ROS Historical Migration Validation

Date: 2026-08-17

Scope

This validates the historical attribution schema, the eight reconstructed records covering the pre-install interval, registry determinism, compatibility, and the live/historical transition boundary. It does not retroactively validate the substantive historical research or prove missing approvals and tests.

Inventory reconciliation

• Historical cutoff: b2877f1cdc87f9ab0df732947b78354b789ba986.
• ROS installation commit: 4f8b816480d1c27780e6f594d846ba33ebe325d3.
• Pre-install commits: 20.
• Historical records: 8.
• Unique commits referenced by records: 20.
• comm comparison between git rev-list b2877f1 and JSONL commit references: no differences.
• Unique attributed path groups: 35.
• Reconstruction confidence: 7 high, 1 medium.
• Tags found: 0.
• Side branch tips found: 0; only main and its remote-tracking references were present.

Commands and results

Historical validator tests

npm run test:ros

Passed 6/6 tests:

1. valid records parse and preserve legacy compatibility metadata;
2. duplicate migration IDs fail;
3. malformed and unknown commit hashes fail with Git history available;
4. invalid semantic states fail; and
5. historical records cannot contain live transition authority fields; and
6. valid history can satisfy dirty-path attribution while invalid history cannot.

Existing repository-integrity tests

python3 -m unittest tools.repository.test_validate_research_integrity

Passed 7/7 tests.

Combined project test command

npm test

Passed: 6 ROS history tests plus 7 repository-integrity tests.

Research Publisher validation

npm run research:validate

Passed. Compatibility-mode metadata diagnostics remain warnings rather than strict failures.

Repository audit

python3 tools/evaluation/evalctl.py repo-audit

Passed. It reported no case collisions, duplicate identifiers, broken explicit local Markdown links, or frontier structure errors. Relationship and metadata coverage diagnostics remain non-blocking and include known legacy gaps.

ROS registries and validation

./ros registry build
./ros registry check
./ros validate
./ros registry build --dry-run

Results:

• registries/decisions.json was rebuilt by ROS to include DF-ROS-HISTORY-2026-0001.
• Registry check passed.
• ROS validation passed.
• The subsequent dry run reported 0 registry file(s) would change, confirming deterministic regeneration for the current tree.

An intermediate ./ros validate --json correctly failed only because the new Decision Record made registries/decisions.json stale before the required registry rebuild. The rebuild resolved that expected finding.

Diff and coverage checks

git diff --check
comm -3 <(git rev-list --reverse b2877f1 | sort) \
  <(jq -r '.commits[]' .ros/history/historical-work.jsonl | sort -u)

Both completed without output. No historical commit was omitted or added, and no whitespace error was found in tracked diffs.

Fresh-agent reconstruction check

The repository now provides a conversation-independent path:

1. root CURRENT-STATE.md for present authority;
2. docs/repository/ros-migration/HISTORICAL-WORK-INDEX.md for the human reconstruction and grouping rationale;
3. .ros/history/historical-work.jsonl for machine attribution;
4. record-level decision and evidence paths for supporting artifacts; and
5. .ros/events/events.jsonl for live ROS-era transitions only.

The index states all known gaps and distinguishes commit observation from inferred work boundaries.

Preserved limitations

• No external issues, pull requests, approvals, or work-item export were found.
• Missing historical test results remain “not found,” never “passed.”
• Generated output is not counted as independent research evidence.
• Git author timestamps do not prove the full duration of human work.
• The state-constrained research sequence remains medium-confidence as a grouped reconstruction because its commits do not expose a complete authoring journal.