fix(openapi): spec property naming now mirrors the runtime Newtonsoft serializer (#198) #201

Merged
timothy merged 4 commits from fix/198-openapi-casing into main 2026-07-09 20:19:21 +02:00
4 Commits
Author SHA1 Message Date
timothyandClaude Fable 5 e0d30c9b7b test(api): cover [JsonProperty] path + fail loudly on rename collision (#198, review)
Build ErsatzTV Image / Build & test (.NET) (pull_request) Successful in 7m59s
Build ErsatzTV Image / Docs update reminder (pull_request) Successful in 14s
Build ErsatzTV Image / EF migration integrity (SQLite + MySql) (pull_request) Successful in 10m31s
Build ErsatzTV Image / Build & push image (amd64) (pull_request) Has been skipped
Adversarial-review follow-ups:

- OpenApiSerializerContractTests: add a fifth case, a fully-populated
  ChannelResponseModel — the only DTO with a [JsonProperty("ffmpegProfile")]
  override, i.e. the attribute path of NewtonsoftSchemaNamingTransformer that
  the existing four cases never exercised.
- NewtonsoftSchemaNamingTransformer: a rename colliding with an existing schema
  key now throws InvalidOperationException (naming the type and keys) instead of
  silently overwriting/dropping a property — the generator must never emit a
  lossy spec.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-09 20:06:26 +02:00
timothyandClaude Fable 5 d9faa8be7b docs: spec-follows-runtime naming decision + casing wart (#198)
- api-conventions.md §5a: runtime Newtonsoft casing vs generated spec, the
  schema transformer that mirrors it, and the contract test guarding it.
- decisions.md: append the "wire format is source of truth; spec follows via the
  real contract resolver" decision.
- spa-conventions.md §4: trust the generated key casing; note the removed
  troubleshooting escape hatch and runtime-cased test mocks.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-09 20:00:48 +02:00
timothyandClaude Fable 5 7ffde6225f fix(web): read runtime-cased ffmpegProfileId/ffmpegProfile keys (#198)
Regenerate v1.d.ts from the aligned spec (fFmpegProfileId->ffmpegProfileId,
fFmpegProfile->ffmpegProfile) and update every SPA reader/writer and test mock
to the runtime casing:

- ChannelEditScreen: read channel.ffmpegProfileId; draft/set/select use
  ffmpegProfileId.
- ChannelBuilder: template reads/writes + ADVANCED_KEYS use ffmpegProfileId.
- App.tsx channel list: read channel.ffmpegProfile (was fFmpegProfile, a latent
  bug that always rendered "Unassigned" since the runtime never sent that key).
- PlaybackTroubleshootingScreen: drop the #198 escape hatch (rawChannel cast +
  dual-read) and read data.channel.ffmpegProfileId directly.
- Test mocks now use runtime casing; pinning comments updated.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-09 20:00:38 +02:00
timothyandClaude Fable 5 f26790aa35 fix(api): mirror runtime Newtonsoft JSON casing in OpenAPI spec (#198)
The OpenAPI document is generated from System.Text.Json metadata, whose
camelCase drifted from the runtime MVC serializer (Newtonsoft via
CustomContractResolver/CustomNamingStrategy): the spec said "fFmpegProfileId"
and "fFmpegProfile" while the wire emits "ffmpegProfileId" (naming-strategy
special case) and "ffmpegProfile" (ChannelResponseModel's [JsonProperty]
override). That fed the SPA the wrong keys.

Add NewtonsoftSchemaNamingTransformer, an OpenAPI schema transformer registered
on all three documents that renames each object schema's Properties (and
Required) keys through the SAME Newtonsoft contract resolver the runtime uses,
so the spec matches the wire format by construction. Regenerate v1.json.

Guard with OpenApiSerializerContractTests: serializes fully-populated DTOs
through the runtime Newtonsoft settings and pins the v1.json schema property
sets to the emitted keys, failing if generation drifts again.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-09 20:00:23 +02:00