chore(deps): Bump strict-kwargs from 2026.7.24 to 2026.8.16

[dependabot]

Bumps strict-kwargs from 2026.7.24 to 2026.8.16.

Release notes

Sourced from strict-kwargs's releases.

Release 2026.8.16

2026.8.16

• Avoid redundant cache-fingerprint walks for nested source roots and unused default Python environments.

• Add error_on_unused_noqa (and --error-on-unused-noqa), which reports a new KW002 error for every # noqa: KW001 comment that suppressed nothing. Blanket # noqa comments, and comments naming only other tools' codes, are never reported.

• Avoid copying every ty-bound source file during no-cache checks.

• Avoid copying ty hover payloads and share cached group responses, reducing cold-check time on repositories with many inferred calls.

• Balance large no-cache checks by the ty hover requests remaining after grouped-answer reuse, reducing whole-repository fallback time.

• Index the embedded typeshed files once, speeding up no-cache checks that resolve many standard-library calls.

• Isolate reused CPython completeness checkouts so parent project metadata cannot change the golden oracle.

• Keep safe self hover reuse in unaffected methods when another method narrows or escapes its receiver, reducing no-cache ty fallback work.

• Reuse identical self hover responses across ordinary methods of the same class, reducing no-cache ty fallback work on test-heavy projects.

• Run large no-cache fix and diff passes across deterministic ty server shards, substantially reducing whole-repository runtime while keeping smaller projects on the lower-overhead serial path.

• Scan projects normally when an ancestor of the project root is a dot-directory, including macOS temporary directories used by generated benchmarks.

• Serialize newly computed cache diagnostics without cloning their strings into a second in-memory representation, cutting cold-cache population medians by about 5% on the pinned Sphinx and CPython benchmarks.

• Share cached ty definition responses between grouped calls instead of deep-cloning their JSON trees.

• Sort whole-project diagnostic results in parallel, cutting cold-check medians by about 8% on CPython and 11% on Sphinx in the pinned repository benchmarks.

• Store warm-cache diagnostics in one bounded manifest, avoiding one cache-file read and write per checked Python file.

• Use eight deterministic ty server shards for uncached whole-project checks. In balanced ty 0.0.64 benchmarks with a fresh strict-kwargs cache, median wall time fell by about 6% on Sphinx and 16% on CPython. The new partition also stops emitting three dynamic-call false positives from the completeness floors.

Changelog

Sourced from strict-kwargs's changelog.

2026.8.16

• Avoid redundant cache-fingerprint walks for nested source roots and unused default Python environments.

• Add error_on_unused_noqa (and --error-on-unused-noqa), which reports a new KW002 error for every # noqa: KW001 comment that suppressed nothing. Blanket # noqa comments, and comments naming only other tools' codes, are never reported.

• Avoid copying every ty-bound source file during no-cache checks.

• Avoid copying ty hover payloads and share cached group responses, reducing cold-check time on repositories with many inferred calls.

• Balance large no-cache checks by the ty hover requests remaining after grouped-answer reuse, reducing whole-repository fallback time.

• Index the embedded typeshed files once, speeding up no-cache checks that resolve many standard-library calls.

• Isolate reused CPython completeness checkouts so parent project metadata cannot change the golden oracle.

• Keep safe self hover reuse in unaffected methods when another method narrows or escapes its receiver, reducing no-cache ty fallback work.

• Reuse identical self hover responses across ordinary methods of the same class, reducing no-cache ty fallback work on test-heavy projects.

• Run large no-cache fix and diff passes across deterministic ty server shards, substantially reducing whole-repository runtime while keeping smaller projects on the lower-overhead serial path.

• Scan projects normally when an ancestor of the project root is a dot-directory, including macOS temporary directories used by generated benchmarks.

• Serialize newly computed cache diagnostics without cloning their strings into a second in-memory representation, cutting cold-cache population medians by about 5% on the pinned Sphinx and CPython benchmarks.

• Share cached ty definition responses between grouped calls instead of deep-cloning their JSON trees.

• Sort whole-project diagnostic results in parallel, cutting cold-check medians by about 8% on CPython and 11% on Sphinx in the pinned repository benchmarks.

• Store warm-cache diagnostics in one bounded manifest, avoiding one cache-file read and write per checked Python file.

• Use eight deterministic ty server shards for uncached whole-project checks. In balanced ty 0.0.64 benchmarks with a fresh strict-kwargs cache, median wall time fell by about 6% on Sphinx and 16% on CPython. The new partition also stops emitting three dynamic-call false positives from the completeness floors.

Commits

• 7e45015 Bump CHANGELOG and version for release
• df79ee8 Merge pull request #356 from adamtheturtle/dependabot/github_actions/CodSpeed...
• 89cccc7 Merge pull request #358 from adamtheturtle/dependabot/github_actions/taiki-e/...
• 0ecb8d3 build(deps): bump taiki-e/install-action from 2.85.10 to 2.85.11
• c056a85 Switch Vale to the upstream ai-tells package (#357)
• 386b311 Merge pull request #355 from adamtheturtle/dependabot/cargo/ignore-0.4.33
• dffe1ae Merge pull request #353 from adamtheturtle/dependabot/cargo/clap-4.6.6
• aea5c01 Merge pull request #354 from adamtheturtle/dependabot/github_actions/taiki-e/...
• 6351273 build(deps): bump CodSpeedHQ/action from 5.0.2 to 5.0.3
• 52e2550 build(deps): bump ignore from 0.4.32 to 0.4.33
• Additional commits viewable in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

---

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)