release: 0.29.3 - #6416
Open
davidhewitt wants to merge 47 commits into
Open
release: 0.29.3#6416davidhewitt wants to merge 47 commits into
davidhewitt wants to merge 47 commits into
Conversation
Bumps [actions/setup-node](https://github.com/actions/setup-node) from 6 to 7. - [Release notes](https://github.com/actions/setup-node/releases) - [Commits](actions/setup-node@v6...v7) --- updated-dependencies: - dependency-name: actions/setup-node dependency-version: '7' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Bumps [actions/setup-python](https://github.com/actions/setup-python) from 6 to 7. - [Release notes](https://github.com/actions/setup-python/releases) - [Commits](actions/setup-python@v6...v7) --- updated-dependencies: - dependency-name: actions/setup-python dependency-version: '7' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Bumps [actions/checkout](https://github.com/actions/checkout) from 7.0.0 to 7.0.1. - [Release notes](https://github.com/actions/checkout/releases) - [Commits](actions/checkout@v7...v7.0.1) --- updated-dependencies: - dependency-name: actions/checkout dependency-version: 7.0.1 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* Preallocate sets when extracting Python set values * Document Python set extraction allocation improvement * Remove ineffective set capacity assertions * Preallocate BTreeSet extraction buffers * Fix Clippy Vec import for set extraction
* update `PySet_GET_SIZE` and similar for free-threaded Python * split `cpython` portion of `setobject.h` off
* internal: simplify module creation conditional branches * further cleanup * fixup test
* ci: fix ffi-check for backport of `Py_CompileStringFlags` symbol * always use latest Python versions
* ci: pin the `careful` job to nightly-2026-08-06
From nightly-2026-08-07 cargo no longer finds the sysroot `cargo careful`
builds for itself when running test binaries. Proc-macro crates link `libstd`
dynamically, so the `pyo3-macros` test binary can't start:
error while loading shared libraries: libstd-<hash>.so: cannot open
shared object file: No such file or directory
That aborts the job before it gets to anything else. Pin to the last nightly
where it works; revert to `@nightly` once
RalfJung/cargo-careful#55 is fixed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N96dd4VafzejRKxEuk1pU4
* ci: fix `wasm32-wasip1` job against CPython 3.14.7
Two problems, both surfaced by CPython 3.14 moving to a new patch release:
* `test-wasm` points the embedded interpreter at `PYTHONPATH=/lib`, which has
never existed in the CPython checkout. It worked because `getpath` found the
build tree itself, via the `Modules/Setup.local` landmark next to the
"executable". CPython 3.14.7 removed that fallback (gh-151544), so the test
binaries now die on startup with "Fatal Python error: Failed to import
encodings module". Pass the stdlib and the WASI build directory explicitly.
* The `.nox/wasi` cache is keyed on `UV_PYTHON` ("3.14"), not on the CPython
version nox builds and then looks for, so a new patch release restores the
previous one's build and skips the build step. `cargo test` then fails
immediately with "failed to search the lib dir at PYO3_CROSS_LIB_DIR=...".
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N96dd4VafzejRKxEuk1pU4
* Update .github/workflows/ci.yml
Co-authored-by: David Hewitt <mail@davidhewitt.dev>
* Update noxfile.py
Co-authored-by: David Hewitt <mail@davidhewitt.dev>
* Update .github/workflows/ci.yml
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: David Hewitt <mail@davidhewitt.dev>
…`Copy` classes (#6309) * fix `clippy::clone_on_copy` firing on `#[pyclass(from_py_object)]` + `Copy` classes Move the clone out of the generated `FromPyObject` impl into a generic helper where the type is only known to be `Clone`, so the lint cannot trigger. Fixes #6308. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * bless UI test snapshot for trimmed `FromPyObject` diagnostic paths The generated `extract` body no longer references `FromPyObject` items inline, which changes how rustc renders the trait path in unrelated diagnostics in the same crate. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * bless `default` revision UI test snapshot as well Same trimmed-path diagnostic drift as the `inspect` revision; this snapshot only runs without `experimental-inspect` enabled. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: David Hewitt <mail@davidhewitt.dev>
Bumps [CodSpeedHQ/action](https://github.com/codspeedhq/action) from 5.0.1 to 5.0.3. - [Release notes](https://github.com/codspeedhq/action/releases) - [Changelog](https://github.com/CodSpeedHQ/action/blob/main/CHANGELOG.md) - [Commits](CodSpeedHQ/action@v5.0.1...v5.0.3) --- updated-dependencies: - dependency-name: CodSpeedHQ/action dependency-version: 5.0.3 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* refactor: move some tests inside the lib * fix abi3 gates * correct std imports * fixup imports * fixup chrono-tz tests * format
Addresses one item from gh-3310's checklist ("String intern!"). The guide already covers extract-vs-cast, Bound::py, calling conventions, Python::detach, and the reference pool, but never mentioned intern!, even though it's one of the simplest, most broadly applicable optimizations (avoiding a repeated PyString allocation at call sites that reuse the same string, e.g. dict keys/attribute names). The example follows the same before/after style as the rest of the guide and is adapted from the intern! macro's own doc comment in src/sync.rs, which is already an existing, passing doctest - not new, unverified example code. Not addressing gh-3310's other checklist items in this PR: - The Vec<u8>/Cow<[u8]> item was confirmed already resolved by the IntoPyObject rework in the issue's own comment thread (by @Icxolu), so documenting it as a live gotcha would now be inaccurate. - The remaining items (conversion overhead in general, #[pyo3(get)] deep-cloning, dictionary dispatch) need more source investigation than fits one PR; leaving those for follow-up.
…6340) * ci: bump wasi builds to use 3.15, read WASI SDK from cpython config * fixup emscripten build
* `experimental-inspect`: cover the `Option`-returning `__next__` / `__anext__` shapes Route A: one wrapper carrying both the conversion and the type hint Review pass on the `IterNextOutput` wrapper pytests: assert the generated `__next__` / `__anext__` hints * Address review - Use `StaticIdent` for the hard-coded wrapper / fallback names instead of `&str` + `format_ident!` (and `TokenGenerator` on the slot side). - Drop the `asyncness` guard around the `__next__` / `__anext__` return type: `async fn` does not compile for *any* slot method today (the slot body hands the future straight to `IntoPyCallbackOutput`), so the guard was unreachable. - Note in `impl_/pymethods.rs` that `am_anext` has no null-without-error convention, so `StopAsyncIteration` has to be raised explicitly. * pytests: skip the async-iter test on GraalPy < 25.1 `PyClassOptionAsyncIter.__anext__` signals exhaustion the way the `am_anext` slot has to: it raises `StopAsyncIteration` synchronously, since the slot has no "returned null, no error set" convention. GraalPy < 25.1 lets such a synchronous raise escape `async for` instead of ending the loop. That is not specific to native classes -- a plain Python class with a non-`async def __anext__` reproduces it -- and it is already fixed in GraalPy 25.1, so gate the test on the version rather than on the implementation alone.
* suppress warning * suppress warning in proc-macro generated code
* Use Cargo workspace publishing Simplify the publishing setup by using https://doc.rust-lang.org/cargo/CHANGELOG.html#cargo-190-2025-09-18 * Remove `cargo publish` for noxfile It's only runnable in CI anyway.
`Arc::<[u8]>::from(py_bytearray.to_vec())` creates a `Vec` copy of the `PyByteArray` contents, then immediately copies that `Vec` into an `Arc<[u8]>`. It should copy the `PyByteArray` contents directly into the `Arc<[u8]>`. This PR inlines `PyByteArrayMethods::to_vec` into `PyBackedBytes::from` without the unnecessary copy. See: <https://github.com/PyO3/pyo3/blob/75507833367e5267dd6fdd7e48de628e4b8e53e0/src/types/bytearray.rs#L291-L300>.
…ectory (#6365) * experimental-inspect: write nested package stubs into their own directory * Correct newsfragement number
* preliminary fixes for PyPy 3.12 * fixup
Bumps [CodSpeedHQ/action](https://github.com/codspeedhq/action) from 5.0.3 to 5.2.1. - [Release notes](https://github.com/codspeedhq/action/releases) - [Changelog](https://github.com/CodSpeedHQ/action/blob/main/CHANGELOG.md) - [Commits](CodSpeedHQ/action@v5.0.3...v5.2.1) --- updated-dependencies: - dependency-name: CodSpeedHQ/action dependency-version: 5.2.1 dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* Add eval frame FFI bindings * Add eval frame FFI bindings * Fix eval code extra index link name
…one` default (#6363) * experimental-inspect: give slot-wrapper trailing arguments their `None` default * test: cover the slot table's optional trailing argument counts * Drop redundant optional_trailing_args test
…6397) * Escape control characters as Python, not as Rust `Display for PyStaticExpr` rendered string constants with `{value:?}`, which uses Rust's escaping rules. Rust writes an unprintable character as `\u{1b}`; Python's `\u` escape takes exactly four hex digits, so that form is a SyntaxError rather than an escape. Any string constant holding a C0 control character other than NUL, tab, newline or carriage return therefore rendered as invalid Python -- an ANSI sequence in a signature default such as `"\x1b[0m"` is the realistic way to hit it. The sibling renderer in `pyo3-introspection/src/stubs.rs` already gets this right, writing `\x1b`. This mirrors its escape table so the two agree. The existing test covered `"\0\t\\\""`, the four characters whose Rust and Python escapes happen to be identical, so the divergence did not show up. * Add newsfragment * Cover the newline, carriage return and literal-character arms
….abc.Buffer` (py312+) (#6395) * Introspection: use `typing_extensions.Buffer` instead of `collections.abc.Buffer` (py312+) * Introspection: only import `Buffer` from `typing_extensions` on `python<3.12` Co-authored-by: Jonas Dedden <mail@jonas-dedden.de> * Introspection: apply review suggestions for the `typing_extensions.Buffer` fix Co-authored-by: Thomas Tanon <thomas@pellissier-tanon.fr> --------- Co-authored-by: Jonas Dedden <mail@jonas-dedden.de> Co-authored-by: Thomas Tanon <thomas@pellissier-tanon.fr>
* Introspection: Escape stub docstrings to prevent invalid syntax. * Introspection: don't escape `"` in docstrings unless necessary Co-authored-by: Thomas Tanon <thomas@pellissier-tanon.fr> --------- Co-authored-by: Thomas Tanon <thomas@pellissier-tanon.fr>
* Record successful attach only after attaching in SuspendAttach * fix limited API builds * add release note * simplify test
Hooray! CodSpeed harness just leveled up!The base and head of this comparison were measured with different runner settings, so their benchmark values are not directly comparable. What changed between base and head:
Re-run the base with the same settings to get a valid performance comparison. Comparing 
Footnotes |
* fix raw-dylib opt-out on windows x86 * newsfragment * correct PyPy build failures * `PyVectorcall_Call` is 3.12+ * workaround incorrect lib name for PyPy with raw-dylib opt out * fix `PyVectorcall_Call` cfg
* ci: update ffi-check for PyPy symbol mangling change * fix formatting * fix ffi check failures on older versions * fix pypy 3.11 link aliases * fmt * try fix windows build * try to fix windows * try switch to allowlist for bindgen * try to make link failures on windows easier to diagnose * fix link name for pypy without raw-dylib * fix missing PyPy link name attributes * only decorate x86 symbols on `raw-dylib` builds * Update link name for PyVectorcall_Call in abstract_.rs * fmt
* internal: new badges * Update README.md * Update README.md
* check static addresses match in `pyo3-ffi-check` * newsfragments * fix private symbols * fmt * fix windows-only pypy symbol
* fix FFI cases where PyPy replaced function with macro * newsfragment
2 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
19 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR contains backfixes for the 0.29 release series, targeting the
release-0.29branch. It's not all possible fixes, just everything which I could cherry-pick relatively cleanly.I hadn't wanted to do this (as CI has been bumpy recently and required a bit of hacking to fix); I was thinking we could get away with going straight for the 0.30 release. Sadly I think given a number of the recent bug discoveries I think this patch release is justified. In particular it seems like we're broken with a few various "supported" interpreters and it seems worth landing fixes for those in a patch release.
Before releasing, I want to merge (and cherry-pick):
Might also be worth landing a fix for #6392 if someone has time to triage & investigate it (sounds potentially problematic).