catalogd: HTML-safe escape fix + decisions tracker entry
Per 2026-05-03 step_7_8_retention_and_parity scrum (opus WARN on
parity_subject_audit.rs:canonical_json):
Go's json.Marshal HTML-escapes < > & to < > & by
default. Rust's serde_json::to_vec keeps them literal. Any audit
row with these chars in any string field would silently produce
different canonical bytes across runtimes → broken HMAC chain.
Latent because no production audit field has carried <>& yet, but
realistic for purpose strings ("error & retry") or trace_id values
("<HTTP-Request-Id>").
Fix: marshalNoEscapeHTML helper wraps json.Encoder.SetEscapeHTML(false)
+ trims trailing newline. Routed through writeCanonical for both
keys and scalar values.
Regression test: TestVerifyChain_HtmlChars_NotEscaped (purpose has &,
trace_id has <>) asserts the canonical bytes contain literal chars,
not escape sequences.
11 unit tests pass including the new one; parity probe still 6/6
byte-identical against live production audit logs.
Decisions tracker: added 2026-05-03 entry for SUBJECT_MANIFESTS_ON_CATALOGD
Steps 1-8 closure + 6th cross-runtime parity probe (was 5).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
root
@@ -59,6 +59,7 @@ Don't:
|
||||
| 2026-05-02 | **extract_json parity probe — 12/12 match across edge cases** | New `scripts/cutover/parity/extract_json_parity.sh` runs identical model-output strings through Rust `gateway::v1::iterate::extract_json` AND Go `validator.ExtractJSON`. 12 fixtures: fenced/unfenced blocks, nested objects, unicode, escaped quotes, top-level array, malformed JSON. Substrate gate: `cargo test -p gateway extract_json` PASS before probe. Result: **12/12 match.** Algorithms genuinely equivalent. Rust side gained `pub` on `extract_json` + new `bin/parity_extract_json` (~30 LOC). |
|
||||
| 2026-05-02 | **Validator wire-format alignment — DONE** | Custom `MarshalJSON`/`UnmarshalJSON` on Go's `validator.ValidationError` emits the Rust serde-tagged-enum shape `{"Schema":{"field":"x","reason":"y"}}`. UnmarshalJSON also accepts the legacy flat shape (migration safety) and rejects unknown variants (drift guard for future Rust enum additions). 4 new pinning tests in `types_test.go`. Re-run validator parity probe: **6/6 match** (was 1/6). |
|
||||
| 2026-05-02 | **Lance backend gauntlet (4-pack + root-cause fix) — DONE** | Lance crate had zero tests + no smoke when audited this morning. Shipped: (a) `sanitize_lance_err` over all 5 routes (search/doc/index/append/migrate) — missing-index now 404 not 500, no `/home/` or `/root/.cargo/` paths leaked; (b) 7 unit tests in `crates/vectord-lance` with synth Parquet helper; (c) 9-probe `scripts/lance_smoke.sh` against live `:3100`; (d) 10M re-bench (`reports/lance_10m_rebench_2026-05-02.md`) — search warm ~20ms, search cold ~46ms median. Bench surfaced doc-fetch p50 ~100ms (300x slower than ADR-019 100K projection); root-caused to lance-bench bypassing IndexMeta → warming auto-build never fired → no `doc_id` btree. **Fix shipped (commit `5d30b3d`)**: `lance_migrate` HTTP handler now auto-builds the btree inline (1.2s on 10M, +269MB), drops doc-fetch to ~5ms (20x). Live verified 9/9 smoke + post-restart doc-fetch 4-15ms. |
|
||||
| 2026-05-03 | **Subject manifests + per-subject HMAC audit log — DONE on Rust + Go** | Local-first compliance substrate per `lakehouse/docs/specs/SUBJECT_MANIFESTS_ON_CATALOGD.md` Steps 1-8. Rust shipped: `SubjectManifest` type + Registry CRUD (`crates/catalogd/src/registry.rs`), `SubjectAuditWriter` with HMAC-SHA256 chain + per-subject Mutex serialization + canonical-JSON via BTreeMap (`subject_audit.rs`), backfill ETL (`bin/backfill_subjects`), gateway tool dispatch + validator decorator wiring, legal-tier `/audit/subject/{id}` endpoint with constant-time-eq token + tampering detection, daily `bin/retention_sweep` (BIPA-aware, idempotent, no auto-mutation). Go shipped: identical `internal/catalogd/subject.go` reader + `VerifyChain` over RAW LINE BYTES (avoids time-precision drift), 11 unit tests. **6th cross-runtime parity probe**: `scripts/cutover/parity/subject_audit_parity.sh` — 6/6 byte-identical assertions across known-answer fixture + 5 real production audit logs. Surfaced + closed three drift classes in authoring loop: (1) Go `omitempty` stripping `trace_id:""`; (2) `time.RFC3339Nano` truncating trailing-zero nanoseconds where chrono AutoSi keeps 9 digits; (3) Go `json.Marshal` HTML-escaping `<>&` where serde keeps literal — fixed via `marshalNoEscapeHTML` + raw-bytes canonicalization. Two cross-lineage scrums caught real bugs each round (chain corruption race, schema-evolution HMAC drift, hardcoded "success" classifier, token min length, chain_root from windowed slice, tampering detection, HTML escape divergence). |
|
||||
| _open_ | Decide Lance vs Parquet+HNSW for primary | Lance verified production-ready at 10M (this morning's gauntlet). HNSW at 10M doesn't fit RAM (~60GB for vectors+graph), so the comparison is between Lance and Parquet+HNSW-with-spilling. Decide once we have a 10M ingest scenario where the Parquet path is bottlenecked. |
|
||||
| _open_ | Pick Go primary vs Rust primary | Both viable. Go has perf edge after today; Rust has production deploy + producer-side completeness. |
|
||||
|
||||
|
||||
Reference in New Issue
Block a user