sequence_hdc: replace the three map-typed fields with key/value slices (typecheck refuses the spec) #6095

Original file list

• specs/vsa/sequence_hdc.t27
• .trinity/seals/SequenceHdc.json
• .trinity/seals/vsa_SequenceHdc.json

Defect

t27c typecheck specs/vsa/sequence_hdc.t27 refuses the spec: struct LanguageDetector declares language_vectors : map[:]const u8, PackedBigInt>, a map type, and t27 has no map type. The parser recovers PackedBigInt as a field with no type (#3225), and gen-rust writes the field as pub a: Xb:Y,. tri misread --list lists the spec under "bare colon inside a type" and under "typecheck refuses". ItemMemory.cache and SequenceMemory.sequences (map<u32, PackedBigInt>, map<u64, PackedBigInt>) are the same shape.

Original scenarios

• Given the fixed spec, When t27c typecheck specs/vsa/sequence_hdc.t27 runs, Then it reports no error that names a struct field.
• Given the fixed spec, When tri misread --list runs, Then specs/vsa/sequence_hdc.t27 is not listed under "bare colon inside a type".

Original requirements

• FR-001: Each map-typed field MUST become a slice of a key/value struct, as git/schema did for env in fix(t27c): refuse a colon inside a field type and @as to a slice type; fix the six silent specs #5971 (env: []EnvVar with struct EnvVar { name: str, value: str }).
• FR-002: The indexed write detector.language_vectors[language] = encoded.vector; MUST become a write the new type supports (append an entry), with a test that asserts the entry is found after it.
• FR-003: The spec MUST stay ASCII-only (L3), and both seals MUST be refreshed with t27c seal specs/vsa/sequence_hdc.t27 --save.

Original checks (kept for history; the Queen does not run these)

• t27c typecheck specs/vsa/sequence_hdc.t27 prints no field error.
• python3 tools/check_seal_currency.py exits 0.
• python3 tools/ci/check_specs_still_parse.py --base origin/master prints ok.

Refs #6092

Queen-runnable restatement (2026-10-05)

This block restates the issue so the Queen's runner reads one boundary and one list of checks; the text above is kept for history and its own file list and checks are superseded by the sections below.

The bee branch queen-6095 had no diff against master. Since c8f4e5b the spec typechecks, but it still declares 3 map< fields (lines 40, 66, 91), assigns into one with detector.language_vectors[language] = ... (line 188), and t27c test-report is BLOCKED, so none of its 17 tests runs. The pattern to follow is specs/git/schema.t27, which replaced a map with env: []EnvVar.

Boundary

• specs/vsa/sequence_hdc.t27
• .trinity/seals/SequenceHdc.json
• .trinity/seals/vsa_SequenceHdc.json

User Scenarios & Testing

• Given sequence_hdc.t27 BLOCKED on its map< fields, when each map becomes a fixed-capacity array of key/value structs with a lookup function, then the spec compiles and its 17 tests run and pass.

Requirements

• FR-001: Every map< field in specs/vsa/sequence_hdc.t27 MUST become an array of key/value structs (plus a count), with lookup and insert written as t27 functions.
• FR-002: The 17 existing tests MUST be kept (adapted to the new shape, not deleted) and pass.
• FR-003: Both seals MUST be refreshed with t27c seal specs/vsa/sequence_hdc.t27 --save.
• FR-004: The change MUST NOT add or modify non-t27 hand-written code (specs/policy/own_language.t27).

Success Criteria

• grep -c 'map<' specs/vsa/sequence_hdc.t27 prints 0 (today 3)
• t27c test-report specs/vsa/sequence_hdc.t27 2>&1 | grep -c BLOCKED prints 0 - a BLOCKED spec never ran a test
• t27c test-report specs/vsa/sequence_hdc.t27 2>&1 | grep -cE '^ FAIL +0$' prints 1 - the generated code compiles and every test passes
• t27c test-report specs/vsa/sequence_hdc.t27 2>&1 | grep -E '^ tests ' | tr -dc 0-9 prints at least 17 - the count of tests that ran
• t27c typecheck specs/vsa/sequence_hdc.t27 prints Typecheck OK (0 errors, 0 warnings)