[00:00:24] will make my ipset and iputils patches redundant xD [00:05:05] I just kicked off another run [00:05:23] YAYYYYYYYYY [00:05:42] Can't wait to fix all the stuff :D [00:26:32] (03CR) 10CI reject: [V: 04-1] build: Updating mediawiki/mediawiki-phan-config to 0.12.0 [integration/docroot] - 10https://gerrit.wikimedia.org/r/840372 (owner: 10Libraryupgrader) [00:26:53] lol [00:26:59] * Reedy looks [00:27:12] ah yes [00:27:59] (03Abandoned) 10Reedy: build: Updating mediawiki/mediawiki-phan-config to 0.12.0 [integration/docroot] - 10https://gerrit.wikimedia.org/r/840372 (owner: 10Libraryupgrader) [00:40:16] extension patches coming now [00:40:37] https://doc.wikimedia.org/cover/mediawiki-libs-XMPReader/index.html [00:40:42] wheee 100% coverage of Validate.php [00:41:42] And no failures to fix so far [00:44:49] (03CR) 10Jforrester: "We intentionally standardised on the nodesource node, and the same Wikimedia-wide version of npm; are we definitely wanting to reverse thi" [integration/config] - 10https://gerrit.wikimedia.org/r/840336 (owner: 10Legoktm) [00:45:14] (03CR) 10Jforrester: "Argh. You can't version bump quibble images like this!" [integration/config] - 10https://gerrit.wikimedia.org/r/840337 (https://phabricator.wikimedia.org/T320332) (owner: 10Legoktm) [00:45:26] sorry :< [00:45:41] * James_F tsks Reedy for merging. [00:45:56] CI didn't mind :P [00:46:02] legoktm: Quibble images (and some others in CI)'s main version string is "version of the software", not the image. [00:46:10] So this means we can never ever release quibble 1.4.7. [00:46:23] er, no? [00:46:25] As the docker registry is now poisoned. [00:46:30] I changed it in a later patchset, it's 1.4.6-3 [00:46:43] Oh. Where? [00:46:47] The merged patch is 1.4.7. [00:46:55] ? [00:47:04] https://gerrit.wikimedia.org/r/c/integration/config/+/840337/5/dockerfiles/quibble-buster-php74-coverage/changelog [00:47:15] >quibble-buster-php74-coverage (1.4.6-3) wikimedia; urgency=medium [00:48:01] https://github.com/wikimedia/integration-config/blob/master/dockerfiles/quibble-buster-php74-coverage/changelog#L1 agrees... [00:48:03] Did you not yet build them? [00:48:09] Yeah? [00:48:14] It's working fine in prod.. [00:48:22] Bleh, Varnishj caches everything, I shouldn't trust it. [00:48:27] heh [00:48:33] ?debug=true [00:48:52] Also gerrit does? [00:49:31] (03CR) 10Legoktm: Use our packages for node16, bringing in npm v8 (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/840336 (owner: 10Legoktm) [00:49:35] https://gerrit.wikimedia.org/r/plugins/gitiles/integration/config/+/84c14054ef28d194b91f6e5eb55a13494aadd4e5%5E%21/#F0 shows it as 1.4.6-3 but https://gerrit.wikimedia.org/r/c/integration/config/+/840337/5/dockerfiles/quibble-buster-php74-coverage/changelog is showing 1.4.7. [00:49:47] Oh, and now I refresh it's 1.4.6-3. [00:49:51] o_0 [00:51:11] (03CR) 10Jforrester: Use our packages for node16, bringing in npm v8 (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/840336 (owner: 10Legoktm) [00:51:18] Hence my worry. [00:51:34] I've had gerrit's "clever" client-side caching hit me before. [00:51:37] Sorry. :-( [00:51:59] It's all good [00:52:08] Especially if nothing is actually going to cause problems later [00:52:24] It's going to irritate me that it's 1.4.6-3 not 1.4.6-s3. [00:52:28] But I'll get over it. :-) [01:06:46] (03PS6) 10Jforrester: [WIP] dockerfiles: Provide PHP 8.2 CI test images [integration/config] - 10https://gerrit.wikimedia.org/r/818239 (https://phabricator.wikimedia.org/T314093) [01:12:02] Oldest claimed PHP support in LibUp's knowledge is PipeEscape which still claims to work in PHP 5.0.0: https://libraryupgrader2.wmcloud.org/r/mediawiki/extensions/PipeEscape [01:12:10] * James_F sighs./ [01:17:33] lol [01:17:53] seems legit [01:18:09] It claims to require MW 1.35+ [01:18:13] So patch inbound. [01:18:34] it absolutely does not work on 5.0.0 [01:19:04] * legoktm remembers the before times when we typed array( ... ) [01:19:48] Do not speak to me of the old PHP, lego. I was there when it was forged. [01:19:55] xD [01:22:42] legoktm: Thoughts about me making a LibUp patch to move the phan command into composer test? [01:23:37] I don't think we should do that in general just because `composer test` is supposed to be specific to the repo and self-contained, while phan depends on what version of mw core you have checked out [01:23:52] Hmm. [01:24:06] That is not unreasonable. [01:24:29] legoktm: Comment on T280990 to that point? [01:24:30] T280990: Run phan as part of composer test, rather than in bespoke CI jobs - https://phabricator.wikimedia.org/T280990 [01:24:36] will do [01:27:56] 10Continuous-Integration-Config, 10LibUp, 10phan, 10Composer: Run phan as part of composer test, rather than in bespoke CI jobs - https://phabricator.wikimedia.org/T280990 (10Legoktm) My other qualm about this is that `composer test` is supposed to be self-contained, just applying to what's in the repo, ex... [02:21:13] >which requires 18 arg(s) defined at [02:21:15] * Reedy blinks [02:22:20] That looks like a bug [02:22:29] its complaining about [02:22:30] parent::__construct( ...func_get_args() ); [07:41:11] I could use some help understanding the Phan SecurityCheck issues on https://gerrit.wikimedia.org/r/c/mediawiki/extensions/CentralAuth/+/840425 [08:43:04] 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) [08:45:41] 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) (I don't love this solution of putting a new README on Gerrit since this will caus...