Independent dynamic Opus 5 review of PR #980 head25ff4870 found a pre-existing native Reject discrepancy: native rejectChanges PhaseA checks paragraph mark ins only, while comparison rejectAllChanges also resolves moveTo paragraph marks.
Synthetic reviewer control (no range markers): an empty moveTo-marked paragraph followed by text B produces two paragraphs [empty,B] natively versus one [B] through comparison AST. A plain ins-mark control agrees. This minimizes the differing predicates, but a repair must also test complete schema-valid move ranges/packages, not rely only on this reduced input.
The phase-A predicate is untouched by PR#980; its note-bearing reader regressions pass. No fix included in this PR. Separate from #982 cosmetic mark cleanup and #983 inherited reader field recalculation. Word remains unverified.
Ref: #941, PR #980 review N2.
Independent dynamic Opus 5 review of PR #980 head25ff4870 found a pre-existing native Reject discrepancy: native rejectChanges PhaseA checks paragraph mark ins only, while comparison rejectAllChanges also resolves moveTo paragraph marks.
Synthetic reviewer control (no range markers): an empty moveTo-marked paragraph followed by text B produces two paragraphs [empty,B] natively versus one [B] through comparison AST. A plain ins-mark control agrees. This minimizes the differing predicates, but a repair must also test complete schema-valid move ranges/packages, not rely only on this reduced input.
The phase-A predicate is untouched by PR#980; its note-bearing reader regressions pass. No fix included in this PR. Separate from #982 cosmetic mark cleanup and #983 inherited reader field recalculation. Word remains unverified.
Ref: #941, PR #980 review N2.