t27b ledger: rows for 9 more specs that landed without one (native ratchet UNLISTED 9) #8645

Part of #8637 (item 1: the ledger steward). Until the steward runs, the loop adds rows by hand. Master's t27b-native-ratchet (run 20624dd on #8638) reports UNLISTED 9:

• pass: seal-stands-in, fpga/trinet_node, queen/drain, queen/review_lanes, tri/actors/metrics, tri/lab/receipt, tri/lanes/lanes, tri/pipeline/spec_writer;
• codegen, reason unimplemented: compiler/lexer, FnDecl(frame size). Its fn next_token holds 3344848 bytes of local aggregates, against at most 1048576 per frame.

max_not_pass 25 -> 26 (the one new non-pass spec).

User Scenarios

• Given the t27b ledger at specs/trinet/ledger.t27 lacks rows for 9 specs reported UNLISTED by native ratchet, when the steward adds those rows, then t27b reports all 9 as listed and max_not_pass reflects 26.
• Given the ledger update includes compiler/lexer as codegen with reason "unimplemented", when the ratchet runs, then that spec is not counted as a pass and the frame-size note is recorded.

Requirements

• FR-001: The ledger file specs/trinet/ledger.t27 MUST contain one row for each of the 9 UNLISTED specs with correct status and reason fields.
• FR-002: The max_not_pass value in the ledger MUST be updated from 25 to 26.
• FR-003: The entry for compiler/lexer MUST have status codegen and reason unimplemented with the frame-size annotation preserved.

Success Criteria

• t27c specs/trinet/ledger.t27 exits 0 and emits no UNLISTED warnings for the 9 specs.
• scripts/tri_loop/t27b.py --check-ledger exits 0 and reports max_not_pass = 26.
• The ledger file contains exactly 9 new rows matching the issue's list.

Boundary

• specs/trinet/ledger.t27
• scripts/tri_loop/t27b.py