The smoke job is a reusable-workflow 'uses:' call; Actions does not support
continue-on-error there, producing a 0-job workflow failure that has failed
every deploy (so v0.5.0 never reached production). Matches the prautoblogger
fix in #147. Smoke is not in deploy-production needs, so it stays non-blocking.
Agent-Session: cto restore-prcore-deploy
- Remove invalid continue-on-error from reusable workflow call
(GitHub Actions doesn't support this property on 'uses:' jobs;
it caused 0-job workflow failures)
- Remove 'smoke' from deploy-production needs — smoke failures (e.g.
Playwright infra issues) must not block valid production deploys.
This matches the pattern already used by the theme repo.
- Smoke still runs and uploads artifacts; it's just not a gate.
- smoke job now has continue-on-error: true so a Playwright infra failure
never silently blocks a valid production deploy
- smoke.yml root cause was fixed in peptide-e2e (da79ee2) — explicit
checkout of peptide-e2e in workflow_call context
- This commit retriggers the CI/deploy pipeline for PR Core v0.5.0
Rsync over SSH on push to main. Runs CI first, then syncs
production files to wp-content/plugins/peptide-repo-core/
on Hostinger. Includes LiteSpeed cache purge and health check.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
