# CONTROL versus DryRun DCF comparison

DBMOD covers only the active drawing. This procedure separately compares
content-bearing project files after identical CONTROL and DryRun sessions.

1. With Plant 3D closed and the Launcher session `RolledBack`, copy the local
   fixture to `<FROZEN_FIXTURE>`.
2. Create `<CONTROL_FIXTURE>` and `<DRYRUN_FIXTURE>` from that same frozen
   state; calculate complete local SHA-256 inventories before each run.
3. Open model and Ortho drawings in CONTROL through the same Launcher DryRun
   launch contract, but invoke no P_Rec_Annot command. Close without saving,
   wait for Launcher rollback, then calculate the after inventory.
4. Repeat with DRYRUN and invoke each supported command twice. Close without
   saving, wait for rollback, then calculate the after inventory.
5. Compare deltas by logical relative path. Classify each content-bearing file
   as `UNCHANGED`, `AUTODESK_BASELINE_CHANGE`, `PLUGIN_ONLY_CHANGE`, or
   `INCONCLUSIVE`.

Content-bearing groups include model DWG, Ortho DWG, `Piping.dcf`, remaining
DCF/database files, `Project.xml` or equivalent project metadata, and Ortho
metadata. Lock, cache, session, log, preview, and temporary files are reported
separately and cannot establish project invariance.

Any `PLUGIN_ONLY_CHANGE` blocks Apply. Full paths and hashes remain local;
repository evidence contains only logical fixture markers, groups, counts, and
classifications.