Skip to content

fix(analytics): make failed refreshes visible and isolate failure tests - #1597

Merged
castrojo merged 1 commit into
mainfrom
fix/contributor-refresh-safety
Oct 25, 2026
Merged

castrojo merged 1 commit into
mainfrom
fix/contributor-refresh-safety

Conversation

@castrojo

@castrojo castrojo commented Oct 25, 2026 •

Copy link
Copy Markdown
Contributor

Scoped review follow-up

Two safety corrections; metric definitions, GitHub queries, contributor counts, chart presentation, and the tracked snapshot are unchanged.

  • The write-failure CLI regression runs unmodified copied modules and a fixture seed in a disposable workspace. No fault-injected child can overwrite tracked production data, regardless of temporary filename changes.
  • The fetcher remains exit 0 for builds. A separate final scheduled-job gate requires a complete, available snapshot generated during the current collection interval; unavailable, retained, stale, or future snapshots make the job red. Genuine measured zero passes.
  • Unavailable state is published for review before the final job gate fails. Last-good history and timestamp remain intact.
  • Relevant procedure updated in this same PR. No new dependencies or credentials.

Verified

  • Actual CLI/workflow-command smoke: unavailable fetcher exit 0; scheduled validation exit 1; last-good total and timestamp retained; verified-zero validation exit 0.
  • All complete affected regression modules pass.
  • just check passes: 1,354 tests, typecheck, and zero lint errors (existing warnings remain).
  • Action 40146255069 used github.token and executed the new Require a verified current refresh step successfully after a real complete collection.
  • Main deployment 40147170574 succeeded for aeaa309.
  • Cache-busted live analytics check: 456 shown, two chart canvases rendered, old merge-rate metrics absent.

Keep the fetcher build-safe while requiring scheduled snapshots to be complete and generated during the current run. Publish unavailable state before failing the scheduled job. Run the write-failure CLI regression in a disposable workspace rather than against tracked data.

Assisted-by: GPT-6.1 Sol via GitHub Copilot
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@castrojo
castrojo added this pull request to the merge queue Oct 25, 2026
Merged via the queue into main with commit aeaa309 Oct 25, 2026
12 checks passed
@castrojo
castrojo deleted the fix/contributor-refresh-safety branch October 25, 2026 00:33
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant