release: 0.29.2 - #6285
Conversation
* Method return type: allow to built it from other traits * Fix stub generation of field getters * Fix UI test --------- Co-authored-by: David Hewitt <mail@davidhewitt.dev>
* provide an opt-out of raw-dylib * newsfragments * remove dead constants * only emit link alias on windows * fixup expected lib name check * use explicit "lib" prefix on mingw * test fixups * Update pyo3-build-config/src/impl_.rs Co-authored-by: chiri <chirizxc@proton.me> --------- Co-authored-by: chiri <chirizxc@proton.me>
|
Also it looks like pydantic/pydantic#13572 (updating pydantic-core to 0.29.1) has hit a PyPy 3.11 issue, I will try to debug that in case there's something broken about PyPy support on 0.29.1 (I wouldn't be surprised if the GC fixes have broken PyPy, which doesn't have a fully working cycle GC mechanism). |
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 |
|
Due to #6294 fixing a regression in PyO3 0.29.1 which will probably block projects from upgrading, I'd like to get as many of the PRs mentioned in the OP merged along with that fix and ship the whole lot today. |
Bumps [CodSpeedHQ/action](https://github.com/codspeedhq/action) from 4 to 5.0.1. - [Release notes](https://github.com/codspeedhq/action/releases) - [Changelog](https://github.com/CodSpeedHQ/action/blob/main/CHANGELOG.md) - [Commits](CodSpeedHQ/action@v4...v5.0.1) --- updated-dependencies: - dependency-name: CodSpeedHQ/action dependency-version: 5.0.1 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>
These were deleted in #6130 since `maturin sdist` was reportedly not following symlinks. However, the chosen licenses require their texts to be distributed with the software, so this commit restores them. Using duplicate copies of the top-level license files rather than symlinks sidesteps any possible publication difficulties.
* fix: pass class to classmethod magic methods * docs: add newsfragment for PR 6283
* fix refcount leak in `initialize_tp_dict` * newsfragment
* fix missing trailing nul on Python 3.9 `#[pyclass]` docstrings * newsfragment * use to_bytes_with_nul Co-authored-by: Lily <code@lilyf.org> --------- Co-authored-by: Lily <code@lilyf.org>
* fix double-decref in PyPy in instance dealloc * newsfragment
|
I have merged / cherry-picked everything which was ready to go today, I've pushed the release commit on top. Assuming CI is green I will put this live later. |
eea8d85 to
ddfc45e
Compare
ddfc45e to
f76da5c
Compare
|
I think CI is likely to be green this time; getting 3.8 to run properly in CI has been a faff, but at least 0e608fa is only hacking in the CI scripts. |
|
(netlify build issue is the same CHANGELOG check failing on the URL for the not yet pushed tag being nonexistent, can clean that up another time.) |
I clearly let 0.29.1 get too big, missed some of the fix cherry-picks off 🤦♂️
It looks like we have some other fixes in the pipeline anyway, so let's get those merged and then push 0.29.2:
experimental-inspect: Iterator__next__not optional #6274experimental-inspect: dedup union members #6273#[pyclass]docstrings #6296initialize_tp_dict#6297