Skip to content

Commit 1ce4481

Browse files
gountharsobolevn
andauthored
Bump cibuildwheel to 4.2.0 to fix musllinux riscv64 wheels (msgspec#1146)
The 0.22.0 release run died on `cp310-musllinux_riscv64` with `Command ['which', 'uv'] failed with code 1`, from my riscv64 matrix entry. All six manylinux riscv64 wheels in that same job built, repaired and tested fine, so it is specific to the musllinux image. pypa/manylinux used to skip installing uv in the musllinux riscv64 image. At the commit the pinned image was built from (`8dbfa8e`, 4 June), `docker/build_scripts/finalize.sh` had: ``` musllinux*_riscv64-uv) continue;; # uv doesn't provide musl riscv64 wheels due to Rust issues ``` That was removed by pypa/manylinux#1969 on 11 July, and their image test suite now runs `uv venv` / `uv pip install` inside musllinux riscv64. The comment was already out of date when it was written: uv has published a `manylinux_2_31_riscv64.musllinux_1_1_riscv64` wheel since 0.11.18 (1 June). The pins differ across cibuildwheel releases: | cibuildwheel | `musllinux_1_2_riscv64` image | uv present | |---|---|---| | 4.1.0 | `2026.06.04-1` | no | | 4.2.0 | `2026.08.04-1` | yes | So `build-frontend = "build[uv]"` has nothing to find on 4.1.0. This bumps both `uvx --from` pins to 4.2.0. 4.2.0 builds CPython 3.15 by default, but `.cibuildwheel.toml` sets an explicit `build = "cp310-* cp311-* cp312-* cp313-* cp314-* cp314t-*"` list, so no new target appears. The rest of that release is Pyodide 314.0.4, a GraalPy pip hold, and download retries. I have not run this. The reasoning is from the failing job log and the manylinux source, not from a green build. CI here will not test it either unless someone adds the `build-riscv64` label, since the riscv64 job is excluded on unlabelled PRs, and even then `CIBW_SKIP` drops musllinux outside a release build. So the honest state is that only a release run proves it. If you would rather not move cibuildwheel right before a release, `musllinux-riscv64-image` in `.cibuildwheel.toml` pins just that one image and leaves everything else alone. Two things worth knowing either way: The riscv64 job is gated behind the `build-riscv64` label on PRs, and `CIBW_SKIP` drops musllinux unless `build-all` is set. So nothing had ever exercised musllinux riscv64 until the release turned `build-all` on. Whatever you decide here, that combination will keep hiding riscv64 musllinux breakage until release day. Dependabot covers `github-actions` and `uv`, but the cibuildwheel version is an inline `uvx --from` argument rather than a tracked dependency, which is why it sat at 4.1.0. Moving it into the uv ecosystem somehow would stop the next one, though I do not have a clean suggestion for how. Refs msgspec#1134, msgspec#987 Signed-off-by: Bruno Verachten <[email protected]> Co-authored-by: sobolevn <[email protected]>
1 parent c24bc70 commit 1ce4481

1 file changed

Lines changed: 2 additions & 2 deletions

File tree

‎.github/workflows/build-wheels.yml‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -91,7 +91,7 @@ jobs:
9191
- name: Build & Test Wheels
9292
if: runner.os != 'Windows'
9393
run: >-
94-
uvx --from cibuildwheel==4.1.0
94+
uvx --from cibuildwheel==4.2.0
9595
cibuildwheel
9696
--config-file .cibuildwheel.toml
9797
--output-dir wheelhouse
@@ -100,7 +100,7 @@ jobs:
100100
- name: Build & Test Wheels
101101
if: runner.os == 'Windows'
102102
run: >-
103-
uvx --from cibuildwheel==4.1.0
103+
uvx --from cibuildwheel==4.2.0
104104
cibuildwheel
105105
--config-file .cibuildwheel.toml
106106
--output-dir wheelhouse

0 commit comments

Comments
 (0)