Experiment Archive Index

v3.8.1–v3.8.2 Experiment Archive

These branches are unreleased and unsupported research archives. Do not merge them into main or treat them as downloadable stable releases.

Why Separate Archive Branches Are Used

• Preserve source, tests, failure analysis, and future research clues.
• Leave the current stable main branch unchanged.
• Create no experimental tags or GitHub Releases.
• Stay outside the releases/latest channel used by automatic updates.
• Exclude build outputs, user configuration, databases, logs, and real Mods.

Archive List

Experiment	Branch and commit	What was demonstrated	What was not demonstrated
v3.8.1 Hot Injection	archive/v3.8.1-hot-injection-lab, d17f7fa	Whole-file snapshots, SHA-256, backups, and atomic replacement; 49/49 tests and a successful WinUI build	Reliable F10 restoration of real in-game state
v3.8.1 Persistent State	archive/v3.8.1-persistent-state-lab, 19f6bab	Scalar global persist discovery, stable sampling, and offline merge; 70/70 tests	Automatic runtime restoration for arbitrary third-party Mods
v3.8.2 EFMI Resident Controller	archive/v3.8.2-efmi-resident-controller, 4886380	Standard EFMI structure analysis, controlled copies, and a profile controller; 73/73 tests and a successful WinUI build	Complete gating for arbitrary INIs, real-game compatibility, and resident-mode performance

All existing v3.8.2 tests passed, but core coverage was 81.3% line and 75.8% branch, below the historical 90%/85% gate. It must not be described as passing the complete quality gate.

Suggested Reading Order

1. Cross-Mod Experiment Overview
2. Detailed v3.8.1–v3.8.2 Development Notes
3. Open EXPERIMENT_STATUS.md on the relevant archive branch.

Current Conclusion

Universal, automatic, and reliable runtime persistence of internal $variables across arbitrary third-party XXMI/EFMI Mods cannot currently be promised as a released feature. Stable v3.9.0 retains only verified file deployment and management capabilities.