fix: match requirement identity case-insensitively

[N4M3Z]

Archive merges compared requirement names byte-for-byte, so a Title Case delta appended a duplicate of its sentence-case canonical twin instead of matching it. Lookups now normalize case and whitespace, an identical case-variant add is idempotent, and a conflicting one fails closed. Equality checks that decide whether an edit applies keep exact comparison, so a case-only rename still lands. Two regression tests pin both paths.

spec:none: the change repairs merge behavior and adds no new contract.

Note

Make requirement name matching case- and whitespace-insensitive

• Adds requirement_identity in model.rs that normalizes names by collapsing whitespace and lowercasing, then uses it in CanonicalSpec.requirement_index and CanonicalSpec.requirement_matches so lookups and equality checks ignore case and whitespace differences.
• Adjusts apply_renamed in apply.rs to allow a rename when the target name resolves to the same requirement index, only erroring when it resolves to a different requirement.
• Adds tests in apply/tests.rs covering rejection of case-variant duplicate adds and no-op behavior for identical-up-to-case adds.
• Behavioral Change: requirement_index and requirement_matches now treat names that differ only by case or whitespace as equal; callers relying on exact-string matching for requirement names will see different results.

^{Macroscope summarized 927ec97.}

[runeseer]

🚫 [runeseer] _{reported by reviewdog 🐶}
High — requirement_matches compares names case-insensitively but still compares content byte-for-byte, and parsed content begins with ### Requirement: <name>, so a genuinely identical case-variant ADD never matches and apply_added returns "cannot add existing requirement" instead of skipping.

• src/spec/model.rs:141 content equality includes the case-bearing header line, so the loosened name check is a no-op on real data.
• src/spec/apply/tests.rs:43 the regression test overwrites both contents with "shared content", a value the parser cannot produce, so it passes without exercising the path.

[runeseer]

🚫 [runeseer] _{reported by reviewdog 🐶}
High — requirement_index returns the first case-insensitive match, so on a spec that already holds both case variants — the duplicate state this commit exists to stop — a REMOVED, MODIFIED, or RENAMED operation naming the Title Case requirement silently mutates the sentence-case one instead.

• src/spec/model.rs:123 position() stops at the first identity match and no caller checks for a second.
• src/spec/apply.rs:111 remove_requirement(element_index) deletes that wrong element with no name reconfirmation.

[macroscopeapp]

Approvability

Verdict: Approved at 927ec97

Macroscope's review found this PR approvable — This is a focused archive-merge bug fix that normalizes requirement-name matching and adds regression coverage for duplicate and conflicting case variants. Its runtime impact is confined to the existing requirement application path, with no schema, deployment, or sensitive-domain changes.

Notes:

• Macroscope's correctness review did not run, so approvability was decided on eligibility alone.

^{You can add or adjust custom eligibility rules. Learn more.}

[runeseer]

Runeseer review — 2 open findings

Risk	Finding	Location
High	First-match lookup mutates wrong element on already-duplicated specs	src/spec/model.rs:122
High	Case-variant identical add still rejected; content compare includes header	src/spec/model.rs:139

Request changes. No new commits landed since the last review, so both previously raised issues in src/spec/model.rs are still open.

• The case-insensitive name match is undone by the byte-for-byte content compare, because content carries the ### Requirement: header line.
• Requirement lookup takes the first case-insensitive hit, so remove, modify, and rename can act on the wrong entry in a spec that already holds both variants.
• Macroscope posted an approvability note only; its correctness review did not run.

---

2 open · Reviewed 927ec97b · review run · $0.42 · 0m57s · 14 turns

[runeseer]

@N4M3Z, the following findings block this head:

• src/spec/model.rs:139: Case-variant identical add still rejected; content compare includes header

• src/spec/model.rs:122: First-match lookup mutates wrong element on already-duplicated specs

Choose one action:

• Fix all findings. Then apply review:runeseer to review the new head.
• Apply ignore:runeseer to accept the complete current verdict.