Repository navigation
Conversation
|
This PR's own run reproduces it: all five Everything the step prints before dying is pip output: no assertion, no Beyond the checks listed in Two ways to move this forward — whichever is easier:
🤖 Generated with opencode |
|
I proposed a simpler solution, but this was useful to diagnose things. |
Confirmed on CauseThe repair step still derives the architecture from the runner label, and arch=$([[ "${{ matrix.os }}" == "ubuntu-latest" ]] && echo x86_64 || echo aarch64) # now always aarch64
auditwheel repair --plat auto -w dist/ dist/*linux_${arch}.whl
rm -f dist/*linux_${arch}.whlSame class of bug as the one that hid the Suggested fix — pick the wheel by name, not by runner pip install auditwheel
- arch=$([[ "${{ matrix.os }}" == "ubuntu-latest" ]] && echo x86_64 || echo aarch64)
- auditwheel repair --plat auto -w dist/ dist/*linux_${arch}.whl
- rm -f dist/*linux_${arch}.whl
+ # Derive the wheel from dist/, not from the runner label: keying this on
+ # matrix.os == "ubuntu-latest" silently became aarch64 when that entry
+ # moved to ubuntu-24.04. "-linux_" cannot also match the "manylinux_"
+ # wheel repair just wrote, so the rm below is safe.
+ auditwheel repair --plat auto -w dist/ dist/*-linux_*.whl
+ rm -f dist/*-linux_*.whl
+ ls dist/ # exactly one wheel per (os, python) leg: what upload/PyPI seeVerified against real filenames — This works for both Linux legs without knowing their labels, and the trailing Two things to keep from getting lost
Side note on the tag 🤖 Generated with opencode |
Why PyPI has no
marinholab.sas.core.guigui(camera + touchscreen style sheets,set_state, the logo) has been onmainsince2026-10-21 (
66e22ca..79d4bdc), but PyPI is still at 26.9.15 (2026-10-02), sopip install marinholab-sas-corecannotimport marinholab.sas.core.guiat all. SPARCS'camera/cv_inference_gui.pyimports exactly those three names, which is how this was noticed.publishrequires everybuildleg, and since 2026-10-21 everyubuntu-latest(x86_64) legdies in the Smoke test with exit 139 (SIGSEGV) → publish is skipped for the whole matrix. The
arm legs are pinned to
ubuntu-24.04and pass, which is why the crash looks "x86_64-only".What I could rule out
All from a clean container, x86_64 under emulation (
dqrobotics 26.4.0a7,numpy 2.5.4):ubuntu:24.04ubuntu:26.04mainbuilt from source,ubuntu:26.04, system cmake/ninjamainbuilt with CI-stylepython -m build --wheel(pip cmake 4.x),ubuntu:26.04auditwheel repair --plat auto(auditwheel 6.8.2; bundles nothing)dqroboticsfirst /_corefirst)nm -Don_corefordq::symbols (interposition hypothesis)So it is not the
guicode, not the packaging, notauditwheel, not dqrobotics symbolinterposition, and it does not reproduce outside GitHub runners — the trace has to come from CI,
which this PR makes it produce.
What this PR does
-X faulthandler, and on a Linux SIGSEGV/SIGABRTre-runs it under
gdb(thread apply all bt) — the log currently says onlyProcess completed with exit code 139.is the only one whose image is unpinned (
ubuntu-latestbegan migrating to Ubuntu 26.04 on2026-10-19, i.e. two days before these failures).
guimodules import and thatLOGO_PATHexists inside the installed wheel. Nothing verified that
gui/gui/assetsship, so thesubpackage could (and did) go unreleased unnoticed.
Verified locally: the step, run verbatim from the YAML, exits 0 on x86_64/26.04 and prints
gui style helpers OK.🤖 Generated with opencode
Proposed
.github/workflows/publish.yml(full file)The PAT used to open this PR has no
workflowscope, so the workflow change could not be pushed;the complete proposed file is below (only the
Smoke teststep differs frommain).The same change as a patch, if easier
🤖 Generated with opencode