# Static Linux SDK regression on 6.4? Duplicate symbols

**URL:** <https://forums.swift.org/t/static-linux-sdk-regression-on-6-4-duplicate-symbols/87685>\
**Category:** Package Manager\
**Tags:** packagemanager, build, static-sdk\
**Created:** [June 25, 2026, 1:04pm UTC](https://forums.swift.org/t/static-linux-sdk-regression-on-6-4-duplicate-symbols/87685 "2026-06-25T13:04:09Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![sebsto](https://sea2.discourse-cdn.com/swift/user_avatar/forums.swift.org/sebsto/32/27165_2.png) [@sebsto](https://forums.swift.org/u/sebsto)\
**Post date:** [June 25, 2026, 1:04pm UTC](https://forums.swift.org/t/static-linux-sdk-regression-on-6-4-duplicate-symbols/87685/1 "2026-06-25T13:04:09Z")

</div>

Hello,

I ran into what looks like a Static Linux SDK regression on 6.4 and filed [swiftlang/swift#90196](https://github.com/swiftlang/swift/issues/90196).

Short version: a package that links two vendored BoringSSL copies (swift-nio-ssl + swift-crypto) fails to statically link against musl with duplicate libc++abi symbols (`set_new_handler`, `__cxa_*`, `terminate`, etc). The same package links fine on 6.3.2, and it fails on both Linux and macOS hosts with 6.4, so it points at the toolchain rather than the libraries. Minimal repro and a 6.3-vs-6.4 matrix are in the issue.

Anyone seen this, or know if it belongs somewhere other than swiftlang/swift?

---

<div class="post-metadata">

**Author:** ![dschaefer2](https://sea2.discourse-cdn.com/swift/user_avatar/forums.swift.org/dschaefer2/32/46516_2.png) [@dschaefer2](https://forums.swift.org/u/dschaefer2)\
**Post date:** [June 25, 2026, 1:51pm UTC](https://forums.swift.org/t/static-linux-sdk-regression-on-6-4-duplicate-symbols/87685/2 "2026-06-25T13:51:50Z")

</div>

Does it work if you switch back to the native build system, i.e. `swift build --build-system native ...`?

---

<div class="post-metadata">

**Author:** ![Finagolfin](https://sea2.discourse-cdn.com/swift/user_avatar/forums.swift.org/finagolfin/32/34388_2.png) [@Finagolfin](https://forums.swift.org/u/Finagolfin)\
**Post date:** [June 25, 2026, 3:45pm UTC](https://forums.swift.org/t/static-linux-sdk-regression-on-6-4-duplicate-symbols/87685/3 "2026-06-25T15:45:01Z")

</div>

He said it was `swift-build` on the linked issue, better to keep the discussion there.

---

<div class="post-metadata">

**Author:** ![dschaefer2](https://sea2.discourse-cdn.com/swift/user_avatar/forums.swift.org/dschaefer2/32/46516_2.png) [@dschaefer2](https://forums.swift.org/u/dschaefer2)\
**Post date:** [June 25, 2026, 6:01pm UTC](https://forums.swift.org/t/static-linux-sdk-regression-on-6-4-duplicate-symbols/87685/4 "2026-06-25T18:01:44Z")

</div>

Right, I'm just wondering if the issue is with swift-build or with the SDK itself. If it works with build system native, then it's the SDK that has regressed.

---

<div class="post-metadata">

**Author:** ![dschaefer2](https://sea2.discourse-cdn.com/swift/user_avatar/forums.swift.org/dschaefer2/32/46516_2.png) [@dschaefer2](https://forums.swift.org/u/dschaefer2)\
**Post date:** [June 25, 2026, 6:06pm UTC](https://forums.swift.org/t/static-linux-sdk-regression-on-6-4-duplicate-symbols/87685/5 "2026-06-25T18:06:24Z")

</div>

And if it's SwiftBuild, this is the first release where it supports Swift SDKs so it maybe just something they've overlooked.

---

<div class="post-metadata">

**Author:** ![sebsto](https://sea2.discourse-cdn.com/swift/user_avatar/forums.swift.org/sebsto/32/27165_2.png) [@sebsto](https://forums.swift.org/u/sebsto)\
**Post date:** [June 25, 2026, 8:32pm UTC](https://forums.swift.org/t/static-linux-sdk-regression-on-6-4-duplicate-symbols/87685/6 "2026-06-25T20:32:56Z")

</div>

Sorry for my late reply folks. It's indeed the new build system causing this error, not a regression on the Static SDK.  
I updated the issue description [Swift Build 6.4: duplicate libc++abi symbols when statically linking two C++ targets · Issue #90196 · swiftlang/swift · GitHub](https://github.com/swiftlang/swift/issues/90196)
