Skip to content

Unify retained pull-request attribution for Scoring and metrics - #1312

Merged
JacobStephens2 merged 2 commits into
mainfrom
issue-1311
Oct 25, 2026
Merged

JacobStephens2 merged 2 commits into
mainfrom
issue-1311

Conversation

@JacobStephens2

@JacobStephens2 JacobStephens2 commented Oct 25, 2026 •

Copy link
Copy Markdown
Owner

Summary

Closes #1311

Scoring samples / Bug origins / Run metrics
  → retained pull-request attribution
    → captured Command ownership tree

Match PR URLs without ASCII case or trailing slashes. Select by parsed RFC 3339 instant, then lexical record path and Command name. Saved local path/name pins select exactly or return no match. Scoring keeps its capture points and persisted schema; metrics reuses its Snapshot and lightweight settlement decoder.

Test seams

  • Retained pull-request selection from a captured Snapshot, including its existing filesystem lookup adapter.
  • Scoring and Bug-origin CLI checkpoints and read-only Run metrics, through the existing fake-Harness fixtures.

Evidence

  • Before: cargo test --bin thirdshift retained_pull_request_lookup_orders_instants_across_offsets failed: selected older.command.json at 12:00+02:00 over newer.command.json at 11:00Z.
    After: the retained regression passes. Shared-interface tests cover normalization, chronological/tied/invalid starts, exact pins and nested/dispatched ownership. Command regressions prove Scoring/metrics agreement, one count across Continuations, unavailable/mismatched-pin refusal, unchanged record bytes and no remote metric reads.
  • cargo check, cargo fmt --check and cargo clippy --all-targets -- -D warnings pass.
  • /usr/bin/time -p scripts/affected-tests main passed: 50 site tests and 3,004 Rust tests passed (66 skipped). Wall time: 1,642.22s including compilation, site checks and Rust execution. Nextest's full Rust execution took 1,454.385s; compilation was included in the selector timing.

Merge Danger

Door: two-way

Blast Radius: attribution

Local Command selection for Scoring and Bug-origin metrics; no retained-data migration.

Unaddressed findings

Standards

None. The single P3 fixture-duplication finding was fixed in 1d3553fb and verified in the reviewer's follow-up.

Spec

None. No findings.

Security

None. Security review (guidance mode) of the attribution consolidation found no introduced findings: the new module only compares locally captured record paths/names and PR URLs, returns locally captured identities, and adds no network, write, or authorization path. No proof-of-concept failure was produced, so no fix was made here.

Security review outcome

No unaddressed introduced findings.

Review coverage

Both independent reviewers inspected all 16 changed files, including generated site/docs/index.html, and the final fixture changes. Generated-output exclusions: none. Unread changed files: none under either axis.

Built with codex · gpt-6.1-sol · xhigh

@JacobStephens2
JacobStephens2 merged commit bd4b3e7 into main Oct 25, 2026
18 checks passed
@JacobStephens2
JacobStephens2 deleted the issue-1311 branch October 25, 2026 11:19
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.

Deepen retained pull request attribution across Scoring and Run metrics

1 participant