Prove whether four production test seams are needed

[metaphorics]

Question

Can the four production test seams for equalprg, resolve_method, has_pending_ctrl_bslash, and test_lvalue_parts be replaced by observable behavior tests that fail against the same defects?

Grounded constraints

• Candidate seams live in crates/ox-editor/src/indent.rs, crates/ox-editor/src/mode.rs, and crates/ox-editor/src/excmd_exec.rs.
• A public replacement must detect the same plausible bug, not merely execute the same area.
• Preserve narrow private access when no stable observable path distinguishes the defect.
• Avoid global-state-sensitive or timing-sensitive replacements.
• Delete each seam independently only after one-for-one mutation proof.

Investigation

For each seam, state the defect class it protects. Seed that defect. Run the current private test and the proposed editor or script scenario. Compare failure signal, isolation, and false-positive risk.

Evidence to produce

• Four seam-to-contract descriptions.
• Four independent mutation results.
• The exact replacement test, if one is stronger or equal.
• A per-seam keep or delete verdict.

Blocked by

• None. Choose the first cut for the crates overhaul #13 records the selected first cut.

Done when

The issue closes with a mutation-backed verdict for all four seams. Any deletion names the observable test that now owns the contract.