chore(deps): bump jsonschema from 0.52.1 to 0.56.0

[dependabot]

Bumps jsonschema from 0.52.1 to 0.56.0.

Release notes

Sourced from jsonschema's releases.

[Python] Release 0.56.0

Added

• canonical.RawReason, what stopped a run from modeling a document, and RawView.reason and RawView.pointer, carrying it beside the subschema the run stopped on.

Changed

• Canonicalization of a oneOf whose branches share a region too wide to expand, which spent the document's whole intersection allowance on that choice and kept the document Raw.

Performance

• Preparing or building a schema that names no base URI parses the default one once for the process instead of once per schema.
• Canonicalization reads the facets no checker covers once per node instead of once per containment query.

[Ruby] Release 0.56.0

Added

• JSONSchema::Canonical::RawReason, what stopped a run from modeling a document, and RawView#reason and RawView#pointer, carrying it beside the subschema the run stopped on.

Changed

• Canonicalization of a oneOf whose branches share a region too wide to expand, which spent the document's whole intersection allowance on that choice and kept the document Raw.

Performance

• Preparing or building a schema that names no base URI parses the default one once for the process instead of once per schema.
• Canonicalization reads the facets no checker covers once per node instead of once per containment query.

[Rust] Release 0.56.0

Added

• canonical::RawReason, what stopped a run from modeling a document, and canonical::RawView, carrying it beside the document and a pointer to the subschema the run stopped on.
• ValidationOptions::is_known_format, whether a format of that name is checked under a draft.

Changed

• BREAKING: canonical::CanonicalView::Raw carries a canonical::RawView in place of the document alone.
• Canonicalization of a oneOf whose branches share a region too wide to expand, which spent the document's whole intersection allowance on that choice and kept the document Raw.

Performance

• Preparing or building a schema that names no base URI parses the default one once for the process instead of once per schema.
• Canonicalizing several subschemas of one prepared document compiles each pattern once for the document instead of once per call.
• Canonicalization reads the facets no checker covers once per node instead of once per containment query.

[Python] Release 0.55.1

Changed

• Canonicalization of an anyOf over an allOf through a $ref to a union, which left that union nested inside the outer one.
• Canonicalization of an anyOf whose branch narrowing runs out of intersections, which left true as a branch inside the union.

... (truncated)

Changelog

Sourced from jsonschema's changelog.

[0.56.0] - 2026-09-10

Added

• canonical::RawReason, what stopped a run from modeling a document, and canonical::RawView, carrying it beside the document and a pointer to the subschema the run stopped on.
• ValidationOptions::is_known_format, whether a format of that name is checked under a draft.

Changed

• BREAKING: canonical::CanonicalView::Raw carries a canonical::RawView in place of the document alone.
• Canonicalization of a oneOf whose branches share a region too wide to expand, which spent the document's whole intersection allowance on that choice and kept the document Raw.

Performance

• Preparing or building a schema that names no base URI parses the default one once for the process instead of once per schema.
• Canonicalizing several subschemas of one prepared document compiles each pattern once for the document instead of once per call.
• Canonicalization reads the facets no checker covers once per node instead of once per containment query.

[0.55.1] - 2026-09-08

Changed

• Canonicalization of an anyOf over an allOf through a $ref to a union, which left that union nested inside the outer one.
• Canonicalization of an anyOf whose branch narrowing runs out of intersections, which left true as a branch inside the union.

[0.55.0] - 2026-09-06

Changed

• BREAKING: canonical::ObjectViolationView gained the PatternValueFails variant, which negation records for a patternProperties entry.
• Canonicalization of not over patternProperties, which kept the entry under not instead of recording the demand a key breaks.
• Canonicalization of unevaluated* beside a conditional reaching no key or index the node already evaluates, which spent a case on it.
• Canonicalization of a conditional split outgrowing its per-node case budget, which kept documents like the Open API 3.2 meta-schema Raw.
• Canonicalization of an anyOf over objects sharing a properties key under patternProperties, which settled only on a second pass.
• Canonicalization of not over an allOf folded through a reference, which settled only on a second pass.
• Canonicalization of a multipleOf divisor past f64 precision, which was not taken to divide itself.
• Canonicalization of an anyOf over a value beside a window pinned to it, which settled only on a second pass.
• Canonicalization of minProperties: 1 beside a not that already demands a key, which settled only on a second pass.

Fixed

• absolute_keyword_location naming the sibling keyword whose validator carries the check: maxLength beside minLength, minItems, maxItems or type beside items, and a two-name required beside properties. #1579
• schemaLocation in evaluate output naming minLength for a maxLength failure. #1579

[0.54.0] - 2026-09-06

Added

• backend = SerdeJson on #[jsonschema::validator], generating a validator that reads serde_json::Value.
• PreparedDocument::unsatisfiable_pointers, listing every subschema no value satisfies.

... (truncated)

Commits

• c5ce51d chore(ruby): Release 0.56.0
• 1e244c9 chore(rust): Release 0.56.0
• 507d66a build: Switch to weekly dependabot
• 7762942 build: Increase timeout for s390x test job
• 7d4b40d test: Update hegeltest to 0.41
• b2622e5 build(deps): bump taiki-e/install-action from 2.87.3 to 2.87.8
• c983788 build(deps): update wasm-bindgen requirement from =0.2.127 to =0.2.128
• 6e7e8f6 build(deps): bump crates/jsonschema/tests/suite
• 3f2c0a0 perf: Preparing or building a schema that names no base URI parses the defaul...
• 2b4f25b perf: More benchmarks
• Additional commits viewable in compare view

[github-actions]

Thanks @dependabot[bot] for taking the time to contribute.

This repository is observing a maintainer-managed PR intake gate in dry-run mode, so this pull request is staying open. This note helps maintainers prepare the allowlist before any enforcement is considered.

Please read CONTRIBUTING.md for the expected contribution shape. A maintainer can grant recurring PR access by commenting /lgtm on a pull request.

[Hmbown]

Brought this branch up to date with main (green since ccbef2c) so CI runs against a base that passes; the earlier red lanes were inherited from the red main this PR was opened on. Merging once the re-run is green.

---

Generated by Claude Code

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no bugs or issues to report.

[claude]

Claude finished @Hmbown's task in 1m 11s —— View job

---

Review: bump jsonschema 0.52.1 → 0.56.0

• Gather context (PR diff, changed files)
• Understand request (dependabot version bump review)
• Inspect diff for correctness / breaking API changes
• Check for provider/model/route safety implications
• Check build/compile against new jsonschema version
• Post final review

Summary: Mechanical dependency bump. Cargo.toml moves the jsonschema requirement from 0.52 to 0.56 (crates/workflow-js is the sole consumer, per the comment on that line); Cargo.lock updates jsonschema/jsonschema-regex/jsonschema-value/referencing in lockstep, plus a batch of unrelated transitive windows-sys (→0.61.2) and getrandom (→0.4.3 for tempfile) bumps that came along for the ride from re-resolution — harmless, matches the reported +27/-26 diffstat, nothing to flag.

API surface actually used (crates/workflow-js/src/schema.rs, crates/tui/src/tools/subagent/tests.rs): jsonschema::validator_for, Validator::iter_errors, jsonschema::options().with_draft(Draft::Draft202012). None of this touches the two breaking items in the 0.53–0.56 changelog range — canonical::CanonicalView::Raw gaining a RawView field, and the absolute_keyword_location/schemaLocation sibling-keyword fix in evaluate output. A repo-wide grep for CanonicalView, canonical::, and is_known_format turns up no hits, so the breaking change doesn't reach this codebase. Validator::iter_errors and validator_for are unaffected per the changelog.

Provider/model/route safety: not applicable — no code in this diff touches model/provider selection or prompt text.

Tests: existing coverage in schema.rs (decodes_plain_json, rejects_schema_violation_with_the_validation_kind, rejects_invalid_schema_before_spawn, etc.) already exercises the exact validator_for + iter_errors path this bump touches. No new tests needed for a version bump; none added, which is correct here (this isn't new behavior).

Not verified: I was unable to run cargo check -p codewhale-workflow-js or cargo test in this session — Bash tool calls require interactive approval that isn't available to this review job. Based on static inspection the used API surface is stable across 0.52→0.56, but a maintainer or CI run of cargo check -p codewhale-workflow-js --offline / the jsonschema decode tests in crates/workflow-js would be the concrete confirmation before merge.

No correctness, safety, or security issues found. Approve pending CI (build/tests) since I couldn't execute them locally.

Branch: dependabot/cargo/jsonschema-0.56.0

[codewhale-agent]

Codewhale review did not run

codewhale review --pr 6345 (provider: deepseek) could not reach the model: Review pass 1/1 request failed: Responses API request failed: LLM error: HTTP 402: Insufficient Balance; no partial review was accepted or posted; publication: n.
This is a provider funding/config problem, not a finding about this PR. The check stays advisory; a maintainer with secret access needs to fund or rotate the review key (see .github/workflows/codewhale-review.yml). Re-run the workflow after that.