GeezerVison is showing github.com — chore(release): 5.1.0 by mikolalysenko · Pull Request #2032 · SocketDev/socket-patch
Conversation
Release 5.1.0
Version: 5.1.0 (minor). The notes have an ### Added section (agent-mode gem native-extension rebuild) and no Breaking or Removed entries. Per the CHANGELOG header rule, that makes this a minor release.
Covered range: v5.0.1..7bde3ba55a84e64281372ebc4fd53e7fab0e35fd: 50 first-parent PRs. That is far fewer than the ~300 commits the routine expected; the count is from git log --first-parent.
This PR changes only CHANGELOG.md and the files scripts/version-sync.sh 5.1.0 stamps (Cargo.toml, Cargo.lock, the 15 npm package.json files, the npm lockfile). Local checks: release-lint.sh --stable-only --tag-check passed, release.py changelog check --version 5.1.0 passed, and cargo check --locked passed.
Headline: Ruby native-extension patching (#1996, #1998)
When agent-mode apply (or scan/get --mode agent) patches a gem whose patch changes native-extension sources (ext/**/*.c, extconf.rb, …), it now rebuilds the extensions with RubyGems' own builder (the gem pristine --extensions step). It uses the Ruby that owns the gem home and rebuilds both the extensions/<platform>/<abi>/ copy and the lib/ copy, so Ruby stops loading the stale install-time build.
- If the build fails, the original sources and build are restored (
native_extension_build_failed).
- A copy that can't be rebuilt in place is refused before anything is written (
native_extensions_unsupported).
vex and apply --check attest the patch only once the rebuild is recorded; otherwise they report native_extension_not_rebuilt.rollback/remove restore the original build.--no-gem-native-rebuild (env SOCKET_NO_GEM_NATIVE_REBUILD) opts out for callers that compile the extension themselves.
These claims were checked against CLI_CONTRACT.md and args.rs.
Left out of the notes
Every merged PR is referenced, but these appear only in two condensed Maintenance bullets because they don't change user-visible behavior:
For a maintainer to double-check
Maintainer checklist
- Approve this PR.
- After merge, confirm nothing else landed:
git log --first-parent 7bde3ba55a84e64281372ebc4fd53e7fab0e35fd..origin/main (if something did, add notes first).
- Actions → Release → Run workflow on
main (optionally a dry-run: true run first).
- Approve the npm stages with 2FA: the 14 platform packages first,
@socketsecurity/socket-patch last.
Generated by a one-off release-staging routine.
Generated by Claude Code
Note Low Risk
Version and changelog-only release prep; no runtime code changes in this diff.
Overview
Cuts socket-patch 5.1.0: bumps the workspace and crate versions in Cargo.toml / Cargo.lock, and stamps 5.1.0 on the main npm package, all platform optionalDependencies, and package-lock.json (lockfile trimmed to root + @types/node only).
Documents the release in CHANGELOG.md with a new ## [5.1.0] — 2026-11-06 section: Added highlights agent-mode RubyGems rebuilds of native extensions (gem pristine–style) with new warnings/codes and --no-gem-native-rebuild; Fixed spans Bundler 4.1 overrides, gemspec parsing, pnpm/yarn/Bun/npm edge cases, PyPI/Pipenv/Poetry, Go replace correctness, Maven 4 implicit reactor, vendoring dry runs and refusals, and misc download/remedy behavior; Maintenance notes internal consolidation and CI/test work.
Reviewed by Cursor Bugbot for commit c6ef370. Configure here.
Generated by Claude Code
Sorry, something went wrong.
Uh oh!
There was an error while loading. Please reload this page.
Roll the notes for the 50 PRs merged since v5.0.1 into a 5.1.0 section
and stamp 5.1.0 into every packaging site. Minor bump: gem native
extension rebuild in agent mode is new behavior (#1996, #1998).
Co-Authored-By: Claude <[email protected]>
Sorry, something went wrong.
Uh oh!
There was an error while loading. Please reload this page.
Sorry, something went wrong.
Uh oh!
There was an error while loading. Please reload this page.
|
CI red that isn't caused by this PR: the e2e (ubuntu-latest, e2e_vex_build, … pdm:: --ignored …) check fails.
pdm lock crashes with [AttributeError]: module 'hishel' has no attribute 'SyncSqliteStorage'. A new hishel 2.x release breaks PDM, and both pdm::*_install_then_manifest_less_vex tests fail with it. This PR changes only CHANGELOG.md and the version stamps, so it can't cause this, and PDM's dependencies aren't pinned, so re-running the job would fail the same way. The same failure is red on other PRs right now.
The fix is open as #2028 (ci-janitor/pdm-hishel-2, pins hishel<2). This release PR is limited to the changelog and version files, so I haven't ported that change here. Once #2028 merges, merge main into release/v5.1.0 (or update the branch) so CI runs again. Auto-merge is still armed.
Generated by a one-off release-staging routine.
Generated by Claude Code
|
Sorry, something went wrong.
Uh oh!
There was an error while loading. Please reload this page.
|
Labeled Ready for review by the burn-down agent.
- Head:
c6ef370ca6ca49ad5b1b4ba66fa3287a73bbad17
- CI: 48/50 check runs green (29 success, 19 skipped). The two red ones come from a failure that also hits main:
e2e (ubuntu-latest, e2e_vex_build, … pdm:: --ignored …) and the ci-ok aggregate that depends on it. pdm lock crashes because hishel 2.x removed SyncSqliteStorage. The pdm e2e is red on main too (e2e-macos … pdm:: at 7bde3ba). Fix pdm e2e red on every PR: pin hishel<2 #2028 fixes it.
- Bugbot: reviewed this head with no findings.
- For the reviewer: it changes only
CHANGELOG.md, the version stamps and the lockfiles. Auto-merge is still armed. If you want this PR fully green before you approve, merge Fix pdm e2e red on every PR: pin hishel<2 #2028 first and update this branch.
Generated by Claude Code
|
Sorry, something went wrong.
Uh oh!
There was an error while loading. Please reload this page.
Final review brief (for the human approver)
What it does: cuts release 5.1.0. It adds a ## [5.1.0] section to CHANGELOG.md and stamps 5.1.0 into Cargo.toml, Cargo.lock and the 15 npm package.json files plus the npm lockfile. No Rust or JS code changes.
Risk: low. The diff touches only version stamps and the changelog. The real risk is in the release workflow you run after merging, not in this diff.
Look here
Verified
- Base is current main (7bde3ba, Rebuild gem native extensions in agent-mode apply (#1996) #1998). No commits have landed on main since then.
- Every PR merged on main since 5.0.1 (in the history I checked) is referenced in the notes. The only unreferenced numbers were issue numbers whose PRs are listed (Drive pdm.lock and poetry.lock rewrites through one lock_fragments driver instead of two mirrored triads #1844→Share one lock_fragments driver for PDM and Poetry (#1844) #1898, Share one scoped env-var test guard: EnvGuard is copy-pasted 11 times across core unit and integration tests #1889→Build: share one test env guard in core (#1889) #1891, Build every tests/apply fixture through apply_fixture: 9 suites in the apply binary re-write the manifest, blob, package and event-lookup helpers it already has #1941→Build tests/apply fixtures through apply_fixture (#1941) #1944, Run PTY test children through one common::pty_io::run: the openpty-spawn-feed-collect runner is copied 7×, and only 2 copies pin TERM #1942→Run PTY test children through one pty_io::run (#1942) #1955).
- Every version stamp is consistent at 5.1.0. CHANGELOG changes are allowed here because this is a
release/v* branch.
- CI: 48/50 green. The two red checks are the pdm e2e leg and the
ci-ok aggregate that depends on it. That failure comes from hishel 2.x breaking pdm lock on every branch, main included, and it isn't caused by this PR. The fix is Fix pdm e2e red on every PR: pin hishel<2 #2028.
- Bugbot: reviewed this head, no findings.
Changes I made: none.
Open questions: none blocking. Merge #2028 first so this PR (and the rest of the open PRs) can go green. Auto-merge isn't armed: this PR doesn't qualify until ci-ok is green.
Generated by Claude Code
|
Sorry, something went wrong.
Uh oh!
There was an error while loading. Please reload this page.
The release PR's only red check was the pdm e2e that failed main-wide;
#2028 fixed it on main.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
|
Final reviewer: merged main (d9c57ae) to pick up the pdm e2e fix from #2028, the only red check. Approval stands; auto-merge is armed, so this enters the merge queue once ci-ok is green.
Generated by Claude Code
|
Sorry, something went wrong.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
This was referenced Nov 6, 2026
Uh oh!
There was an error while loading. Please reload this page.
Assignees
No one assigned
Labels
Ready for review
Agent-verified: mergeable, CI green, Bugbot clean — awaiting human review
Milestone
No milestone
Development
Successfully merging this pull request may close these issues.
Uh oh!
There was an error while loading. Please reload this page.
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
|