[07:19:26] jnuche: good morning, for https://gerrit.wikimedia.org/r/c/mediawiki/tools/scap/+/836288/7/bin/install_local_version.sh [07:19:34] I think I will go back to use a trap on `ERR` [07:20:00] I am not very happy about using flag variables to trigger multiple function calls in `EXIT` trap [07:20:17] and I don't think I need to register an additional `EXIT` trap [07:42:55] (03CR) 10Hashar: install-world: ensure a clear venv (036 comments) [tools/scap] - 10https://gerrit.wikimedia.org/r/836288 (owner: 10Hashar) [07:43:00] (03PS8) 10Hashar: install-world: ensure a clear venv [tools/scap] - 10https://gerrit.wikimedia.org/r/836288 [07:43:07] yup trap on `ERR` should do the job [07:43:09] less code :D [08:27:44] 10Gerrit, 10Release-Engineering-Team: Move Gerrit Javascript plugins from gerrit-theme.js to standalone files in the deploy repository - https://phabricator.wikimedia.org/T319378 (10hashar) [08:39:35] 10Gerrit, 10Release-Engineering-Team: Move Gerrit Javascript plugins from gerrit-theme.js to standalone files in the deploy repository - https://phabricator.wikimedia.org/T319378 (10hashar) The `Gerrit.install()` interface is: ` lang=typescript,name=polygerrit-ui/app/api/gerrit.ts export declare interface Gerr... [09:34:37] (03CR) 10Jaime Nuche: "Tested locally. Thanks for the patch!" [tools/scap] - 10https://gerrit.wikimedia.org/r/836288 (owner: 10Hashar) [09:34:40] (03CR) 10Jaime Nuche: [C: 03+2] install-world: ensure a clear venv [tools/scap] - 10https://gerrit.wikimedia.org/r/836288 (owner: 10Hashar) [09:39:30] (03Merged) 10jenkins-bot: install-world: ensure a clear venv [tools/scap] - 10https://gerrit.wikimedia.org/r/836288 (owner: 10Hashar) [09:39:58] jnuche: thank you so much for your suggestions on that change above ^ [09:40:12] I am quite happy to finally have sorted it with simply an `ERR` trap :] [09:41:34] hashar: sure, thanks for the patch πŸ‘ [09:47:49] (03CR) 10Hashar: "rebasing for merge conflicts. I want the comment to be explicit about requiring both --no-deps and -r requirements.txt to be given." [tools/scap] - 10https://gerrit.wikimedia.org/r/836289 (https://phabricator.wikimedia.org/T316755) (owner: 10Hashar) [09:48:00] (03PS3) 10Hashar: install-world: install only declared requirements [tools/scap] - 10https://gerrit.wikimedia.org/r/836289 (https://phabricator.wikimedia.org/T316755) [09:52:09] (03PS1) 10Hashar: install-world: ensure all requirements are met [tools/scap] - 10https://gerrit.wikimedia.org/r/838756 (https://phabricator.wikimedia.org/T316755) [09:53:08] (03CR) 10Hashar: "That might be overkill since CI guarantee we define requirements exhaustively. But we never know ;)" [tools/scap] - 10https://gerrit.wikimedia.org/r/838756 (https://phabricator.wikimedia.org/T316755) (owner: 10Hashar) [09:53:28] and from there I think we more or less prevent installing extra packages [09:53:55] I will look at adding `--requires-hashes` next [10:19:57] (03PS1) 10Hashar: Require hashes when installing dependencies [tools/scap] - 10https://gerrit.wikimedia.org/r/838758 [10:22:47] (03CR) 10CI reject: [V: 04-1] Require hashes when installing dependencies [tools/scap] - 10https://gerrit.wikimedia.org/r/838758 (owner: 10Hashar) [11:34:54] (03CR) 10Jaime Nuche: [C: 03+2] "Tested in traindev by creating patches A and B, with B containing a `Depends-On` on A" [tools/scap] - 10https://gerrit.wikimedia.org/r/837983 (owner: 10Jeena Huneidi) [11:38:21] (03Merged) 10jenkins-bot: scap backport: check for already merged dependency [tools/scap] - 10https://gerrit.wikimedia.org/r/837983 (owner: 10Jeena Huneidi) [12:37:57] 10Continuous-Integration-Infrastructure, 10Jenkins, 10Release-Engineering-Team: Automate integration Jenkins deployment and config changes - https://phabricator.wikimedia.org/T319406 (10jnuche) [12:45:34] 10Continuous-Integration-Infrastructure, 10Jenkins, 10Release-Engineering-Team: Automate integration Jenkins deployment and config changes - https://phabricator.wikimedia.org/T319406 (10jnuche) [13:06:42] (03CR) 10Jaime Nuche: [C: 03+2] install-world: install only declared requirements [tools/scap] - 10https://gerrit.wikimedia.org/r/836289 (https://phabricator.wikimedia.org/T316755) (owner: 10Hashar) [13:10:44] (03Merged) 10jenkins-bot: install-world: install only declared requirements [tools/scap] - 10https://gerrit.wikimedia.org/r/836289 (https://phabricator.wikimedia.org/T316755) (owner: 10Hashar) [13:11:35] first part of T316755 done! [13:11:36] T316755: Strictly lock scap requirements dependencies - https://phabricator.wikimedia.org/T316755 [13:24:03] yay! [15:12:51] (03PS2) 10Hashar: install-world: ensure all requirements are met [tools/scap] - 10https://gerrit.wikimedia.org/r/838756 (https://phabricator.wikimedia.org/T316755) [15:55:17] (03PS1) 10Mglaser: Remove selenium tests for VisualEditorConnector [integration/config] - 10https://gerrit.wikimedia.org/r/838846 [16:07:18] (03CR) 10Hashar: [C: 03+2] Remove selenium tests for VisualEditorConnector [integration/config] - 10https://gerrit.wikimedia.org/r/838846 (owner: 10Mglaser) [16:09:04] (03Merged) 10jenkins-bot: Remove selenium tests for VisualEditorConnector [integration/config] - 10https://gerrit.wikimedia.org/r/838846 (owner: 10Mglaser) [16:09:38] (03CR) 10Hashar: [C: 03+2] "deployed" [integration/config] - 10https://gerrit.wikimedia.org/r/838846 (owner: 10Mglaser) [16:09:40] 10GitLab, 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ): GitLab integrations: the Gerrit & Jenkins integration catalog - https://phabricator.wikimedia.org/T319359 (10thcipriani) [16:12:47] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10User-brennen, 10User-dduvall: Write a GitLab "Migrating a Project" runbook / manual based on Blubber migration - https://phabricator.wikimedia.org/T307538 (10brennen) a:05brennenβ†’03None [16:18:57] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ): Create repo in GitLab for shared pipeline code - https://phabricator.wikimedia.org/T319322 (10thcipriani) [16:24:37] 10Gerrit, 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10Projects-Cleanup, 10Release Pipeline (Blubber): Archive the Blubber gerrit repo - https://phabricator.wikimedia.org/T317820 (10thcipriani) 05Openβ†’03Stalled Stalling until we get blubber fully migrated [16:24:39] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10Projects-Cleanup, 10User-dduvall: Create Blubber repo on GitLab, archive Gerrit repo - https://phabricator.wikimedia.org/T307533 (10thcipriani) [16:26:33] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10User-dduvall: Migrate Blubber project to GitLab - https://phabricator.wikimedia.org/T301168 (10thcipriani) [16:26:36] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10User-brennen, 10User-dduvall: Write a GitLab "Migrating a Project" runbook / manual based on Blubber migration - https://phabricator.wikimedia.org/T307538 (10thcipriani) 05In progressβ†’03Stalled Stalling until we ge... [16:27:23] (03CR) 10Jeena Huneidi: scap backport: check for already merged dependency (031 comment) [tools/scap] - 10https://gerrit.wikimedia.org/r/837983 (owner: 10Jeena Huneidi) [16:27:26] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ): Build and Run Blubber test variant on GitLab untrusted runners - https://phabricator.wikimedia.org/T307536 (10thcipriani) [16:27:58] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ): Build and Run Blubber test variant on GitLab untrusted runners - https://phabricator.wikimedia.org/T307536 (10thcipriani) [16:28:04] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10User-dduvall: Add a feature to blubber buildkit frontend to run entrypoint at the end of a build - https://phabricator.wikimedia.org/T319323 (10thcipriani) [16:28:17] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ): Build and Run Blubber test variant on GitLab untrusted runners - https://phabricator.wikimedia.org/T307536 (10thcipriani) [16:28:23] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10User-dduvall: Migrate Blubber project to GitLab - https://phabricator.wikimedia.org/T301168 (10thcipriani) [16:28:52] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ): Build and Run Blubber test variant on GitLab untrusted runners - https://phabricator.wikimedia.org/T307536 (10thcipriani) [16:28:58] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10User-dduvall: Add a feature to blubber buildkit frontend to run entrypoint at the end of a build - https://phabricator.wikimedia.org/T319323 (10thcipriani) [16:29:08] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10User-dduvall: Migrate Blubber project to GitLab - https://phabricator.wikimedia.org/T301168 (10thcipriani) [16:29:18] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ): Build and Run Blubber test variant on GitLab untrusted runners - https://phabricator.wikimedia.org/T307536 (10thcipriani) [16:29:28] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10User-dduvall: Add a feature to blubber buildkit frontend to run entrypoint at the end of a build - https://phabricator.wikimedia.org/T319323 (10thcipriani) [16:30:02] 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10Scap: Get scap CI working on GitLab - https://phabricator.wikimedia.org/T319335 (10dancy) [16:30:04] 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10Release Pipeline (Blubber): blubber's buildkit frontend should process .dockerignore - https://phabricator.wikimedia.org/T319337 (10dancy) [16:30:28] 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10Scap: Get scap CI working on GitLab - https://phabricator.wikimedia.org/T319335 (10dancy) [16:30:30] 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10Release Pipeline (Blubber): blubber's buildkit frontend should process .dockerignore - https://phabricator.wikimedia.org/T319337 (10dancy) [16:30:32] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10User-dduvall: Add a feature to blubber buildkit frontend to run entrypoint at the end of a build - https://phabricator.wikimedia.org/T319323 (10dancy) [16:30:53] 10GitLab (Project Migration), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10Projects-Cleanup, 10User-dduvall: Create Blubber repo on GitLab, archive Gerrit repo - https://phabricator.wikimedia.org/T307533 (10thcipriani) [16:31:12] 10Gerrit, 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10Projects-Cleanup, 10Release Pipeline (Blubber): Archive the Blubber gerrit repo - https://phabricator.wikimedia.org/T317820 (10thcipriani) [16:31:20] 10GitLab (Integrations), 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10Wikibugs, 10Patch-For-Review: Connect WikiBugs IRC bot to Wikimedia GitLab - https://phabricator.wikimedia.org/T288381 (10thcipriani) [16:31:26] 10GitLab, 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ): GitLab integrations: the Gerrit & Jenkins integration catalog - https://phabricator.wikimedia.org/T319359 (10thcipriani) [16:33:10] 10GitLab, 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ): GitLab integrations: the Gerrit & Jenkins integration catalog - https://phabricator.wikimedia.org/T319359 (10thcipriani) [16:33:13] 10GitLab (Integrations), 10GitLab-Test, 10Phabricator, 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), and 2 others: Experiment with GitLab-Phabricator integration - https://phabricator.wikimedia.org/T265617 (10thcipriani) [16:33:50] 10GitLab (Integrations), 10GitLab-Test, 10Phabricator, 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), and 2 others: Experiment with GitLab-Phabricator integration - https://phabricator.wikimedia.org/T265617 (10thcipriani) [16:38:25] (03CR) 10Hashar: "Looks like xdebug got enabled at some point in 2017 ( https://gerrit.wikimedia.org/r/c/integration/config/+/382939 )." [integration/config] - 10https://gerrit.wikimedia.org/r/814154 (https://phabricator.wikimedia.org/T269489) (owner: 10Daimona Eaytoy) [16:55:44] (03CR) 10Jeena Huneidi: scap backport: check for already merged dependency (031 comment) [tools/scap] - 10https://gerrit.wikimedia.org/r/837983 (owner: 10Jeena Huneidi) [16:56:43] (03CR) 10Daimona Eaytoy: Enable pcov instead of xdebug in composer-package-phpXX (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/814154 (https://phabricator.wikimedia.org/T269489) (owner: 10Daimona Eaytoy) [16:57:40] hashar: Thanks for looking at the pcov stuff. I admit I'm slightly confused now, and I'm not quite sure what could be using xdebug for code coverage. [16:58:14] Daimona: I have no idea honestly [16:58:25] I guess we can deploy it like that and see what ends up broken [16:58:36] Seems a good plan [16:58:45] iirc those composer-package-phpXX images are mainly used for mediawiki/lib repos [16:59:10] I'm also unsure what may be using this: https://gerrit.wikimedia.org/g/integration/config/+/70b746844e7c262a52d8501c7dce987fa31a95bf/jjb/misc.yaml#46 [17:00:02] the lib repos in zuul/layout.yaml (line 8927+) trigger `phpunit-coverage-php74-docker-publish` [17:00:13] Maybe we could update it at the same time? Load pcov and add '-d pcov.enabled=1' [17:00:15] which is defined by the link you gave [17:00:23] given the explicitly enable xdebug [17:00:31] therefor we don't have to explicitly enable xdebug in the image [17:00:36] which address the concern I had [17:01:38] Got it, so the new plan is to remove the 'phpenmod xdebug' line, do NOT enable pcov, and just update misc.yaml to use pcov? [17:02:03] hmm [17:02:08] yeah that sounds about right [17:02:11] in two different changes [17:02:17] one that remove the phpenmod xdebug [17:03:09] and a second one which indeed replace the job to use pcov [17:03:11] Sure, I'll do that. If something breaks, we can always re-configure it to use pcov [17:03:22] sorry I am resaying what you said above bah [17:03:35] yeah [17:04:17] No, it's a good thing, it means we're on the same page ahah [17:04:18] I haven't touched those images/jobs in age :-( [17:05:10] (03CR) 10Ahmon Dancy: scap backport: check for already merged dependency (031 comment) [tools/scap] - 10https://gerrit.wikimedia.org/r/837983 (owner: 10Jeena Huneidi) [17:10:45] (03CR) 10Ahmon Dancy: "When you're done with this please ensure that something somewhere in this repo says how to generate the requirements.txt with hashes in th" [tools/scap] - 10https://gerrit.wikimedia.org/r/838758 (owner: 10Hashar) [17:11:23] dancy: will do for sure thanks :-] [17:11:48] some tooling to mentioned on the task to assist with generating the hashes, I guess I will look at that [17:12:25] 10Continuous-Integration-Config, 10Patch-For-Review: composer-package-phpXX-docker jobs run with xdebug enabled - https://phabricator.wikimedia.org/T269489 (10Daimona) Update: after talking with @hashar, the new plan is to simply stop enabling xdebug by default, without enabling pcov. The reason is that it's u... [17:15:31] (03PS3) 10Daimona Eaytoy: Stop enabling xdebug by default in composer-package-phpXX [integration/config] - 10https://gerrit.wikimedia.org/r/814154 (https://phabricator.wikimedia.org/T269489) [17:17:19] Daimona: there are some version bumps which are not increasing the version :) [17:17:45] dockerfiles/composer-package-php73/changelog goes from 0.4.0-s3 to 0.4.0, should be 0.5.0 for example [17:17:55] (03PS1) 10Daimona Eaytoy: jjb: Use pcov instead of xdebug in phpunit-coverage-{php}-docker-publish [integration/config] - 10https://gerrit.wikimedia.org/r/838860 (https://phabricator.wikimedia.org/T269489) [17:22:39] (03PS1) 10Daimona Eaytoy: jjb: Use new images without xdebug enabled [integration/config] - 10https://gerrit.wikimedia.org/r/838861 (https://phabricator.wikimedia.org/T269489) [17:23:03] (03PS4) 10Daimona Eaytoy: dockerfiles: Stop enabling xdebug by default in composer-package-phpXX [integration/config] - 10https://gerrit.wikimedia.org/r/814154 (https://phabricator.wikimedia.org/T269489) [17:23:53] Lol, what the heck did I do? I can't count apparently :D [17:24:17] Daimona: Counting is hard. [17:24:34] (03PS5) 10Daimona Eaytoy: dockerfiles: Stop enabling xdebug by default in composer-package-phpXX [integration/config] - 10https://gerrit.wikimedia.org/r/814154 (https://phabricator.wikimedia.org/T269489) [17:24:52] ;D [17:25:23] Yeeaaaaaah this brings back memories from the combinatorics course :O [17:25:41] (03PS2) 10Daimona Eaytoy: jjb: Use pcov instead of xdebug in phpunit-coverage-{php}-docker-publish [integration/config] - 10https://gerrit.wikimedia.org/r/838860 (https://phabricator.wikimedia.org/T269489) [17:26:31] (03CR) 10Hashar: dockerfiles: Stop enabling xdebug by default in composer-package-phpXX (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/814154 (https://phabricator.wikimedia.org/T269489) (owner: 10Daimona Eaytoy) [17:26:41] (03PS2) 10Daimona Eaytoy: jjb: Use new images without xdebug enabled [integration/config] - 10https://gerrit.wikimedia.org/r/838861 (https://phabricator.wikimedia.org/T269489) [17:26:45] Daimona: well done, going to build those images [17:27:01] (03CR) 10Hashar: [C: 03+2] dockerfiles: Stop enabling xdebug by default in composer-package-phpXX [integration/config] - 10https://gerrit.wikimedia.org/r/814154 (https://phabricator.wikimedia.org/T269489) (owner: 10Daimona Eaytoy) [17:27:08] Noice, thank you [17:27:25] and thanks for finding https://gerrit.wikimedia.org/g/integration/config/+/70b746844e7c262a52d8501c7dce987fa31a95bf/jjb/misc.yaml#46 :) [17:28:03] Yup, for some reason I didn't see it when we made the big switch to pcov a couple years ago [17:28:46] (03Merged) 10jenkins-bot: dockerfiles: Stop enabling xdebug by default in composer-package-phpXX [integration/config] - 10https://gerrit.wikimedia.org/r/814154 (https://phabricator.wikimedia.org/T269489) (owner: 10Daimona Eaytoy) [17:29:43] !log Building docker images for https://gerrit.wikimedia.org/r/814154 [17:29:44] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [17:30:54] Once all the changes are deployed, I will verify that the "normal" phan job no longer has xdebug enabled, and that composer-package jobs also don't have it. And let me see if I can find something using the coverage-publish job, just to confirm that it still works. [17:33:09] ^tools/codesniffer will do [17:35:32] (03CR) 10Hashar: [C: 03+2] "Successfully published image docker-registry.discovery.wmnet/releng/mediawiki-phan-php73:0.3.1-s4" [integration/config] - 10https://gerrit.wikimedia.org/r/814154 (https://phabricator.wikimedia.org/T269489) (owner: 10Daimona Eaytoy) [17:36:23] 10Release-Engineering-Team, 10docker-pkg, 10serviceops: docker-pkg / docker downloads all versions of parent image upon building - https://phabricator.wikimedia.org/T310458 (10hashar) I had to build some new images for https://gerrit.wikimedia.org/r/c/integration/config/+/814154/ which I have done on `contin... [17:37:59] Daimona: you can check whether the jenkins jobs are using the latest docker images with `./utils/docker-updates` [17:38:30] jjb/parsoidsvc.yaml:58:docker-registry.wikimedia.org/releng/composer-package-php74:0.3.0-s11 > 0.4.0 [17:38:30] jjb/php.yaml:189:docker-registry.wikimedia.org/releng/composer-package-php74:0.3.0-s11 > 0.4.0 [17:38:32] two missed :) [17:38:50] Ohhhhh thank you, I had never seen that [17:38:51] bah I should comment on the change really [17:39:46] The output looks good for me though [17:39:51] hmm [17:39:53] what did I screw up [17:40:28] Maybe you didn't checkout the latest PS? [17:41:05] oh cause there is another change! [17:41:39] sorry [17:42:09] Hah, yes [17:42:39] (Also, apparently I got the versions right for the jjb change, so maybe I can count sometimes?) [17:42:46] (03CR) 10Hashar: [C: 03+2] "INFO:jenkins_jobs.builder:Number of jobs generated: 3" [integration/config] - 10https://gerrit.wikimedia.org/r/838860 (https://phabricator.wikimedia.org/T269489) (owner: 10Daimona Eaytoy) [17:44:38] (03Merged) 10jenkins-bot: jjb: Use pcov instead of xdebug in phpunit-coverage-{php}-docker-publish [integration/config] - 10https://gerrit.wikimedia.org/r/838860 (https://phabricator.wikimedia.org/T269489) (owner: 10Daimona Eaytoy) [17:45:22] Hmm. [17:45:23] https://integration.wikimedia.org/ci/job/phpunit-coverage-php74-docker-publish/ [17:45:31] Doesn't look very healthy before the change. [17:45:45] Hah :) [17:45:50] (03CR) 10Hashar: [C: 03+2] "jobs updated using:" [integration/config] - 10https://gerrit.wikimedia.org/r/838861 (https://phabricator.wikimedia.org/T269489) (owner: 10Daimona Eaytoy) [17:45:52] I told you we were not configuring xdebug... :D [17:46:04] ;D [17:46:10] All the syncs failing with `00:00:01.398 rsync: [sender] change_dir "/srv/jenkins/workspace/phpunit-coverage-php74-docker-publish/log/coverage" failed: No such file or directory (2)` [17:46:12] all jobs updated [17:46:21] We could try re-triggering them and see what happens. [17:46:35] But I imagine they weren't published for a long while. [17:46:45] Yeah [17:47:18] So... Could any of you please manually trigger mwext-php74-phan-docker for any repo, and phpunit-coverage-php74-docker-publish for tools/codesniffer? [17:47:54] I will do it for codesniffer [17:48:07] Done. [17:48:11] oh [17:48:17] Meanwhile I'm checking composer-package for taint-check, which used to fail hard due to xdebug [17:48:32] (03Merged) 10jenkins-bot: jjb: Use new images without xdebug enabled [integration/config] - 10https://gerrit.wikimedia.org/r/838861 (https://phabricator.wikimedia.org/T269489) (owner: 10Daimona Eaytoy) [17:48:35] https://integration.wikimedia.org/ci/job/phpunit-coverage-php74-docker-publish/22/ [17:48:42] Not for mwext though. [17:49:16] https://integration.wikimedia.org/ci/job/mwext-php74-phan-docker/1443/console is currently running. [17:49:21] Oh, pcov not available it looks like [17:49:34] :-( [17:49:36] Well, still broken as before I guess. [17:49:37] 00:00:17.231 Error: No code coverage driver is available [17:49:46] But now it's differently broken. [17:49:55] docker-registry.wikimedia.org/releng/composer-php74:0.2.0-s9 -d zend_extension=pcov.so -d pcov.enable=1 [17:49:57] I can try and fix it, but before I want to see the other jobs [17:50:07] so yeah hmm mystery [17:50:11] Yes, of course, it's important to always find new and fun ways to break things. [17:50:20] :-) [17:50:25] well that is CI since 2011 :D [17:50:45] Something's working as expected at least https://gerrit.wikimedia.org/r/c/mediawiki/tools/phan/SecurityCheckPlugin/+/793764 [17:51:41] $ php -d zend_extension=pcov.so -m|grep pcov [17:51:41] /usr/lib/php/20190902/pcov.so doesn't appear to be a valid Zend extension [17:51:49] when running the image composer-php74:0.2.0-s9 [17:51:52] And the normal phan job is also no longer restarting due to xdebug, so actually 2 things working [17:52:01] Yay. [17:52:21] it has to be loaded with -d extension=pcov.so [17:52:38] $ php -d extension=pcov.so -m|grep pcov [17:52:38] Ahhhh yes, it's not a zend extension [17:52:38] pcov [17:52:49] Did the same mistake last time [17:52:57] Fool me once, shame on thee... [17:53:04] :] [17:53:16] I gotta shutdown the computer due to some power supply issue [17:53:34] 10Release-Engineering-Team (GitLab II: Wrath of Kahn πŸ‘Ύ), 10Release Pipeline (Blubber): blubber's buildkit frontend should process .dockerignore - https://phabricator.wikimedia.org/T319337 (10jnuche) a:03jnuche [17:53:35] I guess James can update the jobs [17:53:43] Daimona: Our official dev docker images should probably get pcov too. [17:54:06] https://gitlab.wikimedia.org/repos/releng/dev-images [17:54:43] (03PS1) 10Daimona Eaytoy: jjb: pcov is not a zend extension [integration/config] - 10https://gerrit.wikimedia.org/r/838866 [17:54:49] Otherwise we'll break running tests locally. [17:55:21] I'm all for it, but... I would have to set up the gitlab thingies first... [17:55:28] It's fun! [17:55:37] And definitely not hard or anything. [17:55:39] * James_F coughs. [17:56:18] Suuuuure ^_^ [17:56:29] (03CR) 10CI reject: [V: 04-1] jjb: pcov is not a zend extension [integration/config] - 10https://gerrit.wikimedia.org/r/838866 (owner: 10Daimona Eaytoy) [17:57:09] (03PS2) 10Daimona Eaytoy: jjb: pcov is not a zend extension [integration/config] - 10https://gerrit.wikimedia.org/r/838866 [17:57:13] (03PS3) 10Jforrester: jjb: pcov is not a zend extension [integration/config] - 10https://gerrit.wikimedia.org/r/838866 (owner: 10Daimona Eaytoy) [17:58:13] (03CR) 10Jforrester: [C: 03+2] "Deployed." [integration/config] - 10https://gerrit.wikimedia.org/r/838866 (owner: 10Daimona Eaytoy) [17:58:59] (03CR) 10CI reject: [V: 04-1] jjb: pcov is not a zend extension [integration/config] - 10https://gerrit.wikimedia.org/r/838866 (owner: 10Daimona Eaytoy) [17:59:05] Oh dear. [17:59:55] Ah, the classic edit conflict. [18:00:18] (03PS4) 10Jforrester: jjb: Follow-up 3c44647: pcov is not a zend extension [integration/config] - 10https://gerrit.wikimedia.org/r/838866 (owner: 10Daimona Eaytoy) [18:00:24] (03CR) 10Jforrester: "…" [integration/config] - 10https://gerrit.wikimedia.org/r/838866 (owner: 10Daimona Eaytoy) [18:00:28] Gotta love ECs. [18:00:43] At least MW handles them pretty well [18:00:49] * hashar hints at includes/EditPage.php [18:00:53] Yeah yeah. [18:01:04] hashar: Feel free to make includes/EditJJB.php [18:01:10] ;D [18:01:18] I've got out of practice of applying C-2 to patches I hate. [18:01:22] You know what? Someone should write a MW extension that allows you to review code. Seems a killer idea, definitely something that has never been done before. [18:01:23] I'd love the opportunity. :-) [18:01:35] /kick Daimona no. [18:01:38] Oh wait. ;-) [18:02:00] Yup [18:02:45] (03Merged) 10jenkins-bot: jjb: Follow-up 3c44647: pcov is not a zend extension [integration/config] - 10https://gerrit.wikimedia.org/r/838866 (owner: 10Daimona Eaytoy) [18:02:58] OK, so. [18:03:06] Daimona: Did you re-trigger with the new job definition? [18:06:34] No, I have no idea how to trigger jobs manually. [18:07:03] tis complicated [18:07:13] sometime one can do it from the web ui directly in jenkins [18:07:29] some other times one has to manually enqueue the change in Zuul using the `zuul` CLI on the server [18:07:59] Or you can just `recheck` on a patch. [18:08:05] Simples. [18:08:09] ah yeah that is the easiest one :D [18:08:17] Though not for post-merge jobs. [18:08:47] Re-building for codesniffer: https://integration.wikimedia.org/ci/job/phpunit-coverage-php74-docker-publish/24/ [18:09:06] Good grief, it worked. [18:09:53] https://doc.wikimedia.org/cover/mediawiki-tools-codesniffer/ updated 2022-10-05. [18:10:03] I mean, the data isn't very good, but it's now there. [18:10:08] success! [18:11:04] hahah [18:11:24] But it failed for Minify [18:11:32] https://integration.wikimedia.org/ci/job/phpunit-coverage-php74-docker-publish/25/console [18:11:44] 00:00:11.737 PHPUnit\Framework\Exception: PHP Fatal error: Uncaught SebastianBergmann\CodeCoverage\RuntimeException: No code coverage driver available in /src/vendor/phpunit/php-code-coverage/src/CodeCoverage.php:907 [18:11:50] I assume that was pre-existing? [18:12:13] I'm looking at the output, there's something weird [18:12:22] This looks like the parser test problem... a lot of code is covered, but the @covers aren't there [18:12:27] (or can't be) [18:12:44] The xdebug version was running 56 tests, but the new one has 106 [18:13:27] Could be that the pcov presence triggers new tests that were hidden before? [18:13:36] ObjectFactory also worked: https://doc.wikimedia.org/cover/mediawiki-libs-ObjectFactory/ [18:14:06] And parsoid: https://doc.wikimedia.org/cover/mediawiki-services-parsoid/ [18:14:18] Not really, becaue there were 106 tests even when we tried to load pcov with zend_extension [18:14:28] So maybe xdebug was hiding some tests?! Is it even possible? [18:14:33] 10Continuous-Integration-Config, 10Patch-For-Review: composer-package-phpXX-docker jobs run with xdebug enabled - https://phabricator.wikimedia.org/T269489 (10Daimona) 05Openβ†’03Resolved a:03Daimona This was done as explained in T269489#8287922. Some effects of this change: - The MW phan jobs (e.g. mwext-... [18:15:11] No in-repo hits for 'pcov' or 'xdebug' in Minify so not explicitly enabled/disabled for it. [18:16:51] Running locally I get 842 tests on xdebug-enabled and on neither. [18:19:07] Might bumping PHPUnit help? https://gerrit.wikimedia.org/r/c/mediawiki/libs/Minify/+/838869 [18:19:51] 10Continuous-Integration-Config: Investigate potential issues with pcov in phpunit-coverage-php74-docker-publish - https://phabricator.wikimedia.org/T319458 (10Daimona) [18:20:52] Maybe? I guess it'd have to be done at some point anyway, so... [18:21:18] 10Release-Engineering-Team, 10docker-pkg, 10serviceops: docker-pkg / docker downloads all versions of parent image upon building - https://phabricator.wikimedia.org/T310458 (10Clement_Goubert) \o/ [18:21:26] FTR, https://doc.wikimedia.org/cover/mediawiki-tools-codesniffer/ is looking good, same coverage as in July (last successful run) [18:21:34] nice [18:22:25] Ah I think I know what happened with codesniffer [18:23:34] One of the builds was for an older version of the repo, I'm looking at the diff and several composer dependencies are different. So maybe it's just that more tests were added since the last build [18:23:47] No wait [18:24:11] It's just me being an idiot again, one of the builds is not for codesniffer :) [18:25:09] 10Continuous-Integration-Config: Investigate potential issues with pcov in phpunit-coverage-php74-docker-publish - https://phabricator.wikimedia.org/T319458 (10Daimona) [18:25:57] 10Continuous-Integration-Config: Investigate potential issues with pcov in phpunit-coverage-php74-docker-publish - https://phabricator.wikimedia.org/T319458 (10Daimona) First question answered: the first output above is not from codesniffer, but from php-session-serializer. So obviously they run different tests... [18:28:35] 10Continuous-Integration-Config: Investigate potential issues with pcov in phpunit-coverage-php74-docker-publish - https://phabricator.wikimedia.org/T319458 (10Daimona) [18:34:48] 10Continuous-Integration-Config: Investigate potential issues with pcov in phpunit-coverage-php74-docker-publish - https://phabricator.wikimedia.org/T319458 (10Daimona) Second question answered: codesniffer's phpunit.xml.dist had ` enforceTimeLimit="true" ` this was unused, and as noted in the output above, i... [18:39:35] I think the Minify one is caused by the test being run in a separate process https://gerrit.wikimedia.org/g/mediawiki/libs/Minify/+/77d3421b01c68ea50a834ff4096ac77f1607eb27/tests/phpunit/CliTest.php#57 [18:40:14] I'm not quite sure how that option works, maybe pcov is not enabled because the options to the PHP interpreter are not passed through? [18:43:07] Hmm. Could be. [18:43:43] Dates from https://gerrit.wikimedia.org/r/plugins/gitiles/mediawiki/libs/Minify/+/7559fc137d3adcd6f85abf01adf42d1be4d0b143%5E%21/#F2 [18:44:02] 10Continuous-Integration-Config: Investigate potential issues with pcov in phpunit-coverage-php74-docker-publish - https://phabricator.wikimedia.org/T319458 (10Daimona) Third and last question partly answered: testRunJsFile uses the `@runInSeparateProcess` to, well, run in a separate process. I suspect that the... [18:44:06] (03PS1) 10Jeena Huneidi: scap backport: thoroughly check cross-repo deps [tools/scap] - 10https://gerrit.wikimedia.org/r/838873 [18:44:37] (03PS2) 10Jeena Huneidi: scap backport: thoroughly check cross-repo deps [tools/scap] - 10https://gerrit.wikimedia.org/r/838873 [18:45:10] Just posted on phab. I think we should just refactor the test [18:46:26] Ack. [18:47:36] (03CR) 10CI reject: [V: 04-1] scap backport: thoroughly check cross-repo deps [tools/scap] - 10https://gerrit.wikimedia.org/r/838873 (owner: 10Jeena Huneidi) [18:48:21] Yeah, the upgrade to PHPUnit didn't fix anything, but it's not a bad fix to have made. [18:50:25] 10Continuous-Integration-Config, 10Performance-Team, 10Wikimedia-Minify: Re-work Minify's testRunJsFile test to not use forbidden @runInSeparateProcess which breaks CI publishing - https://phabricator.wikimedia.org/T319459 (10Jdforrester-WMF) [18:50:49] 10Continuous-Integration-Config, 10Performance-Team, 10Wikimedia-Minify: Re-work Minify's testRunJsFile test to not use forbidden @runInSeparateProcess which breaks CI publishing - https://phabricator.wikimedia.org/T319459 (10Jdforrester-WMF) [18:50:54] Daimona: Filed. :-) [18:51:02] woo, tech debt [18:51:05] Yup [18:51:39] https://doc.wikimedia.org/cover/mediawiki-libs-Minify/ hasn't been updated since 2021-03-09. [18:52:54] 10Continuous-Integration-Config: Investigate potential issues with pcov in phpunit-coverage-php74-docker-publish - https://phabricator.wikimedia.org/T319458 (10Daimona) 05Openβ†’03Resolved a:03Daimona >>! In T319458#8288231, @Daimona wrote: > Third and last question partly answered: testRunJsFile uses the `@... [18:52:57] I wonder how much of a PITA it would be to add a "last modified" time onto https://doc.wikimedia.org/cover/ [18:53:06] (for each thing listed) [18:53:13] Reedy: First step would be getting Doxygen to output it. [18:53:15] But yes. [18:53:28] It would definitely help [18:53:58] So maybe we'd stop being like "Oh, apparently the coverage hasn't been updated in 17 years for this repo" [18:53:59] * Reedy files a task [18:54:15] 10Continuous-Integration-Config, 10Performance-Team, 10Wikimedia-Minify: Re-work Minify's testRunJsFile test to not use forbidden @runInSeparateProcess which breaks CI publishing - https://phabricator.wikimedia.org/T319459 (10Jdforrester-WMF) This annotation was brought in with [[https://gerrit.wikimedia.org... [18:54:20] And someone could notice within a reasonable time, possibly. [18:55:47] 10Continuous-Integration-Infrastructure: Add "last updated" timestamp to test coverage pages - https://phabricator.wikimedia.org/T319461 (10Reedy) [18:56:02] 10Continuous-Integration-Infrastructure: Add "last updated" timestamp to test coverage index pages - https://phabricator.wikimedia.org/T319461 (10Reedy) [18:59:30] 10Continuous-Integration-Infrastructure, 10Doxygen, 10MediaWiki-Documentation: Upgrade Doxygen to 1.9.1 - https://phabricator.wikimedia.org/T291950 (10Jdforrester-WMF) 05Openβ†’03Resolved a:03hashar This was done in {rCICF8ac830ac} / {rCICF206d5ab3} by Antoine. [18:59:51] (03PS3) 10Jeena Huneidi: scap backport: thoroughly check cross-repo deps [tools/scap] - 10https://gerrit.wikimedia.org/r/838873 [19:00:26] Reedy: https://phabricator.wikimedia.org/T319462 [19:00:43] (For my end.) [19:01:34] 10Continuous-Integration-Infrastructure: Add "last updated" timestamp to test coverage index pages - https://phabricator.wikimedia.org/T319461 (10Reedy) [19:02:10] if we can just output a LAST_UPDATE file with a timestamp... [19:02:34] Ideally not a need that we have to fiddle with each output. [19:02:47] We already generate with qunit, mocha, jest, and doxygen. [19:02:57] I imagine others will get added, not removed. [19:03:01] Can we use disc timestamp? [19:04:29] I.. don't necessarily see why not [19:04:38] If we know what file on disk is *always* updated [19:04:51] like the first index.html or similar [19:04:54] I am off, I am overdue for dinner [19:05:08] if there is something that needs my attention, ping here or on phab and I will look at it tomorrow :] [19:05:19] * hashar waves [19:05:38] Cause the index pages are PHP already, right? [19:06:00] the coverage index pages, not the individual coverage files [19:06:16] * Reedy looks [19:06:37] https://github.com/wikimedia/integration-docroot/blob/master/org/wikimedia/doc/cover/index.php [19:07:33] So "just" a change in https://github.com/wikimedia/integration-docroot/blob/master/shared/CoveragePage.php [19:09:36] Is clover.xml the common thing? [19:10:28] So presumably we can stat it.... [19:15:30] (03PS1) 10Reedy: composer.json: Add ext-simplexml to require [integration/docroot] - 10https://gerrit.wikimedia.org/r/838878 [19:21:06] (03PS1) 10Reedy: CoveragePage: Minor cleanup [integration/docroot] - 10https://gerrit.wikimedia.org/r/838879 [19:21:08] (03PS1) 10Reedy: CoveragePage: Output mtime as last modified [integration/docroot] - 10https://gerrit.wikimedia.org/r/838880 (https://phabricator.wikimedia.org/T319461) [19:21:10] James_F: Simple as that? [19:21:21] (it almost certainly needs formatitng/css type tweaks) [19:22:35] and probably formatting of hte mtime [19:23:07] php > var_dump( stat( '.phpcs.xml' )['mtime'] ); [19:23:07] int(1664997175) [19:23:09] heh, yup [19:27:35] (03PS2) 10Reedy: CoveragePage: Output mtime as last modified [integration/docroot] - 10https://gerrit.wikimedia.org/r/838880 (https://phabricator.wikimedia.org/T319461) [19:43:12] Some CSS sounds nice. [19:43:18] Let's deploy what we've got? [19:43:28] Oh, wait. [19:43:41] (03PS1) 10Reedy: Cleanup some PHP 7.0 compat [integration/docroot] - 10https://gerrit.wikimedia.org/r/838885 [19:43:56] Yup, SimpleXML is installed. [19:43:58] (03CR) 10Jforrester: [C: 03+2] composer.json: Add ext-simplexml to require [integration/docroot] - 10https://gerrit.wikimedia.org/r/838878 (owner: 10Reedy) [19:44:01] heh [19:44:05] it's one of those sillynesses [19:44:34] (03CR) 10Reedy: Cleanup some PHP 7.0 compat (031 comment) [integration/docroot] - 10https://gerrit.wikimedia.org/r/838885 (owner: 10Reedy) [19:44:43] (03CR) 10Jforrester: CoveragePage: Minor cleanup (031 comment) [integration/docroot] - 10https://gerrit.wikimedia.org/r/838879 (owner: 10Reedy) [19:45:28] (03Merged) 10jenkins-bot: composer.json: Add ext-simplexml to require [integration/docroot] - 10https://gerrit.wikimedia.org/r/838878 (owner: 10Reedy) [19:45:55] (03CR) 10Jforrester: Cleanup some PHP 7.0 compat (031 comment) [integration/docroot] - 10https://gerrit.wikimedia.org/r/838885 (owner: 10Reedy) [19:46:05] (03PS3) 10Jforrester: CoveragePage: Output mtime as last modified [integration/docroot] - 10https://gerrit.wikimedia.org/r/838880 (https://phabricator.wikimedia.org/T319461) (owner: 10Reedy) [19:46:13] (03CR) 10Jforrester: [C: 03+2] CoveragePage: Output mtime as last modified [integration/docroot] - 10https://gerrit.wikimedia.org/r/838880 (https://phabricator.wikimedia.org/T319461) (owner: 10Reedy) [19:46:27] (03PS2) 10Jforrester: Cleanup some PHP 7.0 compat [integration/docroot] - 10https://gerrit.wikimedia.org/r/838885 (owner: 10Reedy) [19:46:31] (03CR) 10Jforrester: [C: 03+2] Cleanup some PHP 7.0 compat [integration/docroot] - 10https://gerrit.wikimedia.org/r/838885 (owner: 10Reedy) [19:47:19] (03Merged) 10jenkins-bot: CoveragePage: Output mtime as last modified [integration/docroot] - 10https://gerrit.wikimedia.org/r/838880 (https://phabricator.wikimedia.org/T319461) (owner: 10Reedy) [19:48:02] (03Merged) 10jenkins-bot: Cleanup some PHP 7.0 compat [integration/docroot] - 10https://gerrit.wikimedia.org/r/838885 (owner: 10Reedy) [19:48:54] (03PS2) 10Reedy: CoveragePage: Minor cleanup [integration/docroot] - 10https://gerrit.wikimedia.org/r/838879 [19:49:07] (03PS3) 10Reedy: CoveragePage: Minor cleanup [integration/docroot] - 10https://gerrit.wikimedia.org/r/838879 [19:49:44] Reedy: Oh, hah, I was fixing it for you locally. [19:49:57] (03PS4) 10Jforrester: CoveragePage: Don't set $color or $minWidth and then not use them [integration/docroot] - 10https://gerrit.wikimedia.org/r/838879 (owner: 10Reedy) [19:50:02] Now you have a nicer commit message. :-) [19:50:07] I had done it locally... Then you rebased my other patches :P [19:50:10] lol [19:50:19] does integration-docroot auto deploy? [19:50:28] Nope, manual. [19:50:33] Want to do the honours? [19:50:45] (03CR) 10Jforrester: [C: 03+2] CoveragePage: Don't set $color or $minWidth and then not use them [integration/docroot] - 10https://gerrit.wikimedia.org/r/838879 (owner: 10Reedy) [19:51:05] Reedy: https://wikitech.wikimedia.org/wiki/Doc.wikimedia.org#Runbook [19:51:47] (03Merged) 10jenkins-bot: CoveragePage: Don't set $color or $minWidth and then not use them [integration/docroot] - 10https://gerrit.wikimedia.org/r/838879 (owner: 10Reedy) [19:54:07] (03PS1) 10Reedy: CoveragePageTest: Remove compat for PHPUnit 4 and 6 [integration/docroot] - 10https://gerrit.wikimedia.org/r/838891 [19:54:11] * Reedy cringes [19:54:36] Wow. [19:54:46] WCPGW? [19:55:14] I guess it technically doesn't remove compat for PHPUnit 6... [19:56:28] We already require PHPUnit 8.5 [19:56:36] it's 8.0 in composer.json for that repo :) [19:56:49] I was more meaning the functions in 6 are still there in 8 [19:57:18] Yeah. [19:57:23] (03CR) 10Jforrester: [C: 03+2] CoveragePageTest: Remove compat for PHPUnit 4 and 6 [integration/docroot] - 10https://gerrit.wikimedia.org/r/838891 (owner: 10Reedy) [19:57:32] Oops, too late, I C+2'ed. [19:57:45] meh [19:57:54] :-) [19:58:06] (03Merged) 10jenkins-bot: CoveragePageTest: Remove compat for PHPUnit 4 and 6 [integration/docroot] - 10https://gerrit.wikimedia.org/r/838891 (owner: 10Reedy) [19:59:28] (03PS1) 10Reedy: Fix MediaWiki.Commenting.MissingCovers.MissingCovers [integration/docroot] - 10https://gerrit.wikimedia.org/r/838893 [20:00:04] (03CR) 10CI reject: [V: 04-1] Fix MediaWiki.Commenting.MissingCovers.MissingCovers [integration/docroot] - 10https://gerrit.wikimedia.org/r/838893 (owner: 10Reedy) [20:00:10] aww [20:01:07] (03PS2) 10Reedy: Fix MediaWiki.Commenting.MissingCovers.MissingCovers [integration/docroot] - 10https://gerrit.wikimedia.org/r/838893 [20:02:24] There we go [20:03:12] Gotta wait for some caching to expire I guess [20:08:39] (03PS1) 10Reedy: WIP: Fix more PHPCS excludes [integration/docroot] - 10https://gerrit.wikimedia.org/r/838895 [20:09:06] (03CR) 10CI reject: [V: 04-1] WIP: Fix more PHPCS excludes [integration/docroot] - 10https://gerrit.wikimedia.org/r/838895 (owner: 10Reedy) [20:13:55] (03Abandoned) 10Reedy: WIP: Fix more PHPCS excludes [integration/docroot] - 10https://gerrit.wikimedia.org/r/838895 (owner: 10Reedy) [20:17:01] (03CR) 10Jforrester: [C: 03+2] Fix MediaWiki.Commenting.MissingCovers.MissingCovers [integration/docroot] - 10https://gerrit.wikimedia.org/r/838893 (owner: 10Reedy) [20:17:11] (03PS1) 10Reedy: composer.json: Bump PHP require to >= 7.3.0 [integration/docroot] - 10https://gerrit.wikimedia.org/r/838897 [20:17:26] (03CR) 10Jforrester: [C: 03+2] composer.json: Bump PHP require to >= 7.3.0 [integration/docroot] - 10https://gerrit.wikimedia.org/r/838897 (owner: 10Reedy) [20:18:15] Reedy: A short hitlist for you https://codesearch.wmcloud.org/search/?q=%22php%22%3A%20%22.*7%5C.2%5C.&i=nope&files=%5Ecomposer.json&excludeFiles=&repos= [20:18:18] (03Merged) 10jenkins-bot: Fix MediaWiki.Commenting.MissingCovers.MissingCovers [integration/docroot] - 10https://gerrit.wikimedia.org/r/838893 (owner: 10Reedy) [20:18:46] (03Merged) 10jenkins-bot: composer.json: Bump PHP require to >= 7.3.0 [integration/docroot] - 10https://gerrit.wikimedia.org/r/838897 (owner: 10Reedy) [20:18:49] heh [20:18:59] Note that dropping environmental allowance is a breaking change, so not urgent. :-) [20:19:48] deploy-local failed: {} [20:19:51] well, that's helpful [20:20:23] Err. [20:20:25] In scap? [20:20:30] yeah [20:25:12] I guess we could send a manual purge to Varnish for https://doc.wikimedia.org/cover/ to see what we have wrought? [20:25:52] I did try the echo | purgeList.php [20:25:59] 10Continuous-Integration-Infrastructure, 10Patch-For-Review: Add "last updated" timestamp to test coverage index pages - https://phabricator.wikimedia.org/T319461 (10Jdforrester-WMF) [20:26:03] I don't know if it actually deployed it though [20:26:05] Oh huh, I guess Varnish ignores that for non-MW URLs? [20:26:16] Reedy: Use `lynx` locally. :-) [20:26:39] Oh, hmm, that's not installed in prod? Surprise. [20:27:53] Yeah, it doesn't look to have actually deployed it [20:27:56] James_F: Reedy: hi, need something like this? [20:27:57] [doc1002:~] $ curl https://doc.wikimedia.org/cover/ [20:28:01] -rw-r--r-- 1 deploy-ci-docroot deploy-ci-docroot 813 Aug 19 04:26 composer.json [20:28:18] mutante: Won't that go through Varnish though? [20:28:37] what do you need to know? [20:28:41] I can look at the local files [20:28:46] curl https://doc.wikimedia.org/cover/ | grep mtime doesn't have any hits. [20:29:12] 10Scap: scap deploy errors aren't helpful: deploy-local failed: {} - https://phabricator.wikimedia.org/T319472 (10Reedy) [20:29:13] mutante: Did scap deploy the change we made to output `…` on that page. [20:29:25] James_F: See my composer.json ls result above [20:29:29] I'm 99% sure it hasn't deployed any of it [20:29:35] Looking. [20:29:37] (looking at doc1002) [20:30:12] Yup, agreed, looks untouched. [20:30:46] the original deploy looked like it worked fine [20:30:48] 19:57:01 integration/docroot: fetch stage(s): 100% (in-flight: 0; ok: 4; fail: 0; left: 0) - [20:30:48] 19:57:02 integration/docroot: config_deploy stage(s): 100% (in-flight: 0; ok: 4; fail: 0; left: 0) - [20:30:48] 19:57:04 integration/docroot: promote stage(s): 100% (in-flight: 0; ok: 4; fail: 0; left: 0) - [20:30:52] looked, but not actioned? :P [20:31:20] hmm [20:31:22] the original deploy [20:31:23] commit 09eb565ec72b6080995a11a54f01515ff9e916ff (tag: scap/sync/2022-10-05/0001, tag: scap/sync/2022-08-19/0001) [20:31:34] Is it confused with the tags? [20:31:52] the latest thing that was touched under ./cover/ [20:31:56] 4.0K drwxrwxr-x 4 doc-uploader doc-uploader 4.0K Oct 5 19:06 mediawiki-services-function-schemata [20:32:45] Reedy: Locally it's on scap git tag scap/sync/2022-08-19/0001 [20:32:49] Yeah [20:33:08] Is https://wikitech.wikimedia.org/wiki/Doc.wikimedia.org#Runbook actually right? [20:33:12] Or have we found an actual scap bug? [20:33:17] It was when I wrote it. [20:33:36] Local git on doc1002 is detached HEAD as expected. [20:33:58] 10Scap: scap deploy errors aren't helpful: deploy-local failed: {} - https://phabricator.wikimedia.org/T319472 (10Reedy) [20:34:39] Scap has applied tag: scap/sync/2022-10-05/0006, tag: scap/sync/2022-10-05/0005, tag: scap/sync/2022-10-05/0004, tag: scap/sync/2022-10-05/0003, tag: scap/sync/2022-10-05/0002 to your code. [20:34:47] But not actually synced it? [20:35:04] Indeed [20:35:23] T319472 has what it's giving me when I try (and when I try to rollback too) [20:35:23] T319472: scap deploy errors aren't helpful: deploy-local failed: {} - https://phabricator.wikimedia.org/T319472 [20:35:28] what are the scap targets for doc [20:35:42] doc1002 and doc2001? [20:35:48] :* doc1002.eqiad.wmnet [20:35:48] :* contint1001.wikimedia.org [20:35:48] :* doc2001.codfw.wmnet [20:35:48] :* contint2001.wikimedia.org [20:35:49] Unless it's changed. [20:35:59] Oh yes, and the contint machines. [20:36:14] ci-docroot has 4 machines [20:36:24] yea, doc1002 should be right though [20:36:44] contint2001 hasn't updated either, it's the same state as doc1002. [20:36:57] Which is good, machines aren't inconsistent, except deployment.eqiad [20:37:36] I could just run `scap pull` on each machine? [20:38:32] it'd at least get round the issue for the moment [20:38:55] Well, that didn't work. [20:39:33] dancy: About? I think I've broke scap (of non MW) :D [20:39:47] Apparently scap pull is only for /srv/mediawiki [20:40:06] Which makes sense, TBF. [20:40:15] Hello. What's up? [20:40:51] 10Scap: scap deploy errors aren't helpful: deploy-local failed: {} - https://phabricator.wikimedia.org/T319472 (10Reedy) [20:40:57] dancy: "deploy-local failed: {}" https://phabricator.wikimedia.org/T319472 [20:40:59] dancy: ^ full deploy log on that bug [20:41:30] Ew that's very unpleasant. [20:41:35] * Reedy grins [20:41:41] Sorry to be the bearers of bad news. [20:41:50] i wonder if this is related to the recent changes to checks.yaml format [20:41:50] it's also not an UBN either [20:42:00] the change should have been backwards compatible [20:42:03] restart-php-fpm shouldn't run for non-MW scaps, I guess? [20:42:48] On the bright side, isn't it good we found this for a trivial scap user, not an emergency deploy real-world service? [20:44:06] heh [20:44:22] It's curious why the first deploy seemed to behave [20:45:05] (though, I think I didn't rebase after fetching... even thought I checked the diff) [20:46:01] Did you `git fetch` but not `git pull` and so it was syncing the as-is set? [20:47:37] First time, yeah [20:47:51] which explains... [20:47:51] commit 09eb565ec72b6080995a11a54f01515ff9e916ff (tag: scap/sync/2022-10-05/0001, tag: scap/sync/2022-08-19/0001) [20:47:52] Author: Stang [20:47:52] Date: Tue Jul 5 00:29:59 2022 +0200 [20:47:54] So no changed PHP files so the fpm-restart didn't trigger? [20:48:06] Ah, yes, that would. [20:49:41] Yup: [20:49:47] Started deploy [integration/docroot@09eb565]: T319461 and cleanup [20:49:47] T319461: Add "last updated" timestamp to test coverage index pages - https://phabricator.wikimedia.org/T319461 [20:49:51] Bah. [20:49:56] https://sal.toolforge.org/production?p=0&q=docroot&d= [20:50:01] At least that explains why the code didn't change :D [20:50:04] Git hash matches the old one. [20:50:06] Indeed. [20:54:03] James_F: Regarding the need for php-fpm-restart: T275468 [20:54:04] T275468: Apache on doc1001 does not see updated PHP files for hours/days after deployment - https://phabricator.wikimedia.org/T275468 [20:54:11] Oh huh. [20:54:32] Never mind then [20:55:43] doc1002 - Active: active (running) since Thu 2022-09-15 20:50:04 UTC; 2 weeks 6 days ago [20:55:53] systemctl status php7.3-fpm [20:56:01] running since 2 weeks [20:56:14] conflicts with the comment "We are now restarting php-fpm on the doc hosts. That clears out the opcache and should address the stall content after a deployment." ? [20:56:18] Does the php-fpm-restart script happen to hard-code PHP 7.4 or 7.2 and so not work with 7.3 or something? [20:56:26] that sounds like a winner, honestly [20:56:42] the PHP version changed when doc1001 to doc1002 [20:56:46] (Could we just move the doc hosts to prod PHP 7.4?) [20:56:51] Yeah, from 7.0 to 7.3. [20:57:16] yea, I mean. we can offer bullseye machines [20:57:34] it's just that it's not long ago that we finally switched to buster [20:57:49] Oh, it's just a buster machine? Never mind then. [20:58:23] I guess T291916 would be the parent. [20:58:24] T291916: Tracking task for Bullseye migrations in production - https://phabricator.wikimedia.org/T291916 [20:58:49] should I do a "wmf_auto_restart_php7.3-fpm.service" start? [20:58:53] on doc1002 [20:59:08] mutante: doc1002 didn't get the new files to restart with. [20:59:13] So I don't think it'd help? [20:59:14] ok [21:01:00] James_F: https://phabricator.wikimedia.org/T319477 [21:01:46] (https://phabricator.wikimedia.org/T285158 :p) [21:02:00] <3 [21:02:27] Joy. [21:03:09] so what went wrong with the scap pull? [21:03:41] [21:41:50] i wonder if this is related to the recent changes to checks.yaml format [21:03:54] 10Continuous-Integration-Infrastructure, 10serviceops-collab: migrate doc hosts to bullseye - https://phabricator.wikimedia.org/T319477 (10Dzahn) [21:03:57] I think something related to this seems likely [21:04:04] ack, thx [21:12:01] (03PS1) 10Dduvall: scripts: Log (debug) registration of scripts [tools/scap] - 10https://gerrit.wikimedia.org/r/838909 (https://phabricator.wikimedia.org/T319472) [21:23:14] (03CR) 10Ahmon Dancy: [C: 03+2] scripts: Log (debug) registration of scripts [tools/scap] - 10https://gerrit.wikimedia.org/r/838909 (https://phabricator.wikimedia.org/T319472) (owner: 10Dduvall) [21:25:35] (03Merged) 10jenkins-bot: scripts: Log (debug) registration of scripts [tools/scap] - 10https://gerrit.wikimedia.org/r/838909 (https://phabricator.wikimedia.org/T319472) (owner: 10Dduvall) [21:34:43] (03PS1) 10Ahmon Dancy: Release 4.26.0-1 [tools/scap] - 10https://gerrit.wikimedia.org/r/838914 [21:34:46] (03CR) 10Ahmon Dancy: [C: 03+2] Release 4.26.0-1 [tools/scap] - 10https://gerrit.wikimedia.org/r/838914 (owner: 10Ahmon Dancy) [21:38:46] (03Merged) 10jenkins-bot: Release 4.26.0-1 [tools/scap] - 10https://gerrit.wikimedia.org/r/838914 (owner: 10Ahmon Dancy) [21:50:36] 10Continuous-Integration-Config, 10Performance-Team, 10Wikimedia-Minify, 10Patch-For-Review: Re-work Minify's testRunJsFile test to not use forbidden @runInSeparateProcess which breaks CI publishing - https://phabricator.wikimedia.org/T319459 (10Daimona) If the goal is to reset $expandedStates after CliTes... [21:55:04] (03PS1) 10Daimona Eaytoy: layout: [mediawiki/tools/phan/SecurityCheckPlugin] Generate coverage on postmerge [integration/config] - 10https://gerrit.wikimedia.org/r/838917 (https://phabricator.wikimedia.org/T279423) [21:57:01] (03CR) 10CI reject: [V: 04-1] layout: [mediawiki/tools/phan/SecurityCheckPlugin] Generate coverage on postmerge [integration/config] - 10https://gerrit.wikimedia.org/r/838917 (https://phabricator.wikimedia.org/T279423) (owner: 10Daimona Eaytoy) [21:57:05] (03PS2) 10Daimona Eaytoy: layout: Generate coverage on postmerge for taint-check [integration/config] - 10https://gerrit.wikimedia.org/r/838917 (https://phabricator.wikimedia.org/T279423) [21:58:44] (03PS3) 10Jforrester: layout: [mediawiki/tools/phan/SecurityCheckPlugin] Publish PHP coverage [integration/config] - 10https://gerrit.wikimedia.org/r/838917 (https://phabricator.wikimedia.org/T279423) (owner: 10Daimona Eaytoy) [21:58:48] (03CR) 10Jforrester: [C: 03+2] layout: [mediawiki/tools/phan/SecurityCheckPlugin] Publish PHP coverage [integration/config] - 10https://gerrit.wikimedia.org/r/838917 (https://phabricator.wikimedia.org/T279423) (owner: 10Daimona Eaytoy) [22:00:45] (03Merged) 10jenkins-bot: layout: [mediawiki/tools/phan/SecurityCheckPlugin] Publish PHP coverage [integration/config] - 10https://gerrit.wikimedia.org/r/838917 (https://phabricator.wikimedia.org/T279423) (owner: 10Daimona Eaytoy) [22:03:23] !log layout: [mediawiki/tools/phan/SecurityCheckPlugin] Publish PHP coverage for T279423 [22:03:25] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [22:03:25] T279423: Enable coverage reports for taint-check - https://phabricator.wikimedia.org/T279423 [22:04:11] (03PS1) 10Dduvall: checks: Only validate checks for the current stage [tools/scap] - 10https://gerrit.wikimedia.org/r/838924 (https://phabricator.wikimedia.org/T319472) [22:05:11] (03CR) 10Ahmon Dancy: [C: 03+2] checks: Only validate checks for the current stage [tools/scap] - 10https://gerrit.wikimedia.org/r/838924 (https://phabricator.wikimedia.org/T319472) (owner: 10Dduvall) [22:05:19] Daimona: Want me to C+2 https://gerrit.wikimedia.org/r/c/mediawiki/tools/phan/SecurityCheckPlugin/+/838128 ? [22:05:43] I was about to ask if the first scap patch helped... I guess it helped :) [22:05:58] Clearly. :-) [22:05:59] It did. :-) [22:06:15] woop [22:07:01] Reedy: It occurs to me that given the index is cached we should put the last-generated line on the overall page too. [22:07:15] The thought crossed my mind earlier too [22:07:22] For exactly the same reason [22:07:23] Like minds. [22:07:53] Though I find I don't know off-hand the PHPism for getting the current time. [22:08:10] Oh. time(). Helpful. [22:08:27] yeah, unless you give it parameters... [22:09:40] (03Merged) 10jenkins-bot: checks: Only validate checks for the current stage [tools/scap] - 10https://gerrit.wikimedia.org/r/838924 (https://phabricator.wikimedia.org/T319472) (owner: 10Dduvall) [22:12:28] (03PS1) 10Ahmon Dancy: Release 4.27.0-1 [tools/scap] - 10https://gerrit.wikimedia.org/r/838925 [22:12:30] (03CR) 10Ahmon Dancy: [C: 03+2] Release 4.27.0-1 [tools/scap] - 10https://gerrit.wikimedia.org/r/838925 (owner: 10Ahmon Dancy) [22:16:31] (03Merged) 10jenkins-bot: Release 4.27.0-1 [tools/scap] - 10https://gerrit.wikimedia.org/r/838925 (owner: 10Ahmon Dancy) [22:18:18] Reedy/James_F: All fixed! [22:18:25] woo [22:18:28] Yay. [22:18:33] thanks dancy and dduvall <3 [22:18:41] np! [22:18:49] +∞ [22:19:05] You brave the depths of scap so we can pretend it doesn't exist. [22:19:41] 10Scap, 10Patch-For-Review: scap deploy errors aren't helpful: deploy-local failed: {} - https://phabricator.wikimedia.org/T319472 (10dancy) 05Openβ†’03Resolved a:03dduvall [22:20:10] James_F: yes, please :) [22:20:10] 10Continuous-Integration-Config, 10phan-taint-check-plugin, 10Patch-For-Review: Enable coverage reports for taint-check - https://phabricator.wikimedia.org/T279423 (10Jdforrester-WMF) 05Openβ†’03Resolved This should be fixed, though we'll find out when the next patch merges. [22:20:17] 10Continuous-Integration-Config, 10phan-taint-check-plugin, 10Patch-For-Review: Enable coverage reports for taint-check - https://phabricator.wikimedia.org/T279423 (10Jdforrester-WMF) [22:20:23] But maybe I should fix coverage first [22:20:27] Let me see [22:20:49] Daimona: I'm in the Heisenstate of merging and not merging. [22:21:00] That's the perfect state [22:21:19] But yeah, you can go ahead [22:21:23] Awesome. [22:21:39] So we'll see what I need to tweak for the coverage stuff to actually work [22:21:43] Thank you! [22:21:45] * James_F grins. [22:21:52] Daimona: Thank *you* so much. [22:22:44] There's also a truckload of "PHP Deprecated" warnings on PHP 8.1 which I want to analyze [22:22:48] https://integration.wikimedia.org/ci/job/composer-package-php81-docker/947/consoleFull [22:24:52] Yeah. [22:25:08] Not just in phan stuff, of course. [22:25:16] But tools are a good place to start. [22:28:16] James_F: echo | purgeList works fine if the code is deployed :D [22:28:17] https://doc.wikimedia.org/cover/ [22:28:37] TIL. :-) [22:28:50] it works for all sorts of random services [22:28:55] Yay. [22:29:06] Also we should style those dates now we have them. [22:29:10] Yeah... [22:29:15] I feel like I should add a sort by date too [22:29:48] Yeah. [22:30:05] β€œLibraries, sorted by least loved.” [22:30:09] heh [22:30:19] least loved, and/or most broken CI [22:33:00] Same diff? [22:34:26] Next up: php-ast upgrade https://integration.wikimedia.org/ci/job/phpunit-coverage-php74-docker-publish/30/console [22:38:56] (03PS1) 10Daimona Eaytoy: jjb: Fix typo in image name [integration/config] - 10https://gerrit.wikimedia.org/r/838928 [22:42:17] (03PS1) 10Reedy: CoveragePage: Sort by mtime [integration/docroot] - 10https://gerrit.wikimedia.org/r/838929 [22:43:13] How do we want to style the date? [22:48:39] Hmmmm where is phpunit-coverage-php74-docker-publish taking php-ast from? It uses 1.0.10, but our php-ast image only provides 1.0.14. [22:49:07] Reedy: Y-m-d HH:mm:ss ? [22:50:11] what's wrong with 2022-10-04T06:37:34+00:00 ? :P [22:54:36] Oh, I don't know. I mean, it certainly has nothing to do with readability. [22:54:42] :P [22:54:50] It's accurate :D [22:55:00] I honestly wasn't sure if James meant CSS or date format... [22:56:26] How can it be accurate if it doesn't show the picoseconds? :D [23:01:46] (03PS1) 10Reedy: Minor cleanup [integration/docroot] - 10https://gerrit.wikimedia.org/r/838932 [23:05:15] (03CR) 10CI reject: [V: 04-1] Minor cleanup [integration/docroot] - 10https://gerrit.wikimedia.org/r/838932 (owner: 10Reedy) [23:08:30] Why is phan being stupid there [23:08:59] 10Continuous-Integration-Config, 10phan-taint-check-plugin: Enable coverage reports for taint-check - https://phabricator.wikimedia.org/T279423 (10Daimona) 05Resolvedβ†’03Open One more config change to do [23:09:11] oh [23:09:11] ffs [23:10:44] (03PS2) 10Reedy: Minor cleanup [integration/docroot] - 10https://gerrit.wikimedia.org/r/838932 [23:12:06] aha [23:13:54] (03PS3) 10Reedy: Minor cleanup [integration/docroot] - 10https://gerrit.wikimedia.org/r/838932 [23:17:38] James_F: can we add php-ast to the composer-phpXY images, instead of the composer-package-phpXY ones? The coverage job uses the former and it complains that php-ast is not up-to-date. Same thing happens in Reedy's patch with composer-php74-docker: https://integration.wikimedia.org/ci/job/composer-php74-docker/1324/console (grep "php-ast") [23:22:11] (03PS1) 10Reedy: composer.json: Bump phpunit/phpunit to ^9.5 [integration/docroot] - 10https://gerrit.wikimedia.org/r/838936 [23:29:22] (03PS1) 10Reedy: WmuiPageBase: Add modified timestamp to footer [integration/docroot] - 10https://gerrit.wikimedia.org/r/838937 [23:33:53] (03CR) 10Reedy: WmuiPageBase: Add modified timestamp to footer (032 comments) [integration/docroot] - 10https://gerrit.wikimedia.org/r/838937 (owner: 10Reedy) [23:34:23] (03PS2) 10Reedy: WmuiPageBase: Add modified timestamp to footer [integration/docroot] - 10https://gerrit.wikimedia.org/r/838937 [23:37:48] Daimona: Or just move them into the base? [23:38:30] You mean the phpXY images? I think that'd be even better, but php-ast is not in Wikimedia APT, and that's what those images use right? [23:49:08] https://integration.wikimedia.org/ci/job/phpunit-coverage-php74-docker-publish/30/console has "WARNING: php-ast 1.0.10 is being used with Phan 5. php-ast 1.0.11 or newer is recommended for compatibility with plugins and support for AST version 85." 200+ times [23:49:27] very compressable [23:49:48] Plus install instructions, yeah... [23:51:52] And that's why I was trying to move the enabling of php-ast to a more base image... [23:52:16] heh [23:54:25] https://github.com/wikimedia/integration-config/blob/master/dockerfiles/php74/Dockerfile.template#L4 [23:54:30] so yeah, they use wikimedia apt [23:55:48] php-ast uses sury.. so if we combined them [23:56:34] Unless we can pin the phpXY versions to install php otherwise from wikimedia apt? [23:57:36] or I'm over complicating it