[rjb32]

Why

Every merge to main shows a failed Deploy to Railway check (#17, #19, #20), even though the site deploys fine.

• Railway already deploys main. Its GitHub integration (railway-app[bot]) builds the Dockerfile and deploys every push. For Update the docs for TuringDB 3.0.2 #20 it deployed 4b73823 successfully at 15:52 UTC, and the live site serves the new docs.
• The Actions job was a second deploy of the same commit with railway up. It needs a RAILWAY_TOKEN secret that the repository never had, at repo or environment level. So it stopped at its first step every time.

What

• Workflow: .github/workflows/deploy.yml becomes build.yml, named Build.
  • It keeps the build check on pull requests and pushes to main: npm run gen, tsc --noEmit, the anchor test, npm run build and the page-count check.
  • The check is still called Build, so any required-check setting keeps working.
• Removed: the deploy job, and the index.html artifact upload that only fed its smoke test.
• Concurrency: a newer push now cancels an in-flight check on the same branch, since there is no deploy left to protect.
• README: the Hosting and CI sections say that Railway's GitHub integration deploys main, and drop the RAILWAY_TOKEN instructions.

After merging

• Old failures stay. The failed production deployment entries from earlier runs remain in the repository's deployment history. New merges won't add more.
• Unused environments. The production and staging GitHub environments are no longer used by any workflow and can be deleted in the repository settings.