Skip to content

chore(deps): Update rocksdb-js to 2.6.1 - #2036

Merged
kriszyp merged 2 commits into
mainfrom
update-rocksdb-js
Aug 1, 2026
Merged

chore(deps): Update rocksdb-js to 2.6.1#2036
kriszyp merged 2 commits into
mainfrom
update-rocksdb-js

Conversation

@harperfastbot

@harperfastbot harperfastbot Bot commented Jul 31, 2026

Copy link
Copy Markdown
Contributor

Update rocksdb-js to 2.6.1.

Release notes: https://github.com/HarperFast/rocksdb-js/releases/tag/v2.6.1

@socket-security

socket-security Bot commented Jul 31, 2026

Copy link
Copy Markdown

Review the following changes in direct dependencies. Learn more about Socket for GitHub.

Diff Package Supply Chain
Security
Vulnerability Quality Maintenance License
Updated@​harperfast/​rocksdb-js@​2.5.0 ⏵ 2.6.17710010099 +1100

View full report

Comment thread package.json
"@fastify/static": "^9.1.3",
"@harperfast/extended-iterable": "^1.0.1",
"@harperfast/rocksdb-js": "^2.5.0",
"@harperfast/rocksdb-js": "^2.6.1",

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

High: rocksdb-js 2.6.1 fixes the 2.6.0 table-open crash but introduces a new componentLoader regression — CI is red

2.6.1 does resolve the #2034 breakage: the empty compression object is now treated as unset instead of throwing (rocksdb-js#739), and table open works again (4218 unit tests pass).

However, this bump introduces a new, deterministic regression. Three unitTests/components/componentLoader.test.js cases fail identically on Node v22, v24, and v26 — scope.ensureTable should be exposed, should handle component loading errors gracefully, and the #1462 deploy:start listener-leak lifecycle — and Integration Tests 6/6 (Windows, Node.js v24) also fails.

This PR's base is exactly the current main tip (c0483f5), which is green on rocksdb-js 2.5.0; the only delta is this bump, so a 2.6.x behavior change is the cause. The most likely culprit is 2.6.0's Forbid stores from being shared (rocksdb-js#719) tripping componentLoader's repeated store opens during validation loads.

Do not merge until CI is green — this needs either a follow-up in rocksdb-js or a harper-side adaptation of component load / store-open to the 2.6.x semantics.


Generated by Barber AI

@cb1kenobi

Copy link
Copy Markdown
Member

The unit-test failures here are not flaky — rocksdb-js 2.6.x strictly validates the compression option, and Harper passes LMDB-era values (''/false when disabled) straight through to RocksDB opens. Without a fix, storage.compression: false instances would fail to open tables on 2.6.1. Fix (with details on the root cause) is in #2037 — once that merges, a rebase/rerun here should go green apart from the Windows EPERM: rename harper-config.yaml.tmp integration flake, which is unrelated.

— Claude (Fable 5)

@kriszyp
kriszyp merged commit b58db1b into main Aug 1, 2026
69 of 70 checks passed
@kriszyp
kriszyp deleted the update-rocksdb-js branch August 1, 2026 02:49
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.

2 participants