1] UnusedUseStatementSniff: Add @property-read and @property-write [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596868 (owner: 10Daimona Eaytoy) [08:24:57] 10Diffusion: Can't push to diffusion - https://phabricator.wikimedia.org/T243848 (10Aklapper) 05Open→03Declined Unfortunately closing this Phabricator task as no further information has been provided. @Magnus: After you have provided the information asked for and if this still happens, please set the status... [08:27:15] (03CR) 10Thiemo Kreuz (WMDE): [C: 03+1] "I think this is pretty uncontroversial. This sniff doesn't do much anyway. All it does is checking what's directly after a ?… nullable tok" [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596772 (https://phabricator.wikimedia.org/T252947) (owner: 10Umherirrender) [08:47:51] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (CI & Testing services), 10Release-Engineering-Team-TODO (201908), 10Technical-Debt: Cleanup CI puppet manifests - https://phabricator.wikimedia.org/T225735 (10hashar) [08:47:53] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (CI & Testing services), 10Release-Engineering-Team-TODO (2020-04 to 2020-06 (Q4)): Cleanup CI puppet manifests - https://phabricator.wikimedia.org/T252942 (10hashar) [08:48:02] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (CI & Testing services), 10Release-Engineering-Team-TODO (201908), 10Technical-Debt: Cleanup CI puppet manifests - https://phabricator.wikimedia.org/T225735 (10hashar) 05Resolved→03Open [08:48:07] (03PS2) 10Thiemo Kreuz (WMDE): Fix UnsortedUseStatements sniff failing with leading backslashes [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596997 [08:48:09] (03PS2) 10Thiemo Kreuz (WMDE): More resource friendly UnsortedUseStatements implementation [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/597000 [08:48:11] (03PS2) 10Thiemo Kreuz (WMDE): Add leading empty token test cases for unused "use" sniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/597001 [08:51:09] (03PS2) 10Thiemo Kreuz (WMDE): UnusedUseStatementSniff: Add @property-read and @property-write [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596868 (owner: 10Daimona Eaytoy) [08:51:50] (03CR) 10Thiemo Kreuz (WMDE): [C: 03+2] "PS2 is a trivial rebase only." [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596868 (owner: 10Daimona Eaytoy) [08:52:31] (03Merged) 10jenkins-bot: UnusedUseStatementSniff: Add @property-read and @property-write [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596868 (owner: 10Daimona Eaytoy) [09:14:16] 10Project-Admins: Requests for addition to the #acl*Project-Admins group (in comments) - https://phabricator.wikimedia.org/T706 (10Aklapper) Hi @Jony, sorry for the late reply. Please also see T706#5419426. Is there a //continuous// need to create projects? Also, Wikidata has some subprojects and [sprints](https... [09:17:55] (03PS3) 10Thiemo Kreuz (WMDE): Fix UnsortedUseStatements sniff failing with leading backslashes [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596997 [09:17:57] (03PS3) 10Thiemo Kreuz (WMDE): More resource friendly UnsortedUseStatements implementation [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/597000 [09:17:59] (03PS3) 10Thiemo Kreuz (WMDE): Add leading empty token test cases for unused "use" sniff [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/597001 [10:25:09] (03PS4) 10Thiemo Kreuz (WMDE): Fix unnecessary use statement detection for const in same namespace [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596765 (owner: 10Umherirrender) [10:32:14] (03CR) 10Thiemo Kreuz (WMDE): [C: 03+1] "The addition is absolutely welcome and all. \o/ But I felt the way this was implemented was, well, suboptimal. Two larger loops have been " (031 comment) [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/596765 (owner: 10Umherirrender) [11:38:50] Why is Jenkins not merging https://gerrit.wikimedia.org/r/c/mediawiki/core/+/594267 ? [11:42:40] maybe because it’s based on an older patch set of its parent chang? [11:42:42] *change [11:42:59] might need a rebase (on parent change or on master) [11:43:39] Nikerabbit: ^ [11:47:02] 10Project-Admins, 10User-Zoranzoki21: Project Creation request: Cronos (WMCH) - https://phabricator.wikimedia.org/T252768 (10Nicolas_NALLET) Hello Valerio, I have added some variables on https://meta.wikimedia.org/wiki/Template:Cronos_event In order to manage event on multiple days, I'm thinking about using th... [11:47:58] Lucas_WMDE: gerrit was not showing the parent as out of date, but trying rebase [11:48:05] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (CI & Testing services), 10Release-Engineering-Team-TODO (201908), 10Patch-For-Review, 10Technical-Debt: Cleanup CI puppet manifests - https://phabricator.wikimedia.org/T225735 (10hashar) 05Open→03Resolved All patches ninja merged by... [11:48:26] hm, it showed the “merged” message in orange on my end [11:48:35] which I think usually means that there’s some base commit mismatch [11:55:43] the reason it doesn't merge it is because it detects that page loading becomes slower with it [11:55:46] afaict [11:56:15] time to first paint changes: 164 ms (± 21 ms) | 201 ms (± 21 ms) | [11:56:18] before/after [11:57:54] I thought that is not a voting change [11:58:05] s/change/check/ [12:06:40] hmm. maybe it isn't. seems a significant change though [12:06:58] that being said, that patch is ready to submit now [12:08:41] maybe better to speak with perf-team ? [12:18:44] Nikerabbit: no clue. If the gate pipeline succeeded, that means Zuul has tried to submit the change in Gerrit [12:19:41] Nikerabbit: 'error: fatal: Failed to submit 1 change due to the following problems:\nChange 594267: depends on change that was not submitted\n\nfatal: one or more reviews failed [12:21:30] it depended on https://gerrit.wikimedia.org/r/#/c/mediawiki/core/+/594267/1 [12:21:33] which got rebased as well [12:21:52] well I guess rebase the whole chain, +2 them all and it should work ;) [12:42:43] 10Project-Admins, 10User-Zoranzoki21: Project Creation request: Cronos (WMCH) - https://phabricator.wikimedia.org/T252768 (10Aklapper) @Nicolas_Nallet: Maybe you commented on the wrong ticket? (This is only about creating a tag in Phabricator for Cronos, not to discuss any aspects of Cronos.) [13:07:06] 10Release-Engineering-Team-TODO (2020-04 to 2020-06 (Q4)), 10Release, 10Train Deployments: 1.35.0-wmf.32 deployment blockers - https://phabricator.wikimedia.org/T249964 (10hashar) [13:19:03] 10Release-Engineering-Team-TODO (2020-04 to 2020-06 (Q4)), 10Release, 10Train Deployments: 1.35.0-wmf.32 deployment blockers - https://phabricator.wikimedia.org/T249964 (10hashar) [13:50:37] 10Gerrit, 10Code-Review-Workgroup, 10Discovery, 10User-DannyS712: Cindy-the-browser-test-bot should use verified, not code review - https://phabricator.wikimedia.org/T249981 (10Aklapper) @Ebernhardson: Is there any repo so someone could provide a patch, or is this bot a black box for anyone not having CLI... [13:52:45] 10Phabricator: Change username from Adotchar to Vermont - https://phabricator.wikimedia.org/T253019 (10Adotchar) [13:55:39] 10Phabricator: Change username from Adotchar to Vermont - https://phabricator.wikimedia.org/T253019 (10Aklapper) 05Open→03Resolved a:03Aklapper [14:08:22] 10Phabricator, 10User-Josve05a: My profile picture on Phabricator looks wierd - https://phabricator.wikimedia.org/T235348 (10Josve05a) >>! In T235348#6142529, @Aklapper wrote: > @Josve05a: Does this also happen in Chrome? Have you found any pattern (assuming this still happens)? I mainly use Safari, so a few... [14:51:03] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (CI & Testing services), 10Release-Engineering-Team-TODO (2020-04 to 2020-06 (Q4)): Reduce size of artifacts stored on the CI Jenkins master - https://phabricator.wikimedia.org/T249268 (10hashar) 05Open→03Stalled Will revisit after the m... [15:29:54] Hey longma & releng :D CUrious to know what sorts of things are currently happening in the land of mediawiki and developer environments? :) [15:30:06] anywhere I can look for a general / generic update? [15:33:20] thcipriani: you have reply! [15:33:24] * RhinosF1 is Samuel [15:37:47] 10Continuous-Integration-Config, 10Release-Engineering-Team (CI & Testing services), 10Release-Engineering-Team-TODO (2020-04 to 2020-06 (Q4)), 10doxygen, and 2 others: Update Doxygen in CI to 1.8.17 or greater - https://phabricator.wikimedia.org/T242155 (10hashar) [15:37:49] 10Continuous-Integration-Config, 10Release-Engineering-Team, 10Release-Engineering-Team-TODO (2020-04 to 2020-06 (Q4)), 10doxygen, and 2 others: Update operations/debs/doxygen to 1.8.17 or later and release into apt.wikimedia.org - https://phabricator.wikimedia.org/T250424 (10hashar) [15:39:53] 10Continuous-Integration-Config, 10Two-Column-Edit-Conflict-Merge, 10WMDE-QWERTY-Team, 10WMDE-QWERTY-Sprint-2020-05-13: Add TwoColConflict to the "gated" extensions list in core - https://phabricator.wikimedia.org/T251276 (10Jdforrester-WMF) >>! In T251276#6142825, @awight wrote: >> Per https://integrati... [15:40:48] RhinosF1: :D thanks for the question! [15:42:16] addshore: links that might be helpful: https://www.mediawiki.org/wiki/MediaWiki-Docker + https://phabricator.wikimedia.org/tag/mediawiki-docker/ [15:42:47] hehe, yup I read those, I was thinking more about active iterations / things in the pipeline to be added to it [15:43:06] I ask as last week I started working on a next iteration of my mediawiki-docker-dev thing, and I wonder if we could eventually merge the 2 things [15:46:05] addshore: ooh, that'd be cool. I don't know of the async place for discussion other than tagging stuff on phab ( brennen or longma might correct me). There's a local dev meeting on Thursdays that this might be right to talk about there (I don't attend, so I could be wrong about that :)) [15:46:51] 10Phabricator, 10User-Josve05a: My profile picture on Phabricator looks wierd - https://phabricator.wikimedia.org/T235348 (10Aklapper) 05Stalled→03Declined Hmm, alright. Thanks for the update! [15:51:51] thcipriani: that sounds like a nice meeting to know about, awesome, I'll wait for one of them to poke me! [15:52:22] thcipriani: np, I’m sure I’ll have many as I learn! [15:52:38] addshore: warning, this week is the Engineering Productivity team's (virtual) offsite, so we/they might be slower to respond. But I hope they do! [15:52:50] thanks for the heads up! [15:52:51] (releng, performance, and q&te) [15:57:42] addshore: yeah, about to dive into meetings, but that sounds like probably a good idea and i'd love to discuss. [15:58:27] <3 [16:05:25] addshore: It would be great to see you at the local dev meeting so we can coordinate :) [16:11:56] James_F: While I'm making you sad about browser tests, let's chisel away at the quibble patches :-) Or just let me know when I can help nudge my shoddy craftspersonship over the cliff... [16:12:12] We're all at a virtual offsite this week, sorry. [16:13:05] Well I would never try to insert this issue into the agenda ;-) [16:19:11] awight: I was on train last week with kids at home [16:19:29] awight: but this week I shuold have more time available to cleanup the Quibble current state ;) [16:19:37] I might end up reverting some patches [16:20:01] Sorry about making the mess, I'm available to help in any way. [16:20:03] and I am definitely going to write tests to cover the plan build [16:20:31] I did push a patch which might fix the issue I caused, but of course fine to revert as well. [16:20:35] ie what happens when one uses --skip-zuul , --run=phpunit and the ZUUL_PROJECT=mediawiki/core [16:20:51] I gotta review the patch that merged [16:21:05] and surely if there is a good fix pending review, I will be happy to +2 it ;) [16:21:23] I guess part of the issue is that the quibble behavior are not described / tested [16:21:36] but a bunch of basic tests for the build plan should be enough [16:21:56] The two I'm thinking of are https://gerrit.wikimedia.org/r/#/c/integration/quibble/+/594800/ and https://gerrit.wikimedia.org/r/#/c/integration/quibble/+/594791/ [16:22:24] yeah will check those [16:22:38] +1 testing the plan builder seems like a great idea [16:22:40] also I have noticed the function that creates the build plan does a mkdir(log_dir) ;D [16:22:51] I guess that can be made a command instead [16:23:34] definitely going to add patches for those ;) [16:23:51] + polish up stuff to let us migrate Jenkins/Zuul to buster/contint2001 [16:24:34] I don't have a good grasp on what should be a command, but my instinct is that the "command" granularity should be c. "a step that should be reported out to the user", maybe mkdir(log) is too small? [16:25:24] yup [16:25:38] good point, so maybe the mkdir(log_dir) should be done somewhere else [16:25:51] will try to find a better place for it [16:26:34] long story short, I enjoy working in that repo as you can see, so maybe just guide my enthusiasm into more useful pursuits [16:26:46] hehe [16:27:16] support for Apache as the webserver would bbe rather nice [16:27:27] Kosta benchmarked apache to be way faster than "php -s' [16:27:45] possibly WMDE devs might have cool ideas for it as well [16:27:47] +1 that's going to be required for any of the parallelism. [16:28:03] But IMO it's blocked on a decision about where we launch docker [16:28:15] yeah that is a pain [16:28:23] cause I don't really want Quibble to become a container orchestrator [16:28:33] +1.1 [16:28:56] when helm / docker composer would be able to assemble the environment / backends [16:29:10] the devil is to then pass the configuration bits to the mediawiki installer / LocalSettings.php [16:29:13] I think we should shrink quibble like you say, and do this sort of thing, https://gerrit.wikimedia.org/r/#/c/integration/quibble/+/590651/ [16:29:18] hashar: why not? (beyond it being a pain, of course) [16:29:22] eg ip/port of an elasticsearch [16:29:24] or the services [16:29:40] kostajh: cause I feel that is reinventing the wheel :] [16:29:54] and would make it way more complex [16:30:06] I think passing the config bits won't be too hard, but we should decide if we really want a swamp of commandline options, or a config file, or env vars? [16:30:33] addshore: the way I have loosely been thinking about it is that mw-cli (https://gerrit.wikimedia.org/g/mediawiki/tools/cli) could do what the various bash scripts in mediawiki-docker-dev do, while the docker-compose.yml file in core remains a stable and extendable base [16:30:45] and there is some overlap with the work going on to provide a mediawiki dvelopment environment [16:31:07] kostajh: My thought about the orchestration level is that whatever does that should be decoupled from mediawiki install and phpunit running [16:31:43] hashar: I don't know if this qualifies as "container orchestration" but running `docker-compose up -d` and then waiting for healthchecks to be OK, then starting tests doesn't seem that bad? [16:32:09] kostajh: that would probably be fine yes [16:32:14] addshore: and would love to collaborate on bringing the features/functionality of these two tools together [16:32:41] the tricky part is having docker compose to generate a setting file that is then passed to the mediawiki installation [16:32:45] I like the idea that quibble would be one of the containers started, rather than a master... but interested in hearing the motivation for doing it the other way [16:33:30] I will also be very happy if the end result turns out to phase out Quibble entirely because all of its functionalityies ended up implemented in mediawiki / something else :] [16:33:45] hehe every developer's dream [16:33:51] ;D [16:34:09] so yeah thsi week I am going to be a bit busy (plus thursday is an holiday here) [16:34:33] yeah quibble could be one of the containers, why not [16:34:36] but probably we might want to start having some kind of weekly meetings to speak about those problems [16:34:43] Enjoy! [16:35:37] hashar: That sounds fun, and I'm sure it would quickly get us past a few blockers. [16:35:44] possibly yeah [16:35:46] puppet is disabled on deployment-docker-changeprop01 for about 4 days now. Doesn't say who disabled it, just "Disabled by disable-puppet" [16:35:49] kostajh: it's in Go? :P [16:35:55] that COVID19 confinment stuff has put a tall on most of my projects :-( [16:36:27] kostajh: I think that is the sort of direction I went on the v1 branch of mediawiki-docker-dev. Currently experimenting with what commands make sense etc [16:36:47] +1 I have a zoo at home, it's taking a bite out of all projects [16:36:50] It's got the potential to not be MediaWiki specific too, rather a generic dev environment. [16:37:01] * bd808 needs to figure out where to send the deployment-prep monitoring nags these days with the shinken replacement system [16:38:06] bd808: it is being worked on apparently. From `last` hnowlan was the user logged at that time [16:38:14] addshore: Thanks for the docker work, it's been incredibly useful. [16:38:32] Maybe we should schedule a quick call and I could give you a dmguided tour of what I have so far? I'd plan on writing the README and changeling this week and have some people try it out next week. [16:38:32] bd808: I would assume it is being provisioned / worked on by sre [16:38:49] addshore: re Go, yeah the idea is to ship static binaries and not require eg php on someone’s system [16:38:49] But ideally all of this could live in someone else's domain :D [16:38:52] bd808: yep, that's me. Apologies for the useless message [16:38:58] I'll fix now [16:39:09] hashar: I need a couple of repos back to 'Active' from 'Read Only' for T239429 - Could you assist? [16:39:10] T239429: Delete/Archive Github and Diffusion mirrors of Gerrit repos related wikiba.se - https://phabricator.wikimedia.org/T239429 [16:39:10] magic :] [16:39:29] kostajh: aah, gotcha, I started looking down the direction of having docker be the only dependency, and running the CLI tool in a container too :p [16:39:37] hauskatze: sure [16:39:45] Merci bien Monsieur [16:39:46] thanks hashar and hnowlan. I'll do the needful to ack the outage using the magic tools y'all don't have access to yet anyway ;) [16:40:14] hauskatze: done [16:40:31] 10Diffusion, 10Release-Engineering-Team, 10Projects-Cleanup, 10Wikimedia-GitHub, 10User-MarcoAurelio: Delete/Archive Github and Diffusion mirrors of Gerrit repos related wikiba.se - https://phabricator.wikimedia.org/T239429 (10hashar) I have made both Gerrit repos active again: * https://gerrit.wikimedi... [16:40:43] hashar: thanks! :] [16:40:47] kostajh: so the go thing could come with a PHP binary?, And the CLI thing could a tuslly be written in PHP? ;) [16:41:02] addshore: I would be happy to be given a tour as well :] [16:41:18] addshore: ehh not sure about that [16:41:26] kostajh: hehe [16:41:46] 10Diffusion, 10Release-Engineering-Team, 10Projects-Cleanup, 10Wikimedia-GitHub, 10User-MarcoAurelio: Delete/Archive Github and Diffusion mirrors of Gerrit repos related wikiba.se - https://phabricator.wikimedia.org/T239429 (10MarcoAurelio) Thanks! - I'll try to do something today. [16:41:51] addshore: but have a look at the repo it’s intelligible enough coming from PHP land I think? [16:42:00] I could even do a call now ish if anyone is free? [16:42:20] Yup, it should be fairly easy to follow what's going on in the repo. [16:45:30] offsite for me, but that should not prevent you guys from talking to each others ;D [16:45:49] Maybe i should just record a guide video and send it to you all tommorrow :P [16:47:29] it even has quibble so you can "mwdd mw:quibble " already passing the --skip-zuul --skip-deps --skip-install options. [16:47:45] That is also the reason that I wrote a few quibble tickets at the end of last week hashar [16:48:10] mwdd = mediawiki dungeons & dragons ? ;D [16:48:11] * addshore goes to look at the go repo [16:48:16] where is my sword of +3 bug [16:49:49] anyway I am going to focus on the ongoing meeting I am attending [16:50:13] kostajh: i looked at 2 files, and it would pretty easy I think the migrate all of what I did in my v1 branch in the last week into the cli tool [16:50:14] but definitely willing to attend any demo / calls to chat about the issues ;) [16:51:53] I ended up boiling the "environment" down into a dns service and proxy service which are always quite useful, and then most of the stuff past that is just powered by individual docker-compose files / snippets now [16:51:55] addshore: cool! Very happy to look at patches [16:52:44] https://usercontent.irccloud-cdn.com/file/mm9PB02z/image.png [16:52:45] ^^ first view [16:53:22] :]]] [16:54:03] * kostajh looks at https://github.com/addshore/mediawiki-docker-dev/compare/v1 [16:54:28] its a total rewrite, but only started last wedneday :P [16:54:31] *wednesday [16:56:09] Symfony Console is great [16:56:32] Yeah, I'm not happy about how verbose the list command ends up being, Id rather have some sort of sub lists, but not figured that out yet [16:57:04] * hashar hints at maintenance/Maintenance.php [16:57:17] would be happy to see that one gone in favor of symfony/console [16:57:26] the docker-compose stuff is all in https://github.com/addshore/mediawiki-docker-dev/tree/v1/docker-compose [16:57:39] thcipriani: I’d go over 6 patches backporting at the moment on the tasks I emailed about with 2 version in use so do we just wait for a clear Monday/Thursday before the train [16:57:56] addshore: fwiw, I didn't want to go the route of using PHP for the CLI because of not wanting to make assumptions about host system environments, and I'd like eventually for non technical people to be able to do something like `mw docker start` and `mw patch-get 12345` and be able to see that in their local env. Go seemed like a good choice because of the compilation into a static binary. (Yes I know doable with python, but ...) [16:58:15] +1 to using symfony/console for our maintenance scripts [16:58:43] I like the sounds of go, I have nothing against it :) The only reason I want for php, was because I was trying to get a working version for friday starting on wednesday! [16:59:37] kostajh: the attempt at running the control app within a container too is at https://github.com/addshore/mediawiki-docker-dev/blob/v1/mwdd#L30-L36 and that works for the simple stuff, but I didn't finish it for most of the stuff due to running out of time [17:00:22] It's all just glue really, but there certainly are some lessons I would like to pass on from it for mw-cli [17:00:40] let me try and install mw-cli now! [17:04:38] addshore: cool lmk if you run into issues [17:05:10] setup 1, update go [17:13:37] on go get i get "unrecognized import path "cli/cmd": import path does not begin with hostname" :P [17:13:40] * addshore googles [17:19:28] aaah, it literally has to be in a directory called cli in the go root [17:19:30] addshore: where did you clone the repo? [17:19:31] yeah [17:19:55] * addshore might make a patch switching to a package name prefixed with a domain? :D [17:19:58] i think there are ways to work around that but maybe more trouble than they're worth [17:20:35] * addshore does more googling [17:21:46] 10Release-Engineering-Team, 10Wikimedia-Extension-setup, 10CPT Initiatives (API Gateway): Determine how to host Chameleon skin and Bootstrap extension in gerrit while primary development happens in github - https://phabricator.wikimedia.org/T252463 (10CCicalese_WMF) Seeking guidance from #release-engineering... [17:24:28] addshore: btw if you want to make phab tasks for patches, I think just use #MediaWiki-Docker project for now [17:29:15] (03CR) 10Jforrester: "If this is going to production, it should have the `in-wikimedia-production` template too. Also, really, yet another skin? Oy." [integration/config] - 10https://gerrit.wikimedia.org/r/596797 (https://phabricator.wikimedia.org/T247837) (owner: 10Reedy) [17:30:36] (03CR) 10Reedy: "Schrödinger's skin... It's simultaneously a skin and not a skin..." [integration/config] - 10https://gerrit.wikimedia.org/r/596797 (https://phabricator.wikimedia.org/T247837) (owner: 10Reedy) [17:31:01] * hauskatze rolls eyes [17:34:50] kostajh: still struggling to make it work right, even putting it exactly where the readme say [17:34:59] however, relative paths seem to work? [17:35:47] okay, finally got it, woo! [17:35:50] addshore: sorry :( you are among the first people to test out those instructions [17:36:40] https://usercontent.irccloud-cdn.com/file/FNtTPpup/image.png [17:36:42] ;) [17:38:29] hauskatze: Don't you roll your eyes at me [17:40:30] 🙄🙄🙄 [17:42:03] Reedy: Whose eyes should they roll? [17:42:26] addshore: yay! [17:43:46] One bit of feedback that I already had from my v1 tool is that people don't like all of the magic being hidden, and they would rather see more of what is going on. "mw docker start" looks like it hides lots :P [17:44:18] addshore: +1 on not hiding magic [17:44:36] addshore: one simple thing would be to add a `--verbose` flag that spits out the `docker-compose` commands being run in the background [17:44:51] yup, thats what I was going to add to the v1 thing [17:45:04] the v1 commands end up being pretty long, as they are using multiple docker-compose files [17:45:42] I wonder what your thoughts are on multiple docker-compose files vs 1 override file? [17:47:15] addshore: for now I've promoted the override file (https://www.mediawiki.org/wiki/MediaWiki-Docker#Example_docker-compose.override.yml_file) as it seems fairly straightforward to build an environment for one particular project. Longer term, it would be really nice to have CLI manage weaving various compose files for you [17:48:04] Yeah, okay, thats what I ended up thinking too, as then I could leave the override file / files for users customaizations rather than extra services for the environment [17:50:14] so, after staring at it a bit and trying the commands I think the main difference between the 2 is now that mwdd is not mediawiki.git centered, but the mwcli tool is currently [17:50:59] although, actually the tool should be able to work for other git repos just fine... with a docker-compose file in there and some magic [17:51:40] so perhaps it might be an idea to name the docker-compose created project, rather than just name it by the name of the repo of core. [17:52:32] I ended up specifying the name pretty early on https://github.com/addshore/mediawiki-docker-dev/blob/v1/control/src/Shell/DockerCompose.php#L12 I might make a patch for it. I was going to alter mine to be mwdd-$USER, so might try mwcli-$USER for mw-cli (also opens up the ability for multiple people to run it on one machine [17:52:38] addshore: sorry I'm not following exactly, could you rephrase that please? [17:53:35] so, in theroy I can cd into a directory for another service, such as the wikibase termbox, and run "mw docker-service start" or something like that, and it would load the service defined in that git repo into the same docker-compose project / network etc as the rest of the mw-cli stuff [17:53:39] you can have multiple instances running on the same host machine [17:53:59] ah [17:55:27] that sounds complicated [17:55:49] but useful [18:05:08] Reedy: Why do you assume I roll eyes at you? [18:05:28] :P [18:05:33] :) [18:05:43] You hadn't spoken and it was after James' and my comments [18:06:36] It was meant to be at the "skin which is not a skin" thing [18:06:51] I gathered :P [18:07:35] k then, I don't want any misunderstandings [18:10:28] James_F: what do you think re https://gerrit.wikimedia.org/r/#/c/mediawiki/core/+/542312/ ? I think it's okay but it's complex code for me :-) [18:11:57] hauskatze: I was carefully ignoring that patch, damn you. ;-) [18:12:53] It [18:12:55] Bah. [18:13:08] It's probably fine, but I don't have time to test it right now, sorry. [18:13:13] Np [18:36:47] kostajh: you have a PR ;) [18:36:58] wait, it is on gerrit, you have a patch... [18:37:58] this is exciting, I can use this mwdd v1 thing as fast prototyping area, and try to bake the learnings into the cli tool [18:38:35] Do I need to worry about breaking changes? / should it have a changelog? [18:52:17] 10Continuous-Integration-Config, 10Release-Engineering-Team-TODO (2020-04 to 2020-06 (Q4)), 10CPT Initiatives (API Integration Tests), 10Code-Health, 10MW-1.35-notes (1.35.0-wmf.19; 2020-02-11): Enable API integration tests in CI for MediaWiki core - https://phabricator.wikimedia.org/T236680 (10apaskulin... [19:14:36] 10Gerrit, 10Code-Review-Workgroup, 10Discovery, 10User-DannyS712: Cindy-the-browser-test-bot should use verified, not code review - https://phabricator.wikimedia.org/T249981 (10EBernhardson) Writing a patch is no problem, it's a 1 line change to the bot which i've already tried. The problem is the bot does... [19:18:03] Anyone around who can help this Blubber neophyte with getting a publish build step working for a TypeScript Node service? (see my last comments on https://gerrit.wikimedia.org/r/c/mediawiki/services/push-notifications/+/594352) [19:20:56] Hi greg-g, everyone.... Whom should I ask for guidance on reproducing and understanding an XSS check failure that shows up on CI? Thanks in advance! [19:21:08] (It's for new code not yet deployed.) [19:27:16] 10Gerrit, 10Code-Review-Workgroup, 10Discovery, 10User-DannyS712: Cindy-the-browser-test-bot should use verified, not code review - https://phabricator.wikimedia.org/T249981 (10Reedy) >>! In T249981#6146292, @EBernhardson wrote: > Writing a patch is no problem, it's a 1 line change to the bot which i've al... [19:48:10] 10MediaWiki-Codesniffer, 10MediaWiki-General, 10MW-1.35-notes (1.35.0-wmf.26; 2020-03-31), 10MW-1.35-release, and 2 others: Fix all functions that have required parameters following optional ones - https://phabricator.wikimedia.org/T248078 (10Reedy) [19:58:45] thcipriani: I have messaged RoanKattouw. I am looking at the hour before Evening SWAT any day. [20:01:44] RhinosF1: ack, thanks, let me know if you need any help with schedule or deployment. [20:03:18] AndyRussG: I'm sorry, I'd have to pawn that off to someone like maybe Reedy [20:03:47] greg-g: okok thanks! [20:04:26] bearND: where are you seeing pull access denied message? is that in CI or locally? [20:04:44] 10MediaWiki-Codesniffer, 10MediaWiki-General, 10MW-1.35-notes (1.35.0-wmf.26; 2020-03-31), 10MW-1.35-release, and 3 others: Fix all functions that have required parameters following optional ones - https://phabricator.wikimedia.org/T248078 (10Reedy) [20:04:44] thcipriani: we are doing it tommorow. Thanks for all the help! [20:05:42] RhinosF1: great! thanks for the question on the list and pushing forward :) [20:05:44] thcipriani: I saw that when I ran this locally. [20:05:56] (03PS1) 10Hashar: Add releng as owner to archive the repo [wikibase/wikiba.se-deploy] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/597148 (https://phabricator.wikimedia.org/T239429) [20:06:00] thcipriani: np :) [20:06:06] (03CR) 10Hashar: [V: 03+2 C: 03+2] Add releng as owner to archive the repo [wikibase/wikiba.se-deploy] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/597148 (https://phabricator.wikimedia.org/T239429) (owner: 10Hashar) [20:09:20] thcipriani: basically I added the lines from my comment to blubber.yaml and ran `blubber .pipeline/blubber.yaml production | docker build -f - .` [20:09:46] yep, I ask because it seems like an error with the docker daemon (but I've never seen it before) [20:10:06] I'm reading the gerrit patch now, trying to catchup to what's happening [20:11:10] node: {env: production} should just do: npm install --only=production but that breaks the build? [20:12:00] Well, I want to run the build first somewhere but in order to do that it needs the typescript compiler `tsc`. [20:12:35] For that we need the typescript package which lives in the devDependencies. [20:12:40] 10Diffusion, 10Release-Engineering-Team, 10Projects-Cleanup, 10Wikimedia-GitHub, and 2 others: Delete/Archive Github and Diffusion mirrors of Gerrit repos related wikiba.se - https://phabricator.wikimedia.org/T239429 (10hashar) Merged both changes and made them both read only again. * https://gerrit.wikime... [20:12:59] `npm install --only=production` does not install devDependencies. [20:14:10] So, we basically need the output from the build (`npm run build`) to be copied over to the production image but then get rid of the devDependencies. [20:14:28] ah, ok, so yeah you proposed solution seems like a good one [20:14:38] lemme fiddle locally to see what that gets me [20:15:00] Sounds great. TY! [20:16:38] thcipriani, RoanKattouw: https://wikitech.wikimedia.org/w/index.php?title=Deployments&diff=1866610&oldid=1866598 [20:17:39] bearND: to check my understanding; you need (1) run npm in developer mode to compile typescript (2) copy the compiled javascript to the production image (3) run npm --only production to install production dependencies in the production image? [20:19:58] thcipriani: yes, that would work. Alternatively if we could somehow get to copy the `dist` folder from a previous image then we could skip this complication. Just not sure if we can assume and easily access the build output from another pipeline. [20:22:12] 10Gerrit, 10Code-Review-Workgroup, 10Discovery, 10User-DannyS712: Cindy-the-browser-test-bot should use verified, not code review - https://phabricator.wikimedia.org/T249981 (10EBernhardson) I've started the bot up with the appropriate options for v+1, it should go back through and re-vote on any patch fro... [20:31:51] kostajh: I'm going to try and make a chain of rough patches going in the direction that I have convinced myself is "right" with the intention of putting across what is in my brain for others to see. the chain is going to get a bit WIProgressy now and I wont keep it "super clean" but hopfully the intentions will come across in the code and commit messages [20:34:23] addshore: ok cool! smaller and less scope is always easier to merge, of course :) [20:35:15] yup indeed :) I like the project so will come and drag some bits out of the patches that you approve of (like the docker compose abstraction) [20:36:12] addshore: one thing I'd like to do is ensure that people can still use the docker-compose environment (with the instructions in core's DEVELOPERS.md, which we can modify of course) without needing the cli tool [20:36:52] like, the cli tool should be preferred because it will make things way easier / more automated, but it wouldn't be a hard requirement. that's how I've been thinking about it anyway [20:37:19] yup, I think my chain of patches will detach the CLI too from the DEVELOPERS stuff in core.git and the docker-compose there a bit, but also still keep using the same docker-composer file (I think) [20:37:55] ^ is in context of seeing the patch about moving .env outside of the core directory [20:38:08] my idea would be that if your just using the guide from core.git it would be easy to "upgrade" to the cli tool, but they would be seperate things too [20:38:49] kostajh: yup, but we will see how this all pans out, as I havn't prototyped this in mediawiki-docker-dev yet [20:39:53] addshore: btw, have you looked at https://docs.lando.dev/basics/ ? there are some good ideas in there [20:40:41] oooh, i havn't see that, I want to read it :P [20:42:04] addshore: it's worth downloading / installing to test drive, there's definitely a well developed set of stories there for how to pull together an environment for a project. [20:43:16] alright, it's late, i'm out. i'll keep an eye out for your patches, thanks in advance for submitting them! [20:43:24] o/ [20:49:51] that lando thing looks interesting, and is probably a slightly more complex version of what I think I have in my head [21:09:10] 10Continuous-Integration-Config, 10Release-Engineering-Team (CI & Testing services), 10Release-Engineering-Team-TODO: Enhance debian-glue job packages validation - https://phabricator.wikimedia.org/T158553 (10hashar) [21:15:07] 10Deployments, 10User-RhinosF1: Namespace & Sitename localisations deployment - 19 May 2020 - https://phabricator.wikimedia.org/T253070 (10RhinosF1) [21:17:23] 10Deployments, 10User-RhinosF1: Namespace & Sitename localisations deployment - 19 May 2020 - https://phabricator.wikimedia.org/T253070 (10RhinosF1) [21:17:33] 10Deployments, 10User-RhinosF1: Namespace & Sitename localisations deployment - 19 May 2020 - https://phabricator.wikimedia.org/T253070 (10RhinosF1) p:05Triage→03Medium [21:19:30] 10Continuous-Integration-Config, 10Browser-Tests: Selenium failure notification email has broken links - https://phabricator.wikimedia.org/T154338 (10hashar) 05Open→03Declined Unfortunately looks like we have missed that bug report at the time it has been opened and never went to triage it. The job was us... [21:24:49] 10Continuous-Integration-Config, 10good first task, 10patch-welcome: adjust recheck regex to also act on comment with additional inline comments - https://phabricator.wikimedia.org/T130640 (10hashar) We had some improvement made to the original regex: d5dd9632ec3c55f9b48b3045e36356db83ba7006 which allows to... [21:25:10] 10Continuous-Integration-Config, 10good first task, 10patch-welcome: adjust recheck regex to also act on comment with additional inline comments - https://phabricator.wikimedia.org/T130640 (10hashar) [21:25:12] 10Continuous-Integration-Config, 10Release-Engineering-Team (Backlog), 10Zuul, 10Patch-For-Review: recheck is ignored if there are also inline comments - https://phabricator.wikimedia.org/T171352 (10hashar) [21:29:08] 10Diffusion, 10Release-Engineering-Team, 10Projects-Cleanup, 10Wikimedia-GitHub, and 2 others: Delete/Archive Github and Diffusion mirrors of Gerrit repos related wikiba.se - https://phabricator.wikimedia.org/T239429 (10MarcoAurelio) 05Open→03Resolved C'est fait! [21:29:21] 10Diffusion, 10Release-Engineering-Team, 10Projects-Cleanup, 10Wikimedia-GitHub, 10User-MarcoAurelio: Delete/Archive Github and Diffusion mirrors of Gerrit repos related wikiba.se - https://phabricator.wikimedia.org/T239429 (10MarcoAurelio) [21:30:44] 10Continuous-Integration-Config, 10Release-Engineering-Team-TODO (2020-04 to 2020-06 (Q4)), 10Core Platform Team, 10Quibble, and 2 others: Jenkins no longer fails build if mw-error.log or mw-dberror.log are non-empty - https://phabricator.wikimedia.org/T246358 (10Krinkle) [21:30:49] 10Continuous-Integration-Config, 10Core Platform Team, 10MediaWiki-Core-Testing, 10Quibble, and 2 others: Jenkins build for MediaWiki should fail when "PHP Warning" is emitted - https://phabricator.wikimedia.org/T214552 (10Krinkle) [21:36:24] !log started  puppetdb on deployment-puppetdb03. it was not running [21:36:25] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [22:09:41] Reedy: doing in the morning [22:15:25] 10Deployments, 10User-RhinosF1: Namespace & Sitename localisations deployment - 19 May 2020 - https://phabricator.wikimedia.org/T253070 (10RhinosF1) [23:05:01] 10Release-Engineering-Team-TODO (2020-04 to 2020-06 (Q4)), 10Release, 10Train Deployments: 1.35.0-wmf.32 deployment blockers - https://phabricator.wikimedia.org/T249964 (10Krinkle)