master red after #8415: type name Pair conflicts, and 79 seals drift on gen_hash_zig #8417

#8415 (gen-zig &[_]T{} fix) merged with three non-required checks red. All three are mine.

• Corpus ratchet, "A type name may not gain a second definition". The regression spec specs/compiler/zig_empty_typed_array.t27 declares Pair, and specs/port/tools/verify_multitarget.t27 declares it too. tri types ratchet reads 86 names against a ledger of 85.
• Seal Coverage and spec-guards, "79 seal(s) newly do not hold" [gen-drift]. The fix changes the Zig that gen emits for every spec with an empty typed array, so those seals no longer describe the output. A codegen change has to refresh the seals it moves in the same PR (tri seals drift --fix). I did not.

Acceptance criteria

• tri types ratchet prints RATCHET: CLEAN on master.
• tri seals drift reports no drifted seal on master.
• Corpus Ratchet and Seal Coverage are green on master.

Boundary

specs/compiler/zig_empty_typed_array.t27
.trinity/seals/

User Scenarios

• Given the tri types ratchet command reads 86 type names against a ledger of 85, when the duplicate Pair definition is resolved, then tri types ratchet prints RATCHET: CLEAN.
• Given tri seals drift reports 79 drifted seals after the Zig empty typed array codegen change, when the seals are refreshed for all affected specs, then tri seals drift reports no drifted seals.

Requirements

• FR-001: The Pair type name MUST be defined in exactly one spec file across the corpus.
• FR-002: All seals in .trinity/seals/ that describe Zig output for empty typed arrays MUST be regenerated to match the current codegen.
• FR-003: The tri types ratchet command MUST exit 0 with output RATCHET: CLEAN on master.
• FR-004: The tri seals drift command MUST exit 0 with zero drifted seals reported on master.

Success Criteria

• tri types ratchet exits 0 and prints RATCHET: CLEAN
• tri seals drift exits 0 and reports 0 drifted seals
• Corpus Ratchet check passes (exit 0)
• Seal Coverage check passes (exit 0)