Skip to content

CHANGELOG.md

Latest commit

 

History

History
89 lines (74 loc) · 5.07 KB

CHANGELOG.md

File metadata and controls

89 lines (74 loc) · 5.07 KB

Changelog

See https://docs.meshery.io/project/releases

v1.1.0 — Phase 1 of the identifier-naming migration

Marks the completion of Phase 1 of the identifier-naming migration. Phase 1 is governance- and tooling-only: no wire-format changes are made in this release. The validator, advisory baseline, CI jobs, and docs have been realigned so that every subsequent per-resource rollout (Phase 2 non-breaking alignment, Phase 3 versioned migrations) has a single source of truth and a single rule surface to reason against.

The canonical-casing contract

In one sentence: wire is camelCase everywhere; DB is snake_case; Go fields follow Go idiom; the ORM layer is the sole translation boundary. Concretely, every JSON tag and every OpenAPI schema property — DB-backed or not — is camelCase; the snake_case DB column name lives only in x-oapi-codegen-extra-tags.db. URL path and query parameters are camelCase with an Id suffix ({orgId}, ?userId=...). The earlier conditional rule "JSON tag matches db: tag when the field is DB-backed" has been retired. Legacy resources retain their published wire casing until their next API-version bump; new API versions adopt the canonical contract uniformly.

See docs/identifier-naming-migration.md §1 for the full contract table and AGENTS.md § Casing rules at a glance for the per-layer rules enforced by the validator.

Phase 1 landings

  • 1.A — AGENTS.md identifier-naming amendment (#788): Replaced the "DB-backed = snake on the wire" rule with the canonical camelCase-on-wire contract; converted page_size / total_count from a perpetual exception to a deprecation list; added § Identifier-naming migration pointing at the executable plan.
  • 1.B — Rule 6 inversion / Rule 32 retirement (#791): Rewrote the schema-property casing rule to require camelCase unconditionally; removed the DB-backing exception path and retired the superseded rule that enforced snake-on-DB-backed tags.
  • 1.C — "Partial casing migrations forbidden" rule (Rule 45) (#792): New rule walks every components/schemas/<Entity> and fails any struct that mixes JSON-tag conventions across its properties (camel + snake + ID). Catches the OrgID/workspace_id/user_id drift class on a single struct.
  • 1.D — "Sibling-endpoint parameter parity" rule (Rule 46) (#793): New rule flags GET /api/<entity> siblings that omit orgIdQuery when their peers declare it — the class of drift that produced the workspaces-400 incident.
  • 1.E — Query-parameter casing (Rule 4 extension) (#794): Extended the path-param casing rule to query parameters so user_id, organizationId, and any orgID that slips into parameters: [{in: query}] is flagged at the same severity as path parameters.
  • 1.F — TypeScript consumer auditor (#795): Added validation/consumer_ts.go and wired it into make consumer-audit. Parses RTK Query builder.query / builder.mutation call sites across meshery/meshery/ui, meshery-cloud/ui, and meshery-extensions/meshmap and reports three finding classes (case-flip, snake-case-wrapper, snake-case-param) against the schemas endpoint index.
  • 1.G — Advisory baseline refresh (#796): Regenerated build/validate-schemas.advisory-baseline.txt against the new rule surface so the advisory-audit CI job stays green while Phases 2–3 migrate each resource. The baseline is subtractive — any new violation introduced after this snapshot blocks CI.
  • 1.H — CI workflow wiring (#797): Adds a non-blocking consumer-audit job to .github/workflows/schema-audit.yml that checks out the downstream repos and runs make consumer-audit on every PR, posting divergence counts per downstream repo. Advisory-only; lands alongside this 1.I release.

Dependency on this version

Every Phase 2 agent pins its target repo's @meshery/schemas dependency to this version — no Phase 2 PR runs until this release is available on npm and as a Go module tag. The Phase 3 per-resource rollouts (22 resources in parallel) likewise depend on consumers reaching this version. Downstream repositories pin as follows:

# TypeScript / npm consumers (meshery, meshery-cloud, meshery-extensions UI):
npm install @meshery/schemas@1.1.0

# Go module consumers (meshery server, meshery-cloud server):
go mod edit -require github.com/meshery/schemas@v1.1.0
go mod tidy

After merging this release, downstream repos can begin Phase 2 non-breaking alignment work (handler query-param casing, outbound URL alignment, local-type displacement) and Phase 3 per-resource API-version authoring. See the orchestration DAG in §11 of the plan for the dependency ordering.

Breaking changes

None. Phase 1 is a governance and tooling release; all wire contracts published in prior releases remain accepted. Existing json:"snake_case" tags on DB-backed fields in already-published API versions remain valid until their resource is migrated to a new API version under Phase 3.