Skip to content

fix: distinguish daily source read failures from absent runs - #485

Merged
Pigbibi merged 2 commits into
mainfrom
fix/daily-source-failure-classification
Nov 1, 2026
Merged

Pigbibi merged 2 commits into
mainfrom
fix/daily-source-failure-classification

Conversation

@Pigbibi

@Pigbibi Pigbibi commented Nov 1, 2026

Copy link
Copy Markdown
Contributor

What changed

When archive listing failed, reports could not be read, or the bounded read limit was exceeded, the daily publisher returned no_attributable_runs if no attributable rows remained. Preserve fixed source-failure statuses instead; unknown source errors use a safe fallback without exposing provider text. Genuine empty or unmatched sources retain no_attributable_runs.

All these paths still fail before POST. Partial projections with attributable runs, identity validation, exact acknowledgement matching and the single-request policy are unchanged. Add regression coverage and document the statuses.

This follows #484 and addresses a reproduced diagnostic defect. It does not establish or repair the original candidate-storage incident, and it does not recover live authentication. No deployment or live invocation was performed.

How verified

  • New regression failed against the original source (no_attributable_runs instead of source_list_failed) and passed after the fix.
  • Focused daily/auth/notification checks: 158 passed.
  • Full local suite: 1469 passed and 13 subtests passed; one third-party NumPy deprecation warning.
  • Ruff, dependency compatibility, lock consistency, pinned shared-package revisions/imports, CI-equivalent locked QPK pin check, package sdist/wheel build, and diff whitespace checks passed.
  • Documentation check: zero blocking findings; one existing README-outline advisory.
  • Reviewed all three changed files and ran an added-line credential/account-pattern heuristic.
  • The legacy repository-local pin helper compares the dependency to moving upstream main and reports an existing mismatch. The CI checker uses the locked revision and passes; no dependency files were changed.

Draft only. No labels, merge, deployment, IAM changes, candidate replay or trading.

AI assistance

OpenAI Codex. No delegated agents.

Co-Authored-By: Codex <noreply@openai.com>
@coderabbitai

coderabbitai Bot commented Nov 1, 2026 •

Copy link
Copy Markdown

Important

  • Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

Run configuration
  • Source mode: Auto
  • Repository: Unavailable
  • Branch: Unavailable
  • File: No configuration file from this repository applied
  • Commit: Unavailable
  • Layers: defaults
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: 2b4e6a31-1a4b-4b3e-8955-a2cc7deefbf4
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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.

Share

@Pigbibi Pigbibi added the owner-approved Owner approval for high-risk PRs. AI agents must never apply this label. label Nov 1, 2026
@Pigbibi
Pigbibi marked this pull request as ready for review November 1, 2026 14:02
@Pigbibi
Pigbibi merged commit b8fae78 into main Nov 1, 2026
9 checks passed
@Pigbibi
Pigbibi deleted the fix/daily-source-failure-classification branch November 1, 2026 14:04
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

owner-approved Owner approval for high-risk PRs. AI agents must never apply this label.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant