Skip to content

docs: 2026-07-02 deep review + ISSUES triage (I143–I160) - #90

Merged
Xof merged 1 commit into
mainfrom
docs/deepdive-review-20260702
Jul 2, 2026
Merged

docs: 2026-07-02 deep review + ISSUES triage (I143–I160)#90
Xof merged 1 commit into
mainfrom
docs/deepdive-review-20260702

Conversation

@Xof

@Xof Xof commented Jul 2, 2026

Copy link
Copy Markdown
Collaborator

Summary

Records the fresh-eyes Rust deep review run at 581be36 and triages its findings into ISSUES.md.

  • Report: docs/reviews/review-20260702-001902.md (full, canonical).
  • Triage: new Deep review 2026-07-02 section in ISSUES.md, I143–I160.

What the review found

Method: 8 category reviewers over the full engine + Python binding + bench harness; every BUG/DESIGN finding put through an independent adversarial verifier before inclusion.

  • 18 findings confirmed, 0 refuted — 4 BUG, 13 DESIGN, 1 SMELL (plus 45 SMELL/NIT and 21 open questions in the report).
  • Delta vs the prior two reviews: 0 regressions, 71 findings verified resolved. The recent reviews were acted on almost in full.
  • Declined/deferred items (I141 StagingTxn, I110 MSRV --tests, wheel-gate release-only, etc.) were not re-filed.

Disposition

  • I143–I146 (the BUGs) are fixed in fix: four BUG-severity findings from the 2026-07-02 deep review (I143–I146) #89 (TOCTOU open race, unvalidated stride panic, freemap poison-wrap, README contract) and marked FIXED here.
  • I147–I160 (DESIGN/SMELL) are recorded OPEN for your triage. The themes worth a look: corrupt-but-checksummed pages fail open in the membership/handle-table radixes (vs closed elsewhere), an unauthenticated Argon2-param DoS, per-page temporal replay, the encrypted-series minor write-gate that silently fails on the first (2,1) file, Python finished-transaction aliasing, and the missing encrypted × on-disk-spillway integration test.

Docs only; no code change.

…ES.md

Adds the fresh-eyes review report (docs/reviews/review-20260702-001902.md)
and triages its verified findings as I143–I160 under a new "Deep review
2026-07-02" section in ISSUES.md.

- 18 findings confirmed by adversarial verification (0 refuted): 4 BUG,
  13 DESIGN, 1 SMELL. Prior-review delta: 0 regressions, 71 resolved.
- I143–I146 (the BUGs) are marked FIXED (PR #89).
- I147–I160 (DESIGN/SMELL) recorded OPEN for triage — corrupt-page
  hardening asymmetry, KDF-param DoS, per-page replay, the encrypted-minor
  write-gate off-by-series, Python finished-transaction aliasing, and the
  encrypted on-disk-spillway test gap are the notable ones.

Docs only; no code change.
@Xof
Xof merged commit c6ade70 into main Jul 2, 2026
10 checks passed
@Xof
Xof deleted the docs/deepdive-review-20260702 branch July 2, 2026 01:29
@github-actions

github-actions Bot commented Jul 2, 2026

Copy link
Copy Markdown

🚦 Bench results: PR vs main

⚠️ 1 regression(s) detected across 1 scenario/mode pair(s)

Scenario Mode Δ throughput Worst Δ
ycsb-a sqlite-strict -1.2% p99 +14.0% ⚠️
document-store chisel-mem +1.1%
document-store chisel-strict +6.3%
document-store redb-strict +3.4%
document-store sqlite-strict +0.3%
mutation-log chisel-mem -0.8%
mutation-log chisel-strict +0.1%
mutation-log redb-strict +0.1%
mutation-log sqlite-strict +1.3%
ycsb-a chisel-mem +20.0%
ycsb-a chisel-strict +2.7%
ycsb-a redb-strict +0.1%
ycsb-b chisel-mem -0.2%
ycsb-b chisel-strict +3.7%
ycsb-b redb-strict +0.7%
ycsb-b sqlite-strict -0.9%
Per-scenario detail (4 metrics × cells)

document-store

Mode Throughput p50 p95 p99
chisel-mem 32477 ops/s → 32841 ops/s (+1.1%) 5.3 µs → 5.2 µs (-1.6%) 69.2 µs → 68.7 µs (-0.7%) 305.9 µs → 301.6 µs (-1.4%)
chisel-strict 2812 ops/s → 2990 ops/s (+6.3%) 13.1 µs → 12.7 µs (-3.2%) 1.10 ms → 973.2 µs (-11.2%) 2.49 ms → 2.18 ms (-12.3%)
redb-strict 3996 ops/s → 4134 ops/s (+3.4%) 18.5 µs → 18.3 µs (-1.3%) 608.2 µs → 542.9 µs (-10.7%) 1.59 ms → 1.47 ms (-7.1%)
sqlite-strict 5102 ops/s → 5116 ops/s (+0.3%) 21.6 µs → 21.5 µs (-0.4%) 417.6 µs → 407.2 µs (-2.5%) 1.35 ms → 1.34 ms (-1.1%)

mutation-log

Mode Throughput p50 p95 p99
chisel-mem 49058 ops/s → 48653 ops/s (-0.8%) 26.0 µs → 26.2 µs (+0.9%) 29.0 µs → 29.6 µs (+1.8%) 39.0 µs → 39.1 µs (+0.3%)
chisel-strict 1709 ops/s → 1711 ops/s (+0.1%) 391.9 µs → 403.3 µs (+2.9%) 862.8 µs → 827.9 µs (-4.0%) 11.17 ms → 10.61 ms (-5.0%)
redb-strict 1805 ops/s → 1807 ops/s (+0.1%) 179.0 µs → 180.3 µs (+0.7%) 271.6 µs → 270.9 µs (-0.2%) 31.79 ms → 31.44 ms (-1.1%)
sqlite-strict 4734 ops/s → 4793 ops/s (+1.3%) 123.7 µs → 123.8 µs (+0.0%) 336.0 µs → 336.8 µs (+0.2%) 470.3 µs → 456.5 µs (-2.9%)

ycsb-a

Mode Throughput p50 p95 p99
chisel-mem 35586 ops/s → 42694 ops/s (+20.0%) 48.1 µs → 39.3 µs (-18.2%) 62.6 µs → 53.7 µs (-14.2%) 88.0 µs → 70.2 µs (-20.3%)
chisel-strict 2129 ops/s → 2188 ops/s (+2.7%) 328.4 µs → 325.2 µs (-1.0%) 959.7 µs → 945.2 µs (-1.5%) 2.30 ms → 2.32 ms (+0.9%)
redb-strict 2680 ops/s → 2683 ops/s (+0.1%) 147.5 µs → 148.2 µs (+0.5%) 243.0 µs → 243.7 µs (+0.3%) 376.0 µs → 373.4 µs (-0.7%)
sqlite-strict 148147 ops/s → 146384 ops/s (-1.2%) 7.0 µs → 7.0 µs (+0.1%) 8.7 µs → 8.8 µs (+1.5%) 10.1 µs → 11.5 µs (+14.0%) ⚠️

ycsb-b

Mode Throughput p50 p95 p99
chisel-mem 260930 ops/s → 260340 ops/s (-0.2%) 1.4 µs → 1.4 µs (+0.7%) 48.4 µs → 48.3 µs (-0.3%) 52.1 µs → 52.0 µs (-0.1%)
chisel-strict 20563 ops/s → 21329 ops/s (+3.7%) 2.5 µs → 2.5 µs (-0.8%) 333.5 µs → 333.8 µs (+0.1%) 875.7 µs → 838.5 µs (-4.2%)
redb-strict 26898 ops/s → 27086 ops/s (+0.7%) 2.6 µs → 2.5 µs (-1.6%) 144.8 µs → 144.8 µs (+0.1%) 225.9 µs → 225.9 µs (+0.0%)
sqlite-strict 180309 ops/s → 178761 ops/s (-0.9%) 5.6 µs → 5.6 µs (+1.1%) 7.5 µs → 7.5 µs (-0.8%) 9.8 µs → 9.7 µs (-0.8%)
Generated by chisel-bench-diff at 2026-07-02T01:41:27Z. Compares PR HEAD against main. Never blocks merge — signal, not gate. Thresholds: throughput 5%, p50 5%, p95 10%, p99 10%.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant