[00:32:44] 10Release-Engineering-Team, 10Web Team Visual Regression Framework, 10Readers-Web-Backlog (Needs Prioritization (Tech)): Make visual regression tests run in CI (non-blocking) for the Vector repo - https://phabricator.wikimedia.org/T308194 (10nray) [00:55:39] 10Gerrit, 10Wikimedia-GitHub: Replication to GitHub seems to be broken - https://phabricator.wikimedia.org/T320305 (10Reedy) [00:55:46] 10Gerrit, 10Wikimedia-GitHub: Replication to GitHub seems to be broken - https://phabricator.wikimedia.org/T320305 (10Reedy) p:05Triage→03High [00:57:26] 10Gerrit, 10Release-Engineering-Team, 10Wikimedia-GitHub: Replication to GitHub seems to be broken - https://phabricator.wikimedia.org/T320305 (10Reedy) [00:58:30] 10Gerrit, 10Release-Engineering-Team, 10Wikimedia-GitHub: Replication to GitHub seems to be broken - https://phabricator.wikimedia.org/T320305 (10Reedy) This doesn't look right either... ` % ssh -p 29418 gerrit.wikimedia.org replication start mediawiki/core --wait Replicate mediawiki/core ref ..all.. to ger... [01:01:09] 10Gerrit, 10Release-Engineering-Team, 10Wikimedia-GitHub, 10Regression: Replication to GitHub seems to be broken - https://phabricator.wikimedia.org/T320305 (10Reedy) Timing looks to be around the time of {98879cec2e8d5621b7f8a510bd23b4952238e5ee} [01:04:53] 10Continuous-Integration-Config, 10MediaWiki-Vendor, 10Composer: Upgrade composer to 2.4.x - https://phabricator.wikimedia.org/T319257 (10Reedy) It gives some nice format updates.. ` % php ../composer.phar outdated Color legend: - patch or minor release available - update recommended - major release availab... [03:50:23] 10Phabricator, 10VPS-project-Codesearch: Consider adding a way to query Codesearch from Phabricator - https://phabricator.wikimedia.org/T183608 (10Krinkle) [06:01:49] 10Continuous-Integration-Infrastructure, 10Doxygen, 10MediaWiki-Documentation, 10MediaWiki-libs-ObjectCache, 10Performance-Team: Upgrade Doxygen to 1.9.1 - https://phabricator.wikimedia.org/T291950 (10Krinkle) [06:19:04] > chmod 644 some files that are 755. https://gerrit.wikimedia.org/r/c/mediawiki/core/+/840243 [06:19:13] hm.. this patch includes fixing .php files [06:19:23] isn't minus-X meant to catch this? [06:23:02] 10Gerrit, 10Release-Engineering-Team, 10Wikimedia-GitHub, 10Regression: Replication to GitHub seems to be broken - https://phabricator.wikimedia.org/T320305 (10hashar) a:03hashar [06:45:08] 10Gerrit, 10Release-Engineering-Team, 10Wikimedia-GitHub, 10Regression: Replication to GitHub seems to be broken - https://phabricator.wikimedia.org/T320305 (10hashar) The last log of replication to Github: [2022-10-07 19:00:47,871] Replication to git@github.com:wikimedia/mediawiki-services-function-... [07:22:06] 10Gerrit, 10Release-Engineering-Team, 10Wikimedia-GitHub, 10Regression: Replication to GitHub seems to be broken - https://phabricator.wikimedia.org/T320305 (10hashar) 05Open→03Resolved [07:22:12] 10Gerrit, 10Release-Engineering-Team, 10Wikimedia-GitHub, 10Regression: Replication to GitHub seems to be broken - https://phabricator.wikimedia.org/T320305 (10hashar) Change to the replication configuration are reloaded but ends up breaking the replication destination. We apparently had the same issue pre... [09:57:56] 10Gerrit, 10Release-Engineering-Team, 10Wikimedia-GitHub, 10Regression, 10Upstream: Replication to GitHub seems to be broken - https://phabricator.wikimedia.org/T320305 (10hashar) Filed upstream as https://bugs.chromium.org/p/gerrit/issues/detail?id=16314 [09:58:05] 10Gerrit, 10Release-Engineering-Team, 10Wikimedia-GitHub, 10Regression, 10Upstream: Replication to GitHub seems to be broken - https://phabricator.wikimedia.org/T320305 (10hashar) [17:22:44] replication has nearly caught up again [20:16:19] (03PS1) 10Legoktm: Use our packages for node16, bringing in npm v8 [integration/config] - 10https://gerrit.wikimedia.org/r/840336 [20:25:13] https://phabricator.wikimedia.org/T320332 Uh-oh... [20:25:19] 10Continuous-Integration-Config, 10MediaWiki-Core-Tests, 10Test-Coverage: Extension coverage reports are broken with PHPUnit 9.5; they seem to report MW core not the extension - https://phabricator.wikimedia.org/T320332 (10Legoktm) a:03Legoktm phpunit-suite-edit needs to be updated [20:26:59] Daimona: :popcorn: [20:27:16] Oh but lego saves the day, I see :) [20:34:58] (03PS1) 10Legoktm: Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 [20:35:13] Daimona: it's always fun to rediscover things I worked on 4+ years ago :p [20:35:29] (03PS2) 10Legoktm: Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 [20:35:33] Reedy: ^ [20:36:06] I'm not sure if "fun" is the right word :P [20:36:58] legoktm: Oh, and since you're here... Is there a way to tell LibUp that all the canaries have been updated to the latest phan and that it can start updating it everywhere now? [20:37:06] (03CR) 10CI reject: [V: 04-1] Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (owner: 10Legoktm) [20:37:08] https://libraryupgrader2.wmcloud.org/library/composer/mediawiki/mediawiki-phan-config?branch=main still only shows 3 canaries with the latest version [20:37:59] it's in the middle of a run now, when it gets to those repos it'll update the status [20:38:37] AssertionError: Expect Quibble image version 1.4.6, got 1.4.7 in dockerfiles/quibble-buster-php74-coverage/changelog [20:39:45] (03PS3) 10Legoktm: Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 [20:39:51] (03CR) 10Daimona Eaytoy: "I also see the following in the job output:" [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (owner: 10Legoktm) [20:40:23] (03CR) 10Legoktm: Update phpunit-suite-edit for PHPUnit 9 (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (owner: 10Legoktm) [20:40:51] Thank you, and out of curiosity, is there a dashboard or something where it shows what it's doing right now? [20:41:26] (03CR) 10CI reject: [V: 04-1] Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (owner: 10Legoktm) [20:42:02] just the home page https://libraryupgrader2.wmcloud.org/ [20:43:34] (03PS4) 10Legoktm: Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 [20:46:26] We should upgrade stuff more often ;) [20:47:37] depending on how bold people feel, we should also move coverage to PHP 8 for some speed boost [20:48:21] Probably a good idea when we're a bit more stable (ie something to do very soon) [20:48:51] (03CR) 10Reedy: [C: 03+1] Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (owner: 10Legoktm) [20:49:33] Thanks, that already helps :) [20:49:51] And we definitely should upgrade stuff more often, yes... [20:50:10] Are tests passing everywhere on PHP 8 now? [20:50:21] heh no [20:50:35] I think we're getting there, at least for MW bundled and WMF deployed [20:51:02] there's quite a few failing for other reasons, I've slowly just been picking stuff out of https://gerrit.wikimedia.org/r/q/owner:libraryupgrader+status:open+label:Verified-1 [20:52:57] Yeah, I saw some progress being made but didn't know the current status [20:53:34] I think we're good enough on MW core (as PHP 8 is voting now) [20:53:41] (03CR) 10Daimona Eaytoy: [C: 03+1] Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (owner: 10Legoktm) [20:54:10] probably someone else should deploy ^, I haven't pushed new docker images in a few years... [20:54:22] (03CR) 10Daimona Eaytoy: [C: 04-1] "Actually," [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (owner: 10Legoktm) [20:54:38] oop [20:54:46] (03PS5) 10Legoktm: Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (https://phabricator.wikimedia.org/T320332) [20:54:52] I'd do it if I could :) [20:55:00] I suspect Antoine or James F would be the best placed [20:55:06] (03CR) 10Daimona Eaytoy: [C: 03+1] Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (https://phabricator.wikimedia.org/T320332) (owner: 10Legoktm) [20:55:31] don't you need to edit the jjb job definitions to reference the new tag? [20:56:03] yeah [20:56:15] Don't we usually do that in two steps? Or build pre-merge [20:56:26] not that either is really wrong [20:56:56] https://www.mediawiki.org/wiki/Continuous_integration/Docker#Deploy_images [20:56:59] What could possibly go wrong [20:57:29] (03CR) 10Reedy: [C: 03+2] Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (https://phabricator.wikimedia.org/T320332) (owner: 10Legoktm) [20:59:14] (03Merged) 10jenkins-bot: Update phpunit-suite-edit for PHPUnit 9 [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (https://phabricator.wikimedia.org/T320332) (owner: 10Legoktm) [21:00:04] !log Updating docker-pkg files on contint primary for https://gerrit.wikimedia.org/r/840337 [21:00:06] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [21:00:48] Just watch out, there's no "I broke CI" t-shirt AFAIK [21:00:55] I just tabbed in and it was all red [21:01:06] but seems red is ok [21:01:23] == Step 0: scanning /etc/zuul/wikimedia/dockerfiles == [21:01:23] Will build the following images: [21:01:23] * docker-registry.discovery.wmnet/releng/quibble-buster-php74-coverage:1.4.6-3 [21:01:23] == Step 1: building images == [21:01:23] * Built image docker-registry.discovery.wmnet/releng/quibble-buster-php74-coverage:1.4.6-3 [21:01:24] == Step 2: publishing == [21:01:26] Successfully published image docker-registry.discovery.wmnet/releng/quibble-buster-php74-coverage:1.4.6-3 [21:01:28] == Build done! == [21:01:30] cat: /tmp/docker-pkg-build.log: No such file or directory [21:03:10] (03PS1) 10Reedy: Update mwcore-codehealth to quibble-buster-php74-coverage:1.4.6-3 [integration/config] - 10https://gerrit.wikimedia.org/r/840341 (https://phabricator.wikimedia.org/T320332) [21:03:48] (03CR) 10Reedy: [C: 03+2] Update mwcore-codehealth to quibble-buster-php74-coverage:1.4.6-3 [integration/config] - 10https://gerrit.wikimedia.org/r/840341 (https://phabricator.wikimedia.org/T320332) (owner: 10Reedy) [21:04:34] Daimona: There is always a first time! [21:04:51] tshirt? I thought phab badges were all the rage now [21:05:29] Oh right [21:05:33] (03Merged) 10jenkins-bot: Update mwcore-codehealth to quibble-buster-php74-coverage:1.4.6-3 [integration/config] - 10https://gerrit.wikimedia.org/r/840341 (https://phabricator.wikimedia.org/T320332) (owner: 10Reedy) [21:07:44] I need to fix my JJB... [21:10:47] (03CR) 10Daimona Eaytoy: "This is correct, but it won't fix the coverage. There are some more to update:" [integration/config] - 10https://gerrit.wikimedia.org/r/840341 (https://phabricator.wikimedia.org/T320332) (owner: 10Reedy) [21:11:13] Also, I think we may have missed the "s" before 3 in the version number? [21:11:25] The s has a special meaning [21:11:33] I can't remember what it is [21:11:55] s is for like, when you're bumping it because an image below it was changed [21:12:47] Ahhhh I see, "s" as in "*s*pecial meaning". Makes sense now [21:13:23] (03PS1) 10Reedy: jjb: A few more quibble-buster-php74-coverage image updates [integration/config] - 10https://gerrit.wikimedia.org/r/840342 [21:14:02] (03CR) 10Reedy: [C: 03+2] jjb: A few more quibble-buster-php74-coverage image updates [integration/config] - 10https://gerrit.wikimedia.org/r/840342 (owner: 10Reedy) [21:14:12] I didn't realize we also used "-num" without the s [21:15:55] (03Merged) 10jenkins-bot: jjb: A few more quibble-buster-php74-coverage image updates [integration/config] - 10https://gerrit.wikimedia.org/r/840342 (owner: 10Reedy) [21:18:09] Any more patches to scribunto to merge? :P [21:19:27] I can try and update phan [21:19:36] Reedy: https://gerrit.wikimedia.org/r/c/mediawiki/extensions/Scribunto/+/839039 and https://gerrit.wikimedia.org/r/c/mediawiki/extensions/Scribunto/+/839035 are trivial [21:19:46] I just put a patch up just changing the version in omposer.json to see what breaks [21:21:31] New phan passed immediately, so there's that too [21:22:06] amazing. ship it [21:22:28] And I did it before seeing your message :D [21:27:02] Might be worth kicking off a CI run of all WMF deployed/MW bundled stuff to see what is currently "broken" too [21:28:38] There's definitely something broken by the PHPUnit upgrade... My hope was that those issues could be surfaced as they were discovered [21:29:16] Also, "someone" added a "sort by modified time" option to https://doc.wikimedia.org/cover-extensions/, so it's easier to see which extensions have broken coverage [21:29:25] "someone" did indeed :P [21:29:41] I was more meaning for other phpunit related fails, rather than broken coverage [21:30:02] Yep, that's what I thought could be fixed gradually [21:30:30] Because I think triggering CI for all of them could take a while, and the failures should be generally easy-ish to fix [21:30:49] Antoine quite often did it... [21:30:52] I guess libup will surface them though [21:30:58] OTOH, the coverage job could be run manually for the extensions with broken coverage, because there's only 6 of them [21:31:16] yeah I was about to say, LibUp should start on the phan ones in an hour or two once the current run finishes [21:31:26] oh, that son? [21:31:27] *son [21:31:30] FFS [21:31:31] soon [21:31:37] If that's happening, NBD then [21:31:42] it's running really fast [21:31:49] Yeah, and I'm eagerly waiting for it :P [21:31:50] win [21:32:01] I don't quite understand the failures in some repos though [21:32:02] faster than these browser tests [21:32:05] I think node16/npm8 made a big difference [21:32:26] Daimona: example? [21:33:07] https://libraryupgrader2.wmcloud.org/r/mediawiki/extensions/CampaignEvents?branch=main [21:33:42] ah yeah, I was debugging that the other day but didn't finish. I think `npm audit` no longer runs `npm install`, it's weird [21:34:04] I've seen that in a few places [21:34:22] Yes, I figured it may have something to do with npm audit because it seems to only happen in repos with vulnerable grunt [21:36:50] https://integration.wikimedia.org/ci/job/mwext-phpunit-coverage-docker-publish/66049/ [21:36:51] survey says... [21:37:58] and `npm audit` seems to be complaining about --only=dev, not sure if related [21:39:05] https://doc.wikimedia.org/cover-extensions/Scribunto/ [21:39:10] yay [21:39:12] * Reedy hugs legoktm [21:40:14] And in scribunto.. 1072/2893 is changed to... 1189 / 2523 [21:41:00] 37.05% of lines upto 47.13% of lines [21:41:03] winning [21:41:21] 10Continuous-Integration-Config, 10MediaWiki-Core-Tests, 10Test-Coverage: Extension coverage reports are broken with PHPUnit 9.5; they seem to report MW core not the extension - https://phabricator.wikimedia.org/T320332 (10Reedy) 05Open→03Resolved [21:42:39] Noice! [21:43:07] wheee [21:43:26] taavi: yeah, they stupidly got rid of --only=dev, but still support --only=prod??? [21:43:59] Daimona: I should be able to run rebuild on those extensions... [21:44:29] For coverage? Sure. [21:45:03] zuul-test-repo postmerge ext:{name} I think [21:45:33] I just did it via the jerkins ui [21:45:40] was only ~6 [21:45:42] so easy enough [21:45:53] https://integration.wikimedia.org/ci/job/mwext-phpunit-coverage-docker-publish/ [21:46:37] https://doc.wikimedia.org/cover-extensions/?sort=mtime [21:46:42] that's betterer :) [21:46:47] On a semi-related note, I would really like to work on https://phabricator.wikimedia.org/T270751 at some point. [21:46:47] (ie they're not 0%) [21:47:01] It's pretty stupid that we're still wasting time removing unused suppression manually [21:47:15] heh [21:47:18] Would be nice [21:47:23] 95% must be trivially removeable [21:47:33] 100% of unused suppressions should be [21:47:45] Occasionally people do weird stuff [21:47:47] We've seen it before :P [21:48:02] For other issues it'd be more complex, but unused suppression are usually a large % of new issues [21:48:08] Ah sure [21:48:18] "most" is better than "none" :) [21:48:44] The problem is, phan's autofixer is not as mature as that of phpcs IIRC. [21:49:20] Probably because static analysis issues are generally harder to autofix than whitespace and the like, and so there wasn't a big need for an autofixer [21:49:32] Can we pinch code? [21:50:58] I think phan has a few autofixers that we could copy from [21:51:31] https://github.com/phan/phan/search?q=AutomaticFixCapability [21:54:01] 10Gerrit, 10GitLab (Project Migration), 10Projects-Cleanup: Drop Gerrit repositories related to Wikimedia CH (they migrated to GitLab) - https://phabricator.wikimedia.org/T319534 (10valerio.bozzolan) Thank you for your help, Trust me, you can nuke. Anyway I've added the migration notes in the READMEs. Feel... [21:54:24] https://github.com/phan/phan/blob/6828584666bbc3404b679191029fb0bec5fe4440/.phan/plugins/WhitespacePlugin/fixers.php [21:54:26] regexy [21:54:29] Bah, let me see how easy it'd be [21:54:37] https://github.com/phan/phan/blob/f36b6b9a2f4143a25f35ce94d712ceb0527e9d90/.phan/plugins/PHPDocRedundantPlugin/Fixers.php [21:54:45] :D [21:54:48] ^This was suggested as an example because it targets docblocks [21:55:26] One thing is that there are many ways to suppress an issue -- inline (with -previous, -current, -next, -next-next), docblock, file-level... [21:55:53] But something's better than nothing, so... [21:56:32] 10Gerrit, 10GitLab (Project Migration), 10Projects-Cleanup: Drop (or just set as read-only) Gerrit repositories related to Wikimedia CH (they migrated to GitLab) - https://phabricator.wikimedia.org/T319534 (10valerio.bozzolan) p:05Triage→03Low [21:58:04] ok, interesting behavior change [21:58:31] because we pin grunt "1.5.2", npm audit fix will no longer bump it to 1.5.3 because that doesn't match our version constraint unless --force is used [21:59:46] 10Diffusion: Update Wikimedia CH Diffusion repositories - https://phabricator.wikimedia.org/T319529 (10valerio.bozzolan) [22:03:35] https://github.com/phan/phan/issues/4304 [22:06:13] #patch-welcme [22:06:15] #patch-welcome [22:22:56] I'm on it... And I don't know what I'm doing :P [22:23:49] heh [22:28:21] :D [22:34:16] Yeah, it's normal [22:34:40] It seems possible to fix the suppressions though, because phan stores the line which caused the suppression [22:36:00] that's helpful [22:36:14] https://gerrit.wikimedia.org/r/c/mediawiki/extensions/WikibaseLexeme/+/840348 [22:36:19] ^ phpunit 9 warning blocking other stuffs [22:38:54] +2'd [22:39:11] cheers [22:39:34] Yup, I had a patch up for that one already :) [22:39:40] ohhh [22:39:41] lol [22:40:06] BTW, feel free to add me as reviewer if you find more [22:43:02] oh ffs [22:43:16] reedy@deploy1002:/srv/mediawiki-staging/php-1.40.0-wmf.4/extensions$ grep -R assertRegExp | wc -l [22:43:16] 122 [22:43:20] There's a few more we need to fix :) [22:45:05] https://phabricator.wikimedia.org/T320334 [22:45:23] Most of them are in Wikibase [22:45:46] Daimona: Any more patches outstanding before I make them? :P [22:46:10] Wikibase's already fixed [22:46:23] That's the biggest one [22:46:31] Also, if you want a canned commit message: https://phabricator.wikimedia.org/P35380 [22:47:00] And no, I only had one for WBLexeme [23:04:03] https://doc.wikimedia.org/cover-extensions/OAuth/src/Lib/index.html [23:04:08] maybe the coverage isn't as bad as it looks :P [23:37:23] Daimona: What should I do with "src/IPSet.php:209 PhanSyntaxError Fallback parser diagnostic error: 'Name' expected. (at column 17)" [23:37:32] WTF?! [23:37:38] lol [23:37:47] Link to patch? [23:37:47] * Reedy looks what CI says [23:37:50] Locally [23:38:16] https://gerrit.wikimedia.org/r/840355 [23:38:17] * Reedy waits [23:38:20] Ahhhh, there's a method called match() [23:38:27] A classic [23:38:35] Reserved keyword in PHP 8 [23:38:46] Oh ffs [23:38:52] How have we not fallen over that since? [23:39:48] And apparently no task for it [23:40:12] Oh but wait [23:40:45] Even if it's a reserved keyword, PHP allows it as a method name, but not as class name etc. (https://phabricator.wikimedia.org/T267689#8089651) [23:41:06] So it's parsed just fine, but there's likely a bug in the fallback parser than phan uses [23:41:15] Nonetheless, it would be a good idea to rename it [23:41:45] matchIPAddress or something probably works [23:43:32] Indeed [23:59:10] Daimona: libup is on with phan upgrades :) [23:59:20] Yayyyy