Menu

JSON Schema quality evidence

JSON Schema Test Suite Results for JSONBuddy

JSONBuddy is tested against the official JSON Schema Test Suite. This page presents run-specific totals with their method, provenance, and limitations.

View the official test suite Explore the ValBuddy CLI

Result scope

Five supported drafts

Required-suite results from two identical native standard-validator runs.

Native CLI evidence

Required tests

Draft Selected Passed Failed Skipped Pass rate
Draft 4http://json-schema.org/draft-04/schema# 618 614 4 0 99.35% (614 of 618 selected)
Draft 6http://json-schema.org/draft-06/schema# 841 834 7 0 99.17% (834 of 841 selected)
Draft 7http://json-schema.org/draft-07/schema# 929 922 7 0 99.25% (922 of 929 selected)
Draft 2019-09https://json-schema.org/draft/2019-09/schema 1261 1252 9 0 99.29% (1252 of 1261 selected)
Draft 2020-12https://json-schema.org/draft/2020-12/schema 1301 1294 7 0 99.46% (1294 of 1301 selected)

Optional tests

Not run for the drafts in this result set.

Result publication

Evidence first, totals second

The official suite is a test corpus, not a certification program. Results on this page describe one identified product build, suite revision, dialect, selection, and validation mode; they do not certify JSONBuddy or imply complete support beyond that scope.

Native desktop engine

Candidate provenance

Tested product
JSONBuddy ValidatorBuddy 7.4.1.0
Build
Unicode Release x64
Validation mode
Native CLI, standard validator
Runs
and
Suite revision
f6fd52a0a95472e079cbfc6ef7f089702b80e045

Web API

Reported independently

No Web API suite results are published on this page.

A native valbuddy.exe run does not establish Web API behavior. API results will require their own identified endpoint and implementation version, parity evidence, repeat run, and publication artifact.

Methodology

How the native suite run is produced

The native standard validator runs locally against a clean checkout pinned to the suite commit shown above. The suite's http://localhost:1234/ retrieval URIs are mapped in-process to its remotes directory, so the canonical run does not depend on a web server or uploaded fixtures.

Reproducible command shape

valbuddy.exe -jsts "JSON-Schema-Test-Suite" ^
              -draft <draft-id> ^
              -report-json "artifacts\<draft-id>.json" ^
              -report-junit "artifacts\<draft-id>.junit.xml"

The command is run separately for each displayed draft. The required selection uses the standard validator. Optional tests and streaming mode are off. Optional and streaming results remain separate publication lanes when they are intentionally run and reviewed.

Publication controls

  1. Pin a clean suite revision and the exact product build.
  2. Produce JSON and JUnit reports without runner errors.
  3. Repeat the run and compare every stable case identity and outcome.
  4. Review each failure or skip and link it to a real tracking issue.
  5. Generate the displayed totals from the reviewed JSON report artifact.

Read the suite runner documentation.

Required and optional coverage

Selections remain separate

Required tests exercise the dialect's core suite. Optional tests cover separately selected behavior such as formats. The page never combines those selections into one percentage and never removes a failed or skipped case from the denominator.

Required tests

Required totals appear as their own selected, passed, failed, skipped, and pass-rate values. Runner-error counts remain available in the machine-readable artifact.

Optional tests

Optional totals appear only after the corresponding categories have been deliberately enabled, repeated, and reviewed.

Streaming validation

Streaming-mode evidence remains distinct from standard-validator evidence until semantic parity has been demonstrated.

Limitations

What is not claimed

  • No certification: the JSON Schema Test Suite does not issue certification.
  • No full-compliance statement while any required result remains unresolved.
  • No Web API conclusion from a native CLI run.
  • No claim for optional tests or streaming mode when those lanes were not selected.
  • No hidden exclusions: every published failure or skip must retain its denominator and link to its reviewed explanation.
Current review boundary

The complete latest native counts are shown, including every known failure. Maintainer review and real tracking issues are still required before this becomes a reviewed expectations baseline.