Implement uv_thread_self on POSIX by robobun · Pull Request #29228 · oven-sh/bun

robobun · 2026-04-12T14:04:52Z

What

Implement uv_thread_self on POSIX (Linux/macOS) as a one-line pthread_self() wrapper.

Why

ffi-napi's prebuilt NAPI addon calls uv_thread_self() during module initialization. Bun used to panic with unsupported uv function: uv_thread_self because the symbol was a stubbed libuv function on POSIX — so require('ffi-napi') crashed before any FFI call was made. The issue title says "empty structure" but the underlying cause has nothing to do with FFI structures.

Reproducer from #29223

import ffi from 'ffi-napi';
import ref from 'ref-napi';

const lib = ffi.Library('./lib', { f: [ 'void', [] ] });
lib.f();

Pre-fix output

Bun encountered a crash when running a NAPI module that tried to call
the uv_thread_self libuv function.

panic(main thread): unsupported uv function: uv_thread_self

How

Add uv_thread_self to src/bun.js/bindings/uv-posix-polyfills.c as return pthread_self();. Matches upstream libuv src/unix/thread.c. uv_thread_t is typedefed to pthread_t on POSIX (see src/bun.js/bindings/libuv/uv/unix.h).
Remove uv_thread_self from generate_uv_posix_stubs_constants.ts's symbols[] list so the generator no longer emits a stub for it, alongside the other real polyfills (uv_mutex_*, uv_once, uv_hrtime, uv_os_getpid, …).
Delete the corresponding stub body from uv-posix-stubs.c and the matching test-plugin call from test/napi/uv-stub-stuff/plugin.c (both are generated from the same constants file).

Tests

test/regression/issue/29223.test.ts: builds a minimal NAPI addon that calls uv_thread_self() from its Init function (the exact shape of the ffi-napi crash) and asserts the addon loads cleanly, uv_thread_self is consistent with pthread_self, and Bun exits 0.
test/napi/uv.test.ts: new testThreadSelf case in the shared uv_impl.c NAPI plugin. Verifies two successive uv_thread_self() calls refer to the same pthread via pthread_equal.

Gate

Without the fix (src stashed, tests applied): the regression test fails with Received: "… panic(main thread): unsupported uv function: uv_thread_self …".
With the fix: passes in ~1s.

Fixes #29223

ffi-napi's NAPI addon calls uv_thread_self() during module init. Bun panicked with "unsupported uv function: uv_thread_self" because the symbol was a stubbed libuv function on POSIX, so require('ffi-napi') crashed before any FFI call was made. Implement it as a pthread_self() wrapper, matching upstream libuv src/unix/thread.c. uv_thread_t is pthread_t on POSIX. Fixes #29223

robobun · 2026-04-12T14:05:03Z

^{Updated 8:24 PM PT - Apr 12th, 2026}

❌ @robobun, your commit 2b92865 has 2 failures in Build #45395 (All Failures):

test/js/bun/http/tls-keepalive.test.ts - code 1 on 🍎 13 aarch64
vendor/elysia/test/standard-schema/standalone.test.ts - pid 6495 error: addresssanitizer on 🐧 13 x64-asan

🧪 To try this PR locally:

bunx bun-pr 29228

That installs a local version of the PR into your bun-29228 executable, so you can run:

bun-29228 --bun

github-actions · 2026-04-12T14:07:30Z

Found 2 issues this PR may fix:

Implement libuv functions #18546 - Umbrella tracking issue for implementing missing libuv functions on POSIX; this PR adds uv_thread_self which directly contributes to that effort
TypeError: undefined is not a function (near '...this.TLSClient.request...') #5583 - Reporter uses ffi-napi which crashes due to uv_thread_self being unimplemented; this PR fixes that crash path

If this is helpful, copy the block below into the PR description to auto-close these issues on merge.

Fixes #18546
Fixes #5583

🤖 Generated with Claude Code

github-actions · 2026-04-12T14:09:39Z

This PR may be a duplicate of:

napi: implement simple libuv functions on posix #28900 - Also implements uv_thread_self as a pthread_self() wrapper (as part of a larger set of ~46 libuv function implementations on POSIX), touching the same files and removing the same stub

🤖 Generated with Claude Code

coderabbitai · 2026-04-12T14:10:09Z

Note

Reviews paused

It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the reviews.auto_review.auto_pause_after_reviewed_commits setting.

Use the following commands to manage reviews:

@coderabbitai resume to resume automatic reviews.
@coderabbitai review to trigger a single review.

Use the checkboxes below for quick actions:

▶️ Resume reviews
🔍 Trigger review

Walkthrough

Added a POSIX implementation of uv_thread_self, removed its stub and symbol export, updated generator comments to reference .c, and added N-API + regression tests that call uv_thread_self during module initialization to validate behavior on POSIX systems.

Changes

Cohort / File(s)	Summary
Polyfills / Stubs `src/bun.js/bindings/uv-posix-polyfills.c`, `src/bun.js/bindings/uv-posix-stubs.c`	Added `UV_EXTERN uv_thread_t uv_thread_self(void)` to `uv-posix-polyfills.c`; removed the stubbed `uv_thread_self` implementation from `uv-posix-stubs.c`.
Stub generator `src/bun.js/bindings/libuv/generate_uv_posix_stubs_constants.ts`	Updated comment text from `uv-posix-polyfills.cpp` → `uv-posix-polyfills.c` for several commented symbols; removed `"uv_thread_self"` from the active `symbols` export (left commented with note).
N-API test module `test/napi/uv-stub-stuff/plugin.c`, `test/napi/uv-stub-stuff/uv_impl.c`	Removed `uv_thread_self` dispatch branch in `plugin.c`; added `test_thread_self` in `uv_impl.c` that calls `uv_thread_self()` twice, compares results with `pthread_equal`, and registers `testThreadSelf`.
Tests `test/napi/uv.test.ts`, `test/regression/issue/29223.test.ts`	Added non-Windows unit test invoking `nativeModule.testThreadSelf()` and a POSIX-only regression test that builds a N-API addon calling `uv_thread_self()` during `Init`, asserting the addon loads and prints `OK`.

🚥 Pre-merge checks | ✅ 4

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and specifically describes the main change: implementing uv_thread_self on POSIX platforms.
Description check	✅ Passed	The description thoroughly covers both required sections: 'What does this PR do?' (with detailed implementation and rationale) and 'How did you verify your code works?' (with comprehensive test details).
Linked Issues check	✅ Passed	The PR implements the core requirement from `#29223`: providing a correct POSIX implementation of uv_thread_self to resolve the crash when requiring ffi-napi, with comprehensive tests verifying the implementation.
Out of Scope Changes check	✅ Passed	All changes directly support the primary objective of implementing uv_thread_self on POSIX. The test fixture binding.gyp modification (removing -Wl,--export-dynamic and widening timeouts) is a necessary supporting change.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

coderabbitai

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@test/regression/issue/29223.test.ts`:
- Around line 100-107: Remove the redundant negative panic-string assertion:
delete the line that asserts expect(stderrStr).not.toContain("unsupported uv
function: uv_thread_self") in the test (leaving the const stdoutStr/stderrStr
declarations if still used) so the test only verifies stdoutStr.trim() === "OK"
and exitCode === 0 (i.e., keep the expect(stdoutStr.trim()).toBe("OK") and
expect(exitCode).toBe(0) assertions).

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

Push a commit to this branch (recommended)
Create a new PR with the fixes

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: edd12ede-6729-4a01-accc-4c306d515b70

📥 Commits

Reviewing files that changed from the base of the PR and between 7c75a87 and 73a186b.

📒 Files selected for processing (7)

src/bun.js/bindings/libuv/generate_uv_posix_stubs_constants.ts
src/bun.js/bindings/uv-posix-polyfills.c
src/bun.js/bindings/uv-posix-stubs.c
test/napi/uv-stub-stuff/plugin.c
test/napi/uv-stub-stuff/uv_impl.c
test/napi/uv.test.ts
test/regression/issue/29223.test.ts

💤 Files with no reviewable changes (2)

test/napi/uv-stub-stuff/plugin.c
src/bun.js/bindings/uv-posix-stubs.c

test/regression/issue/29223.test.ts

claude

LGTM — trivially correct one-liner matching upstream libuv, with good test coverage.

Extended reasoning...

Overview

The PR implements uv_thread_self on POSIX by adding a one-line return pthread_self() wrapper to uv-posix-polyfills.c, removes the panic stub from uv-posix-stubs.c, comments it out in the generator constants list (following the same pattern as uv_hrtime, uv_mutex_*, uv_once, uv_os_getpid, etc.), removes the dead test-plugin branch from plugin.c, and adds a regression test plus a unit test.

Security risks

None. This is a POSIX thread-identity function with no auth, crypto, or permission implications.

Level of scrutiny

Low. The implementation is a single-expression function that matches upstream libuv src/unix/thread.c exactly. uv_thread_t is typedefed to pthread_t on POSIX, so there is nothing to get wrong. The bug-hunter reports are all documentation nits (wrong file extension in a comment, a misleading comment referencing uv_thread_equal instead of pthread_equal, a redundant negative stderr assertion, and an explicit beforeAll timeout) — none affect correctness.

Other factors

The change follows an established pattern in the codebase. Test coverage is solid: two independent test paths (inline NAPI plugin and standalone regression binary) both exercise the new function. The change is self-contained and does not touch security-sensitive code.

src/bun.js/bindings/libuv/generate_uv_posix_stubs_constants.ts

test/regression/issue/29223.test.ts

test/napi/uv-stub-stuff/uv_impl.c

test/regression/issue/29223.test.ts

- Drop negative panic-string assertion from the regression test; the stdout == 'OK' + exitCode == 0 checks already fully cover the regression and the negative check violated the 'no absence-of-panic' test rule in CLAUDE.md. - Drop the explicit 120_000 timeout on beforeAll to match the sibling test/napi/uv.test.ts pattern. - Fix the .cpp typo in the new constants-file comment (file is .c). - Fix a stale test comment that mentioned uv_thread_equal; the code uses pthread_equal.

claude

All prior feedback has been addressed. LGTM.

Extended reasoning...

Overview

The PR implements uv_thread_self on POSIX as a one-line pthread_self() wrapper in uv-posix-polyfills.c, removes the crashing stub from uv-posix-stubs.c, and updates the constants file and generated test plugin accordingly. It adds two test paths: a unit test via the existing uv_impl.c NAPI plugin, and a regression test that builds and runs a standalone addon calling uv_thread_self during Init.

Security risks

None. pthread_self() is a trivial thread identity function with no security implications. The only new code in production is a single-line return statement.

Level of scrutiny

Low. The implementation is a mechanical one-liner matching upstream libuv (src/unix/thread.c), and uv_thread_t is already typedef-d to pthread_t on POSIX. The stub removal and symbol list update follow the exact same pattern as the other already-implemented polyfills.

Other factors

My four prior inline comments (wrong .cpp extension in comments, unnecessary 120s beforeAll timeout, misleading uv_thread_equal comment, and the negative panic assertion) were all addressed in commit 7de5514. The one remaining nit — unnecessary async on a synchronous test callback — is already captured as an inline comment and does not affect correctness.

test/regression/issue/29223.test.ts

coderabbitai

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@test/regression/issue/29223.test.ts`:
- Around line 79-80: The test calls tempDirWithFiles("issue-29223", files) which
already creates and populates the tempdir, so the subsequent await
makeTree(tempdir, files) is redundant and should be removed; locate the test
code using tempDirWithFiles and makeTree (variables tempdir and files, functions
tempDirWithFiles and makeTree) and delete the extra makeTree(...) invocation so
the directory is only created/populated once.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

Push a commit to this branch (recommended)
Create a new PR with the fixes

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: 65db2615-a4ff-464c-be61-67331d7a1c21

📥 Commits

Reviewing files that changed from the base of the PR and between c094ab1 and 409b49e.

📒 Files selected for processing (1)

test/regression/issue/29223.test.ts

test/regression/issue/29223.test.ts

claude

LGTM — all prior review feedback addressed; the verbose comment block nit is flagged inline.

Extended reasoning...

Overview

This PR adds a real POSIX implementation of uv_thread_self() as a one-line pthread_self() wrapper in src/bun.js/bindings/uv-posix-polyfills.c, removes its stub from uv-posix-stubs.c and the stub-generator constants, and adds tests in both test/napi/uv.test.ts and a new regression file test/regression/issue/29223.test.ts. Seven files are modified.

Security risks

None. The change is a trivial wrapper (return pthread_self()) that matches the upstream libuv implementation. No auth, crypto, or permissions code is touched.

Level of scrutiny

Low. The core implementation is a single-line function that delegates directly to a POSIX primitive. The pattern is established by the existing polyfills (uv_hrtime, uv_mutex_*, uv_once, etc.). Removing the stub from the generator constants follows the exact same pattern as those prior implementations.

Other factors

All issues raised in my previous review passes have been addressed by the author: the no-panic assertion was removed, the beforeAll timeout was dropped, all six .cpp→.c comment corrections were applied, the misleading uv_thread_equal reference in the comment was fixed, and the spurious async keyword was removed. The one remaining item — the 13-line prose comment block at the top of the regression test — is a style nit flagged in the inline comment. The core correctness of the fix is not in question.

test/regression/issue/29223.test.ts

coderabbitai · 2026-04-12T15:44:25Z

✅ No merge conflicts detected when merging into main.

Your branch is good to go!

coderabbitai

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@test/regression/issue/29223.test.ts`:
- Around line 46-54: The test fails on macOS because binding.gyp contains the
GNU-specific linker flag "-Wl,--export-dynamic" (seen in the ldflags array)
which Apple ld does not accept; remove that flag from the ldflags entry in
binding.gyp or make its inclusion conditional only for non-Darwin platforms
(e.g., guard the addition when process.platform !== 'darwin') so the
describe.if(!isWindows) test can build successfully on macOS.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

Push a commit to this branch (recommended)
Create a new PR with the fixes

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: 2513f54f-6d12-44a8-8e55-7cc48000dd55

📥 Commits

Reviewing files that changed from the base of the PR and between 409b49e and 3db8f82.

📒 Files selected for processing (1)

test/regression/issue/29223.test.ts

test/regression/issue/29223.test.ts

claude

LGTM — all feedback addressed.

Extended reasoning...

Overview

The PR adds a one-line POSIX implementation of uv_thread_self (return pthread_self();) in uv-posix-polyfills.c, removes the corresponding panic stub from uv-posix-stubs.c and the stub generator's symbol list, and adds regression + integration tests. Seven files changed, all tightly scoped to the single missing function.

Security risks

None. This is a read-only thread-identity query (pthread_self()) with no allocation, no synchronization, and no capability escalation.

Level of scrutiny

Low — the change is a single-line wrapper that directly matches upstream libuv (src/unix/thread.c). uv_thread_t is typedefed to pthread_t on POSIX (confirmed in uv/unix.h), so the return type is exact.

Other factors

All four rounds of review feedback were addressed in follow-up commits: the .cpp → .c comment typos (all six occurrences), the prohibited negative panic assertion, the extraneous async keyword, the explicit beforeAll timeout, the inaccurate comment citing uv_thread_equal, the redundant makeTree call, and the verbose prose header in the regression test file. No bugs were found by the automated system. The note about a potential duplicate PR (#28900) is a process concern rather than a code-correctness issue.

darwin-14-x64-test-bun failed on CI. Two plausible causes that this commit removes: 1. `ldflags: ['-Wl,--export-dynamic']` in binding.gyp — GNU-only flag. node-gyp's gyp generator ignores the `ldflags` key entirely when flavor == mac (it only reads `xcode_settings.OTHER_LDFLAGS`), so this is a no-op on darwin at best and an 'unknown option' from ld64 at worst depending on toolchain. Tiny NAPI modules don't need `--export-dynamic` anyway (that's for main executables exposing symbols to dlopen'd plugins, the opposite direction). 2. beforeAll default 5s timeout — cold `bun install` + node-gyp configure/build can easily exceed this on CI. Widen to 60s, matching the pattern used by other fixture-building regression tests.

test/napi/uv-stub-stuff/uv_impl.c

test/regression/issue/29223.test.ts

- uv_impl.c test_thread_self: also compare against pthread_self() so a consistent-but-wrong constant would still fail. - 29223.test.ts: drop the 60s beforeAll timeout to match the sibling uv.test.ts; darwin CI failure on the prior commit turned out to be a canceled-build status rather than an actual timeout.

claude

LGTM — all prior review feedback addressed.

Extended reasoning...

Overview

The PR adds a one-line POSIX implementation of uv_thread_self (returning pthread_self()), removes the stub that previously panicked, updates the generator constants file, and adds both a unit test in test/napi/uv.test.ts and a regression test in test/regression/issue/29223.test.ts.

Security risks

None. This is a trivial wrapper around a standard POSIX syscall with no new attack surface.

Level of scrutiny

Low. The change is a single-line C function matching upstream libuv verbatim. uv_thread_t is typedefed to pthread_t on POSIX (confirmed in uv/unix.h), so return pthread_self() is the correct and complete implementation.

Other factors

Multiple rounds of review caught and addressed: the negative panic-string assertion, the missing pthread_self() comparison in the unit test, the spurious async keyword, the redundant makeTree call, the prose comment block, the .cpp vs .c typos in the generator file, and the beforeAll timeout. The preloaded PR timeline confirms all inline comments are resolved. No bugs were found by the automated bug hunting system.

claude

LGTM — all prior feedback addressed.

Extended reasoning...

Overview

Adds uv_thread_self to uv-posix-polyfills.c as a one-line pthread_self() wrapper, removes it from the stub generator constants and uv-posix-stubs.c, and adds both a unit test in the existing NAPI test module and a standalone regression test for #29223.

Security risks

None. pthread_self() is a straightforward read of the calling thread's ID with no security implications.

Level of scrutiny

Low. This is a single-line polyfill that exactly matches upstream libuv's src/unix/thread.c implementation. The change is mechanical and self-contained: add the real function, remove the panic stub, update the generator comment.

Other factors

All prior review feedback (comment text accuracy, test hygiene, timeout conventions, unnecessary async, weakened thread-identity assertion) was addressed across four follow-up commits. No bugs were found by the automated bug-hunting pass. The fix directly resolves a crash-on-require for ffi-napi users.

robobun · 2026-04-13T03:24:24Z

Handing off to maintainers. The core fix is in place and all review threads are resolved. CI has been red on debian-13-x64-asan-test-bun across the last few commits, but with a different unrelated vendored test failing each time:

d1aa0fba → test/js/node/test/parallel/test-cluster-primary-kill.js timeout
2b928658 → vendor/elysia/test/standard-schema/standalone.test.ts ASAN error

Same lane, same 1-of-9-shards pattern, different third-party test each run. Neither test has any plausible causal connection to a one-line pthread_self() polyfill or the NAPI thread-self fixture — it looks like a pre-existing flaky shard on that lane. Local gate check passes both ways (fails on baseline with panic(main thread): unsupported uv function: uv_thread_self, passes with the fix in ~1s).

robobun · 2026-04-13T14:00:01Z

Superseded by #29261 which implements the full uv_thread_* polyfill family (_self/_equal/_join/_detach/_create/_create_ex) with regression test + nap addon coverage.

robobun mentioned this pull request Apr 12, 2026

Crash: ffi-napi call triggers unsupported libuv function (uv_thread_self) and aborts Bun #29223

Open

github-actions bot added the claude label Apr 12, 2026

coderabbitai bot reviewed Apr 12, 2026

View reviewed changes

test/regression/issue/29223.test.ts Outdated Show resolved Hide resolved

claude bot reviewed Apr 12, 2026

View reviewed changes

robobun added 2 commits April 12, 2026 14:49

fix remaining uv-posix-polyfills.cpp → .c typos in comment block

c094ab1

claude bot reviewed Apr 12, 2026

View reviewed changes

test/regression/issue/29223.test.ts Outdated Show resolved Hide resolved

drop unused async from sync regression test callback

409b49e

coderabbitai bot reviewed Apr 12, 2026

View reviewed changes

test/regression/issue/29223.test.ts Outdated Show resolved Hide resolved

claude bot reviewed Apr 12, 2026

View reviewed changes

test/regression/issue/29223.test.ts Outdated Show resolved Hide resolved

trim regression test: drop prose comment block and redundant makeTree

3db8f82

coderabbitai bot reviewed Apr 12, 2026

View reviewed changes

test/regression/issue/29223.test.ts Show resolved Hide resolved

claude bot reviewed Apr 12, 2026

View reviewed changes

test/napi/uv-stub-stuff/uv_impl.c Show resolved Hide resolved

test/regression/issue/29223.test.ts Outdated Show resolved Hide resolved

claude bot reviewed Apr 12, 2026

View reviewed changes

This was referenced Apr 12, 2026

inspect: show real class name for DOM constructors, not [class Function] #29229

Merged

Fix dynamic import() resolving before TLA module finishes evaluating #29224

Open

ci: retrigger

2b92865

claude bot reviewed Apr 12, 2026

View reviewed changes

github-actions bot mentioned this pull request Apr 13, 2026

napi: polyfill uv_thread_* against pthread #29261

Open

robobun closed this Apr 13, 2026

Conversation

robobun commented Apr 12, 2026

What

Why

Reproducer from #29223

Pre-fix output

How

Tests

Gate

Uh oh!

robobun commented Apr 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Apr 12, 2026

Uh oh!

github-actions bot commented Apr 12, 2026

Uh oh!

coderabbitai bot commented Apr 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Reviews paused

Walkthrough

Changes

Uh oh!

coderabbitai bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

claude bot left a comment

Choose a reason for hiding this comment

Overview

Security risks

Level of scrutiny

Other factors

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

claude bot left a comment

Choose a reason for hiding this comment

Overview

Security risks

Level of scrutiny

Other factors

Uh oh!

Uh oh!

coderabbitai bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

claude bot left a comment

Choose a reason for hiding this comment

Overview

Security risks

Level of scrutiny

Other factors

Uh oh!

Uh oh!

coderabbitai bot commented Apr 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

coderabbitai bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

claude bot left a comment

Choose a reason for hiding this comment

Overview

Security risks

Level of scrutiny

Other factors

Uh oh!

Uh oh!

Uh oh!

claude bot left a comment

Choose a reason for hiding this comment

Overview

Security risks

Level of scrutiny

Other factors

robobun commented Apr 12, 2026 •

edited

Loading

coderabbitai bot commented Apr 12, 2026 •

edited

Loading

coderabbitai bot commented Apr 12, 2026 •

edited

Loading