Conversation
Field help icons point at off-site FieldWorks doc pages that rot when the docs site changes, not when we change code. Runs the existing opt-in `test:manual` help-link check on a weekly schedule (and on demand). On a scheduled run it retries once after a delay so a brief docs-site outage doesn't cry wolf; the job goes red only if a link is still broken hours later. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Repository UI Review profile: CHILL Plan: Advanced Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Proves check-help-links.yaml (upstream PR sillsdev#2650) before it lands: checks out PR sillsdev#2643's branch, reintroduces the broken link, and runs the sleep-then-retry logic with a dispatchable delay input. Confirms the test detects a broken link in CI and that a long on-failure sleep survives under the 6h job cap. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
First attempt / wait / retry are separate steps so the retry shows in CI, with the wait time in the step name. First attempt is continue-on-error so a recovered retry keeps the job green; a manual dispatch fails fast rather than retrying. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
I think we make the workflow more general, then make the job specific to the help tests. While we're at it why don't we put the cloudflare header size test in this workflow too! |
2 participants
[Claude, autonomous]
Stacked on #2643 (base is its branch, not
develop). Retarget todeveloponce #2643 merges.Runs #2643's opt-in
test:manualhelp-link check weekly (and viaworkflow_dispatch). Help links point at an off-site FieldWorks docs site and rot when that site changes, not when we change code, so this runs on a timer rather than per-PR — per-PR would also make PR CI hostage to the external site.A scheduled run retries once after a 3h delay so a brief docs-site outage doesn't cry wolf: the job goes red only if a link is still broken hours later. A manual
workflow_dispatchrun fails fast (no wait). Free public-repo runners, so the sleeping job costs nothing; capped under the 6h job limit.Testing:
schedule/workflow_dispatchonly surface from the default branch, so this can't run from the PR branch — dispatch it manually once merged todevelopto confirm.🤖 Generated with Claude Code