chore: update pre-commit hooks - #321
Merged
Merged
Conversation
github-actions
Bot
added
the
skip-changelog
bjlittle
added a commit
that referenced
this pull request
Sep 14, 2026
The fragment was written as `321.feature.rst` because that was the next free number when the task that wrote it was dispatched. While the work ran, `pre-commit-ci[bot]` opened #321, so this pull request is #322 and the fragment named a number belonging to someone else. `ci-changelog` asks every pull request for a fragment named after its own number, so this would have failed the gate rather than slipped through -- but it would have failed after review rather than before, which is the wrong moment to learn it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h
bjlittle
added a commit
that referenced
this pull request
Sep 14, 2026
* Give each release a changelog title and an anchor The towncrier template emitted no version headings: one release hid the defect, and from the second the assembled CHANGELOG.rst was an undivided run of entries with no anchor for a per-release whatsnew page to link (whatsnew spec §3.3). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Point the plan's copyright constraint at the file that holds it `notice-rgx` is in `pyproject.toml` under `[tool.ruff.lint.flake8-copyright]`, not `.pre-commit-config.yaml`. Task 1's implementer caught it; verified by grep, which matches pyproject.toml alone. The plan is not frozen yet -- docs spec §3.4 freezes a plan when its implementation pull request merges, and that is this branch. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Define the version and build-date substitutions Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Gate the substitutions' built value, not conf.py's text Review found _defined_substitutions() regex-scanned conf.py's source text, a proxy that would pass on a substitution merely mentioned in a comment and fail on a correctly-built rst_epilog written a different way. Execute conf.py via tests/by_path.load_path and read the actual rst_epilog attribute instead. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Give the changelog a preamble and an anchor for its newest release Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Add the what's new section Creates docs/src/reference/whatsnew/ with the section index, the `latest.rst` page being accumulated toward, and the seed template copied into place at release time. Registers the section in the reference toctree ahead of the changelog, and exempts the section index from the reading-time banner since the included `latest.rst` already carries one. Also fixes a bug in the section index's `_pages()` test helper: it globbed every `*.rst` in the directory, including `index.rst` itself, so `test_the_toctree_lists_every_page_in_the_section` compared the toctree against a set that could never match it (the index does not, and should not, list itself). Excludes the index by name, matching how a quadrant landing page's own table never names itself (narrative spec §3.9, whatsnew spec §4). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Gate the two release-time whatsnew mistakes Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Report every offending page/substitution pair, not just the last test_no_frozen_page_carries_the_substitutions keyed its offenders dict on the page name while iterating both substitutions, so a page carrying both reported only the last match. Key on the (page, substitution) pair instead. Also bring _frozen() in line with its docstring: it excluded latest.rst by name but still swept in the section index, which _pages() already excludes by stem for the same reason. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Carry the what's new steps in the release sequence Insert the freeze-the-what's-new-page step after getting on the release branch, and the reseed-on-main step after the merge-back, renumbering The Sequence from 1..13 and updating every step cross-reference on the page. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Add the changelog fragment Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Fix the whatsnew placeholder gate, and tidy its fixtures The placeholder gate asserted against latest.rst, keyed on the installed version reaching the first release. Two failures followed: whatsnew spec §3.5 step 3 reseeds latest.rst from the template -- placeholder and all -- after every release, so the gate read a freshly reseeded file as a defect for the whole of the next cycle; and it read latest.rst unconditionally, so it raised FileNotFoundError through the frozen window between a release's tag and its merge-back, where the file does not exist at all. The rule belongs to the frozen page, not latest.rst: a frozen page carrying TBD is always wrong, needs no release signal, and fires on the release pull request rather than at the tag. Rewritten over _frozen(), the same corpus test_no_frozen_page_carries_the_substitutions already reads, reporting every offending page. FIRST_RELEASE, the pytest.skip, and the now-unused pytest/ Version/tephpy imports are removed. While in the file: _toctree_entries() was a fourth-hand copy of test_docs_landing_pages.toctree_entries(), unbounded past the toctree's end -- replaced with the shared, unit-tested helper. _rendered()'s docstring claimed a context fidelity to towncrier it did not have (versiondata as a SimpleNamespace rather than towncrier's dict, and the wrong underlines) -- made faithful rather than just honest about the gap, since both were possible. _defined_substitutions()'s docstring now names the side effects of executing conf.py: the plot-scratch directory, the sys.path entry, and the sys.modules registration. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Correct the whatsnew spec's placeholder gate, and a stale step number §4's third gate stated the rule about latest.rst and contradicted §3.5 step 3 three paragraphs above it, which reseeds latest.rst from the template -- placeholder and all -- after every release. The gate as written was modelled on start spec §3.7's pre-release note, but that note makes one transition and stays retired, while this state is cyclic. Rewritten to state the rule about the frozen page instead, matching the test fix on this branch. §6 also cited developer/release.rst step 11 for where announcing is left to a person; renumbering moved it to step 13. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Mention what's new in the reference quadrant's opening prose The quadrant's introduction enumerated its pages but never named the new What's New section -- only the toctree gained the entry. whatsnew spec §3.1 places it before changelog because the highlights are what a reader meets first; the prose now routes them past it in the same order. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Link the changelog fragment to the section, not its design spec The fragment linked whatsnew-spec-1, so a reader of the published changelog who clicked "What's New" landed on a developer design specification instead of the section itself. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Warn that step 2 leaves the changelog anchor undefined until step 3 Step 2 has the maintainer repoint the what's new page's changelog link at changelog-vX.Y.Z, but step 3 is what assembles the file that defines that anchor. A pixi run docs between the two steps hit an undefined label with nothing telling the reader it was expected rather than a mistake. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h * Correct three prose defects in the what's new docs - docs-style.rst's reading-time exemption list described the whatsnew index as directive-generated; it is an include of an already-bannered page, not a directive body. - The whatsnew design spec understated the superseded placeholder gate: it did not read nothing through the frozen window, it read latest.rst unconditionally and raised FileNotFoundError there. - The spec's §3.5 step 1 repoint now notes why the undefined changelog-vX.Y.Z label between steps 1 and 3 is acceptable: the release pull request's head carries both edits. * Name the fragment for the pull request it belongs to The fragment was written as `321.feature.rst` because that was the next free number when the task that wrote it was dispatched. While the work ran, `pre-commit-ci[bot]` opened #321, so this pull request is #322 and the fragment named a number belonging to someone else. `ci-changelog` asks every pull request for a fragment named after its own number, so this would have failed the gate rather than slipped through -- but it would have failed after review rather than before, which is the wrong moment to learn it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PDtaSBAWHdYhYfyyWCcC5h --------- Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
updates: - [github.com/astral-sh/ruff-pre-commit: v0.16.6 → v0.16.7](astral-sh/ruff-pre-commit@v0.16.6...v0.16.7) - [github.com/zizmorcore/zizmor-pre-commit: v1.30.0 → v1.30.1](zizmorcore/zizmor-pre-commit@v1.30.0...v1.30.1)
bjlittle
force-pushed
the
pre-commit-ci-update-config
branch
from
September 14, 2026 20:48
52f0da7 to
58af000
Compare
bjlittle
enabled auto-merge (squash)
September 14, 2026 20:51
This branch was successfully deployed
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
1 participant
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.
updates: