You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Evaluate abi3 wheels: one wheel per platform and day-one support for new Python releases #46
CI builds one wheel per (platform × Python version): currently 4 CPython versions across 10 platform targets, growing by 10 wheels every October. Users on a Python that shipped after our last release fall back to the sdist and need a Rust toolchain.
Building against the stable ABI (pyo3/abi3-py311) gives one wheel per platform that works on every CPython ≥ 3.11, including ones released after us.
Trade-offs to measure before switching
Performance. Under abi3, PyO3 loses some fast paths. The one most likely to matter here is datetime: the chrono conversion used for timestamp/duration values goes through the C datetime API on regular builds but through Python-level attribute access under the limited API. size(items) on a 1000-element list and the timestamp cases in examples/performance/compile_execute_benchmark.py are the numbers to compare. If abi3 costs more than ~10% on timestamp-heavy workloads, keep per-version wheels.
Free-threading. abi3 wheels cannot target the free-threaded build, so if we want cp314t wheels (see the GIL issue) we would ship abi3 for the default build plus explicit cp314t wheels.
Reproducibility. The --find-interpreter flag in maturin-action becomes unnecessary; pyproject.toml gains [tool.maturin] features = ["pyo3/extension-module", "pyo3/abi3-py311"].
Suggested approach
Build an abi3 wheel locally, run the benchmark and the full test suite against it, and post the numbers here. Decide based on the datetime cases.
A data point for this issue: CPython 3.15 is final (3.15.2), and 0.10.0 has cp315 wheels for
manylinux x86_64/i686 only. macOS arm64 and manylinux aarch64 users on 3.15 fall back to the
sdist and need Rust. We run CEL in amd64 containers (fine) but develop on macOS arm64.
Until abi3 is decided, would you cut a release with cp315 on every target? Passing explicit
interpreters instead of relying on --find-interpreter would make it deterministic: setup-python with 3.11 … 3.15 on the macOS and Windows jobs and -i python3.11 … python3.15 in maturin's args. The manylinux images updated since 3.15.0 final should carry
3.15 for aarch64 as well. #68's runner move looks like the natural place.
Happy to send a PR for the workflow change if that helps.
Activity
raufer commentedon Nov 16, 2026
A data point for this issue: CPython 3.15 is final (3.15.2), and 0.10.0 has cp315 wheels for
manylinux x86_64/i686 only. macOS arm64 and manylinux aarch64 users on 3.15 fall back to the
sdist and need Rust. We run CEL in amd64 containers (fine) but develop on macOS arm64.
Until abi3 is decided, would you cut a release with cp315 on every target? Passing explicit
interpreters instead of relying on
--find-interpreterwould make it deterministic:setup-pythonwith3.11 … 3.15on the macOS and Windows jobs and-i python3.11 … python3.15in maturin's args. The manylinux images updated since 3.15.0 final should carry3.15 for aarch64 as well.
#68's runner move looks like the natural place.
Happy to send a PR for the workflow change if that helps.