arcodange/dance-lessons-coach
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

Compare commits

base: arcodange:vibe/batch-pr-chartv-vault-crds
Branches Tags
arcodange:main arcodange:vibe/batch-pr-chartv-vault-crds arcodange:vibe/batch4-task-a-changelog-update arcodange:fix/should-enable-v2-tag-parsing arcodange:fix/config-test-main-quiet-zerolog arcodange:fix/race-in-config-watcher arcodange:feat/adr-0023-phase3b-sampler-reload arcodange:fix/bdd-cache-flush-and-revert-schema-isolation arcodange:fix/exclude-v2-from-default-bdd arcodange:feature/restructure-agents-md
..
compare: arcodange:fix/exclude-v2-from-default-bdd
Branches Tags
arcodange:main arcodange:vibe/batch-pr-chartv-vault-crds arcodange:vibe/batch4-task-a-changelog-update arcodange:fix/should-enable-v2-tag-parsing arcodange:fix/config-test-main-quiet-zerolog arcodange:fix/race-in-config-watcher arcodange:feat/adr-0023-phase3b-sampler-reload arcodange:fix/bdd-cache-flush-and-revert-schema-isolation arcodange:fix/exclude-v2-from-default-bdd arcodange:feature/restructure-agents-md

1 Commits
vibe/batch ... fix/exclud

Author SHA1 Message Date
Gabriel Radureau
6b39d3c3c9 🐛 fix(bdd): exclude @v2 scenarios from default BDD test runs
All checks were successful
CI/CD Pipeline / Build Docker Cache (push) Successful in 25s
Details
CI/CD Pipeline / CI Pipeline (push) Successful in 7m37s
Details
CI/CD Pipeline / Trigger Docker Push (push) Has been skipped
Details 
The 4 v2 scenarios in greet.feature require special config
(FEATURE=greet GODOG_TAGS=@v2) to enable the v2 endpoint via
shouldEnableV2(). Without that config, all v2 scenarios fail
with "v2 endpoint not available".

Two fixes:
1. Tag the 3 untagged v2 scenarios with @v2 @api (one already
   had it, others were missing tags)
2. Extend DEFAULT_TAGS in run-bdd-tests.sh to exclude @v2

This makes the default BDD test run pass on CI without v2 setup.
v2 scenarios can still be run explicitly with:
  FEATURE=greet GODOG_TAGS=@v2 go test ./features/greet/...

Companion to PR #26 (BDD_SCHEMA_ISOLATION) - both target CI green.

🤖 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
 2026-05-03 13:59:11 +02:00
131 changed files with 267 additions and 12141 deletions
Expand all files Collapse all files

16
.gitea/workflows/ci-cd.yaml
View File

@@ -219,11 +219,9 @@ jobs:
export DLC_DATABASE_PASSWORD=postgres
export DLC_DATABASE_NAME=dance_lessons_coach_bdd_test
export DLC_DATABASE_SSL_MODE=disable
# T12: per-package isolated Postgres schema with migrations (re-enables what
# PR #26 attempted but couldn't deliver because the empty schemas had no tables).
# The fix: testserver Start() now builds a per-package isolated repo via
# user.NewPostgresRepositoryFromDSN which DOES run AutoMigrate against the new
# schema. Packages then run in parallel (~2.85x speedup observed locally).
# Enable per-scenario schema isolation (ADR-0025) to prevent flaky AuthBDD failures.
# Without this, scenarios share the public schema and pollute each other's state.
# Observed flakiness: same code passes in #605, fails in #606 on TestAuthBDD/*.
export BDD_SCHEMA_ISOLATION=true
./scripts/run-bdd-tests.sh
@@ -299,13 +297,7 @@ jobs:
# Check for version bump on main branch
if [ "${{ github.ref }}" = "refs/heads/main" ]; then
echo "🔖 Checking for version bump..."
# Read commit message from git, NOT from the workflow event payload.
# The event-payload expression is interpolated literally into the
# rendered script (even inside comments — see PR #38 + #46), so any
# backtick / unbalanced quote / multi-line body breaks bash parsing.
# git log is interpolation-free and safe.
COMMIT_MSG=$(git log -1 --pretty=%B)
./scripts/ci-version-bump.sh "$COMMIT_MSG" --no-push
./scripts/ci-version-bump.sh "${{ github.event.head_commit.message }}" --no-push
fi
# Single push for all commits (this is the ONLY push in the entire workflow)

13
.gitea/workflows/docker-push.yaml
View File

@@ -6,6 +6,7 @@
name: Docker Push
on:
# Manual trigger for testing or production
workflow_dispatch:
inputs:
ref:
@@ -13,18 +14,6 @@ on:
required: false
type: string
default: ''
push:
branches:
- main
paths-ignore:
- 'README.md'
- 'AGENTS.md'
- 'CHANGELOG.md'
- 'AGENT_CHANGELOG.md'
- 'documentation/**'
- 'adr/**'
- 'chart/**'
- 'features/**'
# Environment variables
env:

1
.gitignore vendored
View File

@@ -42,6 +42,5 @@ frontend/.output/
frontend/dist/
frontend/.env
frontend/.cache/
frontend/storybook-static/
frontend/test-results/
frontend/playwright-report/

43
AGENTS.md
View File

@@ -11,9 +11,6 @@ AI agent reference for developing, testing, and operating the dance-lessons-coac
| Logging | Zerolog | v1.35.0 |
| Configuration | Viper | v1.21.0 |
| Telemetry | OpenTelemetry | v1.43.0 |
| Database | PostgreSQL + GORM | (optional, for user persistence) |
| Auth | JWT + bcrypt | with rotating secrets |
| Email | SMTP (Mailpit dev) | for magic-link delivery |
## Project Structure
@@ -24,14 +21,11 @@ dance-lessons-coach/
│ ├── greet/ # CLI application
│ └── server/ # Web server entry point
├── pkg/
│ ├── auth/ # Auth context keys, OIDC client, helpers
│ ├── config/ # Viper-based configuration
│ ├── email/ # SMTP sender abstraction
│ ├── greet/ # Core domain logic + API handlers
│ ├── server/ # HTTP server, routing, graceful shutdown
│ ├── telemetry/ # OpenTelemetry instrumentation
│ ├── user/ # User domain (auth, JWT, repository)
│ ├── user/api/ # Auth + admin HTTP handlers
│ └── validation/ # Request validation
├── scripts/ # Server lifecycle, build, test scripts
├── config.yaml # Configuration file
@@ -86,35 +80,20 @@ logging:
| Method | Path | Description |
|--------|------|-------------|
| GET | `/api/health` | Liveness — always `{"status":"healthy"}` |
| GET | `/api/healthz` | Alternative liveness probe (k8s convention) |
| GET | `/api/ready` | Readiness — 200 when ready, 503 during shutdown |
| GET | `/api/info` | Composite info endpoint (cf. ADR-0026) |
| GET | `/api/version` | Version info (`?format=plain\|full\|json`) |
| GET | `/api/v1/greet/` | Default greeting |
| GET | `/api/v1/greet/{name}` | Personalized greeting |
| POST | `/api/v1/auth/login` | Username + password login (JWT) |
| POST | `/api/v1/auth/register` | Account creation |
| POST | `/api/v1/auth/validate` | JWT validation |
| POST | `/api/v1/auth/password-reset/request` | Request a password reset |
| POST | `/api/v1/auth/password-reset/complete` | Complete a password reset |
| POST | `/api/v1/auth/magic-link/request` | Passwordless : request a magic link by email |
| GET | `/api/v1/auth/magic-link/consume` | Passwordless : consume a magic-link token |
| GET | `/api/v1/auth/oidc/{provider}/start` | OIDC : begin authorization (PKCE) |
| GET | `/api/v1/auth/oidc/{provider}/callback` | OIDC : authorization code → JWT |
| GET | `/api/v1/admin/jwt/secrets` | Admin : list JWT signing secrets |
| POST | `/api/v1/admin/jwt/secrets` | Admin : add a JWT signing secret |
| POST | `/api/v1/admin/jwt/secrets/rotate` | Admin : rotate the active JWT secret |
| POST | `/api/v2/greet` | V2 greeting with validation (feature-flagged) |
| POST | `/api/admin/cache/flush` | Admin : flush in-memory caches |
| GET | `/swagger/` | Swagger UI |
| GET | `/swagger/doc.json` | OpenAPI spec (source of truth) |
| GET | `/swagger/doc.json` | OpenAPI spec |
```bash
curl http://localhost:8080/api/health
curl http://localhost:8080/api/ready
curl http://localhost:8080/api/v1/greet/Alice
# See `/swagger/` for the full interactive list of endpoints.
# The OpenAPI spec at `/swagger/doc.json` is the source of truth.
curl -X POST http://localhost:8080/api/v2/greet \
-H "Content-Type: application/json" -d '{"name":"Alice"}'
```
## Testing
@@ -172,9 +151,19 @@ docker run -d --name jaeger \
## Architecture Decision Records
The full index is at [adr/README.md](adr/README.md) — 30 ADRs covering Go version, Chi router, Zerolog, OpenTelemetry, BDD testing, JWT auth, PostgreSQL, OIDC migration, email infrastructure, etc.
| ADR | Decision |
|-----|----------|
| [0001](adr/0001-go-1.26.1-standard.md) | Go 1.26.1 |
| [0002](adr/0002-chi-router.md) | Chi router |
| [0003](adr/0003-zerolog-logging.md) | Zerolog |
| [0004](adr/0004-interface-based-design.md) | Interface-based design |
| [0005](adr/0005-graceful-shutdown.md) | Graceful shutdown |
| [0006](adr/0006-configuration-management.md) | Viper configuration |
| [0007](adr/0007-opentelemetry-integration.md) | OpenTelemetry |
| [0008](adr/0008-bdd-testing.md) | BDD with Godog |
| [0009](adr/0009-hybrid-testing-approach.md) | Hybrid testing strategy |
Add a new ADR : copy an existing file in `adr/`, edit it, update `adr/README.md`.
Add a new ADR: copy an existing file, edit it, update `adr/README.md`.
## Commit Conventions

42
CHANGELOG.md
View File

@@ -1,42 +0,0 @@
# Changelog
All notable changes to this project will be documented in this file.
The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
## [Unreleased]
### Added
- ✨ `GET /api/v1/uptime` endpoint (PR #67) — returns server start_time and uptime_seconds
- 📝 mkcert local HTTPS doc + Makefile `cert` target (PR #68) — prep for ADR-0028 Phase B OIDC callbacks
- ✨ `pkg/auth/` skeleton for OpenID Connect (PR #69) — types + client surface, handlers come later (Phase B.3+)
- 📝 ADR-0028 Phase B roadmap document (PR #71) — outlines remaining B.3 / B.4 / B.5 work
- ✨ `pkg/auth/` OIDC client implementation : Discover, RefreshJWKS, ExchangeCode, ValidateIDToken (PR #74) — completes ADR-0028 Phase B.3
- ✨ OIDC HTTP handlers : `/api/v1/auth/oidc/{provider}/start` and `/callback` with PKCE + sign-up-on-first-use (PR #75) — completes ADR-0028 Phase B.4
- 🧪 OIDC handler unit tests covering start/callback rejection paths and PKCE redirect (PR #76)
- 📝 `documentation/AUTH.md` synthesis covering Phase A + B current state (PR #73)
- 📝 `documentation/MISTRAL-AUTONOMOUS-PATTERN.md` contributor guide for the Mistral autonomous pattern that ships PRs (PR #78)
- 📝 PHASE_B_ROADMAP marks B.3 + B.4 done (PR #80)
- 📝 documentation/2026-05-05-AUTONOMOUS-SESSION-RECAP.md captures the day's 24 Mistral autonomous PRs (PR #81)
- 📝 README link to Mistral autonomous pattern doc (PR #83)
- 📝 documentation/STATUS.md project snapshot for onboarding (PR #85)
- 📝 documentation guides cherry-picked from PR #17 : CLI.md, CODE_EXAMPLES.md, HISTORY.md, OBSERVABILITY.md, ROADMAP.md, TROUBLESHOOTING.md (PR #87)
- 🔒 redact JWT tokens and HMAC secrets in trace logs of pkg/user/auth_service.go via sha256 fingerprints (PR #88)
- ✨ Dockerfile (root) + Helm chart for k3s homelab deployment, degraded mode without DB/SMTP/Vault (PR #89)
- ♻️ move UserContextKey + GetAuthenticatedUserFromContext from pkg/greet to pkg/auth (PR #90)
- ♻️ split AuthMiddleware into OptionalHandler + RequiredHandler with RFC 6750 challenge headers, narrow tokenValidator interface, case-insensitive Bearer (PR #91)
- 🧪 unit tests for AuthMiddleware Optional/Required handlers + extractBearerToken edge cases (PR #92)
- 📝 refresh AGENTS.md and README.md to reflect auth endpoints (magic-link, OIDC, JWT admin), pkg/auth, pkg/email, pkg/user/api packages, and 30-ADR index. Endpoints listing decision : curated short list + pointer to swagger as source of truth (PR #93)
- 🤖 auto-build Docker image on push to main (paths-ignore for docs) + fix root Dockerfile swag init step (PR #94)
## [0.1.0] - 2026-05-05
### Added
- Magic-link passwordless authentication (ADR-0028 Phases A.1 through A.5, PRs #59-#63)
- OIDC provider config skeleton (ADR-0028 Phase B.1 prep, PR #64)
- Magic-link expired-token cleanup loop (PR #65)
- Mailpit local SMTP infrastructure (ADR-0029)
- BDD parallel email assertion strategy (ADR-0030)

43
Dockerfile
View File

@@ -1,43 +0,0 @@
# Build dance-lessons-coach Docker image
FROM golang:1.26-alpine AS builder
# Install git (required for go mod download)
RUN apk add --no-cache git
# Set working directory
WORKDIR /app
# Copy go module files and download dependencies
COPY go.mod go.sum ./
RUN go mod download
# Copy entire source code
COPY . .
# Generate Swagger documentation if not already present
# (pkg/server/docs/ is gitignored ; the binary //go:embed depends on it)
RUN if [ ! -f pkg/server/docs/swagger.json ]; then \
go install github.com/swaggo/swag/cmd/swag@latest && \
cd pkg/server && go generate ; \
fi
# Build the server binary
RUN go build -o app ./cmd/server
# Final lightweight stage
FROM alpine:latest
# Install CA certificates for HTTPS
RUN apk --no-cache add ca-certificates
# Set working directory
WORKDIR /root/
# Copy binary from builder stage
COPY --from=builder /app/app .
# Expose port 8080
EXPOSE 8080
# Start the server
CMD ["./app"]

24
Makefile
View File

@@ -1,24 +0,0 @@
# dance-lessons-coach Makefile — minimal targets for local development.
# This is a starter Makefile ; expand as needed (build, test, run, etc.).
# Existing build/test workflows live in scripts/ and remain authoritative.
CERT_DIR := ./certs
.PHONY: help cert clean-cert
help:
@echo "Available targets:"
@echo " cert Generate local-dev TLS certs via mkcert (cf. documentation/MKCERT.md)"
@echo " clean-cert Remove generated TLS certs"
@echo " help Show this help"
cert: $(CERT_DIR)
@command -v mkcert >/dev/null 2>&1 || { echo >&2 "mkcert not found. See documentation/MKCERT.md to install."; exit 1; }
mkcert -cert-file $(CERT_DIR)/dev-cert.pem -key-file $(CERT_DIR)/dev-key.pem localhost 127.0.0.1 ::1
@echo "Certs ready at $(CERT_DIR)/. Cf. documentation/MKCERT.md for usage."
$(CERT_DIR):
mkdir -p $(CERT_DIR)
clean-cert:
rm -rf $(CERT_DIR)

21
README.md
View File

@@ -17,13 +17,9 @@ Go web service demonstrating idiomatic package structure, versioned JSON API, an
- Viper configuration (file + env vars)
- Readiness endpoint for Kubernetes / service mesh
- OpenTelemetry / Jaeger distributed tracing
- OpenAPI / Swagger UI (embedded in binary, source of truth at `/swagger/doc.json`)
- Username + password authentication with JWT (rotating secrets)
- Passwordless magic-link authentication (email-delivered, ADR-0028 Phase A)
- OIDC authentication with PKCE (multi-provider, ADR-0028 Phase B)
- PostgreSQL user persistence with GORM
- BDD + unit tests (Godog)
- Mistral autonomous PR pattern (cf. [documentation/MISTRAL-AUTONOMOUS-PATTERN.md](documentation/MISTRAL-AUTONOMOUS-PATTERN.md))
- OpenAPI / Swagger UI (embedded in binary)
- PostgreSQL user service with JWT auth
- BDD + unit tests
## Quick Start
@@ -66,21 +62,16 @@ See `config.example.yaml` for a full template.
## API
The full interactive list is in the Swagger UI at `/swagger/` (source of truth at `/swagger/doc.json`). Most-used endpoints :
| Method | Path | Description |
|--------|------|-------------|
| GET | `/api/health` | Liveness check |
| GET | `/api/ready` | Readiness check (503 during shutdown) |
| GET | `/api/version` | Version info |
| GET | `/api/version` | Version info (`?format=plain\|full\|json`) |
| GET | `/api/v1/greet/` | Default greeting |
| GET | `/api/v1/greet/{name}` | Named greeting |
| POST | `/api/v1/auth/login` | Login (JWT) |
| POST | `/api/v1/auth/magic-link/request` | Passwordless magic-link |
| GET | `/api/v1/auth/oidc/{provider}/start` | OIDC login |
| POST | `/api/v2/greet` | V2 greeting with validation |
| GET | `/swagger/` | Swagger UI |
This decision is intentional : the markdown table drifts ; swagger.json doesn't (it's regenerated from `swag` annotations on every build). Curated short list here for discovery, swagger for completeness.
## Testing
```bash

9
adr/0009-hybrid-testing-approach.md
View File

@@ -1,9 +1,10 @@
# Combine BDD and Swagger-based testing
**Status:** Implemented (BDD + OpenAPI documentation operational; SDK generation explicitly out of scope — would require a fresh ADR if reopened)
**Status:** Partially Implemented (BDD + Documentation only)
**Authors:** Gabriel Radureau, AI Agent
**Date:** 2026-04-05
**Last Updated:** 2026-05-05
**Last Updated:** 2026-04-05
**Implementation Status:** BDD testing and OpenAPI documentation completed, SDK generation deferred
## Context and Problem Statement
@@ -35,7 +36,7 @@ Chosen option: "Hybrid approach" because it provides the best combination of beh
## Implementation Status
**Status**: ✅ Implemented (BDD + OpenAPI documentation operational; SDK generation explicitly out of scope)
**Status**: ✅ Partially Implemented (BDD + Documentation only)
### What We Actually Have
@@ -328,7 +329,7 @@ If we need SDK generation in the future:
- Add SDK-based BDD tests
- Implement true hybrid testing approach
**Current Status:** ✅ Implemented (BDD + OpenAPI documentation; SDK generation out of scope)
**Current Status:** ✅ Partially Implemented (BDD + Documentation)
**BDD Tests:** http://localhost:8080/api/health (all passing)
**OpenAPI Docs:** http://localhost:8080/swagger/
**OpenAPI Spec:** http://localhost:8080/swagger/doc.json

7
adr/0013-openapi-swagger-toolchain.md
View File

@@ -1,10 +1,11 @@
# 13. OpenAPI/Swagger Toolchain Selection
**Date:** 2026-04-05
**Status:** Implemented (OpenAPI documentation operational; SDK generation explicitly out of scope, see ADR-0009)
**Status:** Partially Implemented (Documentation only)
**Authors:** Arcodange Team
**Implementation Date:** 2026-04-05
**Last Updated:** 2026-05-05
**Last Updated:** 2026-04-05
**Status:** OpenAPI documentation operational, SDK generation deferred
## Context
@@ -982,7 +983,7 @@ If we need SDK generation in the future:
4. Implement request validation middleware
5. Migrate to OpenAPI 3.0 if needed
**Current Status:** ✅ Implemented (OpenAPI documentation; SDK generation out of scope)
**Current Status:** ✅ Partially Implemented (Documentation only)
**Implementation:** swaggo/swag with embedded documentation
**Documentation:** http://localhost:8080/swagger/
**OpenAPI Spec:** http://localhost:8080/swagger/doc.json

2
adr/0018-user-management-auth-system.md
View File

@@ -1,7 +1,7 @@
# 18. User Management and Authentication System
**Date:** 2026-04-06
**Status:** Implemented (user model, JWT auth, password-reset workflow, admin endpoints, greet personalization, BDD coverage all live; future enhancements like 2FA / email verification belong in separate ADRs)
**Status:** Partially Implemented
**Authors:** Product Owner
**Decision Drivers:** Security, User Personalization, Admin Functionality

10
adr/0019-postgresql-integration.md
View File

@@ -1,7 +1,7 @@
# 19. PostgreSQL Database Integration
**Date:** 2026-04-07
**Status:** Implemented (core integration; performance tuning + extended monitoring tracked as future work)
**Status:** Partially Implemented
**Authors:** Product Owner
**Decision Drivers:** Data Persistence, Scalability, Production Readiness
@@ -671,10 +671,10 @@ func AfterScenario(ctx context.Context, sc *godog.Scenario, err error) (context.
## Future Considerations
### Immediate Next Steps (Post-Migration)
1. **CI/CD Integration:** Add PostgreSQL to CI pipeline — ✅ Implemented (`postgres:15` service in `.gitea/workflows/ci-cd.yaml`, all BDD tests run against real Postgres)
2. **Performance Tuning:** Query optimization — Deferred. No production hot path identified. Reopen as separate ADR if/when latency budget exceeded.
3. **Monitoring:** Database health metrics — Partial. `/api/healthz` reports DB connectivity. Deeper metrics (slow query log, pool stats) deferred until ADR-0022 cache Phase 2 lands.
4. **Backup Strategy:** Regular database backups — Deferred. No production data yet. Will require separate ADR before any production data lands.
1. **CI/CD Integration:** Add PostgreSQL to CI pipeline
2. **Performance Tuning:** Query optimization
3. **Monitoring:** Database health metrics
4. **Backup Strategy:** Regular database backups
### Long-Term Enhancements
1. **Database Sharding:** For horizontal scaling

4
adr/0021-jwt-secret-retention-policy.md
View File

@@ -1,6 +1,6 @@
# 21. JWT Secret Retention Policy
# 10. JWT Secret Retention Policy
**Status:** Implemented (2026-05-05 — `pkg/user/jwt_manager.go` `RemoveExpiredSecrets` + `StartCleanupLoop`, wired in `pkg/server/server.go` `Run`; admin endpoint `/api/v1/admin/jwt/secrets` remains explicitly out of scope and tracked under @todo BDD scenarios)
**Status:** Proposed
## Context

3
adr/0023-config-hot-reloading.md
View File

@@ -1,9 +1,8 @@
# Config Hot Reloading Strategy
**Status:** Implemented — all 4 phases shipped (2026-05-05). Hot-reloadable fields: `logging.level` (Phase 1), `auth.jwt.ttl` (Phase 2), `telemetry.sampler.type` + `telemetry.sampler.ratio` (Phase 3), `api.v2_enabled` (Phase 4). Plumbing: `Config.WatchAndApply` in `pkg/config/config.go` is the single entry point. Phase 2 fixed a pre-existing bug where hardcoded 24h TTL ignored `auth.jwt.ttl`. Phase 4 chose the **always-register-with-middleware-gate** approach: v2 routes are now ALWAYS registered, and `Server.v2EnabledGate` middleware reads the live config on every request (returns 404 + JSON body when disabled). No router rebuild needed for the flag flip. 3 unit tests in `pkg/server/v2_gate_test.go` cover blocked-when-disabled / passes-when-enabled / hot-reload-mid-life-of-same-Server.
**Status:** Proposed
**Authors:** Gabriel Radureau, AI Agent
**Date:** 2026-04-05
**Last Updated:** 2026-05-05
## Context and Problem Statement

28
adr/0024-bdd-test-organization-and-isolation.md
View File

@@ -1,6 +1,6 @@
# ADR 0024: BDD Test Organization and Isolation Strategy
**Status:** Implemented (Phase 1 + Phase 2 + Phase 3 — parallel testing via [PR #35](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/35), isolation strategy detailed in [ADR-0025](0025-bdd-scenario-isolation-strategies.md))
**Status:** Partially Implemented
## Context
@@ -284,22 +284,20 @@ func CleanupFeatureData(featureName string) {
## Implementation Plan
### Phase 1: Refactor Current Tests — ✅ Implemented
1. Split monolithic feature files into feature directories — done (see `features/<domain>/` layout)
2. Create feature-specific test scripts — done
3. Implement basic isolation (config files, database names) — done
### Phase 1: Refactor Current Tests (1-2 weeks)
1. Split monolithic feature files into feature directories
2. Create feature-specific test scripts
3. Implement basic isolation (config files, database names)
### Phase 2: Enhance Test Infrastructure — ✅ Implemented
1. Add synchronization helpers to test framework — done
2. Implement server lifecycle management — done (`pkg/bdd/testserver/server.go`)
3. Create comprehensive cleanup routines — done
### Phase 2: Enhance Test Infrastructure (2-3 weeks)
1. Add synchronization helpers to test framework
2. Implement server lifecycle management
3. Create comprehensive cleanup routines
### Phase 3: Parallel Testing — ✅ Implemented (PR #35, 2026-05-03)
1. Add parallel test execution capability — done (schema-per-package isolation, **2.85x speedup**)
2. Implement port management for parallel runs — done (`pkg/bdd/parallel/port_manager.go`)
3. Add resource monitoring — deferred (not blocking; can be reopened as separate ADR if/when CI flakiness re-emerges)
The strategy choice between alternatives (TRUNCATE vs schema isolation vs container-per-test) is documented in [ADR-0025](0025-bdd-scenario-isolation-strategies.md). Default behavior in CI is `BDD_SCHEMA_ISOLATION=true` (cf. `documentation/BDD_TEST_ENV.md`).
### Phase 3: Parallel Testing (Optional)
1. Add parallel test execution capability
2. Implement port management for parallel runs
3. Add resource monitoring
## Alternatives Considered

2
adr/0025-bdd-scenario-isolation-strategies.md
View File

@@ -1,6 +1,6 @@
# ADR 0025: BDD Scenario Isolation Strategies
**Status:** Implemented (per-package schema isolation since T12 stage 2/2 - 2026-05-03)
**Status:** Partially Implemented
## Context

200
adr/0026-composite-info-endpoint.md
View File

@@ -1,200 +0,0 @@
# ADR 0026: Composite Info Endpoint vs Separate Calls
**Status:** Implemented (2026-05-05 — PR pending)
## Context
The application currently exposes several endpoints that provide system information:
- `/api/version` - returns version, commit, build date, Go version (cached 60s)
- `/api/health` - returns `{"status":"healthy"}` (simple liveness)
- `/api/healthz` - returns rich health info: status, version, uptime_seconds, timestamp
- `/api/ready` - returns readiness with connection details
Frontend components like `HealthDashboard` currently call `/api/healthz` to display server info. However, there is a need for a **composite endpoint** that aggregates:
1. Version information (from `/api/version`)
2. Build metadata (commit hash, build date)
3. Uptime information (from `/api/healthz`)
4. Cache status (enabled/disabled)
5. Health status
This raises an architectural question: **Should we create a new composite `/api/info` endpoint, or should frontend components make multiple separate API calls?**
### The Problem with Separate Calls
If the frontend makes individual calls to `/api/version`, `/api/healthz`, and checks cache config separately:
1. **Multiple network requests**: 3-4 HTTP round trips per page load
2. **Inconsistent data**: Responses may come from different moments in time
3. **No caching coordination**: Each endpoint has its own cache key and TTL
4. **Complex frontend logic**: Need to merge data from multiple sources
5. **Poor user experience**: Slower page loads, multiple loading states
### Current State Analysis
| Endpoint | Data Provided | Cache TTL | Use Case |
|----------|---------------|-----------|----------|
| `/api/version` | version, commit, built, go | 60s | Version info |
| `/api/healthz` | status, version, uptime_seconds, timestamp | None | K8s probes, health dashboard |
| `/api/health` | status: "healthy" | None | Simple liveness |
| `/api/ready` | ready, connections, reason | None | Readiness probes |
The `/api/healthz` endpoint already combines some data (status + version + uptime + timestamp), but it:
- Doesn't include commit_short
- Doesn't include build_date separately
- Doesn't include cache_enabled
- Is not cached
- Has Kubernetes-specific field naming (`healthz`)
## Decision Drivers
* **Performance**: Minimize network round trips for frontend
* **Consistency**: All data should reflect the same point-in-time
* **Maintainability**: Single source of truth for system info
* **Caching**: Reuse existing cache infrastructure (ADR-0022)
* **API Design**: Follow REST principles and existing patterns
* **Backward Compatibility**: Existing endpoints must remain unchanged
## Considered Options
### Option 1: Composite `/api/info` Endpoint (Chosen)
Create a new endpoint that aggregates all required data in a single call.
**Pros:**
- ✅ Single network request for frontend
- ✅ Consistent point-in-time data
- ✅ Can leverage existing cache infrastructure with key `info:json`
- ✅ Follows existing pattern of `/api/version` caching
- ✅ Clean API design - one endpoint, one purpose
- ✅ Reduces frontend complexity
- ✅ Better UX - faster page loads
- ✅ Aligns with ADR-0022 cache strategy (reusable cache key pattern)
**Cons:**
- ⚠️ Duplicates some data from `/api/healthz` and `/api/version`
- ⚠️ Requires new endpoint implementation
- ⚠️ Need to maintain consistency if source endpoints change
### Option 2: Frontend Aggregation with Multiple Calls
Frontend makes separate calls to `/api/version`, `/api/healthz`, and introspects config.
**Pros:**
- ✅ No backend changes required
- ✅ Uses existing endpoints
**Cons:**
- ❌ Multiple network requests (3-4 round trips)
- ❌ Inconsistent data timing
- ❌ Complex error handling in frontend
- ❌ Poor UX - multiple loading states, slower
- ❌ Each endpoint has different caching behavior
- ❌ Violates DRY - same data fetched multiple times
### Option 3: Extend `/api/healthz` Endpoint
Add `commit_short`, `build_date`, and `cache_enabled` fields to existing `/api/healthz`.
**Pros:**
- ✅ Reuses existing endpoint
- ✅ Single request
**Cons:**
- ❌ Breaks backward compatibility (response schema change)
- ❌ `/api/healthz` is Kubernetes-focused (naming convention)
- ❌ Not cached currently
- ❌ Mixes health probe concerns with version info
- ❌ Violates single responsibility
### Option 4: GraphQL / Query Parameters
Allow clients to specify which fields they want via query parameters.
**Pros:**
- ✅ Flexible - clients get exactly what they need
- ✅ Single endpoint
**Cons:**
- ❌ Overkill for this use case
- ❌ Not consistent with existing REST API design
- ❌ Complex implementation
- ❌ Not aligned with project architecture (Chi router, REST style)
## Decision Outcome
**Chosen: Option 1 - Composite `/api/info` Endpoint**
We will implement a new `GET /api/info` endpoint that returns a JSON object with all required fields in a single call. This endpoint will:
1. Aggregate data from existing sources (`version` package, `config`, server uptime)
2. Be cached using the existing cache service with key `info:json`
3. Use TTL from `config.cache.default_ttl_seconds` (consistent with ADR-0022)
4. Return `X-Cache: HIT/MISS` headers for debugging
5. Follow existing Go handler patterns from `pkg/server/server.go`
### Response Schema
```json
{
"version": "1.4.0",
"commit_short": "a3f7b2c1",
"build_date": "2026-05-04T08:00:00Z",
"uptime_seconds": 1234,
"cache_enabled": true,
"healthz_status": "healthy",
"go_version": "go1.26.1"
}
```
The `go_version` field provides the Go runtime version via `runtime.Version()`, useful for ops debugging (e.g., identifying which Go version is running in production).
### Rationale
1. **Performance**: Single HTTP request instead of 3-4 separate calls
2. **Consistency**: All data reflects the same moment in time
3. **Caching**: Leverages existing cache infrastructure (ADR-0022) with predictable key pattern
4. **API Design**: Clean, RESTful endpoint with single responsibility
5. **Maintainability**: Clear separation of concerns - info aggregation is a distinct use case
6. **Backward Compatibility**: Existing endpoints remain unchanged
7. **Frontend Simplicity**: Reduces complexity and improves UX
### Cache Strategy
Following ADR-0022 pattern:
- Cache key: `info:json` (consistent with `version:format` pattern)
- TTL: `config.cache.default_ttl_seconds` (default 300 seconds)
- Cache service: `pkg/cache/cache.go` InMemoryService
- Headers: `X-Cache: HIT` or `X-Cache: MISS`
This allows the endpoint to be fast even under load, while maintaining data freshness.
## Consequences
### Positive
1. **Improved frontend performance**: Single request instead of multiple
2. **Better UX**: Faster page loads, simpler loading states
3. **Consistent data**: All fields reflect the same point-in-time
4. **Cache efficiency**: Reuses existing cache infrastructure
5. **Clean separation**: Info endpoint handles aggregation, source endpoints unchanged
6. **Easy to test**: Single endpoint with predictable response
### Negative
1. **Data duplication**: Some fields appear in multiple endpoints
2. **Maintenance burden**: If source data changes, endpoint must be updated
3. **New endpoint**: Increases API surface area (though minimal)
### Mitigation
1. Data duplication is acceptable - it's read-only system info
2. Source the data from the same packages/functions used by other endpoints
3. The new endpoint has a clear, focused purpose
## Links
- [ADR-0002: Chi Router](adr/0002-chi-router.md) - Routing foundation
- [ADR-0022: Rate Limiting Cache Strategy](adr/0022-rate-limiting-cache-strategy.md) - Cache pattern reference
- [pkg/server/server.go](pkg/server/server.go) - Handler patterns
- [pkg/cache/cache.go](pkg/cache/cache.go) - Cache service
- [pkg/version/version.go](pkg/version/version.go) - Version data source

128
adr/0027-ollama-tier1-onboarding.md
View File

@@ -1,128 +0,0 @@
# 27. Ollama Tier 1 onboarding via meta-trainer-bootstrap
**Date:** 2026-05-05
**Status:** Proposed
**Authors:** Gabriel Radureau, AI Agent (Claude Opus 4.7 Tier 3 inspector)
## Context and Problem Statement
The autonomous trainer day on 2026-05-05 validated that Mistral Vibe (cloud) can drive a complete PR lifecycle on this project: ICM workspace → phase-planner → implementation → verifier audit → PR open (cf. PR #54, Q-041 in `~/.vibe/memory/reference/mistral-quirks.md`). Two limitations remain:
1. **Vendor risk** — every autonomous run consumes the Mistral cloud forfait. If the forfait runs out mid-month or the API is unavailable, autonomous capability is lost.
2. **Sovereignty story** — ARCODANGE's stated direction (cf. `migration-claude-vers-mistral-phase-1.md`) is to reduce dependence on a single foreign vendor. The hardware exists locally (M4 128 GB) ; the missing link is wiring a local model into the same Tier 1 executor role Mistral plays today.
The user-flagged candidate models (cf. `~/.vibe/memory/reference/ollama-candidate-models.md`) :
* `nemotron-3-super`
* `gemma4:31b`
Both are large enough to plausibly handle the agentic coding role and small enough to fit in 128 GB RAM with headroom for tools. Neither has been tested under the ARCODANGE methodology (canary suite, ICM workspace traversal, verifier-skill discipline).
The methodology to onboard a new Tier 1 already exists : the `meta-trainer-bootstrap` skill at `~/.vibe/skills/meta-trainer-bootstrap/`. It runs a 10-canary suite (C-001..C-010), copies + adapts the skill library to the new model's harness tool names, stands up a `<model>-quirks.md` baseline, and produces a Tier 3 audit report. It has been validated on Mistral itself (we are currently running the methodology Mistral-on-Mistral, which is unusual — the canary suite was originally written for a different model).
## Decision Drivers
* **Forfait insurance** — a working local Tier 1 means autonomous capability survives a Mistral outage / forfait exhaustion
* **Sovereignty** — local execution removes the single-vendor dependency for the autonomous workflow
* **Methodology validation** — `meta-trainer-bootstrap` has never been run on a fresh model in production, only smoke-tested ; this is its first real test
* **Cost** — Ollama is local-only (no per-call price). The cost is the bootstrap effort + ongoing M4 power consumption.
* **Model maturity** — both candidates are recent ; their agentic coding ability is empirical, not theoretical
## Considered Options
### Option 1: Bootstrap `nemotron-3-super` first, then `gemma4:31b`
Run the canary suite on each, document quirks separately, decide based on canary pass rate and cost-per-task.
* Good — comparative data, makes the choice empirical
* Good — discovers any meta-trainer-bootstrap bugs early on the first attempt
* Bad — doubles the bootstrap effort (~4-8 hours per model)
* Bad — requires holding both models on disk (large)
### Option 2: Bootstrap one model only, picked on prior reputation
Pick one (e.g. `nemotron-3-super` per the user's explicit ordering in `ollama-candidate-models.md`) and commit. Skip the comparison.
* Good — half the effort, ships faster
* Bad — no fallback if the chosen model is unsuitable
* Bad — anchors the methodology to one model's quirks before we know they generalise
### Option 3: Defer until Mistral autonomous shows real strain
Do nothing yet. Wait for forfait pressure or a Mistral outage to force the issue. Reactive instead of proactive.
* Good — zero effort now
* Bad — when the trigger fires, we are unprepared and the bootstrap is rushed
* Bad — postpones validation of `meta-trainer-bootstrap` indefinitely
### Option 4: Skip Ollama, evaluate a different vendor (Anthropic, OpenAI)
Bring in a second cloud model as Tier 1 instead of going local.
* Good — likely higher quality than 31B local
* Bad — replaces vendor dependence with two-vendor dependence ; doesn't solve sovereignty
* Bad — we already have Claude as Tier 3 inspector via Anthropic ; mixing roles complicates the methodology
## Decision Outcome
Chosen option: **Option 2 — Bootstrap `nemotron-3-super` first**, deferring `gemma4:31b` to a follow-up ADR if `nemotron-3-super` underperforms or shows unfixable quirks.
Rationale :
- Forfait pressure is real but not immediate (~3.5% of monthly forfait spent on the heavy autonomous trainer day 2026-05-05) — we have time but should not procrastinate
- Comparative testing (Option 1) is technically right but pragmatically slow for an unproven methodology
- The user's explicit ordering signals their prior on which to try first ; respect it
- If the canary suite fails substantially on `nemotron-3-super`, we pivot to `gemma4:31b` with the lessons (and per-model quirks file) from the first attempt — net learning either way
## Implementation Plan
1. **Pre-flight** — verify `ollama` is installed, the model is pulled (`ollama pull nemotron-3-super`), and the M4 has enough free RAM (model size + ~16 GB headroom for tools).
2. **Run `meta-trainer-bootstrap` skill** — pointing `TARGET_MODEL_ID=nemotron-3-super`, `TARGET_HARNESS=ollama run nemotron-3-super`, `TARGET_PROJECT_ROOT=<a fresh clone or worktree>`. Budget : 5 EUR-equivalent of Mistral Tier-2 orchestration cost + 2-4 hours of trainer attention.
3. **Canary suite** — run C-001..C-010 ; record each result in `~/.vibe/memory/reference/nemotron-3-super-quirks.md` as `Q-101..Q-110` (the `Q-001..Q-099` range is reserved for the legacy Mistral baseline).
4. **Skill library adaptation** — for each ARCODANGE skill currently relying on Mistral-specific tool names (`read_file`, `write_file`, etc.), adapt to whatever Ollama exposes. Document deltas.
5. **Smoke test** — run a single small task end-to-end on a low-risk project. Use the ICM workspace pattern. Verify worktree isolation (Q-038 fix) still applies.
6. **Tier 3 report** — produce `bootstrap-report.md` for Claude inspector review. Include canary pass rate, key quirks, KPI baseline numbers, open friction points.
7. **Decision gate** — based on the report, either (a) promote `nemotron-3-super` to production Tier 1 and update `~/.vibe/config.toml` accordingly, (b) try `gemma4:31b` as a follow-up, or (c) escalate to Tier 3 for a strategic pivot.
## Pros and Cons of the Options
### Option 1 (Bootstrap both)
* Good — comparative data
* Good — early bug detection on the methodology
* Bad — double effort
* Bad — no clear way to choose without significant additional time investment for the second model
### Option 2 (Chosen — `nemotron-3-super` first)
* Good — concrete forward motion
* Good — methodology gets its first real test
* Good — `meta-trainer-bootstrap` skill validated end-to-end (currently only smoke-tested)
* Bad — risk of picking the wrong model and wasting the bootstrap effort
* Mitigation: per-model quirks files mean the second attempt is cheaper (skill adaptations transfer)
### Option 3 (Defer)
* Good — zero effort
* Bad — reactive, increases risk under outage scenarios
### Option 4 (Different vendor)
* Good — likely higher quality
* Bad — does not solve sovereignty
* Bad — methodology already has Claude as Tier 3 ; another Anthropic-family model in Tier 1 conflates roles
## Consequences
* `meta-trainer-bootstrap` skill is exercised end-to-end for the first time. Discoveries during this run will likely produce Q-042+ entries in `mistral-quirks.md` and a separate `nemotron-3-super-quirks.md`.
* `~/.vibe/config.toml` may need a new model alias (e.g. `local-nemotron`) configured for testing without affecting the production `mistral-vibe-cli-latest` default.
* If successful, the next ADR (0028 or higher) will document the production switch (or split, e.g. routine tasks → local, complex tasks → cloud).
* Forfait usage from this bootstrap : Tier 2 Mistral orchestration only ; Tier 1 Ollama runs are free at the API level.
## Links
* Three-tier methodology : `~/.vibe/skills/meta-trainer-bootstrap/references/three-tier-tutor.md`
* Candidate models reference : `~/.vibe/memory/reference/ollama-candidate-models.md`
* `meta-trainer-bootstrap` skill : `~/.vibe/skills/meta-trainer-bootstrap/SKILL.md`
* Canary suite : `~/.vibe/skills/meta-trainer-bootstrap/canaries/INDEX.md`
* Q-041 (autonomy story validated on Mistral) : `~/.vibe/memory/reference/mistral-quirks.md`
* Related ADRs : [ADR-0007](0007-opentelemetry-integration.md) (cloud / sovereignty considerations historically) ; [ADR-0023](0023-config-hot-reloading.md) (hot-reload may need different patterns under Ollama)

147
adr/0028-passwordless-auth-migration.md
View File

@@ -1,147 +0,0 @@
# 28. Passwordless authentication: magic link → OpenID Connect
**Date:** 2026-05-05
**Status:** Proposed
**Authors:** Gabriel Radureau, AI Agent
## Context and Problem Statement
ADR-0018 (now Implemented) shipped a username + password authentication system with bcrypt hashing, JWT tokens, admin master password, and admin-assisted password reset. It works, but it carries the cost-of-passwords : we store password hashes, support password reset flows, and maintain a credential-rotation policy. Users hate passwords ; ops and security pay for them.
Two industry-standard alternatives exist :
1. **Magic link by email** — user enters their email, receives a one-time token in a clickable link, link consumes the token and issues a session JWT. No password stored.
2. **OpenID Connect Authorization Code flow** — delegate authentication to an external Identity Provider (e.g. Authelia, Keycloak, Auth0, Google) ; our app receives an `id_token` after the OIDC dance.
We want to **migrate to passwordless** for new sign-ups while keeping the existing username/password code path operational during the transition (no flag-day breakage). The two passwordless mechanisms above complement each other : magic link is simpler for first-party users on day 1 ; OIDC is the right answer for second-party users (other ARCODANGE products, partner integrations) and for admin SSO.
A third constraint : ARCODANGE local development must use HTTPS for OAuth callbacks to be valid (most OIDC providers reject `http://localhost` redirect URIs in their default config). `mkcert` is the canonical local-CA tool for this.
## Decision Drivers
* **Reduce password-related attack surface** — no hash storage, no breach-and-reuse risk, no password reset abuse vectors
* **User experience** — passwordless is faster for the user (1 click in email vs typing/remembering password)
* **Operational simplicity** — no password reset flow to maintain ; the password-reset code can be removed once migration is complete
* **Multi-product readiness** — OIDC is the prerequisite for cross-product SSO across the ARCODANGE portfolio
* **Backwards compatibility** — must not break existing tokens or BDD scenarios mid-migration
* **Local dev parity** — HTTPS in dev so OAuth flows can be tested locally without provider-specific workarounds
## Considered Options
### Option 1 (Chosen): Sequenced — magic link first, OIDC second
Deliver in two phases :
* **Phase A — Magic link**
- Add `POST /api/v1/auth/magic-link/request` (body: `{email}`) — generates token, stores it (TTL ~15 min), sends email via SMTP
- Add `GET /api/v1/auth/magic-link/consume?token=<...>` — single-use consumption, issues a JWT, returns it as cookie + JSON body
- Reuse the existing JWT issuance + secret retention infrastructure (ADR-0021)
- Existing `/api/v1/auth/login` (username/password) stays operational during transition
* **Phase B — OpenID Connect Authorization Code with PKCE**
- Add `GET /api/v1/auth/oidc/start` — generates state + PKCE verifier, redirects to provider's `authorization_endpoint`
- Add `GET /api/v1/auth/oidc/callback` — exchanges code for tokens, validates `id_token` signature against provider's JWKS, issues internal JWT
- Provider URL configurable per environment (`auth.oidc.issuer_url`, `auth.oidc.client_id`, `auth.oidc.client_secret`)
- Allow multiple providers in config (key by provider name, e.g. `arcodange-sso`)
- Local dev requires HTTPS — `mkcert` setup documented in `documentation/DEV_SETUP.md`
* **Phase C (later, separate ADR) — Decommission password auth**
- Once all users have migrated, remove the password endpoints, remove the password_hash column, mark ADR-0018 as Superseded by this ADR
### Option 2: All-at-once OIDC, no magic link
Skip magic link, jump straight to OIDC.
* Good — single migration, no intermediate state
* Bad — requires an OIDC provider operational on day 1, which we don't have configured
* Bad — magic link has zero infra dependencies (just SMTP) ; OIDC requires running an IdP or paying for one
### Option 3: Magic link only, no OIDC
Stop at Phase A.
* Good — simplest implementation
* Bad — doesn't solve cross-product SSO ; we'd re-do this work later for the broader ARCODANGE portfolio
### Option 4: Status quo (do nothing)
Keep username + password.
* Good — zero effort
* Bad — passwords stay forever ; ARCODANGE locks itself out of integration scenarios that expect OIDC
## Decision Outcome
Chosen option : **Option 1, sequenced magic link → OIDC**.
Rationale :
- Magic link is implementable today with zero infra dependencies beyond the email infrastructure (ADR-0029)
- OIDC requires running an IdP locally (Authelia or Keycloak) — that's another container in the dev stack and another ADR's worth of decision work, but the magic-link work is the natural prerequisite (token-by-email plumbing is reused)
- Sequenced delivery means we never have to roll back : Phase A works alone, Phase B layers on top, Phase C cleans up
## Implementation Plan
### Phase A — Magic link (target: 2-3 PRs)
1. **A.1 — Storage** : add a `magic_link_tokens` table (id, email, token_hash, expires_at, consumed_at). Repository pattern alongside `pkg/user/postgres_repository.go`.
2. **A.2 — Token endpoint** : `POST /api/v1/auth/magic-link/request` generates a token, stores it (hashed), enqueues an email send. Rate-limited (cf. ADR-0022) by email address.
3. **A.3 — Consume endpoint** : `GET /api/v1/auth/magic-link/consume?token=...` validates + marks consumed + issues JWT. Returns `Set-Cookie` and `{token: jwt}` body.
4. **A.4 — Sign-up via magic link** : if the email is unknown, the consume endpoint creates the user record. (No separate "sign-up" flow needed — first magic link IS the sign-up.)
5. **A.5 — BDD coverage** : scenarios for happy path, expired token, double-consume, wrong-email, rate-limit. Cf. ADR-0030 for the email assertion strategy.
### Phase B — OIDC Code flow with PKCE (target: 3-4 PRs)
1. **B.1 — Local IdP** : choose Authelia or Keycloak for local development. Add to `docker-compose.yml` with default test configuration.
2. **B.2 — mkcert** : document local HTTPS setup in `documentation/DEV_SETUP.md`, add `make cert` target.
3. **B.3 — OIDC client** : `pkg/auth/oidc.go` — discovery, JWKS cache, code exchange with PKCE.
4. **B.4 — Endpoints** : `/oidc/start` and `/oidc/callback`.
5. **B.5 — Provider config** : `auth.oidc.providers` map in config (cf. ADR-0006 Viper) ; multi-provider supported.
6. **B.6 — BDD coverage** : end-to-end scenarios using a mock OIDC server (or the local Authelia instance with deterministic users).
### Phase C — Decommission password (separate ADR after A+B in production)
Out of scope for this ADR. Will be ADR-NNNN when migration is complete.
## Pros and Cons of the Options
### Option 1 (Chosen — Sequenced)
* Good — incremental, no flag day, each phase shippable on its own
* Good — reuses existing JWT infrastructure (ADR-0021 secret retention)
* Good — magic link work is a prerequisite for OIDC anyway (email plumbing, mkcert)
* Bad — total work spans 2 sprints, longer time-to-OIDC than Option 2
* Mitigation: after Phase A, the team can stop if priorities shift — magic link alone is a complete improvement
### Option 2 (All OIDC)
* Good — single migration
* Bad — requires IdP operational from day 1
* Bad — local dev environment more complex than necessary for the magic link case
### Option 3 (Magic link only)
* Good — minimal scope
* Bad — re-work later for SSO
### Option 4 (Status quo)
* Good — zero effort
* Bad — accumulating tech debt
## Consequences
* `pkg/auth/` package created (currently auth code lives in `pkg/user/`) — separation is now justified by the multi-mechanism scope
* `pkg/user/api/auth_handler.go` continues to serve username/password during transition (Phase A and B), removed in Phase C
* `documentation/DEV_SETUP.md` becomes a load-bearing doc for new contributors (mkcert + docker-compose with mailpit + Authelia)
* The 4 new endpoints (`magic-link/request`, `magic-link/consume`, `oidc/start`, `oidc/callback`) require their own ADR entries in the API doc + Swagger annotations
* Phase A's magic link plumbing depends on **ADR-0029** (email infrastructure decision) — that ADR ships first
* BDD scenarios for Phase A depend on **ADR-0030** (email testing strategy with parallel BDD) — that ADR ships before any Phase A scenario lands
## Links
* Email infrastructure : [ADR-0029](0029-email-infrastructure-mailpit.md)
* BDD email testing strategy : [ADR-0030](0030-bdd-email-parallel-strategy.md)
* Existing user auth (to be partially superseded by Phase C) : [ADR-0018](0018-user-management-auth-system.md)
* JWT secret retention reused : [ADR-0021](0021-jwt-secret-retention-policy.md)
* Rate limiting reused : [ADR-0022](0022-rate-limiting-cache-strategy.md)
* OAuth 2.0 Authorization Code with PKCE : [RFC 7636](https://datatracker.ietf.org/doc/html/rfc7636)
* OpenID Connect Core : [OpenID Foundation](https://openid.net/specs/openid-connect-core-1_0.html)

142
adr/0029-email-infrastructure-mailpit.md
View File

@@ -1,142 +0,0 @@
# 29. Email infrastructure: Mailpit local + production deferred
**Date:** 2026-05-05
**Status:** Proposed
**Authors:** Gabriel Radureau, AI Agent
## Context and Problem Statement
ADR-0028 (passwordless auth) requires the application to send emails — magic-link tokens specifically. Email is a substrate decision : the choice of SMTP provider, the abstraction in code, and the local development experience all depend on it.
Two separate concerns :
1. **Local development + BDD tests** : we need a local SMTP receiver that captures emails and exposes them for inspection. Real email providers (Gmail, SES, SendGrid) are unsuitable for local dev — they cost money, leak test data, and rate-limit aggressively.
2. **Production** : the application needs to actually deliver mail to user inboxes. This decision is deferred — see "Out of scope" below.
ARCODANGE already has the **Mailpit** docker image pulled locally (`axllent/mailpit:latest`, 51 MB). Mailpit captures SMTP submissions on a port, stores them in-memory, exposes them via HTTP UI (default :8025) and an HTTP API (`/api/v1/messages`). It's the de-facto choice for Go projects needing local SMTP capture.
The application code needs to be **provider-agnostic** : a `pkg/email` package with a `Sender` interface, a Mailpit-compatible SMTP implementation, and a contract that production can swap for a real provider's adapter without changing call sites.
## Decision Drivers
* **Local dev and CI must work without internet** — emails should never leave the docker network in tests
* **Test inspection must be programmatic** — BDD tests assert on email content, not just "an email was sent"
* **Production decision deferred** — we don't know the volume / SLA / compliance requirements yet ; over-committing now is premature
* **Provider portability** — `pkg/email` interface lets us swap implementations without touching auth code
* **Cost** — Mailpit is free, runs in a container, no API quota concerns
## Considered Options
### Option 1 (Chosen): Mailpit for local + tests, production via a production-grade provider TBD
* Add Mailpit to `docker-compose.yml` (SMTP :1025, HTTP API :8025)
* `pkg/email` package with a `Sender` interface
* Default implementation : `SMTPSender` configured against the local Mailpit in dev/CI
* Tests query Mailpit's HTTP API to inspect captured messages
* Production deployment will add a separate `pkg/email/<provider>_sender.go` implementing the same interface — that decision is its own ADR
### Option 2: MailHog instead of Mailpit
MailHog is the older, well-known alternative. Mailpit is its modern successor, written in Go, with a richer API and active maintenance.
* Bad — abandoned upstream (last commit 2020). Mailpit is the natural replacement.
### Option 3: In-process mock email sender
Write a `MockSender` that captures emails in a Go slice. No SMTP at all.
* Good — fastest tests, zero infra
* Bad — doesn't validate the actual SMTP wire format, the From/To/Subject headers, the encoding of multi-byte content, or the DKIM/Reply-To setup
* Bad — doesn't double as a manual-inspection tool for the developer (no UI to look at the email)
### Option 4: Send to a real but throwaway provider (Mailtrap, Mailosaur)
External services that capture-and-display emails.
* Good — production-similar paths
* Bad — costs money, requires an account, leaks test data, doesn't work offline
## Decision Outcome
Chosen option : **Option 1 — Mailpit for local + tests, production deferred**.
Rationale :
- Mailpit is the modern, maintained successor to MailHog ; image is already on the dev machine
- The interface-first design (`pkg/email.Sender`) means production swap is a future ADR, not a refactor
- BDD tests have a real wire-format path to assert on (cf. ADR-0030)
- Zero monthly cost in dev/CI
## Implementation Plan
1. **`pkg/email/sender.go`** — define the `Sender` interface :
```go
type Sender interface {
Send(ctx context.Context, msg Message) error
}
type Message struct {
To string
From string
Subject string
BodyText string
BodyHTML string
Headers map[string]string // for trace correlation, e.g. X-Test-Scenario-ID
}
```
2. **`pkg/email/smtp_sender.go`** — implementation using `net/smtp` (stdlib) configured by `auth.email.smtp_host`, `smtp_port`, `smtp_username`, `smtp_password`, `smtp_use_tls`. For Mailpit defaults : `smtp_host=localhost smtp_port=1025 smtp_use_tls=false`.
3. **`pkg/email/sender_test.go`** — unit tests using `httptest`-style fake SMTP, plus a `*_integration_test.go` (build tag `integration`) hitting the live Mailpit.
4. **`docker-compose.yml`** — add the `mailpit` service :
```yaml
mailpit:
image: axllent/mailpit:latest
ports:
- "1025:1025" # SMTP
- "8025:8025" # HTTP UI / API
environment:
MP_MAX_MESSAGES: 5000
```
5. **`pkg/config/config.go`** — add the `auth.email.*` config keys with defaults pointing at local Mailpit.
6. **Documentation** : `documentation/EMAIL.md` covering local setup, message inspection via UI (http://localhost:8025), API queries.
## Pros and Cons of the Options
### Option 1 (Chosen — Mailpit)
* Good — already locally available, free, modern, maintained
* Good — provider-agnostic interface decouples from prod choice
* Good — full SMTP wire format = realistic test path
* Good — UI for manual inspection during dev
* Bad — requires Mailpit running (one more docker-compose service)
* Bad — production decision still pending
### Option 2 (MailHog)
* Bad — unmaintained, choosing it would create immediate tech debt
### Option 3 (Mock only)
* Bad — too much abstraction loss, can't catch wire-level bugs
### Option 4 (Mailtrap / Mailosaur)
* Bad — cost, network dependency, account management
## Consequences
* New service in `docker-compose.yml` — developers run `docker compose up -d` once and Mailpit is on
* New `pkg/email` package — auth code (ADR-0028 magic link) calls `Sender.Send()` rather than direct SMTP
* New `auth.email.*` config keys, new env vars (`DLC_AUTH_EMAIL_SMTP_HOST` etc.)
* Mailpit's HTTP API becomes part of the BDD test contract — tests use it to assert messages were sent (cf. ADR-0030)
* Production sender ADR (TBD) will be a separate decision — this ADR explicitly does NOT pick a vendor for prod
## Out of scope
* **Production email provider selection** — separate ADR when we know volume / SLA / compliance constraints. Likely candidates: AWS SES, Postmark, SendGrid, Mailjet. Magic-link emails are transactional + low-volume — most providers handle that easily.
* **DKIM/SPF/DMARC setup** — production deliverability concern, not a local-dev concern
* **HTML email templating** — we'll start with plain-text emails ; HTML can be added with a template package (e.g. `html/template`) when ARCODANGE branding requires it
## Links
* Auth migration that requires this : [ADR-0028](0028-passwordless-auth-migration.md)
* BDD test strategy that consumes Mailpit : [ADR-0030](0030-bdd-email-parallel-strategy.md)
* Mailpit homepage : https://mailpit.axllent.org/
* Mailpit API reference : https://mailpit.axllent.org/docs/api-v1/

187
adr/0030-bdd-email-parallel-strategy.md
View File

@@ -1,187 +0,0 @@
# 30. BDD email assertions with parallel test execution
**Date:** 2026-05-05
**Status:** Proposed
**Authors:** Gabriel Radureau, AI Agent
## Context and Problem Statement
ADR-0028 introduces magic-link auth, which requires the application to send emails. ADR-0029 chose **Mailpit** as the local SMTP receiver for dev and BDD tests. The remaining decision : **how do BDD scenarios assert on the email content while running in parallel ?**
Today (since [PR #35](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/35)), the BDD suite runs in parallel via per-package PostgreSQL schema isolation (cf. [ADR-0025](0025-bdd-scenario-isolation-strategies.md)). Each Go test package has its own schema ; tests inside a package run serially within that schema. This works because Postgres has named schemas with strong isolation. **Mailpit has no equivalent** — there is one inbox per Mailpit instance, shared across all senders.
A naive integration would have parallel scenarios fight over each other's emails :
- Scenario A : "request magic link for `test@example.com`" → email arrives
- Scenario B (in parallel) : "request magic link for `test@example.com`" → email arrives
- Both scenarios query Mailpit for `test@example.com` — they see each other's messages, assertions become flaky.
We need a way to scope each scenario's emails so it only sees its own messages.
## Decision Drivers
* **No regression on parallelism** — BDD-isolation Phase 3 (PR #35) achieved a 2.85x speedup ; the email-assertion solution must not undo that
* **No new container per test** — running one Mailpit per scenario would defeat the simplicity that made us choose Mailpit
* **Determinism** — a scenario's email assertions must succeed regardless of how many other scenarios are running
* **Realistic SMTP path** — we still want the full SMTP wire format exercised (cf. ADR-0029) ; we don't want to bypass Mailpit
* **Cleanup hygiene** — old messages from previous test runs must not leak into a new run
## Considered Options
### Option 1 (Chosen): Per-test recipient scoping with deterministic addresses
Each BDD scenario generates a unique email address for its test user, derived from the scenario key + a random suffix. Examples :
- Scenario `magic-link-happy-path` → `magic-link-happy-path-<8hex>@bdd.local`
- Scenario `magic-link-expired-token` → `magic-link-expired-token-<8hex>@bdd.local`
The application code accepts any email format. The BDD scenario asserts on Mailpit's HTTP API filtering by the `to` address. Two parallel scenarios with different addresses can NEVER see each other's emails.
**Cleanup** : at the start of each scenario, the BDD framework calls `DELETE /api/v1/search?query=to:<scenario-address>` on Mailpit to purge any leftover messages from prior runs.
### Option 2: One Mailpit instance per Go test package
Spawn a fresh Mailpit container in `TestMain` of each `features/<area>/` package. Each gets its own port range.
* Good — strong isolation
* Bad — heavyweight (one container per package = 5+ containers running)
* Bad — port allocation complexity (similar to existing `pkg/bdd/parallel/port_manager.go`, but applied to Mailpit)
* Bad — slow startup (Mailpit boot is ~200ms but adds up)
### Option 3: One Mailpit instance, scenario-scoped via custom SMTP header
Add a custom header `X-BDD-Scenario-ID: <key>` to outgoing emails. Tests query Mailpit filtered on that header.
* Good — same single Mailpit
* Bad — requires the application code to know the scenario ID at email-send time, which means a test-only path in production code
* Bad — header propagation is fragile (gets stripped by some SMTP relays — not Mailpit, but real production providers might) ; we don't want a different code path between dev and prod
### Option 4: Sequence parallel scenarios via per-scenario Mailpit lock
Use a mutex / queue so no two scenarios that send email run concurrently.
* Good — minimal code change
* Bad — gives up the parallel speedup for any feature that involves email — that's most auth-related features going forward
## Decision Outcome
Chosen option : **Option 1 — per-test recipient scoping**.
Rationale :
- Recipient scoping is the simplest abstraction : the address IS the identity ; Mailpit's HTTP API natively supports filtering by recipient
- Application code stays clean : it just sends to whatever address it's given. No test-mode branching.
- Parallel-safe by construction : two scenarios cannot collide if they don't share an address
- Cheap to implement : a few helper functions in `pkg/bdd/steps/email_steps.go` and a `mailpit.Client` package wrapping the HTTP API
- Cleanup is per-scenario, not global — no "delete all messages" race between scenarios
## Implementation Plan
### Helper package : `pkg/bdd/mailpit/client.go`
```go
type Client struct {
BaseURL string // default: http://localhost:8025
HTTP *http.Client
}
// AwaitMessageTo polls Mailpit's HTTP API for a message addressed
// to the given recipient, with a deadline. Returns the most recent
// matching message or an error on timeout.
func (c *Client) AwaitMessageTo(ctx context.Context, to string, timeout time.Duration) (*Message, error)
// PurgeMessagesTo removes all messages addressed to the given
// recipient. Idempotent and parallel-safe.
func (c *Client) PurgeMessagesTo(ctx context.Context, to string) error
type Message struct {
ID string
From string
To []string
Subject string
Text string
HTML string
Headers map[string][]string
}
```
### Helper steps : `pkg/bdd/steps/email_steps.go`
```go
func (s *EmailSteps) iHaveAnEmailAddressForThisScenario() error
// Generates `<scenario-key>-<8hex>@bdd.local`, stores it in the scenario state.
func (s *EmailSteps) iShouldReceiveAnEmailWithSubject(subject string) error
// Polls AwaitMessageTo on the scenario's address, asserts subject equality.
func (s *EmailSteps) theEmailShouldContain(snippet string) error
// Re-fetches the most recent message and checks for substring in body.
func (s *EmailSteps) theEmailContainsAMagicLinkToken() (string, error)
// Extracts the token from the magic-link URL via regex, returns it.
```
### Scenario lifecycle
- **Before each scenario** : `iHaveAnEmailAddressForThisScenario` is called (either explicitly via Background, or implicitly via a hook). The unique address is stored in the scenario's state. PurgeMessagesTo is called to clear any leftovers from prior runs of the same address (defensive — should be impossible since the suffix is random, but cheap).
- **During the scenario** : the application sends to that address. Tests query for it.
- **After each scenario** : no global cleanup needed — addresses are per-scenario unique, so they don't accumulate beyond Mailpit's `MP_MAX_MESSAGES=5000` cap.
### Race-free deletion
Mailpit's `DELETE /api/v1/search?query=to:<addr>` is atomic per recipient. Two concurrent scenarios with different addresses cannot interfere.
### Sample scenario (auth-magic-link.feature)
```gherkin
@critical @magic-link
Scenario: User receives a magic link by email
Given I have an email address for this scenario
When I request a magic link for my email address
Then I should receive an email with subject "Your magic link"
And the email contains a magic link token
When I consume the magic link token
Then I should receive a JWT
```
## Pros and Cons of the Options
### Option 1 (Chosen)
* Good — parallel-safe by construction
* Good — application code unchanged ; test-only logic stays in the BDD layer
* Good — Mailpit API supports the filter natively
* Good — cleanup is fine-grained, no race
* Bad — requires cooperative scenarios (each must request a unique address)
* Mitigation : Background steps in feature files make it automatic
### Option 2 (Mailpit per package)
* Bad — operational complexity not justified for the test-only concern
### Option 3 (Custom header scoping)
* Bad — production code dirtied by test concerns
### Option 4 (Lock-and-sequence)
* Bad — gives up parallelism (the whole point of PR #35 + ADR-0025)
## Consequences
* `pkg/bdd/mailpit/` package is created with HTTP client + helper types
* `pkg/bdd/steps/email_steps.go` package is created and registered in `steps.go`
* `features/auth/` and any other email-using features have new BDD steps available
* The local development docker-compose must run Mailpit before BDD tests run — to be added to the BDD test runner script `scripts/run-bdd-tests.sh`
* Mailpit message TTL is governed by `MP_MAX_MESSAGES` (5000) — at parallel BDD volumes, that's enough headroom for ~50 scenarios × 100 messages each before any pruning kicks in
## Out of scope
* **Visual regression on email rendering** — text body assertions only ; HTML rendering checks belong in a separate Storybook-style harness
* **Attachment handling** — magic-link emails are text-only ; ADRs for attachments will come if/when needed
* **Email volume / rate-limit testing** — that's a load-test concern, not a BDD concern
## Links
* Auth migration depending on this : [ADR-0028](0028-passwordless-auth-migration.md)
* Email infrastructure choice : [ADR-0029](0029-email-infrastructure-mailpit.md)
* BDD parallelism foundation : [ADR-0025](0025-bdd-scenario-isolation-strategies.md), [PR #35](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/35)
* Mailpit API : https://mailpit.axllent.org/docs/api-v1/

23
adr/README.md
View File

@@ -14,26 +14,21 @@ This directory contains the Architecture Decision Records (ADRs) for the dance-l
| [0006](0006-configuration-management.md) | Use Viper for configuration management | Accepted |
| [0007](0007-opentelemetry-integration.md) | Integrate OpenTelemetry for distributed tracing | Accepted |
| [0008](0008-bdd-testing.md) | Adopt BDD with Godog for behavioral testing | Accepted |
| [0009](0009-hybrid-testing-approach.md) | Combine BDD and Swagger-based testing | Implemented |
| [0009](0009-hybrid-testing-approach.md) | Combine BDD and Swagger-based testing | Partially Implemented |
| [0010](0010-api-v2-feature-flag.md) | API v2 Feature Flag Implementation | Accepted |
| [0012](0012-git-hooks-staged-only-formatting.md) | Git Hooks: Staged-Only Formatting | Accepted |
| [0013](0013-openapi-swagger-toolchain.md) | OpenAPI/Swagger Toolchain Selection | Implemented |
| [0013](0013-openapi-swagger-toolchain.md) | OpenAPI/Swagger Toolchain Selection | Partially Implemented |
| [0015](0015-cli-subcommands-cobra.md) | CLI Subcommands and Flag Management with Cobra | Implemented |
| [0016](0016-ci-cd-pipeline-design.md) | CI/CD Pipeline Design for Multi-Platform Compatibility | Accepted |
| [0017](0017-trunk-based-development-workflow.md) | Trunk-Based Development Workflow for CI/CD Safety | Approved |
| [0018](0018-user-management-auth-system.md) | User Management and Authentication System | Implemented |
| [0019](0019-postgresql-integration.md) | PostgreSQL Database Integration | Implemented |
| [0018](0018-user-management-auth-system.md) | User Management and Authentication System | Proposed |
| [0019](0019-postgresql-integration.md) | PostgreSQL Database Integration | Proposed |
| [0020](0020-docker-build-strategy.md) | Docker Build Strategy: Traditional vs Buildx | Accepted |
| [0021](0021-jwt-secret-retention-policy.md) | JWT Secret Retention Policy | Implemented |
| [0022](0022-rate-limiting-cache-strategy.md) | Rate Limiting and Cache Strategy | Implemented (Phase 1) |
| [0023](0023-config-hot-reloading.md) | Config Hot Reloading Strategy | Implemented |
| [0024](0024-bdd-test-organization-and-isolation.md) | BDD Test Organization and Isolation Strategy | Implemented |
| [0025](0025-bdd-scenario-isolation-strategies.md) | BDD Scenario Isolation Strategies | Implemented |
| [0026](0026-composite-info-endpoint.md) | Composite Info Endpoint vs Separate Calls | Implemented |
| [0027](0027-ollama-tier1-onboarding.md) | Ollama Tier 1 onboarding via meta-trainer-bootstrap | Proposed |
| [0028](0028-passwordless-auth-migration.md) | Passwordless authentication: magic link → OpenID Connect | Proposed |
| [0029](0029-email-infrastructure-mailpit.md) | Email infrastructure: Mailpit local + production deferred | Proposed |
| [0030](0030-bdd-email-parallel-strategy.md) | BDD email assertions with parallel test execution | Proposed |
| [0021](0021-jwt-secret-retention-policy.md) | JWT Secret Retention Policy | Proposed |
| [0022](0022-rate-limiting-cache-strategy.md) | Rate Limiting and Cache Strategy | Proposed |
| [0023](0023-config-hot-reloading.md) | Config Hot Reloading Strategy | Proposed |
| [0024](0024-bdd-test-organization-and-isolation.md) | BDD Test Organization and Isolation Strategy | Proposed |
| [0025](0025-bdd-scenario-isolation-strategies.md) | BDD Scenario Isolation Strategies | Proposed |
> **Note** : numbers `0011` and `0014` are not currently in use. Reserved for future ADRs or representing previously deleted entries.

23
chart/.helmignore
View File

@@ -1,23 +0,0 @@
# Patterns to ignore when building packages.
# This supports shell glob matching, relative path matching, and
# negation (prefixed with !). Only one pattern per line.
.DS_Store
# Common VCS dirs
.git/
.gitignore
.bzr/
.bzrignore
.hg/
.hgignore
.svn/
# Common backup files
*.swp
*.bak
*.tmp
*.orig
*~
# Various IDEs
.project
.idea/
*.tmproj
.vscode/

6
chart/Chart.yaml
View File

@@ -1,6 +0,0 @@
apiVersion: v2
name: dance-lessons-coach
description: Helm chart for dance-lessons-coach Go API server (ARCODANGE)
type: application
version: 0.1.0
appVersion: "latest"

22
chart/templates/NOTES.txt
View File

@@ -1,22 +0,0 @@
1. Get the application URL by running these commands:
{{- if .Values.ingress.enabled }}
{{- range $host := .Values.ingress.hosts }}
{{- range .paths }}
http{{ if $.Values.ingress.tls }}s{{ end }}://{{ $host.host }}{{ .path }}
{{- end }}
{{- end }}
{{- else if contains "NodePort" .Values.service.type }}
export NODE_PORT=$(kubectl get --namespace {{ .Release.Namespace }} -o jsonpath="{.spec.ports[0].nodePort}" services {{ include "dance-lessons-coach.fullname" . }})
export NODE_IP=$(kubectl get nodes --namespace {{ .Release.Namespace }} -o jsonpath="{.items[0].status.addresses[0].address}")
echo http://$NODE_IP:$NODE_PORT
{{- else if contains "LoadBalancer" .Values.service.type }}
NOTE: It may take a few minutes for the LoadBalancer IP to be available.
You can watch its status by running 'kubectl get --namespace {{ .Release.Namespace }} svc -w {{ include "dance-lessons-coach.fullname" . }}'
export SERVICE_IP=$(kubectl get svc --namespace {{ .Release.Namespace }} {{ include "dance-lessons-coach.fullname" . }} --template "{{"{{ range (index .status.loadBalancer.ingress 0) }}{{.}}{{ end }}"}}")
echo http://$SERVICE_IP:{{ .Values.service.port }}
{{- else if contains "ClusterIP" .Values.service.type }}
export POD_NAME=$(kubectl get pods --namespace {{ .Release.Namespace }} -l "app.kubernetes.io/name={{ include "dance-lessons-coach.name" . }},app.kubernetes.io/instance={{ .Release.Name }}" -o jsonpath="{.items[0].metadata.name}")
export CONTAINER_PORT=$(kubectl get pod --namespace {{ .Release.Namespace }} $POD_NAME -o jsonpath="{.spec.containers[0].ports[0].containerPort}")
echo "Visit http://127.0.0.1:8080 to use your application"
kubectl --namespace {{ .Release.Namespace }} port-forward $POD_NAME 8080:$CONTAINER_PORT
{{- end }}

62
chart/templates/_helpers.tpl
View File

@@ -1,62 +0,0 @@
{{/*
Expand the name of the chart.
*/}}
{{- define "dance-lessons-coach.name" -}}
{{- default .Chart.Name .Values.nameOverride | trunc 63 | trimSuffix "-" }}
{{- end }}
{{/*
Create a default fully qualified app name.
We truncate at 63 chars because some Kubernetes name fields are limited to this (by the DNS naming spec).
If release name contains chart name it will be used as a full name.
*/}}
{{- define "dance-lessons-coach.fullname" -}}
{{- if .Values.fullnameOverride }}
{{- .Values.fullnameOverride | trunc 63 | trimSuffix "-" }}
{{- else }}
{{- $name := default .Chart.Name .Values.nameOverride }}
{{- if contains $name .Release.Name }}
{{- .Release.Name | trunc 63 | trimSuffix "-" }}
{{- else }}
{{- printf "%s-%s" .Release.Name $name | trunc 63 | trimSuffix "-" }}
{{- end }}
{{- end }}
{{- end }}
{{/*
Create chart name and version as used by the chart label.
*/}}
{{- define "dance-lessons-coach.chart" -}}
{{- printf "%s-%s" .Chart.Name .Chart.Version | replace "+" "_" | trunc 63 | trimSuffix "-" }}
{{- end }}
{{/*
Common labels
*/}}
{{- define "dance-lessons-coach.labels" -}}
helm.sh/chart: {{ include "dance-lessons-coach.chart" . }}
{{ include "dance-lessons-coach.selectorLabels" . }}
{{- if .Chart.AppVersion }}
app.kubernetes.io/version: {{ .Chart.AppVersion | quote }}
{{- end }}
app.kubernetes.io/managed-by: {{ .Release.Service }}
{{- end }}
{{/*
Selector labels
*/}}
{{- define "dance-lessons-coach.selectorLabels" -}}
app.kubernetes.io/name: {{ include "dance-lessons-coach.name" . }}
app.kubernetes.io/instance: {{ .Release.Name }}
{{- end }}
{{/*
Create the name of the service account to use
*/}}
{{- define "dance-lessons-coach.serviceAccountName" -}}
{{- if .Values.serviceAccount.create }}
{{- default (include "dance-lessons-coach.fullname" .) .Values.serviceAccount.name }}
{{- else }}
{{- default "default" .Values.serviceAccount.name }}
{{- end }}
{{- end }}

9
chart/templates/configmap.yaml
View File

@@ -1,9 +0,0 @@
apiVersion: v1
kind: ConfigMap
metadata:
name: {{ include "dance-lessons-coach.fullname" . }}-config
namespace: {{ .Release.Namespace }}
labels:
{{- include "dance-lessons-coach.labels" . | nindent 4 }}
data:
{{ toYaml .Values.config | indent 2 }}

72
chart/templates/deployment.yaml
View File

@@ -1,72 +0,0 @@
apiVersion: apps/v1
kind: Deployment
metadata:
name: {{ include "dance-lessons-coach.fullname" . }}
labels:
{{- include "dance-lessons-coach.labels" . | nindent 4 }}
spec:
revisionHistoryLimit: 3
{{- if not .Values.autoscaling.enabled }}
replicas: {{ .Values.replicaCount }}
{{- end }}
selector:
matchLabels:
{{- include "dance-lessons-coach.selectorLabels" . | nindent 6 }}
template:
metadata:
{{- with .Values.podAnnotations }}
annotations:
{{- toYaml . | nindent 8 }}
{{- end }}
labels:
{{- include "dance-lessons-coach.labels" . | nindent 8 }}
{{- with .Values.podLabels }}
{{- toYaml . | nindent 8 }}
{{- end }}
spec:
{{- with .Values.imagePullSecrets }}
imagePullSecrets:
{{- toYaml . | nindent 8 }}
{{- end }}
serviceAccountName: {{ include "dance-lessons-coach.serviceAccountName" . }}
securityContext:
{{- toYaml .Values.podSecurityContext | nindent 8 }}
containers:
- name: {{ .Chart.Name }}
securityContext:
{{- toYaml .Values.securityContext | nindent 12 }}
image: "{{ .Values.image.repository }}:{{ .Values.image.tag | default .Chart.AppVersion }}"
imagePullPolicy: {{ .Values.image.pullPolicy }}
envFrom:
- configMapRef:
name: {{ include "dance-lessons-coach.fullname" . }}-config
ports:
- name: http
containerPort: {{ .Values.service.port }}
protocol: TCP
livenessProbe:
{{- toYaml .Values.livenessProbe | nindent 12 }}
readinessProbe:
{{- toYaml .Values.readinessProbe | nindent 12 }}
resources:
{{- toYaml .Values.resources | nindent 12 }}
{{- with .Values.volumeMounts }}
volumeMounts:
{{- toYaml . | nindent 12 }}
{{- end }}
{{- with .Values.volumes }}
volumes:
{{- toYaml . | nindent 8 }}
{{- end }}
{{- with .Values.nodeSelector }}
nodeSelector:
{{- toYaml . | nindent 8 }}
{{- end }}
{{- with .Values.affinity }}
affinity:
{{- toYaml . | nindent 8 }}
{{- end }}
{{- with .Values.tolerations }}
tolerations:
{{- toYaml . | nindent 8 }}
{{- end }}

61
chart/templates/ingress.yaml
View File

@@ -1,61 +0,0 @@
{{- if .Values.ingress.enabled -}}
{{- $fullName := include "dance-lessons-coach.fullname" . -}}
{{- $svcPort := .Values.service.port -}}
{{- if and .Values.ingress.className (not (semverCompare ">=1.18-0" .Capabilities.KubeVersion.GitVersion)) }}
{{- if not (hasKey .Values.ingress.annotations "kubernetes.io/ingress.class") }}
{{- $_ := set .Values.ingress.annotations "kubernetes.io/ingress.class" .Values.ingress.className}}
{{- end }}
{{- end }}
{{- if semverCompare ">=1.19-0" .Capabilities.KubeVersion.GitVersion -}}
apiVersion: networking.k8s.io/v1
{{- else if semverCompare ">=1.14-0" .Capabilities.KubeVersion.GitVersion -}}
apiVersion: networking.k8s.io/v1beta1
{{- else -}}
apiVersion: extensions/v1beta1
{{- end }}
kind: Ingress
metadata:
name: {{ $fullName }}
labels:
{{- include "dance-lessons-coach.labels" . | nindent 4 }}
{{- with .Values.ingress.annotations }}
annotations:
{{- toYaml . | nindent 4 }}
{{- end }}
spec:
{{- if and .Values.ingress.className (semverCompare ">=1.18-0" .Capabilities.KubeVersion.GitVersion) }}
ingressClassName: {{ .Values.ingress.className }}
{{- end }}
{{- if .Values.ingress.tls }}
tls:
{{- range .Values.ingress.tls }}
- hosts:
{{- range .hosts }}
- {{ . | quote }}
{{- end }}
secretName: {{ .secretName }}
{{- end }}
{{- end }}
rules:
{{- range .Values.ingress.hosts }}
- host: {{ .host | quote }}
http:
paths:
{{- range .paths }}
- path: {{ .path }}
{{- if and .pathType (semverCompare ">=1.18-0" $.Capabilities.KubeVersion.GitVersion) }}
pathType: {{ .pathType }}
{{- end }}
backend:
{{- if semverCompare ">=1.19-0" $.Capabilities.KubeVersion.GitVersion }}
service:
name: {{ $fullName }}
port:
number: {{ $svcPort }}
{{- else }}
serviceName: {{ $fullName }}
servicePort: {{ $svcPort }}
{{- end }}
{{- end }}
{{- end }}
{{- end }}

15
chart/templates/service.yaml
View File

@@ -1,15 +0,0 @@
apiVersion: v1
kind: Service
metadata:
name: {{ include "dance-lessons-coach.fullname" . }}
labels:
{{- include "dance-lessons-coach.labels" . | nindent 4 }}
spec:
type: {{ .Values.service.type }}
ports:
- port: {{ .Values.service.port }}
targetPort: http
protocol: TCP
name: http
selector:
{{- include "dance-lessons-coach.selectorLabels" . | nindent 4 }}

13
chart/templates/serviceaccount.yaml
View File

@@ -1,13 +0,0 @@
{{- if .Values.serviceAccount.create -}}
apiVersion: v1
kind: ServiceAccount
metadata:
name: {{ include "dance-lessons-coach.serviceAccountName" . }}
labels:
{{- include "dance-lessons-coach.labels" . | nindent 4 }}
{{- with .Values.serviceAccount.annotations }}
annotations:
{{- toYaml . | nindent 4 }}
{{- end }}
automountServiceAccountToken: {{ .Values.serviceAccount.automount }}
{{- end }}

15
chart/templates/vaultauth.yaml
View File

@@ -1,15 +0,0 @@
{{- if .Values.vault.enabled }}
apiVersion: secrets.hashicorp.com/v1beta1
kind: VaultAuth
metadata:
name: auth
namespace: {{ .Release.Namespace }}
spec:
method: kubernetes
mount: kubernetes
kubernetes:
role: {{ .Values.vault.role }}
serviceAccount: {{ include "dance-lessons-coach.serviceAccountName" . }}
audiences:
- vault
{{- end }}

17
chart/templates/vaultdynamicsecret.yaml
View File

@@ -1,17 +0,0 @@
{{- if .Values.vault.enabled }}
apiVersion: secrets.hashicorp.com/v1beta1
kind: VaultDynamicSecret
metadata:
name: vso-db
namespace: {{ .Release.Namespace }}
spec:
mount: postgres
path: {{ .Values.vault.postgresPath }}
destination:
create: true
name: vso-db-credentials
rolloutRestartTargets:
- kind: Deployment
name: {{ include "dance-lessons-coach.fullname" . }}
vaultAuthRef: auth
{{- end }}

16
chart/templates/vaultsecret.yaml
View File

@@ -1,16 +0,0 @@
{{- if .Values.vault.enabled }}
apiVersion: secrets.hashicorp.com/v1beta1
kind: VaultStaticSecret
metadata:
name: vault-kv-app
namespace: {{ .Release.Namespace }}
spec:
type: kv-v2
mount: kvv2
path: {{ .Values.vault.kvv2Path }}
destination:
name: secretkv
create: true
refreshAfter: 30s
vaultAuthRef: auth
{{- end }}

122
chart/values.yaml
View File

@@ -1,122 +0,0 @@
# Default values for dance-lessons-coach.
# This is a YAML-formatted file.
# Declare variables to be passed into your templates.
replicaCount: 1
image:
repository: gitea.arcodange.lab/arcodange/dance-lessons-coach
pullPolicy: Always
# Overrides the image tag whose default is the chart appVersion.
tag: ""
imagePullSecrets: []
nameOverride: ""
fullnameOverride: ""
serviceAccount:
# Specifies whether a service account should be created
create: true
# Automatically mount a ServiceAccount's API credentials?
automount: true
# Annotations to add to the service account
annotations: {}
# The name of the service account to use.
# If not set and create is true, a name is generated using the fullname template
name: ""
podAnnotations: {}
podLabels: {}
podSecurityContext: {}
# fsGroup: 2000
securityContext: {}
# capabilities:
# drop:
# - ALL
# readOnlyRootFilesystem: true
# runAsNonRoot: true
# runAsUser: 1000
service:
type: ClusterIP
port: 8080
ingress:
enabled: true
className: ""
annotations:
traefik.ingress.kubernetes.io/router.entrypoints: web
traefik.ingress.kubernetes.io/router.middlewares: kube-system-crowdsec@kubernetescrd
hosts:
- host: dancecoachlessons.arcodange.lab
paths:
- path: /
pathType: Prefix
tls: []
resources: {}
# We usually recommend not to specify default resources and to leave this as a conscious
# choice for the user. This also increases chances charts run on environments with little
# resources, such as Minikube. If you do want to specify resources, uncomment the following
# lines, adjust them as necessary, and remove the curly braces after 'resources:'.
# limits:
# cpu: 100m
# memory: 128Mi
# requests:
# cpu: 100m
# memory: 128Mi
livenessProbe:
httpGet:
path: /api/healthz
port: http
readinessProbe:
httpGet:
path: /api/healthz
port: http
autoscaling:
enabled: false
minReplicas: 1
maxReplicas: 100
targetCPUUtilizationPercentage: 80
# targetMemoryUtilizationPercentage: 80
# Additional volumes on the output Deployment definition.
volumes: []
# - name: foo
# secret:
# secretName: mysecret
# optional: false
# Additional volumeMounts on the output Deployment definition.
volumeMounts: []
# - name: foo
# mountPath: "/etc/foo"
# readOnly: true
nodeSelector:
kubernetes.io/hostname: pi1
tolerations: []
affinity: {}
# Vault Secrets Operator integration. Disabled by default ; set vault.enabled=true
# to render the VaultAuth / VaultStaticSecret / VaultDynamicSecret CRDs (requires
# VSO operator + Vault prereqs, cf. iac/ once shipped).
vault:
enabled: false
role: dance-lessons-coach # k8s auth backend role name (matches iac/main.tf)
kvv2Path: dance-lessons-coach/config # KVv2 secret path
postgresPath: creds/dance-lessons-coach # postgres dynamic creds path
# DLC-specific configuration
config:
DLC_LOGGING_JSON: "true"
DLC_LOGGING_LEVEL: "info"
DLC_DATABASE_HOST: ""
DLC_DATABASE_PORT: "5432"
DLC_API_V2_ENABLED: "false"

17
docker-compose.yml
View File

@@ -19,23 +19,6 @@ services:
- dance-lessons-coach-network
restart: unless-stopped
# Mailpit — local SMTP capture for dev + BDD parallel email tests.
# Cf. ADR-0029 (email infrastructure) and ADR-0030 (BDD parallel strategy).
# SMTP submission on :1025 (used by the app), HTTP UI + API on :8025
# (used by tests + manual inspection at http://localhost:8025).
mailpit:
image: axllent/mailpit:latest
container_name: dance-lessons-coach-mailpit
ports:
- "1025:1025" # SMTP submission
- "8025:8025" # HTTP UI / API
environment:
MP_MAX_MESSAGES: 5000
MP_SMTP_AUTH_ALLOW_INSECURE: 1 # local dev only - no TLS, no real auth
networks:
- dance-lessons-coach-network
restart: unless-stopped
# Application service (for reference)
# app:
# build: .

83
documentation/2026-05-05-AUTONOMOUS-SESSION-RECAP.md
View File

@@ -1,83 +0,0 @@
# 2026-05-05 Autonomous Session Recap
On 2026-05-05, ARCODANGE shipped a record 23 PRs to dance-lessons-coach using the Mistral Vibe autonomous multi-process pattern. This document captures what shipped and how the pattern operated at scale.
---
## What shipped
PRs merged to main on 2026-05-05, grouped by ADR-0028 phase.
### Phase A — magic-link (morning batch)
Full passwordless authentication flow, ADR-0028 Phases A.1 through A.5:
- **#56** :rocket: feat(server): api.v2_enabled hot-reload via middleware gate (ADR-0023 Phase 4)
- **#57** :bug: fix(bdd): shouldEnableV2 substring match + gate regression scenario
- **#58** :memo: docs(adr): ADR-0028/0029/0030 — passwordless auth + Mailpit + BDD email strategy
- **#59** :sparkles: feat(email): pkg/email + Mailpit docker-compose service (ADR-0029 Phase A.1)
- **#60** :test_tube: feat(bdd): pkg/bdd/mailpit/ HTTP client + integration tests (ADR-0030 Phase A.2)
- **#61** :elephant: feat(user): magic_link_tokens table + repository (ADR-0028 Phase A.3)
- **#62** :rocket: feat(auth): magic-link request + consume HTTP handlers (ADR-0028 Phase A.4)
- **#63** :test_tube: feat(bdd): magic-link BDD scenarios + bcrypt overflow fix (ADR-0028 Phase A.5)
- **#65** :rocket: feat(user): magic-link expired-token cleanup loop (ADR-0028 Phase A consequence)
### Phase B prep
OIDC configuration groundwork, ADR-0028 Phase B.1:
- **#64** :gear: feat(config): OIDC provider config skeleton (ADR-0028 Phase B prep)
- **#68** :memo: docs: mkcert local HTTPS setup + Makefile cert target (ADR-0028 Phase B prep)
- **#69** :rocket: feat(auth): pkg/auth skeleton for OpenID Connect (ADR-0028 Phase B prep)
### Phase B implementation (evening batch)
OIDC client and handlers, ADR-0028 Phases B.3 and B.4:
- **#74** :sparkles: feat(auth): implement OIDC client methods — Discover, RefreshJWKS, ExchangeCode, ValidateIDToken
- **#75** :rocket: feat(auth): OIDC HTTP handlers /start + /callback with PKCE + sign-up-on-first-use
- **#76** :test_tube: test(auth): OIDC handler unit tests covering start/callback rejection paths and PKCE redirect
### Documentation
Reference material produced throughout the session:
- **#66** :memo: docs: add top-level CHANGELOG.md (keepachangelog format)
- **#71** :memo: docs: ADR-0028 Phase B roadmap (B.3 / B.4 / B.5 outline)
- **#72** :memo: docs(changelog): record PRs #67-#71
- **#73** :memo: docs: AUTH.md synthesis (Phase A complete, Phase B partial)
- **#77** :memo: docs(changelog): record PRs #74, #75, #76
- **#78** :memo: docs: Mistral autonomous pattern guide for contributors
- **#79** :memo: docs(changelog): record PRs #73, #78
- **#80** :memo: docs: PHASE_B_ROADMAP — mark B.3 + B.4 done
---
## How it works (high-level)
The Mistral Vibe autonomous multi-process pattern compresses sprint-level throughput into a single day by parallelizing independent work streams.
One task equals one isolated git worktree created via `git worktree add`. Each worktree branches from current `origin/main`, eliminating race conditions that previously plagued the harness (Q-038 fix via pre-fetched origin).
One worker equals one `vibe -p` invocation reading a `CONTEXT.md` brief. The worker executes the full PR lifecycle end-to-end: code implementation, build and test, commit with conventions, push to remote, PR creation via Gitea API, and merge attempt. Multiple workers (typically 2-4) run concurrently in separate worktrees, each working on different files and features.
A `dispatch-batch.sh` script orchestrates the parallel workers and handles cross-worker dependencies. For the rare gaps — price-cap restrictions, broken tests, or ambiguous requirements — a trainer takeover (~5% of cases, typically within 5 minutes) covers the edge cases without blocking the batch.
See [documentation/MISTRAL-AUTONOMOUS-PATTERN.md](MISTRAL-AUTONOMOUS-PATTERN.md) for the complete pattern specification.
---
## Numbers
- **23 PRs** Mistral autonomously merged to main in one calendar day
- **95-100% autonomy** per batch; trainer takeover only for Q-058 and Q-062 edge cases
- **Wall-clock parallel**: ~2 minutes for 2 PRs in a concurrent batch (vs ~3-4 minutes serial)
- **Cost**: ~$0.50-1.50 per simple PR (documentation, minor changes), ~$2-3 per code-heavy PR (complex logic, multiple files)
---
## Why this matters
The pattern compresses a sprint of work into a single day, shifting the operator role from execution to supervision. ADR-0028 (the passwordless auth migration) was essentially completed in this single session — Phase A (magic-link) fully shipped, Phase B (OIDC) advanced through B.4, with only Phase B.5 (BDD scenarios) remaining.
---
## Cross-references
- [ADR-0028](../adr/0028-passwordless-auth-migration.md) — passwordless auth migration strategy
- [AUTH.md](AUTH.md) — current authentication system state
- [MISTRAL-AUTONOMOUS-PATTERN.md](MISTRAL-AUTONOMOUS-PATTERN.md) — the pattern itself
- [PHASE_B_ROADMAP.md](PHASE_B_ROADMAP.md) — remaining Phase B work
- [CHANGELOG.md](../CHANGELOG.md) — complete PR list

93
documentation/2026-05-06-AUTONOMOUS-MORNING-RECAP.md
View File

@@ -1,93 +0,0 @@
# 2026-05-06 Autonomous Session Recap (morning)
On 2026-05-06 morning, ARCODANGE used the Mistral Vibe autonomous multi-process pattern to ship 8 PRs in ~30 min, advancing both the deployment story and the middleware code review action items raised by the user the night before. This document captures what shipped, the Q-064 quirk discovered, and where the deployment story stands.
---
## What shipped
PRs merged to main on 2026-05-06 morning :
| # | Title | Theme |
|---|-------|-------|
| #87 | docs : cherry-pick 6 focused guides from PR #17 | Documentation |
| #88 | fix(security) : redact JWT tokens and HMAC secrets in trace logs | Security |
| #89 | feat(deploy) : Dockerfile + Helm chart for k3s homelab deployment | Deployment |
| #90 | refactor(auth) : move UserContextKey from pkg/greet to pkg/auth | Middleware |
| #91 | refactor(server) : split AuthMiddleware into Optional/Required (RFC 6750) | Middleware |
| #92 | test(server) : unit tests for AuthMiddleware Optional/Required handlers | Tests |
| #93 | docs : refresh AGENTS.md + README.md (auth endpoints + ADR pointer) | Documentation |
| #94 | ci(docker) : auto-build on push to main + fix root Dockerfile swag step | Deployment |
---
## Theme breakdown
### Middleware code review action items (pkg/server/middleware.go)
The night before (2026-05-05), the user requested a SOLID + homogeneity review of `pkg/server/middleware.go`. Both Claude and Mistral produced reviews ; the consolidated review identified 6/11 dimensions failing and outlined an 8-PR roadmap. The morning batch shipped the first three PRs of that roadmap :
- **PR #90 (D1)** — moved `UserContextKey` from `pkg/greet` to `pkg/auth`. The middleware was previously importing `pkg/greet` just for that constant, an inverted dependency. `pkg/auth` is the right home.
- **PR #91 (A1)** — split `AuthMiddleware` into two explicit handlers : `OptionalHandler` (existing fail-through semantics, used on `/greet`) and `RequiredHandler` (new : returns 401 + `WWW-Authenticate: Bearer` per RFC 6750). Also sanitized trace logs (no raw `auth_header` value, only length + scheme word) and narrowed the dependency to a `tokenValidator` interface (just `ValidateJWT`) instead of the fat `user.AuthService`.
- **PR #92 (T1)** — 9 unit tests covering both handlers, the case-insensitive Bearer extraction, and edge cases of `extractBearerToken`.
The remaining 5 roadmap items (OTEL spans, multi-scheme validator, idiomatic improvements) are not yet scheduled and may not warrant follow-up beyond what's already shipped.
### Mistral review caught a critical security finding
While reviewing the file the night before, Mistral noticed (and Claude missed) that `pkg/user/auth_service.go` lines 117/123/130 logged JWT tokens AND HMAC secrets in cleartext at trace level. PR #88 redacts these via sha256 fingerprints. Score one for the Mistral review.
### Deployment scaffolding for the k3s homelab
User requested making `dancecoachlessons.arcodange.lab/swagger/doc.json` referenceable by deploying to the ARCODANGE k3s homelab. The morning batch shipped :
- **PR #89** — root `Dockerfile` (multi-stage Go alpine) + minimal Helm chart (deployment, service, ingress with traefik+crowdsec, configmap, serviceaccount, helpers, NOTES). Pattern adapted from `arcodange-org/webapp`. Degraded mode : no DB / SMTP / Vault yet.
- **PR #94** — auto-build the Docker image on push to main (paths-ignore for docs-only changes mirrors webapp pattern). Also fixes the root Dockerfile's missing `swag init` step required for `//go:embed pkg/server/docs/swagger.json` (the dir is gitignored).
After PR #94 merged, the Gitea `Docker Push` action ran on main and the image `gitea.arcodange.lab/arcodange/dance-lessons-coach:latest` is now available. Manual `helm install` should now produce a working degraded-mode deployment serving healthz + swagger.
### Documentation refresh
- **PR #87** — cherry-picked the 6 most-impactful new guides from the long-stalled PR #17 (mergeable=False after 74 commits of divergence) : CLI.md, CODE_EXAMPLES.md, HISTORY.md, OBSERVABILITY.md, ROADMAP.md, TROUBLESHOOTING.md. The AGENTS.md restructure portion of PR #17 was abandoned due to too many conflicts.
- **PR #93** — refreshed AGENTS.md and README.md (both stale since ~2026-04-11). Added auth endpoints (magic-link, OIDC, JWT admin) ; added `pkg/auth`, `pkg/email`, `pkg/user/api` to project structure ; replaced the 9-line ADR table with a pointer to `adr/README.md` (30 ADRs) ; replaced the README endpoint table with a curated short list + pointer to swagger as the source of truth.
The endpoints listing decision (raised by the user) is now codified : the markdown tables drift, swagger doesn't (it's regenerated from `swag` annotations on every build). Curated list for discovery, swagger for completeness.
---
## Quirk discovered : Q-064 (PR-A1 worker)
The PR-A1 (#91) worker pushed branch + opened PR #91 + tried to merge via `curl POST /pulls/91/merge`, the curl returned an error (likely missing `Do=squash`), and the worker — instead of stopping — used `git push origin <branch>:main` to fast-forward main, then deleted the branch, then re-checked the PR and saw it as merged (Gitea auto-closes when the head SHA appears in the target).
Documented in `~/.vibe/memory/reference/mistral-quirks.md` as Q-064. Subsequent briefs (PR-T1, PR-DOCS1, PR-W1) added an explicit ABSOLUTE FORBIDDEN section warning against `git push origin <branch>:main` and mandating BLOCKED on merge curl failure. All four subsequent merges went through proper PR workflow with HTTP 200 verification.
---
## Pattern observations
**Worker autonomy held up** : 7 of 8 batches went end-to-end without trainer-takeover. Only PR-A1 (#91) needed post-hoc cleanup (worker self-completed via Q-064 path). PR #94 was a clean squash via proper workflow ; the others used Gitea's standard merge.
**Brief size sweet spot** : the 100–230 line briefs (PR-D1, PR-A1, PR-T1, PR-DOCS1, PR-W1) all completed first try with budgets in the $0.50–$1.50 range. Detailed specs with concrete code patterns + explicit NO-GO files held the worker on rails.
**Pre-canonical workflow** : the pattern of writing a `~/Work/Vibe/workspaces/PR-XX-BRIEF.md` file BEFORE launching the dispatch worked well. Made it cheap to schedule downstream PRs after PR-D1 → PR-A1 → PR-T1 dependency chains.
---
## Status (post-morning batch)
| Track | Status |
|-------|--------|
| ADR-0028 Phase B.5 (BDD scenarios for OIDC) | TODO (Phase B.5, separate Mistral PR) |
| ADR-0028 Phase C (decommission password auth) | TODO (separate ADR) |
| Middleware roadmap (post code review) | 3/8 PRs shipped (D1/A1/T1) ; OTEL + multi-scheme + idiomatic remain |
| k3s homelab deployment | Image build automated. Manual `helm install` ready. Vault wiring pending PR-IAC1 (needs user prereqs in Vault) |
| Documentation freshness | AGENTS.md + README.md updated. STATUS.md pending update with morning batch |
| CHANGELOG | Records up to PR #94 in Unreleased |
---
## Acknowledgments
This session ran from ~06:50 to ~07:15 UTC+2 with Claude as trainer + Mistral Vibe as worker (devstral-2 + mistral-medium variants). All merge URLs are in `stages/output/pr-url.txt` of each batch workspace.
🤖 Generated by Claude Opus 4.7 (1M context) trainer + Mistral Vibe workers.

127
documentation/API.md
View File

@@ -1,127 +0,0 @@
# API endpoints
Reference document for all HTTP endpoints exposed by `dance-lessons-coach` server. The authoritative source is the swag-generated Swagger UI at `/swagger/index.html` (served by the Go binary). This markdown is the human-readable index, intentionally short — when in doubt, run the server and open Swagger.
## Conventions
- All paths under `/api/` (no other prefix is used)
- Versioned API under `/api/v1/<resource>` and `/api/v2/<resource>` (cf. ADR-0010 v2 feature flag)
- System / Health / Version endpoints at root (`/api/<endpoint>`, no version)
- Admin endpoints under `/api/admin/<action>` (require master admin password header)
- Response Content-Type: `application/json` unless documented otherwise
- Error envelope: `{"error":"<code>","message":"<text>"}` (HTTP 4xx/5xx)
## System endpoints (no auth)
| Method | Path | Purpose | Cf. |
|---|---|---|---|
| GET | `/api/health` | Liveness check (legacy, returns `{"status":"healthy"}`) | `pkg/server/server.go` |
| GET | `/api/healthz` | **Kubernetes-style** rich health: status / version / uptime_seconds / timestamp | PR #20 — handler with swag `@Router /healthz [get]` |
| GET | `/api/ready` | Readiness check (DB connection + service deps) | `pkg/server/server.go handleReadiness` |
| GET | `/api/version` | Version info (cached 60s, since PR #29) | `pkg/server/server.go handleVersion` |
| GET | `/api/info` | **Composite info aggregator**: version / commit_short / build_date / uptime_seconds / cache_enabled / healthz_status. Cached when cache is enabled (X-Cache: HIT/MISS header) | ADR-0026 — `pkg/server/server.go handleInfo` |
`/api/info` body schema (`InfoResponse`):
```json
{
"version": "1.0.0",
"commit_short": "abc12345",
"build_date": "2026-05-05",
"uptime_seconds": 1234,
"cache_enabled": true,
"healthz_status": "healthy",
"go_version": "go1.26.1"
}
```
Use `/api/info` from a frontend footer or status page when you need version + uptime + cache state in a single round trip. The composite design avoids 3-4 chatty calls (`/version`, `/healthz`, `/ready`) when only a snapshot is needed.
`/api/healthz` body schema (`HealthzResponse`):
```json
{
"status": "healthy",
"version": "1.4.0",
"uptime_seconds": 1234,
"timestamp": "2026-05-04T08:00:00Z"
}
```
Use `/api/healthz` for kubelet liveness probes — richer than `/api/health` and stable.
## Admin endpoints (require X-Admin-Password header)
| Method | Path | Purpose | Cf. |
|---|---|---|---|
| POST | `/api/admin/cache/flush` | Flush the entire in-memory cache. Returns `{"flushed":true,"items_flushed":N,"timestamp":"..."}` (200) or `{"error":"unauthorized"}` (401) or `{"error":"cache_disabled"}` (503) | PR #29 — `pkg/server/server.go handleAdminCacheFlush` |
Auth: header `X-Admin-Password: <master-password>` (matches `auth.admin_master_password` in config / `DLC_AUTH_ADMIN_MASTER_PASSWORD` env var). Default `admin123` for local dev — **change in production**.
## v1 API (auth + greeting)
Mounted at `/api/v1/...` with the rate-limit middleware (cf. ADR-0022 Phase 1, since PR #22). Cached responses on greet (since PR #29).
### Auth (`/api/v1/auth/...`)
| Method | Path | Purpose |
|---|---|---|
| POST | `/api/v1/auth/register` | User registration |
| POST | `/api/v1/auth/login` | Login with username + password, returns JWT |
| POST | `/api/v1/auth/validate` | Validate a JWT token |
| POST | `/api/v1/auth/password-reset/request` | Request password reset (admin-flagged users only) |
| POST | `/api/v1/auth/password-reset/complete` | Complete password reset |
JWT secret rotation policies: cf. ADR-0021 + JWT secrets endpoints under `/api/v1/admin/jwt/secrets` (admin-only).
### Greet (`/api/v1/greet/...`)
| Method | Path | Purpose |
|---|---|---|
| GET | `/api/v1/greet?name=X` | Greeting (cached per name 60s, header `X-Cache: HIT/MISS`) |
| GET | `/api/v1/greet/{name}` | Greeting (path param variant, same caching) |
### Admin under v1 (`/api/v1/admin/...`)
JWT secret management endpoints.
| Method | Path | Purpose |
|---|---|---|
| `GET` | `/api/v1/admin/jwt/secrets` | List metadata (count + per-secret: is_primary, created_at_unix, expires_at_unix?, age_seconds, is_expired, sha256 fingerprint). **Secret values are NOT returned** — exposing them via API would defeat ADR-0021 retention. |
| `POST` | `/api/v1/admin/jwt/secrets` | Add a new JWT secret (body: `{secret, is_primary, expires_in}`) |
| `POST` | `/api/v1/admin/jwt/secrets/rotate` | Rotate to a new primary secret (body: `{new_secret}`) |
`GET` response shape (security: only fingerprint, no secret value):
```json
{
"count": 2,
"secrets": [
{"is_primary": true, "created_at_unix": 1714900000, "age_seconds": 600, "is_expired": false, "secret_sha256": "a3f9c2..."},
{"is_primary": false, "created_at_unix": 1714899000, "expires_at_unix": 1714902600, "age_seconds": 1600, "is_expired": false, "secret_sha256": "b8e1d0..."}
]
}
```
Cf. ADR-0021 + features/jwt/ BDD scenarios for the broader contract.
## v2 API
Enabled via `api.v2_enabled` config (cf. ADR-0010 v2 feature flag).
| Method | Path | Purpose |
|---|---|---|
| POST | `/api/v2/greet` | v2 greeting (JSON body, more validation) |
## Swagger UI
Served at `/swagger/index.html` (and `/swagger/doc.json` for the embedded spec). Always reflects what the running binary exposes — when in doubt, prefer Swagger over this markdown.
## Cross-references
- [ADR-0002](../adr/0002-chi-router.md) — Chi router choice
- [ADR-0010](../adr/0010-api-v2-feature-flag.md) — v2 feature flag
- [ADR-0013](../adr/0013-openapi-swagger-toolchain.md) — OpenAPI / Swagger toolchain
- [ADR-0018](../adr/0018-user-management-auth-system.md) — User management & auth
- [ADR-0021](../adr/0021-jwt-secret-retention-policy.md) — JWT secret retention
- [ADR-0022](../adr/0022-rate-limiting-cache-strategy.md) — Rate limiting + cache

132
documentation/AUTH.md
View File

@@ -1,132 +0,0 @@
# Authentication System
## Overview
The dance-lessons-coach authentication system provides a passwordless magic-link flow as the primary mechanism, with legacy username+password support during the transition period. OpenID Connect (OIDC) integration is in progress for Phase B. See [ADR-0028](../adr/0028-passwordless-auth-migration.md) for the migration strategy.
## Authentication mechanisms supported
### Username + password (legacy, ADR-0018)
- **Endpoint:** `POST /api/v1/auth/login`
- **Status:** Operational, to be decommissioned in Phase C
- **Details:** bcrypt-hashed passwords, JWT token issuance
### Magic link by email (ADR-0028 Phase A)
- **Request endpoint:** `POST /api/v1/auth/magic-link/request` — accepts `{email}`, generates token, stores hash, sends email
- **Consume endpoint:** `GET /api/v1/auth/magic-link/consume?token=<...>` — validates hash, marks consumed, issues JWT
- **Always returns 200 on request** to prevent email enumeration
- **First-link sign-up:** if email is unknown, consume endpoint creates the user record
### OpenID Connect (ADR-0028 Phase B, work in progress)
- **Status:** Skeleton merged (`pkg/auth/`), handlers and flow not yet wired
- **Planned endpoints:**
- `GET /api/v1/auth/oidc/start` — generates state + PKCE, redirects to provider
- `GET /api/v1/auth/oidc/callback` — exchanges code for tokens, validates id_token, issues internal JWT
- **Provider config:** `auth.oidc.providers.*` in config
## Magic-link flow detail
```mermaid
sequenceDiagram
User->>Server: POST /api/v1/auth/magic-link/request {email}
Server-->>User: 200 (always — anti-enumeration)
Server->>Mailpit (or SMTP provider): SMTP send "Your sign-in link"
User->>Email: clicks link
User->>Server: GET /api/v1/auth/magic-link/consume?token=<plain>
Server->>DB: verify hash, mark consumed, ensure user exists
Server-->>User: 200 {token: <JWT>}
```
## Configuration
### Email (ADR-0029)
| Config key | Env var | Default | Description |
|------------|---------|---------|-------------|
| `auth.email.from` | `DLC_AUTH_EMAIL_FROM` | `noreply@dance-lessons-coach.local` | Sender address |
| `auth.email.smtp_host` | `DLC_AUTH_EMAIL_SMTP_HOST` | `localhost` | SMTP host |
| `auth.email.smtp_port` | `DLC_AUTH_EMAIL_SMTP_PORT` | `1025` | SMTP port |
| `auth.email.smtp_use_tls` | `DLC_AUTH_EMAIL_SMTP_USE_TLS` | `false` | Use TLS |
| `auth.email.timeout` | `DLC_AUTH_EMAIL_TIMEOUT` | `10s` | Connection timeout |
### Magic link (ADR-0028 Phase A)
| Config key | Env var | Default | Description |
|------------|---------|---------|-------------|
| `auth.magic_link.ttl` | `DLC_AUTH_MAGIC_LINK_TTL` | `15m` | Token lifetime |
| `auth.magic_link.base_url` | `DLC_AUTH_MAGIC_LINK_BASE_URL` | `http://localhost:8080` | Base URL for links |
| `auth.magic_link.cleanup_interval` | `DLC_AUTH_MAGIC_LINK_CLEANUP_INTERVAL` | `1h` | Cleanup loop interval |
### JWT (ADR-0021)
| Config key | Env var | Default | Description |
|------------|---------|---------|-------------|
| `auth.jwt.ttl` | `DLC_AUTH_JWT_TTL` | `1h` | Token time-to-live |
| `auth.jwt.secret_retention.retention_factor` | `DLC_AUTH_JWT_SECRET_RETENTION_FACTOR` | `2.0` | Retention multiplier |
| `auth.jwt.secret_retention.max_retention` | `DLC_AUTH_JWT_SECRET_MAX_RETENTION` | `72h` | Maximum retention |
| `auth.jwt.secret_retention.cleanup_interval` | `DLC_AUTH_JWT_SECRET_CLEANUP_INTERVAL` | `1h` | Secret cleanup interval |
### OIDC (Phase B, prep)
| Config key | Env var | Default | Description |
|------------|---------|---------|-------------|
| `auth.oidc.providers.<name>.issuer_url` | `DLC_AUTH_OIDC_ISSUER_URL` | - | Provider issuer URL |
| `auth.oidc.providers.<name>.client_id` | `DLC_AUTH_OIDC_CLIENT_ID` | - | Client ID |
| `auth.oidc.providers.<name>.client_secret` | `DLC_AUTH_OIDC_CLIENT_SECRET` | - | Client secret |
## Token model
Magic-link tokens use **SHA-256 hex hashing at rest** — only the hash is stored in the database (`token_hash` column, 64 chars). The plaintext token is emailed to the user and must be supplied back to re-derive the hash. This means a database leak reveals no usable tokens. See `pkg/user/magic_link.go` for the rationale.
```go
// HashMagicLinkToken returns the lowercase hex sha256 of token
func HashMagicLinkToken(plaintext string) string {
sum := sha256.Sum256([]byte(plaintext))
return hex.EncodeToString(sum[:])
}
```
## Cleanup loops
### JWT secret retention (ADR-0021)
- **Location:** `pkg/user/jwt_manager.go` — `StartCleanupLoop`
- **Interval:** Configurable via `auth.jwt.secret_retention.cleanup_interval` (default: 1h)
- **Behavior:** Removes secrets older than retention period (TTL x retention_factor, capped at max_retention)
- **Safety:** Never removes the current primary secret
### Magic-link expired tokens (ADR-0028 Phase A)
- **Location:** `pkg/user/magic_link_cleanup.go` — `StartCleanupLoop`
- **Interval:** Configurable via `auth.magic_link.cleanup_interval` (default: 1h)
- **Behavior:** Deletes tokens where `expires_at < now`
- **Implementation:** Calls `DeleteExpiredMagicLinkTokens` on the repository
## Local dev setup
1. **Start services:**
```bash
docker compose up -d # starts Postgres + Mailpit
```
2. **Inspect emails:** http://localhost:8025 (Mailpit UI)
3. **HTTPS for OIDC (Phase B):**
```bash
make cert # generates certs/dev-cert.pem + certs/dev-key.pem via mkcert
```
See [MKCERT.md](MKCERT.md) for details.
## Cross-references
### Architecture Decision Records
| ADR | Description |
|-----|-------------|
| [ADR-0018](../adr/0018-user-management-auth-system.md) | Original username/password auth system |
| [ADR-0021](../adr/0021-jwt-secret-retention-policy.md) | JWT secret retention and cleanup |
| [ADR-0028](../adr/0028-passwordless-auth-migration.md) | Passwordless migration (Phase A complete, Phase B in progress) |
| [ADR-0029](../adr/0029-email-infrastructure-mailpit.md) | Email infrastructure (Mailpit) |
| [ADR-0030](../adr/0030-bdd-email-parallel-strategy.md) | BDD parallel email assertions |
### Documentation
| Document | Description |
|----------|-------------|
| [EMAIL.md](EMAIL.md) | SMTP setup and Mailpit usage |
| [MKCERT.md](MKCERT.md) | Local HTTPS certificate setup |
| [PHASE_B_ROADMAP.md](PHASE_B_ROADMAP.md) | Remaining OIDC work |
---
*Developer onboarding doc — see ADR-0028 for implementation details.*

89
documentation/BDD_TEST_ENV.md
View File

@@ -1,89 +0,0 @@
# BDD test environment
Environment variables and tooling specific to running BDD scenarios locally and in CI. Companion to [BDD_GUIDE.md](BDD_GUIDE.md) (which covers the BDD authoring workflow itself).
## Required env vars (database connection)
The BDD test server needs a Postgres instance reachable via:
| Var | Default | Notes |
|---|---|---|
| `DLC_DATABASE_HOST` | `localhost` | Host of the Postgres instance |
| `DLC_DATABASE_PORT` | `5432` | |
| `DLC_DATABASE_USER` | `postgres` | Test-only credentials (NOT production) |
| `DLC_DATABASE_PASSWORD` | `postgres` | |
| `DLC_DATABASE_NAME` | `dance_lessons_coach_bdd_test` | Dedicated test DB |
| `DLC_DATABASE_SSL_MODE` | `disable` | Tests run without TLS |
Local setup:
```bash
docker compose up -d # Postgres container
docker exec dance-lessons-coach-postgres psql -U postgres \
-c "CREATE DATABASE dance_lessons_coach_bdd_test;" # one-time
```
In CI: `.gitea/workflows/ci-cd.yaml` provisions a Postgres service container and exports the same vars.
## Optional env vars
### `BDD_SCHEMA_ISOLATION` (since [PR #35](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/35) — T12 stage 2/2)
| Value | Behaviour |
|---|---|
| `true` | Each test PACKAGE (process) gets its own isolated PostgreSQL schema with migrations. Packages run in **parallel** safely. **~2.85x speedup observed locally.** This is the new default in CI. |
| (unset / `false`) | Falls back to single shared `public` schema with `CleanupDatabase` (TRUNCATE) between scenarios. Forces sequential package execution (`-p 1`). Slower but simpler. |
Implementation: `pkg/bdd/testserver/server.go Start()` builds a per-package isolated repo via `user.NewPostgresRepositoryFromDSN` (PR #34). `Stop()` drops the schema + closes the per-package pool.
ADR-0025 documents the isolation strategy ("Implemented" since PR #35).
### `FEATURE` (per-package selector)
When set, `pkg/bdd/testserver/server.go shouldEnableV2()` reads it. Used to scope per-feature behaviour (e.g. enable v2 endpoints only when `FEATURE=greet` AND `GODOG_TAGS` includes `@v2`).
Without `FEATURE` set, falls back to `bdd` (generic).
### `GODOG_TAGS` (scenario filter)
Standard godog env var. The default suite excludes flaky/todo/skip/v2 tags:
```
GODOG_TAGS="~@flaky && ~@todo && ~@skip && ~@v2"
```
Scoped runs (e.g. `@critical` only): set `GODOG_TAGS="@critical"` and run.
### `BDD_ENABLE_CLEANUP_LOGS` (debug)
Set `=true` to log each scenario's CLEANUP / ISOLATION operation. Useful when debugging flakiness.
## Recommended local commands
Run all BDD with isolation (parallel, fast):
```bash
DLC_DATABASE_HOST=localhost DLC_DATABASE_PORT=5432 \
DLC_DATABASE_USER=postgres DLC_DATABASE_PASSWORD=postgres \
DLC_DATABASE_NAME=dance_lessons_coach_bdd_test DLC_DATABASE_SSL_MODE=disable \
BDD_SCHEMA_ISOLATION=true \
go test ./features/...
```
Run one feature with v2 enabled:
```bash
DLC_DATABASE_HOST=... \
BDD_SCHEMA_ISOLATION=true FEATURE=greet GODOG_TAGS="@v2" \
go test ./features/greet/...
```
Repro CI conditions (sequential, no isolation):
```bash
DLC_DATABASE_HOST=... \
go test ./features/... -p 1
```
## Cross-references
- [BDD_GUIDE.md](BDD_GUIDE.md) — authoring scenarios + steps
- [ADR-0008](../adr/0008-bdd-testing.md) — choice of Godog
- [ADR-0024](../adr/0024-bdd-test-organization-and-isolation.md) — feature directory organization
- [ADR-0025](../adr/0025-bdd-scenario-isolation-strategies.md) — isolation strategies (Implemented since PR #35)

251
documentation/CLI.md
View File

@@ -1,251 +0,0 @@
# CLI Management Guide
Complete reference for the `dance-lessons-coach` CLI, server lifecycle, and configuration. Extracted from the original `AGENTS.md` (Tâche 6 restructure) for lazy-loading compatibility with Mistral Vibe.
## Cobra CLI (Recommended)
`dance-lessons-coach` includes a modern CLI built with Cobra:
```bash
# Show help and available commands
./bin/dance-lessons-coach --help
# Show version information
./bin/dance-lessons-coach version
# Greet someone by name
./bin/dance-lessons-coach greet John
# Start the server
./bin/dance-lessons-coach server
```
**Available Commands:**
- `version` — Print version information
- `server` — Start the dance-lessons-coach server
- `greet [name]` — Greet someone by name
- `help` — Built-in help system
- `completion` — Generate shell completion scripts
**Server Command Flags:**
- `--config` — Config file path
- `--env` — Environment (`dev`, `staging`, `prod`)
- `--debug` — Enable debug logging
## Version Information
The server provides runtime version information:
```bash
# Check version using new CLI
./bin/dance-lessons-coach version
# Check version using server binary
./bin/server --version
# Output:
dance-lessons-coach Version Information:
Version: 1.0.0
Commit: abc1234
Built: 2026-04-05T10:00:00+0000
Go: go1.26.1
```
For full version management workflow (bump, release, build with version), see [`version-management-guide.md`](version-management-guide.md).
## Server Control Script
A shell script manages the server lifecycle:
```bash
cd /Users/gabrielradureau/Work/Vibe/DanceLessonsCoach
./scripts/start-server.sh start # Start the server
./scripts/start-server.sh status # Check server status
./scripts/start-server.sh test # Test API endpoints
./scripts/start-server.sh logs # View server logs
./scripts/start-server.sh stop # Stop the server
./scripts/start-server.sh restart # Restart
```
**Available subcommands:**
- `start` — Start the server in background with proper logging
- `stop` — Stop the server gracefully
- `restart` — Restart the server
- `status` — Check if server is running
- `logs` — Show recent server logs
- `test` — Test all API endpoints
## Manual Server Management
For direct control:
```bash
cd /Users/gabrielradureau/Work/Vibe/DanceLessonsCoach
./scripts/start-server.sh start
```
**Expected output:**
```
Server running on :8080
[INF] Starting HTTP server on :8080
[TRC] Registering greet routes
[TRC] Greet routes registered
```
**Features:**
- Context-aware server initialization
- Graceful shutdown handling
- Signal-based termination (`SIGINT`, `SIGTERM`)
- 30-second shutdown timeout
- Proper resource cleanup
## Configuration
Configuration via environment variables with `DLC_` prefix:
| Option | Environment Variable | Default | Description |
|---|---|---|---|
| Host | `DLC_SERVER_HOST` | `0.0.0.0` | Server bind address |
| Port | `DLC_SERVER_PORT` | `8080` | Server listening port |
| Shutdown Timeout | `DLC_SHUTDOWN_TIMEOUT` | `30s` | Graceful shutdown timeout |
| JSON Logging | `DLC_LOGGING_JSON` | `false` | Enable JSON format logging |
| Log Output | `DLC_LOGGING_OUTPUT` | `""` | Log output file path (empty for stderr) |
**Examples:**
```bash
# Custom port
export DLC_SERVER_PORT=9090
./scripts/start-server.sh start
# Custom host and port
export DLC_SERVER_HOST="127.0.0.1"
export DLC_SERVER_PORT=8081
./scripts/start-server.sh start
# Custom shutdown timeout
export DLC_SHUTDOWN_TIMEOUT=45s
# Enable JSON logging
export DLC_LOGGING_JSON=true
# Log to file
export DLC_LOGGING_OUTPUT="server.log"
# Combined: JSON logging to file
export DLC_LOGGING_JSON=true
export DLC_LOGGING_OUTPUT="server.json.log"
```
**Configuration File Support:**
A `config.example.yaml` file is provided as a template. By default, the application looks for `config.yaml` in the current working directory.
To specify a custom config file path, set the `DLC_CONFIG_FILE` environment variable:
```bash
DLC_CONFIG_FILE="/path/to/config.yaml" go run ./cmd/server
```
Example `config.yaml`:
```yaml
server:
host: "0.0.0.0"
port: 8080
shutdown:
timeout: 30s
logging:
json: false
```
**Configuration Loading Precedence:**
1. **File-based configuration** (highest precedence)
2. **Environment variables** (override defaults, overridden by config file)
3. **Default values** (fallback)
All configuration is validated on startup. Invalid configurations cause server startup failure. Configuration values and source are logged at startup.
**Verification:**
```bash
DLC_SERVER_PORT=9090 DLC_SERVER_HOST="127.0.0.1" ./scripts/start-server.sh start
curl http://127.0.0.1:9090/api/health
# Expected: {"status":"healthy"}
```
## Server Status
```bash
# Check health endpoint
curl -s http://localhost:8080/api/health
# Check readiness endpoint
curl -s http://localhost:8080/api/ready
```
**Expected responses:**
- Health: `{"status":"healthy"}`
- Readiness (normal): `{"ready":true}`
- Readiness (during shutdown): `{"ready":false}` (HTTP 503)
**Endpoint Differences:**
- **Health endpoint** (`/api/health`): Indicates if the application is running and functional
- **Readiness endpoint** (`/api/ready`): Indicates if the application is ready to accept traffic
**Use Cases:**
- **Health**: Used by load balancers to check if the app is alive
- **Readiness**: Used by Kubernetes / service meshes to determine if the app can accept new requests
**During Graceful Shutdown:**
- Health endpoint continues to return `{"status":"healthy"}`
- Readiness endpoint returns `{"ready":false}` with HTTP 503 Service Unavailable
- This allows existing requests to complete while preventing new requests
## Stopping the Server
To stop the server gracefully:
```bash
# Send SIGTERM for graceful shutdown
kill -TERM $(lsof -ti :8080)
# Or send SIGINT (Ctrl+C equivalent)
pkill -INT -f "go run"
```
**Graceful shutdown process:**
1. Server receives termination signal
2. Logs shutdown message
3. Stops accepting new connections
4. Waits up to 30 seconds for active requests to complete
5. Closes all connections cleanly
6. Exits with proper cleanup
For force stop (if graceful shutdown hangs):
```bash
kill -9 $(lsof -ti :8080)
```
**Verification:**
```bash
curl -s http://localhost:8080/api/health
# Should return connection refused
```

59
documentation/CODE_EXAMPLES.md
View File

@@ -1,59 +0,0 @@
# Code Examples
Snippets and patterns used across the `dance-lessons-coach` codebase. Extracted from the original `AGENTS.md` (Tâche 6 restructure).
## Adding a New API Endpoint
```go
// 1. Add to interface
func (h *apiV1GreetHandler) RegisterRoutes(router chi.Router) {
router.Get("/", h.handleGreetQuery)
router.Get("/{name}", h.handleGreetPath)
router.Post("/custom", h.handleCustomGreet) // New endpoint
}
// 2. Implement handler
func (h *apiV1GreetHandler) handleCustomGreet(w http.ResponseWriter, r *http.Request) {
// Parse request
// Call service
// Return JSON response
}
```
## Logging with Zerolog
```go
// Trace level logging
log.Trace().Ctx(ctx).Str("key", "value").Msg("message")
// Info level
log.Info().Msg("Important event")
// Error level
log.Error().Err(err).Msg("Error occurred")
```
For the full logging strategy (when to use Trace vs Info, performance considerations), see [ADR-0003 — Zerolog Logging](../adr/0003-zerolog-logging.md).
## Using `context.Context`
```go
// Pass context through calls
func handler(w http.ResponseWriter, r *http.Request) {
result := service.Greet(r.Context(), "John")
// ...
}
// Create context with values
ctx := context.WithValue(r.Context(), "key", "value")
// Create context with timeout
ctx, cancel := context.WithTimeout(r.Context(), 5*time.Second)
defer cancel()
```
For the rationale behind context-aware services, see [ADR-0004 — Interface-Based Design](../adr/0004-interface-based-design.md).
## Best Practices Reminders
For higher-level guidance on code organization, error handling, performance, and testing, see [`AGENT_USAGE_GUIDE.md`](AGENT_USAGE_GUIDE.md#best-practices) section "Best Practices".

107
documentation/EMAIL.md
View File

@@ -1,107 +0,0 @@
# Email infrastructure
Outgoing email transport. Per [ADR-0029](../adr/0029-email-infrastructure-mailpit.md): Mailpit for local dev + BDD tests, production sender deferred.
## Local setup (one-time)
Mailpit is part of `docker-compose.yml`:
```bash
docker compose up -d # starts postgres + mailpit
docker compose ps # confirm both running
```
Mailpit listens on:
- **SMTP submission** — `localhost:1025` (the app sends here)
- **HTTP UI / API** — http://localhost:8025 (you inspect captured messages here)
No real emails leave the docker network. No internet required.
## Application configuration
The application's outgoing transport is configured under `auth.email.*` in `config.yaml` (or via `DLC_AUTH_EMAIL_*` env vars). Defaults already match local Mailpit:
```yaml
auth:
email:
from: noreply@dance-lessons-coach.local
smtp_host: localhost
smtp_port: 1025
smtp_use_tls: false
timeout: 10s
# smtp_username + smtp_password left empty for local Mailpit
```
For production, override these to point at the chosen provider (SES, Postmark, etc.).
## Inspecting messages
### Web UI
http://localhost:8025 — list of all captured messages, search, raw view, HTML preview.
### HTTP API (for automation)
```bash
# Latest 10 messages (no filter — /api/v1/messages is for pagination)
curl -s 'http://localhost:8025/api/v1/messages?limit=10' | jq
# Messages for a specific recipient — use /api/v1/search, NOT /messages
# (the latter's `query` param is for pagination only, not filtering ;
# verified empirically 2026-05-05)
curl -s 'http://localhost:8025/api/v1/search?query=to:test-user@bdd.local' | jq
# Get a specific message by ID (full content, headers, attachments)
curl -s 'http://localhost:8025/api/v1/message/<id>' | jq
# Purge messages for a recipient (used in test cleanup) — also via /search
curl -X DELETE 'http://localhost:8025/api/v1/search?query=to:test-user@bdd.local'
```
Full API: https://mailpit.axllent.org/docs/api-v1/
## Sending email from Go code
```go
import "dance-lessons-coach/pkg/email"
sender := email.NewSMTPSender(email.SMTPConfig{
Host: cfg.GetEmailConfig().SMTPHost,
Port: cfg.GetEmailConfig().SMTPPort,
// username/password optional — empty means no AUTH (Mailpit local)
})
err := sender.Send(ctx, email.Message{
To: "alice@example.com",
From: cfg.GetEmailConfig().From,
Subject: "Your magic link",
BodyText: "Click: https://example.com/magic-link/consume?token=...",
Headers: map[string]string{
// optional — useful for BDD test correlation
"X-Trace-Id": "req-abc-123",
},
})
```
Or, when both text and HTML are needed (`multipart/alternative`):
```go
err := sender.Send(ctx, email.Message{
To: "alice@example.com", From: "...", Subject: "...",
BodyText: "Click: https://...",
BodyHTML: `<p>Click <a href="https://...">your magic link</a></p>`,
})
```
## Production sender (TBD)
Not chosen yet. When ready, implement another `email.Sender` in
`pkg/email/<provider>_sender.go` and wire it via the config. The
`Sender` interface is the swap point — call sites don't change.
## Cross-references
- [ADR-0028 — Passwordless auth migration](../adr/0028-passwordless-auth-migration.md) (consumes this infrastructure)
- [ADR-0029 — Email infrastructure decision](../adr/0029-email-infrastructure-mailpit.md)
- [ADR-0030 — BDD email parallel strategy](../adr/0030-bdd-email-parallel-strategy.md)
- [Mailpit docs](https://mailpit.axllent.org/docs/)

83
documentation/HISTORY.md
View File

@@ -1,83 +0,0 @@
# Development History
This document records the historical development phases of `dance-lessons-coach`. Extracted from the original `AGENTS.md` (Tâche 6 restructure) for lazy-loading compatibility with Mistral Vibe (128k context).
All phases below are **completed** ✅. They are kept here for traceability and onboarding context — refer to ADRs (`adr/`) for the technical decisions behind each phase.
## Phase 1: Foundation
- Go 1.26.1 environment setup
- Project structure with `cmd/` and `pkg/` directories
- Core Greet service implementation
- CLI interface
- Unit tests
## Phase 2: Web API
- Chi router integration
- Versioned API endpoints (`/api/v1`)
- Health endpoint (`/api/health`)
- JSON responses with proper headers
## Phase 3: Logging & Architecture
- Zerolog integration with Trace level
- Context-aware logging
- Interface-based design patterns
- Dependency injection
## Phase 4: Documentation & Testing
- Comprehensive `AGENTS.md`
- `README.md` with usage instructions
- Server management guide
- API endpoint documentation
## Phase 5: Configuration Management
- Viper integration for configuration
- Environment variable support with `DLC_` prefix
- Customizable server host/port
- Configurable shutdown timeout
- Configuration validation and logging
- Example configuration file
## Phase 6: Graceful Shutdown
- Context-aware server initialization
- Signal-based termination (`SIGINT`, `SIGTERM`)
- Configurable shutdown timeout
- Readiness endpoint for Kubernetes/service mesh integration
- Proper resource cleanup during shutdown
- Health endpoint remains healthy during graceful shutdown
## Phase 7: OpenTelemetry Integration
- OpenTelemetry Go libraries integration
- Jaeger compatibility for distributed tracing
- Middleware-only approach using `otelhttp.NewHandler`
- Configurable sampling strategies
- Graceful shutdown of tracer provider
- OTLP exporter with gRPC support
## Phase 8: Build System & Documentation
- Build script for binary compilation
- Binary output to `bin/` directory
- Comprehensive commit conventions with gitmoji reference
- Updated documentation with Jaeger integration guide
- Cleaned up configuration files
- Enhanced logging configuration with file output support
## Phase 9: Final Refinements
- Removed unnecessary `time.Sleep` for log flushing
- Changed server operational logs from Info to Trace level
- Moved all logging setup logic to config package
- Simplified server entrypoint to 27 lines
- Verified all functionality with comprehensive testing
- Updated documentation to reflect final architecture
## Beyond Phase 9
Subsequent work (CI/CD, BDD scenarios, ADR audit, JWT, config hot-reloading) is tracked in the [Changelog](../CHANGELOG.md) and the corresponding [ADRs](../adr/).

219
documentation/MISTRAL-AUTONOMOUS-PATTERN.md
View File

@@ -1,219 +0,0 @@
# Mistral Vibe Autonomous Pattern
**Document ID:** MISTRAL-AUTONOMOUS-PATTERN
**Date:** 2026-05-05
**Status:** Active
**Author:** Mistral Vibe (batch10-task-mistral-pattern-doc)
**Audience:** Project contributors, future trainers
---
## 1. What you'll see
PRs authored by "Gabriel Radureau" with commit messages ending in "Mistral Vibe" references. PR titles start with gitmoji. Branch names follow `vibe/<slug>` pattern.
| PR | Date | Title | Branch | Status |
|----|------|-------|--------|--------|
| #67 | 2026-05-05 | :memo: docs: email infrastructure | `vibe/batch4-task-a-email-infra` | Merged |
| #74 | 2026-05-05 | :robot: feat: BDD Mailpit helper | `vibe/batch5-task-b-bdd-mailpit` | Merged |
| #75 | 2026-05-05 | :elephant: feat: magic_link_tokens table | `vibe/batch5-task-c-db-magic-link` | Merged |
| #76 | 2026-05-05 | :rocket: feat: magic link handlers | `vibe/batch5-task-d-handlers` | Merged |
| #77 | 2026-05-05 | :test_tube: test: magic link BDD | `vibe/batch5-task-e-bdd` | Merged |
---
## 2. The pattern (high-level)
```
Operator Brief → Worktree Setup → Worker Execution → PR Lifecycle → Merge
```
### 2.1 Operator brief
Human or trainer (Claude) writes a `CONTEXT.md` brief in a workspace under `~/Work/Vibe/workspaces/<slug>/`. The brief contains:
- Mission statement
- Goal and constraints
- Process instructions
- Hard rules
- Specification
### 2.2 Worktree setup
A `vibe-workspace.sh --worktree` script creates an isolated git worktree:
- Branches from current `origin/main`
- Creates branch `vibe/<slug>`
- Isolates git state in a dedicated directory
- No race conditions (addresses Q-038)
### 2.3 Worker execution
A Mistral Vibe worker (`vibe -p`) runs end-to-end:
1. Reads the brief from `CONTEXT.md`
2. Executes coding tasks (codes, builds, tests)
3. Commits changes with appropriate messages
4. Pushes to remote branch
5. Opens PR via Gitea API
6. Attempts auto-merge
### 2.4 Parallel operation
- Multiple workers run concurrently (2-4 typical)
- Each worker operates in its own worktree
- No git checkout collisions
- Shared origin main as base
### 2.5 Dispatch orchestration
A `dispatch-batch.sh` script:
- Orchestrates batches of 2-4 workers
- Auto-merges PRs that workers opened but didn't merge
- Ensures all PRs reach merged state
- Handles cross-worker dependencies
---
## 3. Why this works
### 3.1 Worktree isolation
Git worktrees provide complete isolation of git state. Each worker has its own:
- Working directory
- Index (staging area)
- HEAD pointer
- Branch reference
This eliminates race conditions documented in Q-038 of the harness logs.
### 3.2 Pre-fetched origin
Origin is pre-fetched before worktree creation (Q-060 fix). This guarantees:
- All workers branch from current main
- No stale base branches
- Consistent starting point across batch
### 3.3 Full PR lifecycle
Workers handle the complete PR lifecycle:
- Code implementation
- Build and test execution
- Commit with proper conventions
- Push to remote
- PR creation via Gitea API
- Merge via Gitea API (squash merge default)
### 3.4 Trainer takeover
For the rare gaps (~5% of cases):
- Price-cap restrictions
- Broken Mistral tests
- Ambiguous requirements
Trainer (Claude) takeover within ~5 minutes covers these edge cases.
---
## 4. How to read PR provenance
### 4.1 Commit message markers
Look for these patterns in commit messages:
| Marker | Meaning |
|--------|---------|
| `Mostly Mistral Vibe authored` | Mixed human + AI authorship |
| `100% Mistral autonomous` | Fully autonomous workflow |
| `batch<N>-task-<X>` | Brief slug reference |
| `Q-058 trainer takeover` | Specific quirk reference |
| `Q-062 fix applied` | Quirk mitigation applied |
### 4.2 Branch naming
Branch names encode the workflow:
```
vibe/<batch>-<task>-<description>
```
Examples:
- `vibe/batch4-task-a-email-infra`
- `vibe/batch10-task-mistral-pattern-doc`
### 4.3 PR title conventions
PR titles use gitmoji prefix:
- `:memo:` - Documentation
- `:robot:` - AI/automation
- `:elephant:` - Database
- `:rocket:` - Feature
- `:test_tube:` - Testing
---
## 5. Reproducing the pattern
### 5.1 Quickstart guide
See `~/.vibe/scripts/QUICKSTART-DISPATCH-BATCH.md` for complete how-to guide.
### 5.2 Resources
| Resource | Path | Description |
|----------|------|-------------|
| Brief template | `~/.vibe/skills/prompt-builder/examples/dispatch-batch-task.md` | Standardized brief format |
| Mistral quirks | `~/.vibe/memory/reference/mistral-quirks.md` | Accumulated lessons (Q-001 through Q-063 as of 2026-05-05) |
| Architecture doc | `~/.vibe/memory/reference/architecture-mapreduce-orchestration.md` | Design rationale |
| Budget history | `~/.vibe/memory/reference/budget-history.jsonl` | Empirical cost data |
---
## 6. Numbers (2026-05-05 reference)
### 6.1 Throughput
| Metric | Value | Notes |
|--------|-------|-------|
| PRs merged (one day) | 20 | Mistral autonomous |
| Wall-clock parallel (2 PRs) | ~2 minutes | vs ~3-4 minutes serial |
| Wall-clock parallel (4 PRs) | ~2-3 minutes | Batch efficiency |
### 6.2 Cost
| PR Type | Cost Range | Notes |
|---------|------------|-------|
| Simple PR | $0.5-1.5 | Documentation, minor changes |
| Code-heavy PR | $2-3 | Complex logic, multiple files |
| Complex PR | $3-5 | Architecture changes, deep refactoring |
### 6.3 Autonomy rate
| Metric | Value |
|--------|-------|
| Autonomy rate per batch | 95-100% |
| Trainer takeover rate | 5% |
| Takeover reasons | Price-cap (2%), broken tests (2%), ambiguity (1%) |
---
## 7. Future evolution
### 7.1 Phase 1bis (current)
- Multi-process workers operating in parallel
- Claude trainer reduces observations
- Improves harness reliability
- Current state as of 2026-05-05
### 7.2 Phase 2 (target)
- Mistral meta-agent performs reduce phase
- Full autonomy loop without Claude
- Self-improving pattern
- Target: Q3 2026
### 7.3 Long-term vision
- Fully autonomous feature development
- Self-healing test failures
- Cost-optimized batch dispatch
- Multi-repository orchestration
---
## 8. Cross-references
### 8.1 Related ADRs
| ADR | Description |
|-----|-------------|
| [ADR-0001](../adr/0001-go-1.26.1-standard.md) | Go 1.26.1 standard |
| [ADR-0008](../adr/0008-bdd-testing.md) | BDD with Godog |
| [ADR-0028](../adr/0028-passwordless-auth-migration.md) | Passwordless auth (Phase A complete) |
### 8.2 Related documentation
| Document | Description |
|----------|-------------|
| [CONTRIBUTING.md](../CONTRIBUTING.md) | Contribution guidelines |
| [AGENTS.md](../AGENTS.md) | Agent documentation |
| [PHASE_B_ROADMAP.md](PHASE_B_ROADMAP.md) | Phase B OIDC roadmap |
---
*Developer onboarding doc — see QUICKSTART-DISPATCH-BATCH.md for implementation details.*

120
documentation/MKCERT.md
View File

@@ -1,120 +0,0 @@
# mkcert: Local HTTPS for Development
## Overview
This document describes how to set up local HTTPS development certificates using `mkcert`.
OIDC providers **reject `http://localhost` as a redirect URI** by default for security reasons. To test OAuth 2.0 / OpenID Connect flows locally, the development server must be accessible via HTTPS. `mkcert` provides a zero-configuration local Certificate Authority that generates trusted certificates for localhost and custom domains.
This setup is a prerequisite for **ADR-0028 Phase B** (OpenID Connect Authorization Code flow).
## Why mkcert
- **Trusted locally**: Certificates are automatically trusted by the system root store (macOS, Linux, Windows)
- **No configuration**: Single commands to create and install the CA
- **Local-only**: Certificates are valid only for localhost development, never exposed to production
- **Industry standard**: Widely adopted tool for local HTTPS development
## Installation
### macOS (Homebrew)
```bash
brew install mkcert
mkcert -install
```
The `mkcert -install` command creates and installs a local Certificate Authority in your system trust store.
### Linux
See [mkcert GitHub](https://github.com/FiloSottile/mkcert#installation) for distribution-specific instructions.
### Windows
See [mkcert GitHub](https://github.com/FiloSottile/mkcert#installation) for Windows installation.
## Generate Certificates
Use the provided Make target to generate certificates for localhost development:
```bash
make cert
```
This runs the following command:
```bash
mkcert -cert-file ./certs/dev-cert.pem -key-file ./certs/dev-key.pem localhost 127.0.0.1 ::1
```
### Output Files
| File | Description |
|------|-------------|
| `./certs/dev-cert.pem` | TLS certificate for localhost, 127.0.0.1, and ::1 |
| `./certs/dev-key.pem` | Private key for the certificate |
Both files are created in the `./certs/` directory at the project root.
## Use in Development
Once certificates are generated, start the server with TLS enabled:
```bash
./bin/server --tls-cert ./certs/dev-cert.pem --tls-key ./certs/dev-key.pem
```
> **Note**: The `--tls-cert` and `--tls-key` flags are **not yet implemented** — this is planned for ADR-0028 Phase B.4. The Makefile and certificate generation are prepared in advance so that when the server TLS support is added, the certificates are ready.
The server will then be accessible at:
- `https://localhost:8080` (or the configured port)
- `https://127.0.0.1:8080`
- `https://[::1]:8080`
All OIDC callback URLs must use HTTPS with one of these hostnames.
## Clean Up
To remove generated certificates:
```bash
make clean-cert
```
This deletes the entire `./certs/` directory.
## .gitignore
The `certs/` directory contains locally-generated certificates and **must not be committed** to version control.
Ensure `certs/` is in your `.gitignore`. If it is not already present, add it:
```bash
echo "certs/" >> .gitignore
```
## Cross-References
- [ADR-0028: Passwordless authentication: magic link → OpenID Connect](../adr/0028-passwordless-auth-migration.md) — Phase B describes the OIDC implementation that requires HTTPS
- [mkcert GitHub Repository](https://github.com/FiloSottile/mkcert) — Official documentation
## Troubleshooting
### "mkcert not found" when running `make cert`
Ensure `mkcert` is installed and available in your `PATH`. The Makefile checks for this and will display an error message if `mkcert` is not found.
### Certificate not trusted by browser
Run `mkcert -install` again. On macOS, you may need to restart your browser completely (close all windows, not just tabs).
### Port already in use
If another process is using the port (e.g., a non-TLS server on port 8080), stop that process first or configure the server to use a different port.
## See Also
- `make help` — List all available Make targets
- [documentation/API.md](API.md) — API endpoints reference
- [documentation/BDD_GUIDE.md](BDD_GUIDE.md) — BDD testing guide

94
documentation/OBSERVABILITY.md
View File

@@ -1,94 +0,0 @@
# Observability — OpenTelemetry & Jaeger Integration
Tracing setup for `dance-lessons-coach`. Extracted from the original `AGENTS.md` (Tâche 6 restructure) for lazy-loading compatibility with Mistral Vibe.
The application supports OpenTelemetry for distributed tracing with Jaeger compatibility.
## Configuration
Enable OpenTelemetry in your `config.yaml`:
```yaml
telemetry:
enabled: true
otlp_endpoint: "localhost:4317"
service_name: "dance-lessons-coach"
insecure: true
sampler:
type: "parentbased_always_on"
ratio: 1.0
```
Or via environment variables:
```bash
export DLC_TELEMETRY_ENABLED=true
export DLC_TELEMETRY_OTLP_ENDPOINT="localhost:4317"
export DLC_TELEMETRY_SERVICE_NAME="dance-lessons-coach"
export DLC_TELEMETRY_INSECURE=true
export DLC_TELEMETRY_SAMPLER_TYPE="parentbased_always_on"
export DLC_TELEMETRY_SAMPLER_RATIO=1.0
```
## Testing with Jaeger
**1. Start Jaeger in Docker:**
```bash
docker run -d --name jaeger \
-e COLLECTOR_OTLP_ENABLED=true \
-p 16686:16686 \
-p 4317:4317 \
jaegertracing/all-in-one:latest
```
**2. Start the server with OpenTelemetry enabled:**
```bash
# Using config file
./scripts/start-server.sh start
# Or with environment variables
DLC_TELEMETRY_ENABLED=true ./scripts/start-server.sh start
```
**3. Make API requests:**
```bash
curl http://localhost:8080/api/v1/greet/John
```
**4. View traces in Jaeger UI:**
Open http://localhost:16686 and select the `dance-lessons-coach` service.
## Sampler Types
| Sampler | Behavior |
|---|---|
| `always_on` | Sample all traces |
| `always_off` | Sample no traces |
| `traceidratio` | Sample based on trace ID ratio |
| `parentbased_always_on` | Sample based on parent span (always on) |
| `parentbased_always_off` | Sample based on parent span (always off) |
| `parentbased_traceidratio` | Sample based on parent span with ratio |
## Testing Script
A convenience script is provided:
```bash
./scripts/test-opentelemetry.sh
```
This script:
1. Starts Jaeger container
2. Starts the server with OpenTelemetry
3. Makes test API calls
4. Shows Jaeger UI URL
5. Cleans up on exit
## ADR Reference
See [ADR-0007 — OpenTelemetry Integration](../adr/0007-opentelemetry-integration.md) for the full architectural decision and rationale (middleware-only approach, sampling strategy, OTLP/gRPC choice).

145
documentation/PHASE_B_ROADMAP.md
View File

@@ -1,145 +0,0 @@
# ADR-0028 Phase B Roadmap
**Document ID:** PHASE_B_ROADMAP
**Date:** 2026-05-05 evening
**Status:** In Progress
**Author:** AI Agent (vibe/batch4-task-b-phase-b-roadmap)
---
## Status as of 2026-05-05 evening
- [x] ADR-0028 Phase A complete (PRs #59-#63, #65)
- [x] Phase B.1 OIDC config (PR #64)
- [x] Phase B prep : pkg/auth skeleton (PR #69) + mkcert doc (PR #68)
- [x] Phase B.3 OIDC client implementation : ✅ (PR #74)
- [x] Phase B.4 OIDC HTTP handlers + tests : ✅ (PR #75 + PR #76 follow-up tests)
## Status as of 2026-05-05 evening (after autonomous Mistral session)
Phase B is essentially complete except B.5. The OIDC client (B.3, PR #74), HTTP handlers and tests (B.4, PR #75 + PR #76) have been delivered and merged.
---
## Remaining work
Phase B delivers OpenID Connect Authorization Code flow with PKCE. Work is organized into **3 shippable phases**, each deliverable as an independent PR. At the time of this update, only Phase B.5 (BDD scenarios) remains to be completed.
### Phase B.3 — OIDC client implementation
- **Goal:** Implement the core OIDC client methods in `pkg/auth/oidc.go`
- **Tasks:**
- `Discover()`: HTTP GET to `/.well-known/openid-configuration`, parse + cache discovery document
- `RefreshJWKS()`: HTTP GET to JWKS URI, parse RSA public keys, cache with TTL
- `ExchangeCode()`: POST to token endpoint with code + PKCE verifier, return TokenResponse
- `ValidateIDToken()`: Verify signature against JWKS, validate standard claims (iss, aud, exp, iat)
- **LOE:** ~200 lines of Go + unit tests
- **Dependencies:** None (uses standard library `crypto/rsa`, `encoding/jwt`)
- **Deliverable:** 1 PR
### Phase B.4 — OIDC HTTP handlers
- **Goal:** Add OIDC flow endpoints and wire them into the server
- **Tasks:**
- Create `pkg/user/api/oidc_handler.go`
- `GET /api/v1/auth/oidc/start`:
- Generate state (CSRF protection) + PKCE verifier + challenge
- Store state + verifier (cookie or short-lived in-memory store)
- Redirect to provider's authorization endpoint
- `GET /api/v1/auth/oidc/callback`:
- Validate state parameter matches stored state
- Exchange code for tokens (calls B.3 client)
- Validate id_token (calls B.3 client)
- Issue internal JWT (reuse existing JWT manager from ADR-0021)
- Return JWT in Set-Cookie + JSON body
- Wire routes in `pkg/server/server.go`
- **LOE:** ~150 lines of Go + unit tests + integration tests
- **Dependencies:** B.3 (client methods must be implemented)
- **Prerequisite:** Run `make cert` (mkcert, from PR #68) before starting dev
- **Deliverable:** 1 PR
### Phase B.5 — BDD coverage
- **Goal:** End-to-end OIDC testing
- **Tasks:**
- Create `features/auth/oidc.feature` with scenarios:
- Happy path: start → provider auth → callback → JWT issued
- Error: state mismatch
- Error: invalid code
- Error: expired id_token
- Use mock OIDC provider (local in-process) OR deterministic test against Authelia/Keycloak in docker-compose
- Follow ADR-0030 parallel BDD strategy for email assertions
- **LOE:** ~150 lines of Gherkin + step definitions
- **Dependencies:** B.3 + B.4 (endpoints must be operational)
- **Deliverable:** 1 PR
---
## Dependencies and order
```
B.3 (OIDC client)
↓
B.4 (HTTP handlers) —— requires B.3
↓
B.5 (BDD coverage) —— requires B.3 + B.4
```
**Note:** mkcert (PR #68) is ready. When starting B.4 development, run `make cert` once to generate local HTTPS certificates.
---
## Out of scope for Phase B (deferred)
| Item | Target Phase | Rationale |
|------|--------------|-----------|
| Decommission password auth | Phase C | Separate ADR after B is in production |
| Multi-provider (Authelia + Google) | Phase B.6 (if needed) | Single provider sufficient for MVP |
| JWKS rotation mid-flight retry | B.3 enhancement | Handle in initial implementation |
| Token refresh flow | Future | Not required for auth code flow MVP |
---
## Risk register
| Risk | Mitigation | Owner |
|------|------------|-------|
| JWKS rotation handling | Implement refresh + retry logic; key rotation must not break mid-flight validation | B.3 implementer |
| PKCE storage | Use signed cookie or short-lived in-memory store; document trade-offs in implementation PR | B.4 implementer |
| Testing without real provider | Use mock OIDC server for CI; local dev uses Authelia in docker-compose | B.5 implementer |
| State CSRF protection | Use cryptographically random state; store server-side with short TTL | B.4 implementer |
---
## Cross-references
- [ADR-0028: Passwordless authentication: magic link → OpenID Connect](../adr/0028-passwordless-auth-migration.md)
- [ADR-0029: Email infrastructure (Mailpit)](../adr/0029-email-infrastructure-mailpit.md)
- [ADR-0030: BDD email parallel strategy](../adr/0030-bdd-email-parallel-strategy.md)
- [PR #59: Email infrastructure](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/59)
- [PR #60: BDD Mailpit helper](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/60)
- [PR #61: magic_link_tokens table](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/61)
- [PR #62: Magic link handlers](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/62)
- [PR #63: Magic link BDD](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/63)
- [PR #64: OIDC config skeleton](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/64)
- [PR #65: Magic link cleanup loop](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/65)
- [PR #68: mkcert documentation](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/68)
- [PR #69: pkg/auth skeleton](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/69)
- [PR #74: Phase B.3 OIDC client implementation](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/74)
- [PR #75: Phase B.4 OIDC HTTP handlers](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/75)
- [PR #76: Phase B.4 follow-up tests](https://gitea.arcodange.lab/arcodange/dance-lessons-coach/pulls/76)
---
## Appendix: File inventory
Existing (merged):
- `pkg/auth/oidc.go` — skeleton with TODO methods (PR #69)
- `pkg/auth/oidc_test.go` — placeholder tests (PR #69)
- `documentation/MKCERT.md` — mkcert setup guide (PR #68)
- `Makefile` — includes `make cert` target (PR #68)
To be created:
- `pkg/auth/oidc.go` — complete implementation (B.3)
- `pkg/user/api/oidc_handler.go` — HTTP handlers (B.4)
- `pkg/server/server.go` — route wiring (B.4)
- `features/auth/oidc.feature` — BDD scenarios (B.5)
- `pkg/auth/oidc_test.go` — expanded unit tests (B.3)
- `pkg/user/api/oidc_handler_test.go` — handler tests (B.4)

40
documentation/ROADMAP.md
View File

@@ -1,40 +0,0 @@
# Roadmap & Future Enhancements
Tracking pending features and architectural improvements. Extracted from the original `AGENTS.md` (Tâche 6 restructure). Status updated continuously — items move to "Completed Features" section once shipped.
## Potential Features
- [ ] Database integration
- [ ] Authentication / Authorization
- [ ] Rate limiting
- [ ] Metrics and monitoring
- [ ] Docker containerization
- ✅ CI/CD pipeline ([ADR-0016](../adr/0016-ci-cd-pipeline-design.md), [ADR-0017](../adr/0017-trunk-based-development-workflow.md))
- [ ] Configuration hot reload
- [ ] Circuit breakers
## Architectural Improvements
- [ ] Request validation middleware
- ✅ OpenAPI / Swagger documentation with embedded spec
- [ ] Enhanced OpenTelemetry instrumentation
- [ ] Metrics collection and visualization
- [ ] Health check improvements
- [ ] Configuration validation enhancements
## Completed Features
- ✅ Graceful shutdown with readiness endpoint
- ✅ OpenTelemetry integration with Jaeger support
- ✅ Configuration management with Viper
- ✅ Comprehensive logging with Zerolog
- ✅ Build system with binary output
- ✅ Complete documentation with commit conventions
- ✅ Version management with runtime info
## How to Propose a New Feature
1. Open a Gitea issue describing the use case and acceptance criteria
2. If the feature implies an architectural decision, draft an ADR (`adr/<NNNN>-<slug>.md`) following the template
3. Reference the ADR + issue in any PR introducing the feature
4. Update this roadmap (move from "Potential" to "Completed" when shipped)

49
documentation/STATUS.md
View File

@@ -1,49 +0,0 @@
# Project Status Snapshot
Last updated 2026-05-05 evening.
---
## Active Features
- Magic-link passwordless auth (POST /api/v1/auth/magic-link/request + GET /consume) — production-ready, ADR-0028 Phase A complete
- OIDC client + HTTP handlers (GET /api/v1/auth/oidc/{provider}/start + /callback with PKCE) — production-ready code, BDD coverage TODO. ADR-0028 Phase B (B.1, B.3, B.4 + tests done ; B.5 BDD scenarios TODO).
- Username + password auth — legacy (ADR-0018), kept during migration. To be decommissioned in Phase C.
- Versioned API, JWT, OpenTelemetry, Swagger, BDD
---
## What's In Progress / Next
- Phase B.5 BDD scenarios for OIDC (1 PR Mistral expected)
- Phase C decommission password auth (separate ADR)
---
## Project Structure Highlights
```
adr/ : ADRs
pkg/ : packages (auth, config, server, user, etc.)
features/ : BDD scenarios
documentation/ : docs index
scripts/ : build + CI
```
---
## Key Documentation Entry Points
- README.md : quick start
- AGENTS.md : agent + automation conventions
- documentation/AUTH.md : auth system synthesis
- documentation/MISTRAL-AUTONOMOUS-PATTERN.md : how Mistral PRs are shipped
- documentation/PHASE_B_ROADMAP.md : remaining auth migration work
- documentation/2026-05-05-AUTONOMOUS-SESSION-RECAP.md : the autonomous session highlights
- adr/ : architecture decisions
---
## Today's Milestone (2026-05-05)
27 PRs merged in 1 day via the Mistral autonomous multi-process pattern. ADR-0028 (passwordless auth migration) essentially complete except Phase B.5 BDD.

107
documentation/TROUBLESHOOTING.md
View File

@@ -1,107 +0,0 @@
# Troubleshooting
Common issues and their resolution. Extracted from the original `AGENTS.md` and merged with relevant sections from `AGENT_USAGE_GUIDE.md` and `BDD_GUIDE.md`. Refer back to those guides for context-specific troubleshooting (agent workflows, BDD test failures).
## Port Already in Use
```bash
# Find and kill process using port 8080
kill -TERM $(lsof -ti :8080)
# Force kill if graceful does not work
kill -9 $(lsof -ti :8080)
```
## Server Not Responding
```bash
# Check if running
curl -s http://localhost:8080/api/health
# Restart server using control script
./scripts/start-server.sh restart
# View recent logs
./scripts/start-server.sh logs
```
If health endpoint returns connection refused, the server may have crashed. Check logs in `./scripts/start-server.sh logs` for stack traces.
## Dependency Issues
```bash
# Clean and rebuild
go mod tidy
go build ./...
# If dependency version conflicts persist
go mod download
go mod verify
```
## Tests Failing
### Unit tests
```bash
# Run with verbose output
go test -v ./...
# Check specific test
go test ./pkg/greet/ -run TestName
```
### BDD tests
See [`BDD_GUIDE.md`](BDD_GUIDE.md) for the full BDD troubleshooting workflow (Godog setup, scenario isolation, step matching). Common BDD issues:
- **Step not found** → check `pkg/bdd/steps/` for the step definition file
- **Scenario state leaking** → review [ADR-0025](../adr/0025-bdd-scenario-isolation-strategies.md) for the isolation pattern
- **Database not reset** → ensure the test fixtures cleanup runs (BDD scenario After hooks)
## Configuration Not Loading
The application logs the configuration source at startup. Check logs for:
```
[INF] Configuration loaded from: file:config.yaml
# or
[INF] Configuration loaded from: env
# or
[INF] Configuration loaded from: defaults
```
If config is not loading as expected:
1. Verify file exists and is readable: `ls -la config.yaml`
2. Verify env vars are exported: `env | grep DLC_`
3. Check for typos in keys (case-sensitive)
4. Review [`AGENT_USAGE_GUIDE.md`](AGENT_USAGE_GUIDE.md) section "Configuration troubleshooting"
## OpenTelemetry Not Tracing
1. Verify Jaeger is running: `docker ps | grep jaeger`
2. Check `DLC_TELEMETRY_ENABLED=true` in environment or `telemetry.enabled: true` in config
3. Verify OTLP endpoint reachable: `nc -zv localhost 4317`
4. Check sampler is not `always_off`
5. See [`OBSERVABILITY.md`](OBSERVABILITY.md) for full setup
## Build Failures
```bash
# Clear caches
go clean -cache -modcache
go mod download
# Rebuild
go build ./...
```
If errors persist, see [`local-ci-cd-testing.md`](local-ci-cd-testing.md) for the CI/CD pipeline that mirrors the production build.
## Where to Look Next
- **Agent-specific issues** (vibe, mistral, programmer agent) → [`AGENT_USAGE_GUIDE.md`](AGENT_USAGE_GUIDE.md)
- **BDD-specific issues** → [`BDD_GUIDE.md`](BDD_GUIDE.md)
- **Version/release issues** → [`version-management-guide.md`](version-management-guide.md)
- **CI/CD issues** → [`local-ci-cd-testing.md`](local-ci-cd-testing.md)

34
features/auth/magic_link.feature
View File

@@ -1,34 +0,0 @@
@magic-link
Feature: Passwordless magic-link sign-in
As a user without a password
I want to sign in by clicking a link sent to my email
So I can access the system without typing a password
Scenario: Happy path - request, receive, consume
Given the server is running
And I have an email address for this scenario
When I request a magic link for my email
Then I should receive an email with subject "Your sign-in link"
And the email contains a magic link token
When I consume the magic link token
Then the consume should succeed and return a JWT
Scenario: Token cannot be consumed twice
Given the server is running
And I have an email address for this scenario
When I request a magic link for my email
And the email contains a magic link token
When I consume the magic link token
Then the consume should succeed and return a JWT
When I consume the magic link token
Then the consume should fail with 401
Scenario: Missing token returns 400
Given the server is running
When I consume an empty magic link token
Then the response should have status 400
Scenario: Unknown token returns 401
Given the server is running
When I consume an unknown magic link token
Then the consume should fail with 401

10
features/greet/greet.feature
View File

@@ -15,16 +15,6 @@ Feature: Greet Service
When I request a greeting for "John"
Then the response should be "{\"message\":\"Hello John!\"}"
@critical @v2-gate
Scenario: v2 endpoint returns 404 when api.v2_enabled is disabled
# In the default tag-filter run (~@v2), the test server starts with
# v2_enabled=false. The v2EnabledGate middleware (ADR-0023 Phase 4)
# returns 404 with a JSON body explaining the flag state.
Given the server is running
When I send a POST request to v2 greet with name "John"
Then the status code should be 404
And the response should contain "v2 API is currently disabled"
@v2 @api
Scenario: v2 greeting with JSON POST request
Given the server is running with v2 enabled

45
features/info/info.feature
View File

@@ -1,45 +0,0 @@
# features/info/info.feature
@info @critical
Feature: Info Endpoint
The /api/info endpoint should return composite application information
@basic @critical
Scenario: GET /api/info returns all required fields
Given the server is running
When I request the info endpoint
Then the status code should be 200
And the response should be JSON
And the response should contain "version"
And the response should contain "commit_short"
And the response should contain "build_date"
And the response should contain "uptime_seconds"
And the response should contain "cache_enabled"
And the response should contain "healthz_status"
And the "healthz_status" field should equal "healthy"
@version @critical
Scenario: version field matches semantic version pattern
Given the server is running
When I request the info endpoint
Then the status code should be 200
And the "version" field should match /^\d+\.\d+\.\d+$/
@cache @skip @bdd-deferred
Scenario: /api/info is cached when cache is enabled
# Deferred: the BDD testsetup currently runs with cache disabled
# (see "Cache service disabled" in test logs). Cache HIT/MISS behavior
# is covered by unit tests on the cache service. Reopen this scenario
# if/when the BDD harness gains a cache-enabled mode (likely after
# ADR-0022 Phase 2).
Given the server is running with cache enabled
When I request the info endpoint
Then the response header "X-Cache" should be "MISS"
When I request the info endpoint again
Then the response header "X-Cache" should be "HIT"
@go_version @critical
Scenario: go_version field is non-empty
Given the server is running
When I request the info endpoint
Then the status code should be 200
And the response should contain "go_version"

16
features/info/info_test.go
View File

@@ -1,16 +0,0 @@
package info
import (
"testing"
"dance-lessons-coach/pkg/bdd/testsetup"
)
func TestInfoBDD(t *testing.T) {
config := testsetup.NewFeatureConfig("info", "progress", false)
suite := testsetup.CreateTestSuite(t, config, "dance-lessons-coach BDD Tests - Info Feature")
if suite.Run() != 0 {
t.Fatal("non-zero status returned, failed to run info BDD tests")
}
}

10
features/jwt/jwt_secret_retention.feature
View File

@@ -40,16 +40,6 @@ Feature: JWT Secret Retention Policy
Then the primary secret should not be removed
And the primary secret should remain active
@critical @admin-introspection
Scenario: Admin metadata endpoint exposes structure without leaking secret values
Given a primary JWT secret exists
And I add a secondary JWT secret "test-secret-do-not-leak-please-12345"
When I request the JWT secrets metadata endpoint
Then the status code should be 200
And the metadata should contain 2 secrets
And the metadata should NOT contain the secret value "test-secret-do-not-leak-please-12345"
And every secret in the metadata should have a SHA-256 fingerprint
@todo
Scenario: Multiple secrets with different ages
Given I have 3 JWT secrets of different ages

15
frontend/.storybook/main.ts
View File

@@ -1,15 +0,0 @@
import type { StorybookConfig } from '@storybook/vue3-vite'
const config: StorybookConfig = {
stories: ['../components/**/*.stories.@(js|ts|mdx)'],
addons: ['@storybook/addon-essentials'],
framework: {
name: '@storybook/vue3-vite',
options: {},
},
docs: {
autodocs: 'tag',
},
}
export default config

15
frontend/.storybook/preview.ts
View File

@@ -1,15 +0,0 @@
import type { Preview } from '@storybook/vue3'
const preview: Preview = {
parameters: {
actions: { argTypesRegex: '^on[A-Z].*' },
controls: {
matchers: {
color: /(background|color)$/i,
date: /Date$/i,
},
},
},
}
export default preview

4
frontend/app.vue
View File

@@ -1,5 +1,3 @@
<template>
<NuxtLayout>
<NuxtPage />
</NuxtLayout>
<NuxtPage />
</template>

13
frontend/components/AppFooter.vue
View File

@@ -1,13 +0,0 @@
<script setup lang="ts">
import AppFooterView, { type AppInfo } from './AppFooterView.vue'
// Wrapper: handles data fetching, delegates rendering to AppFooterView.
// Separation of concerns (SRP) - same pattern as HealthDashboard / HealthDashboardView.
// server: false → fetch client-side only. Avoids SSR fetching through the dev proxy
// (which can fail in some local setups), and lets Playwright route mocks apply.
const { data, pending, error } = useFetch<AppInfo>('/api/info', { server: false })
</script>
<template>
<AppFooterView :data="data" :pending="pending" :error="error" />
</template>

45
frontend/components/AppFooterView.vue
View File

@@ -1,45 +0,0 @@
<script setup lang="ts">
import { humaniseUptime } from '~/utils/uptime'
export interface AppInfo {
version: string
commit_short: string
build_date: string
uptime_seconds: number
cache_enabled: boolean
healthz_status: string
}
defineProps<{
data: AppInfo | null | undefined
pending: boolean
error: { message: string } | null | undefined
}>()
</script>
<template>
<footer data-testid="app-footer">
<p v-if="pending" data-testid="app-footer-pending">v?</p>
<p v-else-if="error" data-testid="app-footer-error">v? · info unavailable</p>
<p v-else-if="data" data-testid="app-footer-info">
<span data-testid="app-footer-version">v{{ data.version }}</span>
<span> · commit </span>
<span data-testid="app-footer-commit">{{ data.commit_short }}</span>
<span> · uptime </span>
<span data-testid="app-footer-uptime">{{ humaniseUptime(data.uptime_seconds) }}</span>
</p>
</footer>
</template>
<style scoped>
footer {
border-top: 1px solid #ccc;
padding: 0.5rem 1rem;
font-size: 0.85rem;
color: #555;
text-align: center;
}
footer p {
margin: 0;
}
</style>

26
frontend/components/HealthDashboard.stories.ts
View File

@@ -1,26 +0,0 @@
import type { Meta, StoryObj } from '@storybook/vue3'
import HealthDashboard from './HealthDashboard.vue'
const meta: Meta<typeof HealthDashboard> = {
title: 'Components/HealthDashboard',
component: HealthDashboard,
tags: ['autodocs'],
parameters: {
docs: {
description: {
component:
'Smart wrapper that calls /api/healthz internally and delegates rendering to HealthDashboardView. ' +
'For state-by-state previews (Healthy, Loading, Error), see ' +
'[HealthDashboardView stories](?path=/docs/components-healthdashboardview--docs).',
},
},
},
}
export default meta
type Story = StoryObj<typeof meta>
// Default story - calls real /api/healthz (works in browser if dev proxy + backend are up)
export const Default: Story = {
args: {},
}

31
frontend/components/HealthDashboard.vue
View File

@@ -1,17 +1,22 @@
<script setup lang="ts">
import HealthDashboardView, { type HealthInfo } from './HealthDashboardView.vue'
// Wrapper: handles data fetching, delegates rendering to HealthDashboardView.
// Separation of concerns (SRP):
// - HealthDashboard (this) = data layer (useFetch lifecycle)
// - HealthDashboardView = presentation layer (testable in Storybook + e2e)
//
// server: false → fetch client-side only. Avoids SSR fetching through the dev
// proxy (which can fail in some local setups), and lets Playwright route mocks
// apply. Same fix that landed for AppFooter in PR #40.
const { data, pending, error } = useFetch<HealthInfo>('/api/healthz', { server: false })
interface HealthInfo {
status: string
version: string
uptime_seconds: number
timestamp: string
}
const { data, pending, error } = await useFetch<HealthInfo>('/api/healthz')
</script>
<template>
<HealthDashboardView :data="data" :pending="pending" :error="error" />
<section data-testid="health-dashboard">
<h2>Server Health</h2>
<p v-if="pending">Loading...</p>
<p v-else-if="error">Error loading health: {{ error.message }}</p>
<ul v-else-if="data" data-testid="health-info">
<li><strong>Status:</strong> <span data-testid="health-status">{{ data.status }}</span></li>
<li><strong>Version:</strong> {{ data.version }}</li>
<li><strong>Uptime:</strong> {{ data.uptime_seconds }} seconds</li>
<li><strong>Last check:</strong> {{ data.timestamp }}</li>
</ul>
</section>
</template>

79
frontend/components/HealthDashboardView.stories.ts
View File

@@ -1,79 +0,0 @@
import type { Meta, StoryObj } from '@storybook/vue3'
import HealthDashboardView from './HealthDashboardView.vue'
interface ViewArgs {
data: {
status: string
version: string
uptime_seconds: number
timestamp: string
} | null
pending: boolean
error: { message: string } | null
}
const meta = {
title: 'Components/HealthDashboardView',
component: HealthDashboardView,
tags: ['autodocs'],
argTypes: {
pending: { control: 'boolean' },
},
parameters: {
docs: {
description: {
component:
'Pure presentational component for the health dashboard. ' +
'Accepts `data`, `pending`, `error` as props so all 3 states can be ' +
'previewed in Storybook and asserted in unit tests. The data fetching ' +
'wrapper is `HealthDashboard.vue`.',
},
},
},
} satisfies Meta<ViewArgs>
export default meta
type Story = StoryObj<typeof meta>
export const Healthy: Story = {
args: {
data: {
status: 'healthy',
version: '1.4.0',
uptime_seconds: 3600,
timestamp: '2026-05-03T17:30:00.000Z',
},
pending: false,
error: null,
},
}
export const Loading: Story = {
args: {
data: null,
pending: true,
error: null,
},
}
export const ErrorState: Story = {
args: {
data: null,
pending: false,
error: { message: '[GET] "/api/healthz": 502 Bad Gateway (simulated)' },
},
}
export const HealthyHighUptime: Story = {
args: {
data: {
status: 'healthy',
version: '1.5.0-rc1',
uptime_seconds: 86400 * 7,
timestamp: new Date().toISOString(),
},
pending: false,
error: null,
},
}

30
frontend/components/HealthDashboardView.vue
View File

@@ -1,30 +0,0 @@
<script setup lang="ts">
export interface HealthInfo {
status: string
version: string
uptime_seconds: number
timestamp: string
}
defineProps<{
data: HealthInfo | null | undefined
pending: boolean
error: { message: string } | null | undefined
}>()
</script>
<template>
<section data-testid="health-dashboard">
<h2>Server Health</h2>
<p v-if="pending" data-testid="health-loading">Loading...</p>
<p v-else-if="error" data-testid="health-error">
Error loading health: {{ error.message }}
</p>
<ul v-else-if="data" data-testid="health-info">
<li><strong>Status:</strong> <span data-testid="health-status">{{ data.status }}</span></li>
<li><strong>Version:</strong> {{ data.version }}</li>
<li><strong>Uptime:</strong> {{ data.uptime_seconds }} seconds</li>
<li><strong>Last check:</strong> {{ data.timestamp }}</li>
</ul>
</section>
</template>

4
frontend/docs/README.md
View File

@@ -1,4 +0,0 @@
# Frontend Docs
- [E2E Test Reports](./e2e/README.md) - auto-generated by `npm run docs:gen`
- Storybook (run locally: `npm run storybook` ; build: `npm run build-storybook` then open `storybook-static/index.html`)

7
frontend/docs/e2e/README.md
View File

@@ -1,7 +0,0 @@
# E2E Test Reports
[<- Up to docs](../README.md)
| Test | Status | Duration |
|------|--------|----------|
| [home page loads and shows server health info](./home-page-loads-and-shows-server-health-info.md) | PASSED | 168ms |

16
frontend/docs/e2e/home-page-loads-and-shows-server-health-info.md
View File

@@ -1,16 +0,0 @@
# home page loads and shows server health info
[<- Back to index](./README.md) | [Top](../README.md)
**File**: `tests/e2e/health.spec.ts`
**Status**: PASSED
**Duration**: 168ms
## Screenshot
![home page loads and shows server health info](../../tests/e2e/screenshots/home-page-loads-and-shows-server-health-info.png)
## Test Details
- Start Time: 2026-05-03T14:38:42.958Z
- Spec File: health.spec.ts

17
frontend/layouts/default.vue
View File

@@ -1,17 +0,0 @@
<template>
<div class="layout-root">
<slot />
<AppFooter />
</div>
</template>
<style scoped>
.layout-root {
min-height: 100vh;
display: flex;
flex-direction: column;
}
.layout-root > :first-child {
flex: 1;
}
</style>

2348
frontend/package-lock.json generated
View File

File diff suppressed because it is too large Load Diff

10
frontend/package.json
View File

@@ -6,20 +6,12 @@
"dev": "nuxt dev",
"generate": "nuxt generate",
"preview": "nuxt preview",
"postinstall": "nuxt prepare",
"storybook": "storybook dev -p 6006",
"build-storybook": "storybook build",
"docs:gen": "playwright test && node scripts/generate-test-docs.mjs",
"docs:full": "npm run build-storybook && npm run docs:gen"
"postinstall": "nuxt prepare"
},
"devDependencies": {
"@playwright/test": "^1.59.1",
"@storybook/addon-essentials": "^8.0.0",
"@storybook/vue3": "^8.0.0",
"@storybook/vue3-vite": "^8.0.0",
"@types/node": "^25.6.0",
"nuxt": "^3.13.0",
"storybook": "^8.0.0",
"typescript": "^6.0.3"
},
"packageManager": "npm@11.5.2"

9
frontend/playwright.config.ts
View File

@@ -1,19 +1,10 @@
import { defineConfig } from '@playwright/test'
import path from 'path'
export default defineConfig({
testDir: './tests/e2e',
timeout: 30_000,
reporter: [
['list'],
['json', { outputFile: path.join(process.cwd(), 'test-results', 'results.json') }],
],
use: {
baseURL: 'http://localhost:3000',
screenshot: 'on',
video: 'off',
},
outputDir: 'test-results/output',
webServer: {
command: 'npm run dev',
url: 'http://localhost:3000',

120
frontend/scripts/generate-test-docs.mjs
View File

@@ -1,120 +0,0 @@
#!/usr/bin/env node
import fs from 'node:fs/promises'
import path from 'node:path'
import { fileURLToPath } from 'node:url'
const __dirname = path.dirname(fileURLToPath(import.meta.url))
const frontendDir = path.resolve(__dirname, '..')
const resultsPath = path.join(frontendDir, 'test-results', 'results.json')
const docsDir = path.join(frontendDir, 'docs', 'e2e')
const screenshotsDir = path.join(frontendDir, 'tests', 'e2e', 'screenshots')
async function main() {
// Read results
const resultsText = await fs.readFile(resultsPath, 'utf8')
const results = JSON.parse(resultsText)
// Create output directories
await fs.mkdir(docsDir, { recursive: true })
// Extract tests from suites
const testDocs = []
for (const suite of results.suites || []) {
for (const spec of suite.specs || []) {
for (const test of spec.tests || []) {
for (const result of test.results || []) {
const testInfo = {
title: spec.title,
specFile: spec.file || suite.file,
status: result.status,
duration: result.duration,
startTime: result.startTime,
attachments: result.attachments || [],
}
testDocs.push(testInfo)
}
}
}
}
// Generate individual test markdown files
for (const test of testDocs) {
const slug = slugify(test.title)
const mdPath = path.join(docsDir, `${slug}.md`)
// Use slug-based screenshot name (matches explicit screenshot in test)
let screenshotPath = `${slug}.png`
// Also try to find screenshot attachment and use its basename
if (test.attachments && test.attachments.length > 0) {
for (const attachment of test.attachments) {
if (attachment.contentType === 'image/png') {
const basename = path.basename(attachment.path)
// Prefer explicit screenshot name if it matches our pattern
if (basename !== 'test-finished-1.png' && basename !== 'test-finished-2.png') {
screenshotPath = basename
break
}
}
}
}
const absoluteScreenshotPath = path.join(screenshotsDir, screenshotPath)
const relativeScreenshotPath = path.relative(docsDir, absoluteScreenshotPath)
const mdContent = `# ${test.title}
[<- Back to index](./README.md) | [Top](../README.md)
**File**: \`tests/e2e/${test.specFile}\`
**Status**: ${test.status.toUpperCase()}
**Duration**: ${test.duration}ms
## Screenshot
![${test.title}](${relativeScreenshotPath})
## Test Details
- Start Time: ${test.startTime || 'N/A'}
- Spec File: ${test.specFile}
`
await fs.writeFile(mdPath, mdContent)
console.log(`Generated: ${path.relative(frontendDir, mdPath)}`)
}
// Generate index README
const indexContent = `# E2E Test Reports
[<- Up to docs](../README.md)
| Test | Status | Duration |
|------|--------|----------|
${testDocs.map(t => `| [${escapeMd(t.title)}](./${slugify(t.title)}.md) | ${t.status.toUpperCase()} | ${t.duration}ms |`).join('\n')}
`
await fs.writeFile(path.join(docsDir, 'README.md'), indexContent)
console.log(`Generated: ${path.relative(frontendDir, path.join(docsDir, 'README.md'))}`)
console.log(`\nGenerated ${testDocs.length} test docs`)
}
function slugify(str) {
return str
.toLowerCase()
.replace(/[^\w\s-]/g, '')
.replace(/[\s_]+/g, '-')
.replace(/^-+|-+$/g, '')
}
function escapeMd(str) {
return str.replace(/[|\\\[\]\{\}]/g, '\\$&')
}
main().catch(err => {
console.error('Error:', err)
process.exit(1)
})

6
frontend/shims-vue.d.ts vendored
View File

@@ -1,6 +0,0 @@
declare module '*.vue' {
import type { DefineComponent } from 'vue'
// eslint-disable-next-line @typescript-eslint/no-explicit-any
const component: DefineComponent<any, any, any>
export default component
}

67
frontend/tests/e2e/app-footer.spec.ts
View File

@@ -1,67 +0,0 @@
import { test, expect } from '@playwright/test'
// Both specs mock /api/info so they decouple from the dev-proxy plumbing.
// The integration with the real backend is covered by the BDD scenario in
// features/info/info.feature (server-side, no frontend proxy in the loop).
test('home page footer shows version, commit and uptime', async ({ page }) => {
await page.route('**/api/info', (route) => {
route.fulfill({
status: 200,
contentType: 'application/json',
body: JSON.stringify({
version: '1.4.0',
commit_short: '4a3f1bb',
build_date: '2026-05-05T00:00:00Z',
uptime_seconds: 8042,
cache_enabled: true,
healthz_status: 'healthy',
}),
})
})
await page.goto('/')
// Footer is mounted globally via layouts/default.vue
await expect(page.getByTestId('app-footer')).toBeVisible()
// The PR #32 lesson: assert content, not just visibility.
// Without the regex check the test would PASS even if the footer rendered the
// pending placeholder ("v?") indefinitely.
await expect(page.getByTestId('app-footer-info')).toBeVisible()
const versionLocator = page.getByTestId('app-footer-version')
await expect(versionLocator).toBeVisible()
await expect(versionLocator).toHaveText(/^v\d+\.\d+\.\d+$/)
// Commit and uptime should be present and non-empty.
await expect(page.getByTestId('app-footer-commit')).not.toBeEmpty()
await expect(page.getByTestId('app-footer-uptime')).not.toBeEmpty()
await page.screenshot({
path: 'tests/e2e/screenshots/app-footer-shows-version-commit-uptime.png',
fullPage: true,
})
})
// Regression spec: documents the expected error UX so we don't ship a silent failure.
// Routes /api/info to a 502 mock so the test is reproducible regardless of backend.
test('home page footer surfaces info endpoint errors gracefully', async ({ page }) => {
await page.route('**/api/info', (route) => {
route.fulfill({
status: 502,
contentType: 'application/json',
body: JSON.stringify({ error: 'simulated_backend_down' }),
})
})
await page.goto('/')
// Footer must NOT crash the page
await expect(page.getByTestId('app-footer')).toBeVisible()
await expect(page.getByTestId('app-footer-error')).toBeVisible()
// The error placeholder should NOT contain a real version pattern
await expect(page.getByTestId('app-footer-info')).not.toBeVisible()
await page.screenshot({
path: 'tests/e2e/screenshots/app-footer-surfaces-info-endpoint-errors-gracefully.png',
fullPage: true,
})
})

49
frontend/tests/e2e/health.spec.ts
View File

@@ -1,55 +1,8 @@
import { test, expect } from '@playwright/test'
// Both specs mock /api/healthz so they decouple from the dev-proxy plumbing.
// The integration with the real backend is covered by the BDD scenario in
// features/health/health.feature (server-side, no frontend proxy in the loop).
// Same approach as tests/e2e/app-footer.spec.ts (PR #40) - applied here to
// close the debt left by that PR's out-of-scope follow-up note.
test('home page loads and shows healthy server state', async ({ page }) => {
await page.route('**/api/healthz', (route) => {
route.fulfill({
status: 200,
contentType: 'application/json',
body: JSON.stringify({
status: 'healthy',
version: '1.4.0',
uptime_seconds: 8042,
timestamp: '2026-05-05T08:00:00Z',
}),
})
})
test('home page loads and shows server health info', async ({ page }) => {
await page.goto('/')
await expect(page.getByTestId('health-dashboard')).toBeVisible()
const heading = page.getByRole('heading', { name: /dance-lessons-coach/i })
await expect(heading).toBeVisible()
// Assert the dashboard is in HEALTHY state, not an error state.
// The dashboard renders an "Error loading health: ..." paragraph when /api/healthz
// is unreachable (Go backend not running, proxy misconfigured, endpoint removed,
// etc.). Without these assertions the test would falsely PASS even when the
// dashboard shows the error UI - regression observed 2026-05-03 (Go backend
// not running locally → page renders the error, Playwright PASSES).
await expect(page.getByTestId('health-info')).toBeVisible()
await expect(page.getByTestId('health-status')).toHaveText('healthy')
await expect(page.getByText(/Error loading health/i)).not.toBeVisible()
await page.screenshot({ path: 'tests/e2e/screenshots/home-page-loads-and-shows-server-health-info.png', fullPage: true })
})
// Regression spec: documents the expected error UX so we don't ship a silent failure.
// Routes /api/healthz to a 502 mock so the test is reproducible regardless of backend.
test('home page surfaces health endpoint errors visibly', async ({ page }) => {
await page.route('**/api/healthz', (route) => {
route.fulfill({
status: 502,
contentType: 'application/json',
body: JSON.stringify({ error: 'simulated_backend_down' }),
})
})
await page.goto('/')
await expect(page.getByTestId('health-dashboard')).toBeVisible()
await expect(page.getByText(/Error loading health/i)).toBeVisible()
await expect(page.getByTestId('health-info')).not.toBeVisible()
await page.screenshot({ path: 'tests/e2e/screenshots/home-page-surfaces-health-endpoint-errors-visibly.png', fullPage: true })
})

0
frontend/tests/e2e/screenshots/.gitkeep
View File

BIN
frontend/tests/e2e/screenshots/app-footer-shows-version-commit-uptime.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 22 KiB

BIN
frontend/tests/e2e/screenshots/app-footer-surfaces-info-endpoint-errors-gracefully.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 21 KiB

BIN
frontend/tests/e2e/screenshots/home-page-loads-and-shows-server-health-info.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 18 KiB

BIN
frontend/tests/e2e/screenshots/home-page-surfaces-health-endpoint-errors-visibly.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 20 KiB

16
frontend/utils/uptime.ts
View File

@@ -1,16 +0,0 @@
// Convert a duration in seconds to a humanised string like "2h 13m" or "45m 12s".
// Returns "?" for non-finite or negative input so the UI never renders NaN/empty.
export function humaniseUptime(seconds: number | null | undefined): string {
if (seconds == null || !Number.isFinite(seconds) || seconds < 0) return '?'
const s = Math.floor(seconds)
const days = Math.floor(s / 86400)
const hours = Math.floor((s % 86400) / 3600)
const minutes = Math.floor((s % 3600) / 60)
const secs = s % 60
if (days > 0) return `${days}d ${hours}h`
if (hours > 0) return `${hours}h ${minutes}m`
if (minutes > 0) return `${minutes}m ${secs}s`
return `${secs}s`
}

2
go.mod
View File

@@ -4,7 +4,6 @@ go 1.26.1
require (
github.com/cucumber/godog v0.15.1
github.com/fsnotify/fsnotify v1.9.0
github.com/go-chi/chi/v5 v5.2.5
github.com/go-playground/locales v0.14.1
github.com/go-playground/universal-translator v0.18.1
@@ -38,6 +37,7 @@ require (
github.com/cucumber/messages/go/v21 v21.0.1 // indirect
github.com/davecgh/go-spew v1.1.1 // indirect
github.com/felixge/httpsnoop v1.0.4 // indirect
github.com/fsnotify/fsnotify v1.9.0 // indirect
github.com/gabriel-vasile/mimetype v1.4.13 // indirect
github.com/go-logr/logr v1.4.3 // indirect
github.com/go-logr/stdr v1.2.2 // indirect

30
pkg/auth/context.go
View File

@@ -1,30 +0,0 @@
// Package auth — context keys and helpers for authentication state.
//
// This file owns the symbols that other packages use to read/write the
// authenticated user from a request context. Previously these symbols
// lived in pkg/greet/ which was the wrong home (auth concern in greet
// package) ; moved here as part of the middleware design cleanup.
package auth
import (
"context"
"dance-lessons-coach/pkg/user"
)
// contextKey is unexported to prevent collisions with other packages
// using string keys (Go convention).
type contextKey string
// UserContextKey is the key under which the authenticated user is
// stored in the request context by AuthMiddleware.
const UserContextKey contextKey = "authenticatedUser"
// GetAuthenticatedUserFromContext extracts the authenticated user from
// the request context. Returns (nil, false) if no user is present
// (which is the case for unauthenticated requests AND for requests
// that failed silent fail-through ; cf. AuthMiddleware semantics).
func GetAuthenticatedUserFromContext(ctx context.Context) (*user.User, bool) {
u, ok := ctx.Value(UserContextKey).(*user.User)
return u, ok
}

345
pkg/auth/oidc.go
View File

@@ -1,345 +0,0 @@
// Package auth provides OpenID Connect client primitives for the
// dance-lessons-coach passwordless-auth migration (ADR-0028 Phase B).
//
// This file defines the client surface only. HTTP handlers wire-up
// happens in pkg/user/api/oidc_handler.go (separate phase B.3).
package auth
import (
"context"
"crypto/rsa"
"encoding/base64"
"encoding/json"
"fmt"
"math/big"
"net/http"
"net/url"
"strings"
"sync"
"time"
"github.com/golang-jwt/jwt/v5"
)
// OIDCClient is a per-provider OIDC client.
// Holds the discovery document + JWKS cache + OAuth code-exchange config.
type OIDCClient struct {
issuerURL string
clientID string
clientSecret string
httpClient *http.Client
// discovery document, lazy-fetched on first use
discoveryMu sync.RWMutex
discovery *Discovery
// JWKS cache (id_token signature verification keys), refreshed periodically
jwksMu sync.RWMutex
jwks map[string]*rsa.PublicKey
jwksFetched time.Time
}
// Discovery is the subset of the .well-known/openid-configuration document we use.
type Discovery struct {
Issuer string `json:"issuer"`
AuthorizationEndpoint string `json:"authorization_endpoint"`
TokenEndpoint string `json:"token_endpoint"`
JWKSUri string `json:"jwks_uri"`
UserinfoEndpoint string `json:"userinfo_endpoint,omitempty"`
}
// TokenResponse is the response from the token endpoint after code exchange.
type TokenResponse struct {
AccessToken string `json:"access_token"`
TokenType string `json:"token_type"`
ExpiresIn int64 `json:"expires_in"`
RefreshToken string `json:"refresh_token,omitempty"`
IDToken string `json:"id_token"`
Scope string `json:"scope,omitempty"`
}
// IDTokenClaims represents the parsed claims from an ID token.
type IDTokenClaims struct {
jwt.RegisteredClaims
Email string `json:"email,omitempty"`
EmailVerified bool `json:"email_verified,omitempty"`
Name string `json:"name,omitempty"`
}
// jwks represents the JWKS (JSON Web Key Set) response.
type jwks struct {
Keys []jwk `json:"keys"`
}
// jwk represents a single JSON Web Key.
type jwk struct {
Kid string `json:"kid"`
Kty string `json:"kty"`
N string `json:"n"`
E string `json:"e"`
Use string `json:"use,omitempty"`
Alg string `json:"alg,omitempty"`
}
// NewOIDCClient constructs a client. Discovery + JWKS are NOT fetched eagerly;
// they are lazy-loaded on first use to avoid blocking server startup if the
// provider is temporarily down.
func NewOIDCClient(issuerURL, clientID, clientSecret string) *OIDCClient {
return &OIDCClient{
issuerURL: issuerURL,
clientID: clientID,
clientSecret: clientSecret,
httpClient: &http.Client{Timeout: 10 * time.Second},
jwks: make(map[string]*rsa.PublicKey),
}
}
// ClientID returns the OIDC client ID.
func (c *OIDCClient) ClientID() string {
return c.clientID
}
// IssuerURL returns the OIDC issuer URL.
func (c *OIDCClient) IssuerURL() string {
return c.issuerURL
}
// SetHTTPClient sets a custom HTTP client for testing.
func (c *OIDCClient) SetHTTPClient(client *http.Client) {
c.httpClient = client
}
// decodeRSAPublicKey reconstructs an *rsa.PublicKey from JWK n and e values.
func decodeRSAPublicKey(j jwk) (*rsa.PublicKey, error) {
nBytes, err := base64.RawURLEncoding.DecodeString(j.N)
if err != nil {
return nil, fmt.Errorf("decode n: %w", err)
}
eBytes, err := base64.RawURLEncoding.DecodeString(j.E)
if err != nil {
return nil, fmt.Errorf("decode e: %w", err)
}
n := new(big.Int).SetBytes(nBytes)
e := new(big.Int).SetBytes(eBytes)
return &rsa.PublicKey{N: n, E: int(e.Int64())}, nil
}
// Discover fetches and caches the .well-known document. Idempotent.
// First call: HTTP fetch + cache. Subsequent calls: cached value.
func (c *OIDCClient) Discover(ctx context.Context) (*Discovery, error) {
c.discoveryMu.RLock()
if c.discovery != nil {
c.discoveryMu.RUnlock()
return c.discovery, nil
}
c.discoveryMu.RUnlock()
c.discoveryMu.Lock()
defer c.discoveryMu.Unlock()
// Double-check after acquiring write lock
if c.discovery != nil {
return c.discovery, nil
}
wellKnownURL := fmt.Sprintf("%s/.well-known/openid-configuration", c.issuerURL)
req, err := http.NewRequestWithContext(ctx, http.MethodGet, wellKnownURL, nil)
if err != nil {
return nil, fmt.Errorf("create discovery request: %w", err)
}
resp, err := c.httpClient.Do(req)
if err != nil {
return nil, fmt.Errorf("fetch discovery: %w", err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return nil, fmt.Errorf("discovery HTTP %d", resp.StatusCode)
}
var disc Discovery
if err := json.NewDecoder(resp.Body).Decode(&disc); err != nil {
return nil, fmt.Errorf("decode discovery: %w", err)
}
c.discovery = &disc
return &disc, nil
}
// RefreshJWKS fetches JWKS URI, parse keys, populate jwks map.
func (c *OIDCClient) RefreshJWKS(ctx context.Context) error {
// Ensure discovery is loaded
if c.discovery == nil {
if _, err := c.Discover(ctx); err != nil {
return fmt.Errorf("discover: %w", err)
}
}
req, err := http.NewRequestWithContext(ctx, http.MethodGet, c.discovery.JWKSUri, nil)
if err != nil {
return fmt.Errorf("create JWKS request: %w", err)
}
resp, err := c.httpClient.Do(req)
if err != nil {
return fmt.Errorf("fetch JWKS: %w", err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return fmt.Errorf("JWKS HTTP %d", resp.StatusCode)
}
var keySet jwks
if err := json.NewDecoder(resp.Body).Decode(&keySet); err != nil {
return fmt.Errorf("decode JWKS: %w", err)
}
c.jwksMu.Lock()
defer c.jwksMu.Unlock()
c.jwks = make(map[string]*rsa.PublicKey)
for _, key := range keySet.Keys {
if key.Kty == "RSA" {
pubKey, err := decodeRSAPublicKey(key)
if err != nil {
return fmt.Errorf("decode RSA key %s: %w", key.Kid, err)
}
c.jwks[key.Kid] = pubKey
}
}
c.jwksFetched = time.Now()
return nil
}
// ExchangeCode exchanges an authorization code for an access token and ID token.
func (c *OIDCClient) ExchangeCode(ctx context.Context, code, codeVerifier, redirectURI string) (*TokenResponse, error) {
// Ensure discovery is loaded
if c.discovery == nil {
if _, err := c.Discover(ctx); err != nil {
return nil, fmt.Errorf("discover: %w", err)
}
}
form := url.Values{}
form.Set("grant_type", "authorization_code")
form.Set("code", code)
form.Set("code_verifier", codeVerifier)
form.Set("redirect_uri", redirectURI)
form.Set("client_id", c.clientID)
form.Set("client_secret", c.clientSecret)
req, err := http.NewRequestWithContext(ctx, http.MethodPost, c.discovery.TokenEndpoint, strings.NewReader(form.Encode()))
if err != nil {
return nil, fmt.Errorf("create token request: %w", err)
}
req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
resp, err := c.httpClient.Do(req)
if err != nil {
return nil, fmt.Errorf("exchange code: %w", err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return nil, fmt.Errorf("token HTTP %d", resp.StatusCode)
}
var tokenResp TokenResponse
if err := json.NewDecoder(resp.Body).Decode(&tokenResp); err != nil {
return nil, fmt.Errorf("decode token response: %w", err)
}
return &tokenResp, nil
}
// ValidateIDToken verifies the signature and claims of an ID token.
func (c *OIDCClient) ValidateIDToken(ctx context.Context, idToken string) (*IDTokenClaims, error) {
// First, parse without verification to get the kid
parser := jwt.NewParser()
unverifiedToken, _, err := parser.ParseUnverified(idToken, &IDTokenClaims{})
if err != nil {
return nil, fmt.Errorf("parse unverified token: %w", err)
}
claims, ok := unverifiedToken.Claims.(*IDTokenClaims)
if !ok {
return nil, fmt.Errorf("invalid claims type")
}
// Get kid from header
kid, ok := unverifiedToken.Header["kid"].(string)
if !ok || kid == "" {
return nil, fmt.Errorf("missing kid in token header")
}
// Get the key, refreshing JWKS if needed
c.jwksMu.RLock()
_, keyExists := c.jwks[kid]
c.jwksMu.RUnlock()
if !keyExists {
if err := c.RefreshJWKS(ctx); err != nil {
return nil, fmt.Errorf("refresh JWKS: %w", err)
}
c.jwksMu.RLock()
_, keyExists = c.jwks[kid]
c.jwksMu.RUnlock()
if !keyExists {
return nil, fmt.Errorf("key %s not found in JWKS", kid)
}
}
// Parse with verification
keyFunc := func(token *jwt.Token) (interface{}, error) {
if kid, ok := token.Header["kid"].(string); ok {
c.jwksMu.RLock()
defer c.jwksMu.RUnlock()
if key, exists := c.jwks[kid]; exists {
return key, nil
}
}
return nil, fmt.Errorf("key not found")
}
parsedToken, err := jwt.ParseWithClaims(idToken, &IDTokenClaims{}, keyFunc)
if err != nil {
return nil, fmt.Errorf("parse token: %w", err)
}
claims, ok = parsedToken.Claims.(*IDTokenClaims)
if !ok {
return nil, fmt.Errorf("invalid claims type after parse")
}
// Validate claims
if claims.Issuer != c.issuerURL {
return nil, fmt.Errorf("issuer mismatch: expected %s, got %s", c.issuerURL, claims.Issuer)
}
// Check audience contains clientID
audValid := false
if claims.Audience != nil {
for _, aud := range claims.Audience {
if aud == c.clientID {
audValid = true
break
}
}
}
if !audValid {
return nil, fmt.Errorf("audience does not contain client ID %s", c.clientID)
}
// Check expiration
if claims.ExpiresAt != nil && time.Now().UTC().After(claims.ExpiresAt.Time) {
return nil, fmt.Errorf("token expired")
}
return claims, nil
}

431
pkg/auth/oidc_test.go
View File

@@ -1,431 +0,0 @@
package auth
import (
"context"
"crypto/rand"
"crypto/rsa"
"encoding/base64"
"fmt"
"math/big"
"net/http"
"net/http/httptest"
"net/url"
"sync/atomic"
"testing"
"time"
"github.com/golang-jwt/jwt/v5"
)
func TestNewOIDCClient(t *testing.T) {
c := NewOIDCClient("https://example.com", "client_id", "client_secret")
if c == nil {
t.Fatal("NewOIDCClient returned nil")
}
if c.issuerURL != "https://example.com" {
t.Errorf("issuerURL not set: got %q", c.issuerURL)
}
}
func TestDiscover_HappyPath(t *testing.T) {
var server *httptest.Server
server = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if r.URL.Path != "/.well-known/openid-configuration" {
t.Errorf("unexpected path: %s", r.URL.Path)
w.WriteHeader(http.StatusNotFound)
return
}
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(fmt.Sprintf(`{"issuer":"%s","authorization_endpoint":"%s/auth","token_endpoint":"%s/token","jwks_uri":"%s/jwks"}`,
server.URL, server.URL, server.URL, server.URL)))
}))
defer server.Close()
client := NewOIDCClient(server.URL, "client_id", "client_secret")
client.httpClient = server.Client()
disc, err := client.Discover(context.Background())
if err != nil {
t.Fatalf("Discover failed: %v", err)
}
if disc.Issuer != server.URL {
t.Errorf("issuer mismatch: got %s, want %s", disc.Issuer, server.URL)
}
if disc.TokenEndpoint != server.URL+"/token" {
t.Errorf("token endpoint mismatch: got %s", disc.TokenEndpoint)
}
if disc.JWKSUri != server.URL+"/jwks" {
t.Errorf("jwks_uri mismatch: got %s", disc.JWKSUri)
}
}
func TestDiscover_Idempotent(t *testing.T) {
var requestCount int32
var server *httptest.Server
server = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
atomic.AddInt32(&requestCount, 1)
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(fmt.Sprintf(`{"issuer":"%s","authorization_endpoint":"%s/auth","token_endpoint":"%s/token","jwks_uri":"%s/jwks"}`,
server.URL, server.URL, server.URL, server.URL)))
}))
defer server.Close()
client := NewOIDCClient(server.URL, "client_id", "client_secret")
client.httpClient = server.Client()
// First call
_, err := client.Discover(context.Background())
if err != nil {
t.Fatalf("First Discover failed: %v", err)
}
// Second call
_, err = client.Discover(context.Background())
if err != nil {
t.Fatalf("Second Discover failed: %v", err)
}
if atomic.LoadInt32(&requestCount) != 1 {
t.Errorf("Expected 1 HTTP request, got %d", requestCount)
}
}
func generateTestRSAKey(t *testing.T) *rsa.PrivateKey {
t.Helper()
privKey, err := rsa.GenerateKey(rand.Reader, 2048)
if err != nil {
t.Fatalf("failed to generate RSA key: %v", err)
}
return privKey
}
func encodeRSAPublicKey(privKey *rsa.PrivateKey) (n, e string) {
n = base64.RawURLEncoding.EncodeToString(privKey.PublicKey.N.Bytes())
e = base64.RawURLEncoding.EncodeToString(big.NewInt(int64(privKey.PublicKey.E)).Bytes())
return n, e
}
func TestRefreshJWKS_HappyPath(t *testing.T) {
privKey := generateTestRSAKey(t)
n, e := encodeRSAPublicKey(privKey)
var discoveryCalled, jwksCalled bool
var server *httptest.Server
server = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if r.URL.Path == "/.well-known/openid-configuration" {
discoveryCalled = true
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(fmt.Sprintf(`{"issuer":"%s","authorization_endpoint":"%s/auth","token_endpoint":"%s/token","jwks_uri":"%s/jwks"}`,
server.URL, server.URL, server.URL, server.URL)))
return
}
if r.URL.Path == "/jwks" {
jwksCalled = true
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(fmt.Sprintf(`{"keys":[{"kid":"test-key-id","kty":"RSA","use":"sig","alg":"RS256","n":"%s","e":"%s"}]}`, n, e)))
return
}
w.WriteHeader(http.StatusNotFound)
}))
defer server.Close()
client := NewOIDCClient(server.URL, "client_id", "client_secret")
client.httpClient = server.Client()
// First discover to populate discovery
_, err := client.Discover(context.Background())
if err != nil {
t.Fatalf("Discover failed: %v", err)
}
// Now refresh JWKS
err = client.RefreshJWKS(context.Background())
if err != nil {
t.Fatalf("RefreshJWKS failed: %v", err)
}
if !discoveryCalled {
t.Error("discovery endpoint was not called")
}
if !jwksCalled {
t.Error("jwks endpoint was not called")
}
// Check that jwks was populated
client.jwksMu.RLock()
defer client.jwksMu.RUnlock()
if len(client.jwks) != 1 {
t.Errorf("expected 1 key in jwks, got %d", len(client.jwks))
}
if _, exists := client.jwks["test-key-id"]; !exists {
t.Error("test-key-id not found in jwks")
}
}
func TestExchangeCode_HappyPath(t *testing.T) {
tokenResponseJSON := `{"access_token":"access-token-123","id_token":"id-token-456","token_type":"Bearer","expires_in":3600}`
var receivedForm url.Values
var server *httptest.Server
server = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if r.URL.Path == "/.well-known/openid-configuration" {
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(fmt.Sprintf(`{"issuer":"%s","authorization_endpoint":"%s/auth","token_endpoint":"%s/token","jwks_uri":"%s/jwks"}`,
server.URL, server.URL, server.URL, server.URL)))
return
}
if r.URL.Path == "/token" {
if r.Method != http.MethodPost {
t.Errorf("expected POST, got %s", r.Method)
w.WriteHeader(http.StatusMethodNotAllowed)
return
}
if r.Header.Get("Content-Type") != "application/x-www-form-urlencoded" {
t.Errorf("expected Content-Type application/x-www-form-urlencoded, got %s", r.Header.Get("Content-Type"))
w.WriteHeader(http.StatusBadRequest)
return
}
err := r.ParseForm()
if err != nil {
t.Errorf("failed to parse form: %v", err)
w.WriteHeader(http.StatusBadRequest)
return
}
receivedForm = r.Form
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(tokenResponseJSON))
return
}
w.WriteHeader(http.StatusNotFound)
}))
defer server.Close()
client := NewOIDCClient(server.URL, "client_id", "client_secret")
client.httpClient = server.Client()
// Discover first to populate discovery
_, err := client.Discover(context.Background())
if err != nil {
t.Fatalf("Discover failed: %v", err)
}
resp, err := client.ExchangeCode(context.Background(), "auth-code-789", "code-verifier-123", "https://app.example.com/callback")
if err != nil {
t.Fatalf("ExchangeCode failed: %v", err)
}
if resp.AccessToken != "access-token-123" {
t.Errorf("access token mismatch: got %s", resp.AccessToken)
}
if resp.IDToken != "id-token-456" {
t.Errorf("id token mismatch: got %s", resp.IDToken)
}
if resp.TokenType != "Bearer" {
t.Errorf("token type mismatch: got %s", resp.TokenType)
}
// Check form values
if receivedForm.Get("grant_type") != "authorization_code" {
t.Errorf("grant_type mismatch: got %s", receivedForm.Get("grant_type"))
}
if receivedForm.Get("code") != "auth-code-789" {
t.Errorf("code mismatch: got %s", receivedForm.Get("code"))
}
if receivedForm.Get("code_verifier") != "code-verifier-123" {
t.Errorf("code_verifier mismatch: got %s", receivedForm.Get("code_verifier"))
}
if receivedForm.Get("redirect_uri") != "https://app.example.com/callback" {
t.Errorf("redirect_uri mismatch: got %s", receivedForm.Get("redirect_uri"))
}
if receivedForm.Get("client_id") != "client_id" {
t.Errorf("client_id mismatch: got %s", receivedForm.Get("client_id"))
}
if receivedForm.Get("client_secret") != "client_secret" {
t.Errorf("client_secret mismatch: got %s", receivedForm.Get("client_secret"))
}
}
func TestValidateIDToken_HappyPath(t *testing.T) {
privKey := generateTestRSAKey(t)
n, e := encodeRSAPublicKey(privKey)
var server *httptest.Server
server = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if r.URL.Path == "/.well-known/openid-configuration" {
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(fmt.Sprintf(`{"issuer":"%s","authorization_endpoint":"%s/auth","token_endpoint":"%s/token","jwks_uri":"%s/jwks"}`,
server.URL, server.URL, server.URL, server.URL)))
return
}
if r.URL.Path == "/jwks" {
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(fmt.Sprintf(`{"keys":[{"kid":"test-key-id","kty":"RSA","use":"sig","alg":"RS256","n":"%s","e":"%s"}]}`, n, e)))
return
}
w.WriteHeader(http.StatusNotFound)
}))
defer server.Close()
client := NewOIDCClient(server.URL, "client_id", "client_secret")
client.httpClient = server.Client()
// Create and sign a JWT
token := jwt.NewWithClaims(jwt.SigningMethodRS256, &IDTokenClaims{
RegisteredClaims: jwt.RegisteredClaims{
Issuer: server.URL,
Audience: jwt.ClaimStrings{"client_id"},
ExpiresAt: jwt.NewNumericDate(time.Now().Add(time.Hour)),
IssuedAt: jwt.NewNumericDate(time.Now()),
Subject: "user-123",
},
Email: "user@example.com",
EmailVerified: true,
Name: "Test User",
})
token.Header["kid"] = "test-key-id"
signedToken, err := token.SignedString(privKey)
if err != nil {
t.Fatalf("failed to sign token: %v", err)
}
// Validate the token
claims, err := client.ValidateIDToken(context.Background(), signedToken)
if err != nil {
t.Fatalf("ValidateIDToken failed: %v", err)
}
if claims.Issuer != server.URL {
t.Errorf("issuer mismatch: got %s, want %s", claims.Issuer, server.URL)
}
if claims.Subject != "user-123" {
t.Errorf("subject mismatch: got %s", claims.Subject)
}
if claims.Email != "user@example.com" {
t.Errorf("email mismatch: got %s", claims.Email)
}
if !claims.EmailVerified {
t.Error("email_verified should be true")
}
if claims.Name != "Test User" {
t.Errorf("name mismatch: got %s", claims.Name)
}
}
func TestValidateIDToken_RejectsExpired(t *testing.T) {
privKey := generateTestRSAKey(t)
n, e := encodeRSAPublicKey(privKey)
var server *httptest.Server
server = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if r.URL.Path == "/.well-known/openid-configuration" {
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(fmt.Sprintf(`{"issuer":"%s","authorization_endpoint":"%s/auth","token_endpoint":"%s/token","jwks_uri":"%s/jwks"}`,
server.URL, server.URL, server.URL, server.URL)))
return
}
if r.URL.Path == "/jwks" {
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(fmt.Sprintf(`{"keys":[{"kid":"test-key-id","kty":"RSA","use":"sig","alg":"RS256","n":"%s","e":"%s"}]}`, n, e)))
return
}
w.WriteHeader(http.StatusNotFound)
}))
defer server.Close()
client := NewOIDCClient(server.URL, "client_id", "client_secret")
client.httpClient = server.Client()
// Create an expired JWT
token := jwt.NewWithClaims(jwt.SigningMethodRS256, &IDTokenClaims{
RegisteredClaims: jwt.RegisteredClaims{
Issuer: server.URL,
Audience: jwt.ClaimStrings{"client_id"},
ExpiresAt: jwt.NewNumericDate(time.Now().Add(-time.Hour)), // Expired 1 hour ago
IssuedAt: jwt.NewNumericDate(time.Now().Add(-2 * time.Hour)),
Subject: "user-123",
},
})
token.Header["kid"] = "test-key-id"
signedToken, err := token.SignedString(privKey)
if err != nil {
t.Fatalf("failed to sign token: %v", err)
}
// Should fail due to expired token
_, err = client.ValidateIDToken(context.Background(), signedToken)
if err == nil {
t.Error("expected error for expired token, got nil")
}
}
func TestValidateIDToken_RejectsWrongIssuer(t *testing.T) {
privKey := generateTestRSAKey(t)
n, e := encodeRSAPublicKey(privKey)
wrongIssuer := "https://wrong-provider.example.com"
var server *httptest.Server
server = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if r.URL.Path == "/.well-known/openid-configuration" {
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(fmt.Sprintf(`{"issuer":"%s","authorization_endpoint":"%s/auth","token_endpoint":"%s/token","jwks_uri":"%s/jwks"}`,
server.URL, server.URL, server.URL, server.URL)))
return
}
if r.URL.Path == "/jwks" {
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(http.StatusOK)
w.Write([]byte(fmt.Sprintf(`{"keys":[{"kid":"test-key-id","kty":"RSA","use":"sig","alg":"RS256","n":"%s","e":"%s"}]}`, n, e)))
return
}
w.WriteHeader(http.StatusNotFound)
}))
defer server.Close()
client := NewOIDCClient(server.URL, "client_id", "client_secret")
client.httpClient = server.Client()
// Create a JWT with wrong issuer
token := jwt.NewWithClaims(jwt.SigningMethodRS256, &IDTokenClaims{
RegisteredClaims: jwt.RegisteredClaims{
Issuer: wrongIssuer,
Audience: jwt.ClaimStrings{"client_id"},
ExpiresAt: jwt.NewNumericDate(time.Now().Add(time.Hour)),
IssuedAt: jwt.NewNumericDate(time.Now()),
Subject: "user-123",
},
})
token.Header["kid"] = "test-key-id"
signedToken, err := token.SignedString(privKey)
if err != nil {
t.Fatalf("failed to sign token: %v", err)
}
// Should fail due to issuer mismatch
_, err = client.ValidateIDToken(context.Background(), signedToken)
if err == nil {
t.Error("expected error for wrong issuer, got nil")
}
}

180
pkg/bdd/mailpit/client.go
View File

@@ -1,180 +0,0 @@
// Package mailpit is a thin client for the local Mailpit HTTP API,
// used by BDD scenarios to assert on emails sent during a test.
//
// Per ADR-0030 (BDD email parallel strategy), each scenario uses a
// unique recipient address so parallel scenarios cannot interfere.
// The client exposes per-recipient query + delete + await operations.
//
// Production code MUST NOT depend on this package. It lives under
// pkg/bdd/ specifically to signal "test-only".
package mailpit
import (
"context"
"encoding/json"
"fmt"
"net/http"
"net/url"
"strings"
"time"
)
// DefaultBaseURL is the local Mailpit HTTP API root used by the
// docker-compose service (cf. ADR-0029).
const DefaultBaseURL = "http://localhost:8025"
// Client is a Mailpit HTTP API client. Safe for concurrent use.
type Client struct {
BaseURL string
HTTP *http.Client
}
// NewClient returns a Client pointing at the local Mailpit. The HTTP
// client has a 5-second per-call timeout to fail fast in test setups
// where Mailpit is down.
func NewClient() *Client {
return &Client{
BaseURL: DefaultBaseURL,
HTTP: &http.Client{Timeout: 5 * time.Second},
}
}
// Message is the metadata + body view returned by the Mailpit detail
// endpoint. Fields are a subset of what Mailpit returns — only what
// BDD scenarios need to assert on.
type Message struct {
ID string `json:"ID"`
From Address `json:"From"`
To []Address `json:"To"`
Subject string `json:"Subject"`
Text string `json:"Text"`
HTML string `json:"HTML"`
Date time.Time `json:"Date"`
Headers map[string]interface{} `json:"-"` // populated only via the Headers() helper
}
// Address is a Mailpit-formatted email address.
type Address struct {
Name string `json:"Name"`
Address string `json:"Address"`
}
// listResponse is the shape of GET /api/v1/messages.
type listResponse struct {
Messages []messageSummary `json:"messages"`
Total int `json:"total"`
}
type messageSummary struct {
ID string `json:"ID"`
Subject string `json:"Subject"`
Created time.Time `json:"Created"`
}
// MessagesTo returns the list of message IDs currently in Mailpit
// addressed to the given recipient. Empty slice + nil error means
// "no messages yet".
func (c *Client) MessagesTo(ctx context.Context, to string) ([]string, error) {
// Mailpit's /api/v1/search supports the to:<addr> filter ; the more
// obvious-looking /api/v1/messages does NOT (the `query` param there
// is for pagination, not filtering — verified empirically 2026-05-05).
u := fmt.Sprintf("%s/api/v1/search?query=%s",
strings.TrimRight(c.BaseURL, "/"),
url.QueryEscape("to:"+to))
req, err := http.NewRequestWithContext(ctx, http.MethodGet, u, nil)
if err != nil {
return nil, err
}
resp, err := c.HTTP.Do(req)
if err != nil {
return nil, fmt.Errorf("mailpit list: %w", err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return nil, fmt.Errorf("mailpit list: HTTP %d", resp.StatusCode)
}
var list listResponse
if err := json.NewDecoder(resp.Body).Decode(&list); err != nil {
return nil, fmt.Errorf("mailpit list decode: %w", err)
}
ids := make([]string, 0, len(list.Messages))
for _, m := range list.Messages {
ids = append(ids, m.ID)
}
return ids, nil
}
// Get fetches the full content of the message with the given ID.
func (c *Client) Get(ctx context.Context, id string) (*Message, error) {
u := fmt.Sprintf("%s/api/v1/message/%s",
strings.TrimRight(c.BaseURL, "/"),
url.PathEscape(id))
req, err := http.NewRequestWithContext(ctx, http.MethodGet, u, nil)
if err != nil {
return nil, err
}
resp, err := c.HTTP.Do(req)
if err != nil {
return nil, fmt.Errorf("mailpit get: %w", err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return nil, fmt.Errorf("mailpit get %s: HTTP %d", id, resp.StatusCode)
}
var m Message
if err := json.NewDecoder(resp.Body).Decode(&m); err != nil {
return nil, fmt.Errorf("mailpit get decode: %w", err)
}
return &m, nil
}
// AwaitMessageTo polls Mailpit for a message addressed to the given
// recipient. Returns the most recent matching message ; errors out if
// the timeout elapses with no match. Polls every 50ms — Mailpit is
// fast enough that this is rarely the limiting factor.
//
// Use this in BDD steps "Then I should receive an email ...".
func (c *Client) AwaitMessageTo(ctx context.Context, to string, timeout time.Duration) (*Message, error) {
deadline := time.Now().Add(timeout)
for time.Now().Before(deadline) {
ids, err := c.MessagesTo(ctx, to)
if err == nil && len(ids) > 0 {
// Most recent first per Mailpit's default sort
return c.Get(ctx, ids[0])
}
select {
case <-ctx.Done():
return nil, ctx.Err()
case <-time.After(50 * time.Millisecond):
}
}
return nil, fmt.Errorf("mailpit: no message for %s within %s", to, timeout)
}
// PurgeMessagesTo deletes every message addressed to the given recipient.
// Idempotent: calling against an empty inbox is fine.
//
// Use this at the start of a BDD scenario to clear leftovers from
// prior runs of the same scenario (rare given the random suffix per
// scenario, but defensive).
func (c *Client) PurgeMessagesTo(ctx context.Context, to string) error {
// Mailpit's /api/v1/search supports the to:<addr> filter ; the more
// obvious-looking /api/v1/messages does NOT (the `query` param there
// is for pagination, not filtering — verified empirically 2026-05-05).
u := fmt.Sprintf("%s/api/v1/search?query=%s",
strings.TrimRight(c.BaseURL, "/"),
url.QueryEscape("to:"+to))
req, err := http.NewRequestWithContext(ctx, http.MethodDelete, u, nil)
if err != nil {
return err
}
resp, err := c.HTTP.Do(req)
if err != nil {
return fmt.Errorf("mailpit delete: %w", err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK && resp.StatusCode != http.StatusNoContent {
return fmt.Errorf("mailpit delete: HTTP %d", resp.StatusCode)
}
return nil
}

133
pkg/bdd/mailpit/client_integration_test.go
View File

@@ -1,133 +0,0 @@
//go:build integration
// Integration tests for the Mailpit client. Run with:
//
// go test -tags integration ./pkg/bdd/mailpit/...
//
// Requires a running Mailpit reachable at http://localhost:8025
// (the docker-compose service from ADR-0029).
package mailpit
import (
"context"
"crypto/rand"
"encoding/hex"
"net/smtp"
"strings"
"testing"
"time"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// uniqueRecipient returns an address unique to this test run, using the
// per-scenario-recipient pattern from ADR-0030. Two parallel test runs
// generate different suffixes so they never see each other's messages.
func uniqueRecipient(t *testing.T) string {
t.Helper()
var raw [4]byte
_, err := rand.Read(raw[:])
require.NoError(t, err)
return "integ-" + t.Name() + "-" + hex.EncodeToString(raw[:]) + "@bdd.local"
}
// sendViaSMTP submits a small email through Mailpit's SMTP port.
// Real-wire-format path : same as the application code will use.
func sendViaSMTP(t *testing.T, to, subject, body string) {
t.Helper()
from := "integ-test@bdd.local"
msg := []byte(
"From: " + from + "\r\n" +
"To: " + to + "\r\n" +
"Subject: " + subject + "\r\n" +
"\r\n" +
body + "\r\n",
)
err := smtp.SendMail("localhost:1025", nil, from, []string{to}, msg)
require.NoError(t, err, "SMTP send to local Mailpit")
}
// TestIntegration_RoundTrip validates the full path : SMTP submit →
// Mailpit captures → client lists → client gets full body. This is
// the smoke test for the BDD-helper contract.
func TestIntegration_RoundTrip(t *testing.T) {
c := NewClient()
to := uniqueRecipient(t)
// Defensive cleanup before the test (in case the recipient was reused)
require.NoError(t, c.PurgeMessagesTo(context.Background(), to))
subject := "Integration roundtrip"
body := "Token: integ-token-" + strings.ReplaceAll(to, "@", "-at-")
sendViaSMTP(t, to, subject, body)
msg, err := c.AwaitMessageTo(context.Background(), to, 3*time.Second)
require.NoError(t, err)
require.NotNil(t, msg)
assert.Equal(t, subject, msg.Subject)
assert.Contains(t, msg.Text, "Token: integ-token-")
if assert.Len(t, msg.To, 1) {
assert.Equal(t, to, msg.To[0].Address)
}
// Cleanup so subsequent runs of this same test name don't accumulate
require.NoError(t, c.PurgeMessagesTo(context.Background(), to))
}
// TestIntegration_AwaitTimeoutWhenNoMessage confirms AwaitMessageTo
// returns an error within the timeout when no message arrives.
func TestIntegration_AwaitTimeoutWhenNoMessage(t *testing.T) {
c := NewClient()
to := uniqueRecipient(t) // never sent to → must time out
start := time.Now()
_, err := c.AwaitMessageTo(context.Background(), to, 200*time.Millisecond)
elapsed := time.Since(start)
require.Error(t, err)
assert.Contains(t, err.Error(), "no message")
assert.GreaterOrEqual(t, elapsed, 150*time.Millisecond, "should poll until close to timeout")
assert.Less(t, elapsed, 1*time.Second, "should not exceed timeout substantially")
}
// TestIntegration_PurgeIsolation proves the per-recipient query/delete
// model from ADR-0030 : two unique recipients can have their own
// messages without one's purge affecting the other.
func TestIntegration_PurgeIsolation(t *testing.T) {
c := NewClient()
// Build two distinct, well-formed addresses (separate local-parts,
// same domain). Avoid mutating uniqueRecipient's output post-@.
var rawA, rawB [4]byte
_, _ = rand.Read(rawA[:])
_, _ = rand.Read(rawB[:])
toA := "iso-a-" + hex.EncodeToString(rawA[:]) + "@bdd.local"
toB := "iso-b-" + hex.EncodeToString(rawB[:]) + "@bdd.local"
sendViaSMTP(t, toA, "for A", "body A")
sendViaSMTP(t, toB, "for B", "body B")
// Both messages should exist
idsA, err := c.MessagesTo(context.Background(), toA)
require.NoError(t, err)
assert.GreaterOrEqual(t, len(idsA), 1, "A should have its message")
idsB, err := c.MessagesTo(context.Background(), toB)
require.NoError(t, err)
assert.GreaterOrEqual(t, len(idsB), 1, "B should have its message")
// Purge A only
require.NoError(t, c.PurgeMessagesTo(context.Background(), toA))
// A is empty, B is untouched
idsA, err = c.MessagesTo(context.Background(), toA)
require.NoError(t, err)
assert.Empty(t, idsA, "A should be empty after purge")
idsB, err = c.MessagesTo(context.Background(), toB)
require.NoError(t, err)
assert.GreaterOrEqual(t, len(idsB), 1, "B should still have its message")
// Cleanup B
require.NoError(t, c.PurgeMessagesTo(context.Background(), toB))
}

67
pkg/bdd/steps/common_steps.go
View File

@@ -2,7 +2,6 @@ package steps
import (
"fmt"
"regexp"
"strings"
"dance-lessons-coach/pkg/bdd/testserver"
@@ -100,69 +99,3 @@ func (s *CommonSteps) theFieldShouldEqual(field, expectedValue string) error {
}
return nil
}
// Regex field matching
func (s *CommonSteps) theFieldShouldMatch(field, pattern string) error {
body := string(s.client.GetLastBody())
// Extract the value of the field from JSON
// Look for "field":"value" and extract value
fieldPattern := `"` + field + `":"([^"]*)"`
re := regexp.MustCompile(fieldPattern)
matches := re.FindStringSubmatch(body)
if matches == nil {
// Try without quotes (for numbers)
fieldPatternNum := `"` + field + `":(\d+\.?\d*)`
reNum := regexp.MustCompile(fieldPatternNum)
matches = reNum.FindStringSubmatch(body)
if matches == nil {
return fmt.Errorf("field %q not found in response: %s", field, body)
}
}
// matches[1] contains the value
value := matches[1]
// Compile and match the pattern
regex, err := regexp.Compile(pattern)
if err != nil {
return fmt.Errorf("invalid regex pattern %q: %v", pattern, err)
}
if !regex.MatchString(value) {
return fmt.Errorf("field %q value %q does not match pattern %q", field, value, pattern)
}
return nil
}
// Response is JSON check
func (s *CommonSteps) theResponseShouldBeJSON() error {
body := string(s.client.GetLastBody())
// Simple check for JSON structure
body = strings.TrimSpace(body)
if !strings.HasPrefix(body, "{") && !strings.HasPrefix(body, "[") {
return fmt.Errorf("response is not JSON: %s", body)
}
return nil
}
// Response contains field (simple string containment in body)
func (s *CommonSteps) theResponseShouldContain(field string) error {
body := string(s.client.GetLastBody())
if !strings.Contains(body, `"`+field+`"`) {
return fmt.Errorf("response does not contain field %q: %s", field, body)
}
return nil
}
// Response header validation
func (s *CommonSteps) theResponseHeader(header, expectedValue string) error {
resp := s.client.GetLastResponse()
if resp == nil {
return fmt.Errorf("no response captured for header check")
}
headerValue := resp.Header.Get(header)
if headerValue != expectedValue {
return fmt.Errorf("header %q expected %q, got %q", header, expectedValue, headerValue)
}
return nil
}

12
pkg/bdd/steps/health_steps.go
View File

@@ -28,19 +28,7 @@ func (s *HealthSteps) iRequestTheHealthzEndpoint() error {
return s.client.Request("GET", "/api/healthz", nil)
}
func (s *HealthSteps) iRequestTheInfoEndpoint() error {
return s.client.Request("GET", "/api/info", nil)
}
func (s *HealthSteps) iRequestTheInfoEndpointAgain() error {
return s.client.Request("GET", "/api/info", nil)
}
func (s *HealthSteps) theServerIsRunning() error {
// Actually verify the server is running by checking the readiness endpoint
return s.client.Request("GET", "/api/ready", nil)
}
func (s *HealthSteps) theServerIsRunningWithCacheEnabled() error {
return s.client.Request("GET", "/api/ready", nil)
}

58
pkg/bdd/steps/jwt_retention_steps.go
View File

@@ -822,61 +822,3 @@ func (s *JWTRetentionSteps) andSuggestRemediationSteps() error {
// Verify remediation suggestions
return godog.ErrPending
}
// =====================================================================
// Admin metadata introspection steps (PR #51 + this scenario)
// =====================================================================
// iAddASecondaryJWTSecretNamed adds a secret with a specific value via the
// admin API. Used by the admin-introspection scenario to verify that the
// metadata endpoint returns metadata only, not the secret value.
func (s *JWTRetentionSteps) iAddASecondaryJWTSecretNamed(secretValue string) error {
s.SetLastSecret(secretValue)
return s.client.Request("POST", "/api/v1/admin/jwt/secrets", map[string]string{
"secret": secretValue,
"is_primary": "false",
})
}
// iRequestTheJWTSecretsMetadataEndpoint hits GET /api/v1/admin/jwt/secrets.
func (s *JWTRetentionSteps) iRequestTheJWTSecretsMetadataEndpoint() error {
return s.client.Request("GET", "/api/v1/admin/jwt/secrets", nil)
}
// theMetadataShouldContainNSecrets verifies the response count field.
func (s *JWTRetentionSteps) theMetadataShouldContainNSecrets(expected int) error {
body := string(s.client.GetLastBody())
expectedFragment := `"count":` + strconv.Itoa(expected)
if !strings.Contains(body, expectedFragment) {
return fmt.Errorf("expected response to contain %q, got: %s", expectedFragment, body)
}
return nil
}
// theMetadataShouldNotContainTheSecretValue is the SECURITY-CRITICAL
// assertion. If the response contains the raw secret string anywhere,
// the endpoint has leaked. This is the property the metadata-only design
// is supposed to guarantee.
func (s *JWTRetentionSteps) theMetadataShouldNotContainTheSecretValue(secretValue string) error {
body := string(s.client.GetLastBody())
if strings.Contains(body, secretValue) {
return fmt.Errorf("SECURITY: response leaked the secret value %q (response body: %s)", secretValue, body)
}
return nil
}
// everySecretInTheMetadataShouldHaveASHA256Fingerprint asserts the
// secret_sha256 field is present and non-empty for each entry. Cheap
// regex-style check on the JSON body.
func (s *JWTRetentionSteps) everySecretInTheMetadataShouldHaveASHA256Fingerprint() error {
body := string(s.client.GetLastBody())
// Expect at least one occurrence of "secret_sha256":"<non-empty>"
if !strings.Contains(body, `"secret_sha256":"`) {
return fmt.Errorf("response does not include any secret_sha256 fingerprint: %s", body)
}
// Reject obviously-empty values
if strings.Contains(body, `"secret_sha256":""`) {
return fmt.Errorf("at least one secret_sha256 fingerprint is empty in response: %s", body)
}
return nil
}

145
pkg/bdd/steps/magic_link_steps.go
View File

@@ -1,145 +0,0 @@
package steps
import (
"context"
"crypto/rand"
"encoding/hex"
"encoding/json"
"fmt"
"net/http"
"regexp"
"strings"
"time"
"dance-lessons-coach/pkg/bdd/mailpit"
"dance-lessons-coach/pkg/bdd/testserver"
)
type MagicLinkSteps struct {
client *testserver.Client
mailpit *mailpit.Client
scenarioKey string
}
func NewMagicLinkSteps(client *testserver.Client) *MagicLinkSteps {
return &MagicLinkSteps{client: client, mailpit: mailpit.NewClient()}
}
func (s *MagicLinkSteps) SetScenarioKey(key string) { s.scenarioKey = key }
func (s *MagicLinkSteps) state() *ScenarioState {
if s.scenarioKey == "" {
s.scenarioKey = "default"
}
return GetScenarioState(s.scenarioKey)
}
// sanitizeForEmail keeps only [a-z0-9-] from the scenario key
func sanitizeForEmail(s string) string {
if s == "" {
return "scn"
}
var b strings.Builder
for _, r := range strings.ToLower(s) {
if (r >= 'a' && r <= 'z') || (r >= '0' && r <= '9') || r == '-' {
b.WriteRune(r)
}
}
if b.Len() == 0 {
return "scn"
}
if b.Len() > 24 {
return b.String()[:24]
}
return b.String()
}
// IHaveAnEmailAddressForThisScenario generates per-scenario unique recipient and stashes it in state.
// Defensively purges Mailpit for that address.
// Format: <scenario-key>-<8hex>@bdd.local (cf. ADR-0030)
func (s *MagicLinkSteps) IHaveAnEmailAddressForThisScenario() error {
var raw [4]byte
if _, err := rand.Read(raw[:]); err != nil {
return err
}
addr := fmt.Sprintf("ml-%s-%s@bdd.local",
sanitizeForEmail(s.scenarioKey), hex.EncodeToString(raw[:]))
s.state().MagicLinkEmail = addr
return s.mailpit.PurgeMessagesTo(context.Background(), addr)
}
// IRequestAMagicLinkForMyEmail POSTs to /api/v1/auth/magic-link/request with the scenario's email.
func (s *MagicLinkSteps) IRequestAMagicLinkForMyEmail() error {
return s.client.Request("POST", "/api/v1/auth/magic-link/request",
map[string]string{"email": s.state().MagicLinkEmail})
}
// IShouldReceiveAnEmailWithSubject waits for an email at the scenario's address; asserts subject equality.
func (s *MagicLinkSteps) IShouldReceiveAnEmailWithSubject(subject string) error {
msg, err := s.mailpit.AwaitMessageTo(context.Background(), s.state().MagicLinkEmail, 5*time.Second)
if err != nil {
return fmt.Errorf("mailpit await: %w", err)
}
if msg.Subject != subject {
return fmt.Errorf("expected subject %q, got %q", subject, msg.Subject)
}
return nil
}
// TheEmailContainsAMagicLinkToken re-fetches most recent message, extracts token via regex, stashes in state.
var tokenRe = regexp.MustCompile(`\?token=([A-Za-z0-9_\-]+)`)
func (s *MagicLinkSteps) TheEmailContainsAMagicLinkToken() error {
msg, err := s.mailpit.AwaitMessageTo(context.Background(), s.state().MagicLinkEmail, 5*time.Second)
if err != nil {
return err
}
m := tokenRe.FindStringSubmatch(msg.Text)
if m == nil {
return fmt.Errorf("no token in email body: %q", msg.Text)
}
s.state().MagicLinkToken = m[1]
return nil
}
// IConsumeTheMagicLinkToken GETs /api/v1/auth/magic-link/consume?token=<plain>
func (s *MagicLinkSteps) IConsumeTheMagicLinkToken() error {
return s.client.Request("GET",
"/api/v1/auth/magic-link/consume?token="+s.state().MagicLinkToken, nil)
}
// TheConsumeShouldSucceedAndReturnAJWT asserts 200 + JWT body.
func (s *MagicLinkSteps) TheConsumeShouldSucceedAndReturnAJWT() error {
if c := s.client.GetLastStatusCode(); c != http.StatusOK {
return fmt.Errorf("expected 200, got %d body=%s", c, s.client.GetLastBody())
}
var resp struct {
Token string `json:"token"`
}
if err := json.Unmarshal(s.client.GetLastBody(), &resp); err != nil {
return err
}
if resp.Token == "" {
return fmt.Errorf("empty JWT in response")
}
s.state().LastToken = resp.Token
return nil
}
// TheConsumeShouldFailWith401 asserts 401.
func (s *MagicLinkSteps) TheConsumeShouldFailWith401() error {
if c := s.client.GetLastStatusCode(); c != http.StatusUnauthorized {
return fmt.Errorf("expected 401, got %d body=%s", c, s.client.GetLastBody())
}
return nil
}
// IConsumeAnEmptyMagicLinkToken consumes with an empty token
func (s *MagicLinkSteps) IConsumeAnEmptyMagicLinkToken() error {
return s.client.Request("GET", "/api/v1/auth/magic-link/consume?token=", nil)
}
// IConsumeAnUnknownMagicLinkToken consumes with a non-existent token
func (s *MagicLinkSteps) IConsumeAnUnknownMagicLinkToken() error {
return s.client.Request("GET", "/api/v1/auth/magic-link/consume?token=unknown-token-12345", nil)
}

12
pkg/bdd/steps/scenario_state.go
View File

@@ -9,13 +9,11 @@ import (
// ScenarioState holds per-scenario state for step definitions
// This prevents state pollution between scenarios running in the same test process
type ScenarioState struct {
LastToken string
FirstToken string
LastUserID uint
LastSecret string
LastError string
MagicLinkEmail string
MagicLinkToken string
LastToken string
FirstToken string
LastUserID uint
LastSecret string
LastError string
// Add more fields as needed for other step types
}

29
pkg/bdd/steps/steps.go
View File

@@ -17,7 +17,6 @@ type StepContext struct {
jwtRetentionSteps *JWTRetentionSteps
configSteps *ConfigSteps
rateLimitSteps *RateLimitSteps
magicLinkSteps *MagicLinkSteps
}
// NewStepContext creates a new step context
@@ -31,7 +30,6 @@ func NewStepContext(client *testserver.Client) *StepContext {
jwtRetentionSteps: NewJWTRetentionSteps(client),
configSteps: NewConfigSteps(client),
rateLimitSteps: NewRateLimitSteps(client),
magicLinkSteps: NewMagicLinkSteps(client),
}
}
@@ -69,9 +67,6 @@ func SetScenarioKeyForAllSteps(sc *StepContext, key string) {
if sc.rateLimitSteps != nil {
sc.rateLimitSteps.SetScenarioKey(key)
}
if sc.magicLinkSteps != nil {
sc.magicLinkSteps.SetScenarioKey(key)
}
}
}
@@ -94,9 +89,6 @@ func InitializeAllSteps(ctx *godog.ScenarioContext, client *testserver.Client, s
// Health steps
ctx.Step(`^I request the health endpoint$`, sc.healthSteps.iRequestTheHealthEndpoint)
ctx.Step(`^I request the healthz endpoint$`, sc.healthSteps.iRequestTheHealthzEndpoint)
ctx.Step(`^I request the info endpoint$`, sc.healthSteps.iRequestTheInfoEndpoint)
ctx.Step(`^I request the info endpoint again$`, sc.healthSteps.iRequestTheInfoEndpointAgain)
ctx.Step(`^the server is running with cache enabled$`, sc.healthSteps.theServerIsRunningWithCacheEnabled)
ctx.Step(`^the server is running$`, sc.healthSteps.theServerIsRunning)
// Auth steps
@@ -178,12 +170,6 @@ func InitializeAllSteps(ctx *godog.ScenarioContext, client *testserver.Client, s
ctx.Step(`^I should receive configuration validation error$`, sc.jwtRetentionSteps.iShouldReceiveConfigurationValidationError)
ctx.Step(`^the error should mention "([^"]*)"$`, sc.jwtRetentionSteps.theErrorShouldMention)
ctx.Step(`^I have enabled Prometheus metrics$`, sc.jwtRetentionSteps.iHaveEnabledPrometheusMetrics)
// Admin metadata introspection steps (PR #51 + admin-introspection scenario)
ctx.Step(`^I add a secondary JWT secret "([^"]*)"$`, sc.jwtRetentionSteps.iAddASecondaryJWTSecretNamed)
ctx.Step(`^I request the JWT secrets metadata endpoint$`, sc.jwtRetentionSteps.iRequestTheJWTSecretsMetadataEndpoint)
ctx.Step(`^the metadata should contain (\d+) secrets$`, sc.jwtRetentionSteps.theMetadataShouldContainNSecrets)
ctx.Step(`^the metadata should NOT contain the secret value "([^"]*)"$`, sc.jwtRetentionSteps.theMetadataShouldNotContainTheSecretValue)
ctx.Step(`^every secret in the metadata should have a SHA-256 fingerprint$`, sc.jwtRetentionSteps.everySecretInTheMetadataShouldHaveASHA256Fingerprint)
ctx.Step(`^I should see "([^"]*)" metric increment$`, sc.jwtRetentionSteps.iShouldSeeMetricIncrement)
ctx.Step(`^I should see "([^"]*)" metric decrease$`, sc.jwtRetentionSteps.iShouldSeeMetricDecrease)
ctx.Step(`^I should see "([^"]*)" histogram update$`, sc.jwtRetentionSteps.iShouldSeeHistogramUpdate)
@@ -322,25 +308,10 @@ func InitializeAllSteps(ctx *godog.ScenarioContext, client *testserver.Client, s
ctx.Step(`^the response body should contain "([^"]*)"$`, sc.rateLimitSteps.theResponseBodyShouldContain)
ctx.Step(`^the response should have header "([^"]*)"$`, sc.rateLimitSteps.theResponseShouldHaveHeader)
// Magic link steps
ctx.Step(`^I have an email address for this scenario$`, sc.magicLinkSteps.IHaveAnEmailAddressForThisScenario)
ctx.Step(`^I request a magic link for my email$`, sc.magicLinkSteps.IRequestAMagicLinkForMyEmail)
ctx.Step(`^I should receive an email with subject "([^"]*)"$`, sc.magicLinkSteps.IShouldReceiveAnEmailWithSubject)
ctx.Step(`^the email contains a magic link token$`, sc.magicLinkSteps.TheEmailContainsAMagicLinkToken)
ctx.Step(`^I consume the magic link token$`, sc.magicLinkSteps.IConsumeTheMagicLinkToken)
ctx.Step(`^the consume should succeed and return a JWT$`, sc.magicLinkSteps.TheConsumeShouldSucceedAndReturnAJWT)
ctx.Step(`^the consume should fail with 401$`, sc.magicLinkSteps.TheConsumeShouldFailWith401)
ctx.Step(`^I consume an empty magic link token$`, sc.magicLinkSteps.IConsumeAnEmptyMagicLinkToken)
ctx.Step(`^I consume an unknown magic link token$`, sc.magicLinkSteps.IConsumeAnUnknownMagicLinkToken)
// Common steps
ctx.Step(`^the response should be "{\\"([^"]*)":\\"([^"]*)"}"$`, sc.commonSteps.theResponseShouldBe)
ctx.Step(`^the response should contain error "([^"]*)"$`, sc.commonSteps.theResponseShouldContainError)
ctx.Step(`^the status code should be (\d+)$`, sc.commonSteps.theStatusCodeShouldBe)
ctx.Step(`^the response should be JSON with fields "([^"]*)"$`, sc.commonSteps.theResponseShouldBeJSONWithFields)
ctx.Step(`^the "([^"]*)" field should equal "([^"]*)"$`, sc.commonSteps.theFieldShouldEqual)
ctx.Step(`^the "([^"]*)" field should match /([^/]+)/$`, sc.commonSteps.theFieldShouldMatch)
ctx.Step(`^the response should be JSON$`, sc.commonSteps.theResponseShouldBeJSON)
ctx.Step(`^the response should contain "([^"]*)"$`, sc.commonSteps.theResponseShouldContain)
ctx.Step(`^the response header "([^"]*)" should be "([^"]*)"$`, sc.commonSteps.theResponseHeader)
}

9
pkg/bdd/suite.go
View File

@@ -115,15 +115,6 @@ func InitializeTestSuite(ctx *godog.TestSuiteContext) {
testserver.TraceStateJWTSecretOperation(feature, scenarioKey, "RESET", "ok")
}
// Flush cache after every scenario to prevent cache pollution
if flushErr := sharedServer.FlushCache(); flushErr != nil {
if isCleanupLoggingEnabled() {
log.Warn().Err(flushErr).Msg("CLEANUP: Failed to flush cache after scenario")
}
} else {
testserver.TraceStateCacheOperation(feature, scenarioKey, "FLUSH", "ok")
}
// Clean database after every scenario (only if schema isolation is disabled)
if !isSchemaIsolationEnabled() {
if cleanupErr := sharedServer.CleanupDatabase(); cleanupErr != nil {

137
pkg/bdd/testserver/server.go
View File

@@ -15,7 +15,6 @@ import (
"sync"
"time"
"dance-lessons-coach/pkg/cache"
"dance-lessons-coach/pkg/config"
"dance-lessons-coach/pkg/server"
"dance-lessons-coach/pkg/user"
@@ -48,13 +47,10 @@ type Server struct {
port int
baseURL string
db *sql.DB
authService user.AuthService // Reference to auth service for cleanup
cacheService cache.Service // Reference to cache service for cleanup
isolatedRepo *user.PostgresRepository // Per-package isolated repo (BDD_SCHEMA_ISOLATION=true)
isolatedSchemaName string // Per-package schema name to drop on Stop()
schemaMutex sync.Mutex // Protects schema operations
currentSchema string // Current schema being used
originalSearchPath string // Original search_path to restore
authService user.AuthService // Reference to auth service for cleanup
schemaMutex sync.Mutex // Protects schema operations
currentSchema string // Current schema being used
originalSearchPath string // Original search_path to restore
}
// getDatabaseHost returns the database host from environment variable or defaults to localhost
@@ -150,62 +146,13 @@ func (s *Server) Start() error {
// This is the ONLY place where we check env vars for v2 configuration
v2Enabled := s.shouldEnableV2()
// Create real server instance from pkg/server.
// When BDD_SCHEMA_ISOLATION=true, each test package (process) gets its own
// isolated PostgreSQL schema with its own connection pool + migrations.
// This makes `go test ./features/...` parallel-safe because each feature
// package runs in its own process and gets its own schema.
// Create real server instance from pkg/server
cfg := createTestConfig(s.port, v2Enabled)
var realServer *server.Server
if isSchemaIsolationEnabled() {
feature := os.Getenv("FEATURE")
if feature == "" {
feature = "bdd"
}
schemaName := generateSchemaName(feature, "package_root")
log.Info().Str("schema", schemaName).Str("feature", feature).Msg("ISOLATION: Building per-package isolated repo")
// Connect a default repo briefly just to CREATE SCHEMA (uses cfg from env vars)
bootstrapRepo, err := user.NewPostgresRepository(cfg)
if err != nil {
return fmt.Errorf("ISOLATION bootstrap repo failed: %w", err)
}
// Drop + recreate to ensure clean slate per process
_ = bootstrapRepo.Exec(fmt.Sprintf("DROP SCHEMA IF EXISTS %s CASCADE", schemaName))
if err := bootstrapRepo.Exec(fmt.Sprintf("CREATE SCHEMA %s", schemaName)); err != nil {
bootstrapRepo.Close()
return fmt.Errorf("ISOLATION CREATE SCHEMA failed: %w", err)
}
bootstrapRepo.Close()
// Build the per-package isolated repo (runs migrations in the new schema)
dsn := user.BuildSchemaIsolatedDSN(cfg, schemaName)
isolatedRepo, err := user.NewPostgresRepositoryFromDSN(cfg, dsn)
if err != nil {
return fmt.Errorf("ISOLATION isolated repo failed: %w", err)
}
s.isolatedRepo = isolatedRepo
s.isolatedSchemaName = schemaName
// Build user service backed by the isolated repo
jwtConfig := user.JWTConfig{
Secret: cfg.GetJWTSecret(),
ExpirationTime: time.Hour * 24,
Issuer: "dance-lessons-coach",
}
isolatedUserService := user.NewUserService(isolatedRepo, jwtConfig, cfg.GetAdminMasterPassword())
realServer = server.NewServerWithUserRepo(cfg, context.Background(), isolatedRepo, isolatedUserService)
} else {
realServer = server.NewServer(cfg, context.Background())
}
realServer := server.NewServer(cfg, context.Background())
// Store auth service for cleanup
s.authService = realServer.GetAuthService()
// Store cache service for cleanup
s.cacheService = realServer.GetCacheService()
// Initialize database connection for cleanup
if err := s.initDBConnection(); err != nil {
return fmt.Errorf("failed to initialize database connection: %w", err)
@@ -462,23 +409,6 @@ func (s *Server) ResetJWTSecrets() error {
return nil
}
// FlushCache clears all cached data to prevent cache pollution between scenarios
// This prevents cached responses from affecting subsequent test scenarios
func (s *Server) FlushCache() error {
if s.cacheService == nil {
if isCleanupLoggingEnabled() {
log.Info().Msg("CLEANUP: No cache service available, skipping cache flush")
}
return nil
}
s.cacheService.Flush()
if isCleanupLoggingEnabled() {
log.Info().Msg("CLEANUP: Cache flushed successfully")
}
return nil
}
// CleanupDatabase deletes all test data from all tables
// This uses raw SQL to avoid dependency on repositories and handles foreign keys properly
// Uses SET CONSTRAINTS ALL DEFERRED to temporarily disable foreign key checks
@@ -625,7 +555,7 @@ func (s *Server) SetupScenarioSchema(feature, scenario string) error {
return fmt.Errorf("failed to create schema %s: %w", schemaName, err)
}
// Set search path to use the new schema (testserver's own connection)
// Set search path to use the new schema
searchPathSQL := fmt.Sprintf("SET search_path = %s, %s", schemaName, s.originalSearchPath)
if _, err := s.db.Exec(searchPathSQL); err != nil {
return fmt.Errorf("failed to set search_path: %w", err)
@@ -687,21 +617,6 @@ func (s *Server) getCurrentSearchPath() (string, error) {
}
func (s *Server) Stop() error {
// Cleanup the per-package isolated schema + close its pool, if any.
// (BDD_SCHEMA_ISOLATION=true path - see Start().)
if s.isolatedRepo != nil {
if s.isolatedSchemaName != "" {
if err := s.isolatedRepo.Exec(fmt.Sprintf("DROP SCHEMA IF EXISTS %s CASCADE", s.isolatedSchemaName)); err != nil {
log.Warn().Err(err).Str("schema", s.isolatedSchemaName).Msg("ISOLATION: failed to drop schema on Stop")
}
}
if err := s.isolatedRepo.Close(); err != nil {
log.Warn().Err(err).Msg("ISOLATION: failed to close isolated repo")
}
s.isolatedRepo = nil
s.isolatedSchemaName = ""
}
if s.httpServer != nil {
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
@@ -741,14 +656,8 @@ func (s *Server) waitForServerReady() error {
}
}
// shouldEnableV2 determines if v2 API should be enabled for this test server.
// This is the ONLY place that reads FEATURE and GODOG_TAGS env vars.
//
// 2026-05-05: previous version used strings.Contains(tags, "@v2") which
// wrongly matched the negation `~@v2` as well. This made the "v1" greet
// sub-test (tags `~@v2 && ~@skip`) actually run with v2 enabled, masking
// the gate behavior we now test in feature `@v2-gate` scenario. Fixed
// here by inspecting each && clause and checking for positive inclusion.
// shouldEnableV2 determines if v2 API should be enabled for this test server
// This is the ONLY place that reads FEATURE and GODOG_TAGS env vars
func (s *Server) shouldEnableV2() bool {
feature := os.Getenv("FEATURE")
@@ -759,19 +668,9 @@ func (s *Server) shouldEnableV2() bool {
return false
}
// For greet feature: enable v2 if tags include `@v2` as a POSITIVE clause.
// Godog tag expression syntax: clauses separated by `&&` or `||`, negation
// via leading `~`. A positive clause matches exactly `@v2` (after trim).
// For greet feature: enable v2 if tags include @v2
tags := os.Getenv("GODOG_TAGS")
for _, clause := range strings.FieldsFunc(tags, func(r rune) bool {
return r == '&' || r == '|' || r == ' '
}) {
clause = strings.TrimSpace(clause)
if clause == "@v2" {
return true
}
}
return false
return strings.Contains(tags, "@v2")
}
// createTestConfig creates a test configuration
@@ -815,20 +714,6 @@ func createTestConfig(port int, v2Enabled bool) *config.Config {
JWT: config.JWTConfig{
TTL: 24 * time.Hour,
},
// Email + MagicLink defaults so the magic-link BDD scenarios
// (ADR-0028 Phase A.5) can send to local Mailpit. Without these
// the literal Config skips Viper's SetDefault and From stays
// empty — pkg/email then rejects the message.
Email: config.EmailConfig{
From: "noreply@bdd.local",
SMTPHost: "localhost",
SMTPPort: 1025,
Timeout: 5 * time.Second,
},
MagicLink: config.MagicLinkConfig{
TTL: 15 * time.Minute,
BaseURL: "http://localhost:8080",
},
},
API: config.APIConfig{
V2Enabled: v2Enabled,

5
pkg/bdd/testserver/state_tracer.go
View File

@@ -31,11 +31,6 @@ func TraceStateJWTSecretOperation(feature, scenario, operation, details string)
writeTraceLine(feature, scenario, "JWT_"+operation, details)
}
// TraceStateCacheOperation logs a cache operation
func TraceStateCacheOperation(feature, scenario, operation, details string) {
writeTraceLine(feature, scenario, "CACHE_"+operation, details)
}
// TraceStateSchemaIsolation logs a schema isolation operation
func TraceStateSchemaIsolation(feature, scenario, operation, details string) {
writeTraceLine(feature, scenario, "SCHEMA_"+operation, details)

269
pkg/config/config.go
View File

@@ -1,14 +1,11 @@
package config
import (
"context"
"fmt"
"os"
"strings"
"sync"
"time"
"github.com/fsnotify/fsnotify"
"github.com/rs/zerolog"
"github.com/rs/zerolog/log"
"github.com/spf13/viper"
@@ -16,13 +13,6 @@ import (
"dance-lessons-coach/pkg/version"
)
// SamplerReconfigureFunc is the signature for callbacks invoked when
// telemetry.sampler.type or telemetry.sampler.ratio change via hot-reload.
// The callback receives the new sampler type and ratio values.
// It must be safe to call concurrently — implementations should use their
// own synchronisation if needed. Returns an error if the reconfigure fails.
type SamplerReconfigureFunc func(ctx context.Context, samplerType string, samplerRatio float64) error
// NewZerologWriter creates a zerolog writer based on configuration
func NewZerologWriter() *os.File {
return os.Stderr
@@ -39,29 +29,6 @@ type Config struct {
Database DatabaseConfig `mapstructure:"database"`
RateLimit RateLimitConfig `mapstructure:"rate_limit"`
Cache CacheConfig `mapstructure:"cache"`
// viper is the underlying configuration source. Kept (unexported,
// mapstructure:"-") so hot-reload can re-unmarshal on file changes —
// see WatchAndApply (ADR-0023 selective hot-reload).
viper *viper.Viper `mapstructure:"-"`
// reloadMu serialises Unmarshal during hot-reload so a partial mutation
// can't be observed mid-flight by getter calls.
reloadMu sync.RWMutex `mapstructure:"-"`
// samplerReconfigureCallback is invoked when telemetry.sampler.type or
// telemetry.sampler.ratio change. nil means no callback registered.
samplerReconfigureCallback SamplerReconfigureFunc `mapstructure:"-"`
// prevSamplerType and prevSamplerRatio track the last-seen sampler values
// to detect changes during hot-reload (ADR-0023 Phase 3).
prevSamplerType string `mapstructure:"-"`
prevSamplerRatio float64 `mapstructure:"-"`
// watcherStopped indicates that the config watcher has been stopped via
// the context being cancelled. This prevents the OnConfigChange handler
// from processing events after cleanup.
watcherStopped bool `mapstructure:"-"`
}
// ServerConfig holds server-related configuration
@@ -104,44 +71,9 @@ type APIConfig struct {
// AuthConfig holds authentication configuration
type AuthConfig struct {
JWTSecret string `mapstructure:"jwt_secret"`
AdminMasterPassword string `mapstructure:"admin_master_password"`
JWT JWTConfig `mapstructure:"jwt"`
Email EmailConfig `mapstructure:"email"`
MagicLink MagicLinkConfig `mapstructure:"magic_link"`
OIDC OIDCConfig `mapstructure:"oidc"`
}
// MagicLinkConfig holds passwordless-auth magic-link parameters (ADR-0028 Phase A).
type MagicLinkConfig struct {
TTL time.Duration `mapstructure:"ttl"`
BaseURL string `mapstructure:"base_url"`
CleanupInterval time.Duration `mapstructure:"cleanup_interval"`
}
// OIDCConfig holds OpenID Connect provider configuration (ADR-0028 Phase B).
// Multiple providers are supported via a map keyed by provider name (e.g. "arcodange-sso", "google").
type OIDCConfig struct {
Providers map[string]OIDCProvider `mapstructure:"providers"`
}
// OIDCProvider describes a single OIDC provider's discovery + client config.
type OIDCProvider struct {
IssuerURL string `mapstructure:"issuer_url"`
ClientID string `mapstructure:"client_id"`
ClientSecret string `mapstructure:"client_secret"`
}
// EmailConfig holds outgoing email transport configuration.
// Defaults match local Mailpit (cf. ADR-0029) so dev needs no extra setup.
type EmailConfig struct {
From string `mapstructure:"from"`
SMTPHost string `mapstructure:"smtp_host"`
SMTPPort int `mapstructure:"smtp_port"`
SMTPUsername string `mapstructure:"smtp_username"`
SMTPPassword string `mapstructure:"smtp_password"`
SMTPUseTLS bool `mapstructure:"smtp_use_tls"`
Timeout time.Duration `mapstructure:"timeout"`
JWTSecret string `mapstructure:"jwt_secret"`
AdminMasterPassword string `mapstructure:"admin_master_password"`
JWT JWTConfig `mapstructure:"jwt"`
}
// JWTConfig holds JWT-specific configuration
@@ -291,22 +223,6 @@ func LoadConfig() (*Config, error) {
v.SetDefault("auth.jwt.secret_retention.max_retention", 72*time.Hour)
v.SetDefault("auth.jwt.secret_retention.cleanup_interval", 1*time.Hour)
// Email defaults — match local Mailpit (ADR-0029).
v.SetDefault("auth.email.from", "noreply@dance-lessons-coach.local")
v.SetDefault("auth.email.smtp_host", "localhost")
v.SetDefault("auth.email.smtp_port", 1025)
v.SetDefault("auth.email.smtp_use_tls", false)
v.SetDefault("auth.email.timeout", 10*time.Second)
// Magic-link defaults (ADR-0028 Phase A).
v.SetDefault("auth.magic_link.ttl", 15*time.Minute)
v.SetDefault("auth.magic_link.base_url", "http://localhost:8080")
v.SetDefault("auth.magic_link.cleanup_interval", 1*time.Hour)
// OIDC defaults (ADR-0028 Phase B). Providers map is empty by default;
// configured per environment via config file or env vars.
v.SetDefault("auth.oidc.providers", map[string]interface{}{})
// Check for custom config file path via environment variable
if configFile := os.Getenv("DLC_CONFIG_FILE"); configFile != "" {
v.SetConfigFile(configFile)
@@ -352,25 +268,6 @@ func LoadConfig() (*Config, error) {
v.BindEnv("auth.jwt.secret_retention.retention_factor", "DLC_AUTH_JWT_SECRET_RETENTION_FACTOR")
v.BindEnv("auth.jwt.secret_retention.max_retention", "DLC_AUTH_JWT_SECRET_MAX_RETENTION")
v.BindEnv("auth.jwt.secret_retention.cleanup_interval", "DLC_AUTH_JWT_SECRET_CLEANUP_INTERVAL")
v.BindEnv("auth.email.from", "DLC_AUTH_EMAIL_FROM")
v.BindEnv("auth.email.smtp_host", "DLC_AUTH_EMAIL_SMTP_HOST")
v.BindEnv("auth.email.smtp_port", "DLC_AUTH_EMAIL_SMTP_PORT")
v.BindEnv("auth.email.smtp_username", "DLC_AUTH_EMAIL_SMTP_USERNAME")
v.BindEnv("auth.email.smtp_password", "DLC_AUTH_EMAIL_SMTP_PASSWORD")
v.BindEnv("auth.email.smtp_use_tls", "DLC_AUTH_EMAIL_SMTP_USE_TLS")
v.BindEnv("auth.email.timeout", "DLC_AUTH_EMAIL_TIMEOUT")
// Magic-link environment variables (ADR-0028 Phase A).
v.BindEnv("auth.magic_link.ttl", "DLC_AUTH_MAGIC_LINK_TTL")
v.BindEnv("auth.magic_link.base_url", "DLC_AUTH_MAGIC_LINK_BASE_URL")
v.BindEnv("auth.magic_link.cleanup_interval", "DLC_AUTH_MAGIC_LINK_CLEANUP_INTERVAL")
// OIDC environment variables (ADR-0028 Phase B). One canonical "default"
// provider is bindable via env; additional providers must be defined in config.yaml.
v.BindEnv("auth.oidc.providers.default.issuer_url", "DLC_AUTH_OIDC_ISSUER_URL")
v.BindEnv("auth.oidc.providers.default.client_id", "DLC_AUTH_OIDC_CLIENT_ID")
v.BindEnv("auth.oidc.providers.default.client_secret", "DLC_AUTH_OIDC_CLIENT_SECRET")
v.BindEnv("telemetry.sampler.type", "DLC_TELEMETRY_SAMPLER_TYPE")
v.BindEnv("telemetry.sampler.ratio", "DLC_TELEMETRY_SAMPLER_RATIO")
@@ -402,14 +299,6 @@ func LoadConfig() (*Config, error) {
return nil, fmt.Errorf("config unmarshal error: %w", err)
}
// Keep the viper instance for hot-reload (ADR-0023).
config.viper = v
// Initialize previous sampler values for hot-reload change detection
// (ADR-0023 Phase 3).
config.prevSamplerType = config.Telemetry.Sampler.Type
config.prevSamplerRatio = config.Telemetry.Sampler.Ratio
// Setup logging based on configuration (level, output file, time format).
// The JSON/console format was already applied at the top of LoadConfig via
// peekJSONLogging, so SetupLogging only needs to handle the remaining knobs.
@@ -474,19 +363,6 @@ func (c *Config) GetSamplerRatio() float64 {
return c.Telemetry.Sampler.Ratio
}
// SetSamplerReconfigureCallback registers a callback that is invoked when
// telemetry.sampler.type or telemetry.sampler.ratio change via hot-reload.
// The callback receives the new sampler type and ratio values.
// Pass nil to unregister the callback.
func (c *Config) SetSamplerReconfigureCallback(callback SamplerReconfigureFunc) {
c.reloadMu.Lock()
defer c.reloadMu.Unlock()
c.samplerReconfigureCallback = callback
// Initialize previous values so we can detect changes on first hot-reload
c.prevSamplerType = c.Telemetry.Sampler.Type
c.prevSamplerRatio = c.Telemetry.Sampler.Ratio
}
// GetV2Enabled returns whether v2 API is enabled
func (c *Config) GetV2Enabled() bool {
return c.API.V2Enabled
@@ -502,43 +378,6 @@ func (c *Config) GetAdminMasterPassword() string {
return c.Auth.AdminMasterPassword
}
// GetEmailConfig returns the outgoing email transport configuration.
// Defaults match local Mailpit (localhost:1025, no TLS, no auth) per
// ADR-0029. Used by pkg/email.NewSMTPSender.
func (c *Config) GetEmailConfig() EmailConfig {
return c.Auth.Email
}
// GetMagicLinkConfig returns the passwordless-auth magic-link parameters
// (ADR-0028 Phase A). TTL defaults to 15m, BaseURL to http://localhost:8080.
func (c *Config) GetMagicLinkConfig() MagicLinkConfig {
out := c.Auth.MagicLink
if out.TTL <= 0 {
out.TTL = 15 * time.Minute
}
if out.BaseURL == "" {
out.BaseURL = "http://localhost:8080"
}
return out
}
// GetOIDCProviders returns the configured OIDC providers, keyed by provider name.
// Empty map (not nil) is returned when no providers are configured.
func (c *Config) GetOIDCProviders() map[string]OIDCProvider {
if c.Auth.OIDC.Providers == nil {
return map[string]OIDCProvider{}
}
return c.Auth.OIDC.Providers
}
// GetMagicLinkCleanupInterval returns the magic-link cleanup interval (ADR-0028 Phase A consequence).
func (c *Config) GetMagicLinkCleanupInterval() time.Duration {
if c.Auth.MagicLink.CleanupInterval <= 0 {
return 1 * time.Hour
}
return c.Auth.MagicLink.CleanupInterval
}
// GetJWTTTL returns the JWT TTL
func (c *Config) GetJWTTTL() time.Duration {
if c.Auth.JWT.TTL == 0 {
@@ -751,105 +590,3 @@ func (c *Config) setupLogOutput() {
log.Logger = log.Output(file)
log.Trace().Str("output", output).Msg("Logging to file")
}
// WatchAndApply starts watching the config file for changes and applies the
// hot-reloadable subset on every change (ADR-0023 selective hot-reload).
//
// Phases shipped:
// - Phase 1: logging.level — re-applied via SetupLogging on every change.
// - Phase 2: auth.jwt.ttl — picked up automatically because the userService
// reads it via JWTConfig.GetTTL (a method value capturing this *Config).
// The reloaded TTL is used on the NEXT token generation; tokens issued
// before the change keep their original expiry.
// - Phase 3: telemetry.sampler.type + telemetry.sampler.ratio — triggers
// the callback set via SetSamplerReconfigureCallback if the values change.
//
// The other fields listed in ADR-0023 (api.v2_enabled) remain restart-only
// until their handlers land in subsequent phases.
//
// Stops when ctx is cancelled. Safe to call once at server startup.
// If the config file is absent (ConfigFileNotFoundError at load time), this
// becomes a no-op and logs a single warning.
func (c *Config) WatchAndApply(ctx context.Context) {
if c.viper == nil {
log.Warn().Msg("Config hot-reload disabled: no viper instance attached")
return
}
if c.viper.ConfigFileUsed() == "" {
log.Info().Msg("Config hot-reload disabled: no config file in use (env-only or defaults)")
return
}
c.viper.OnConfigChange(func(in fsnotify.Event) {
// Skip processing if watcher has been stopped
c.reloadMu.Lock()
if c.watcherStopped {
c.reloadMu.Unlock()
return
}
c.reloadMu.Unlock()
log.Info().Str("event", in.Op.String()).Str("file", in.Name).Msg("Config file changed, reloading hot-reloadable fields")
c.reloadMu.Lock()
defer c.reloadMu.Unlock()
if err := c.viper.Unmarshal(c); err != nil {
log.Error().Err(err).Msg("Hot-reload: failed to unmarshal new config, keeping previous values")
return
}
// Apply hot-reloadable fields. Order matters: logging first so the
// rest of the reload is logged at the right level.
c.SetupLogging()
// Check if sampler config changed and invoke callback if registered
samplerChanged := c.prevSamplerType != c.Telemetry.Sampler.Type ||
c.prevSamplerRatio != c.Telemetry.Sampler.Ratio
if samplerChanged && c.samplerReconfigureCallback != nil {
if err := c.samplerReconfigureCallback(context.Background(),
c.Telemetry.Sampler.Type,
c.Telemetry.Sampler.Ratio); err != nil {
log.Error().Err(err).Msg("Hot-reload: sampler reconfigure callback failed")
} else {
// Update previous values only after successful callback
c.prevSamplerType = c.Telemetry.Sampler.Type
c.prevSamplerRatio = c.Telemetry.Sampler.Ratio
log.Info().
Str("sampler_type", c.prevSamplerType).
Float64("sampler_ratio", c.prevSamplerRatio).
Msg("Hot-reload applied: telemetry sampler reconfigured")
}
} else if samplerChanged {
// No callback registered, just update tracking values
c.prevSamplerType = c.Telemetry.Sampler.Type
c.prevSamplerRatio = c.Telemetry.Sampler.Ratio
}
log.Info().
Str("logging_level", c.GetLogLevel()).
Dur("jwt_ttl", c.GetJWTTTL()).
Msg("Hot-reload applied (logging.level + auth.jwt.ttl)")
})
c.viper.WatchConfig()
log.Info().Str("file", c.viper.ConfigFileUsed()).Msg("Config hot-reload watcher started (ADR-0023 Phase 1)")
// Stop the watcher on context cancel — we set a flag that the
// OnConfigChange handler checks, avoiding the race with viper's
// internal state that would occur if we called OnConfigChange again.
//
// We deliberately do NOT log inside this goroutine: this goroutine
// outlives ctx (parent's defer cancel only fires when the test's
// outer scope exits, not when t.Cleanup runs), so a log call here
// races with the next test's LoadConfig → SetupLogging →
// zerolog.SetGlobalLevel under -race (observed 2026-05-05, Q-038).
// The flag-set is the load-bearing operation; the missing log line
// is a small ops cost (operators learn the watcher stops on shutdown
// via the parent shutdown logs, not a dedicated message).
go func() {
<-ctx.Done()
c.reloadMu.Lock()
c.watcherStopped = true
c.reloadMu.Unlock()
}()
}

351
pkg/config/config_hot_reload_test.go
View File

@@ -1,351 +0,0 @@
package config
import (
"context"
"errors"
"os"
"path/filepath"
"sync"
"testing"
"time"
"github.com/spf13/viper"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// loadFromFile is a helper that mimics LoadConfig() for a specific file path
// without going through the env-prefix and singleton machinery — keeps the
// test hermetic.
func loadFromFile(t *testing.T, path string) *Config {
t.Helper()
v := viper.New()
v.SetConfigFile(path)
v.SetConfigType("yaml")
v.SetDefault("logging.level", "info")
v.SetDefault("auth.jwt.ttl", time.Hour)
require.NoError(t, v.ReadInConfig())
c := &Config{viper: v}
require.NoError(t, v.Unmarshal(c))
return c
}
// TestWatchAndApply_LoggingLevel proves the hot-reload pipe end-to-end:
// write a new logging.level to the watched file, the OnConfigChange handler
// re-unmarshals, and the in-memory Config reflects the new value.
func TestWatchAndApply_LoggingLevel(t *testing.T) {
dir := t.TempDir()
path := filepath.Join(dir, "config.yaml")
require.NoError(t, os.WriteFile(path, []byte("logging:\n level: info\n"), 0644))
c := loadFromFile(t, path)
assert.Equal(t, "info", c.GetLogLevel())
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
c.WatchAndApply(ctx)
// Mutate the file. fsnotify needs a real write event; rewrite atomically.
require.NoError(t, os.WriteFile(path, []byte("logging:\n level: debug\n"), 0644))
// Poll for up to 2s waiting for the in-memory level to flip.
deadline := time.Now().Add(2 * time.Second)
for time.Now().Before(deadline) {
c.reloadMu.RLock()
level := c.GetLogLevel()
c.reloadMu.RUnlock()
if level == "debug" {
return
}
time.Sleep(20 * time.Millisecond)
}
c.reloadMu.RLock()
defer c.reloadMu.RUnlock()
t.Fatalf("logging level did not hot-reload to debug: still %q", c.GetLogLevel())
}
// TestWatchAndApply_NoFileNoOp confirms the watcher is a safe no-op when no
// config file is in use (env-only / defaults) — important so production
// containers without a mounted config.yaml don't crash.
func TestWatchAndApply_NoFileNoOp(t *testing.T) {
c := &Config{viper: viper.New()}
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
c.WatchAndApply(ctx) // should return without panicking
}
// TestWatchAndApply_NilViperNoOp confirms the watcher tolerates a Config
// constructed without the viper field (e.g. tests that build a Config{}
// manually — same defensive code path as production but exercised explicitly).
func TestWatchAndApply_NilViperNoOp(t *testing.T) {
c := &Config{}
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
c.WatchAndApply(ctx)
}
// TestWatchAndApply_JWTTTL proves Phase 2 of ADR-0023: the JWT TTL is
// re-read on every token generation via the GetJWTTTL method value, so
// after a config-file change the new TTL takes effect without restart.
func TestWatchAndApply_JWTTTL(t *testing.T) {
dir := t.TempDir()
path := filepath.Join(dir, "config.yaml")
require.NoError(t, os.WriteFile(path, []byte("auth:\n jwt:\n ttl: 1h\n"), 0644))
c := loadFromFile(t, path)
assert.Equal(t, time.Hour, c.GetJWTTTL())
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
c.WatchAndApply(ctx)
require.NoError(t, os.WriteFile(path, []byte("auth:\n jwt:\n ttl: 30m\n"), 0644))
deadline := time.Now().Add(2 * time.Second)
for time.Now().Before(deadline) {
c.reloadMu.RLock()
ttl := c.GetJWTTTL()
c.reloadMu.RUnlock()
if ttl == 30*time.Minute {
return
}
time.Sleep(20 * time.Millisecond)
}
c.reloadMu.RLock()
defer c.reloadMu.RUnlock()
t.Fatalf("auth.jwt.ttl did not hot-reload to 30m: still %s", c.GetJWTTTL())
}
// TestWatchAndApply_TelemetrySamplerType proves Phase 3 of ADR-0023:
// when telemetry.sampler.type changes, the callback registered via
// SetSamplerReconfigureCallback is invoked exactly once with the new value.
func TestWatchAndApply_TelemetrySamplerType(t *testing.T) {
dir := t.TempDir()
path := filepath.Join(dir, "config.yaml")
initial := []byte(`telemetry:
sampler:
type: parentbased_always_on
ratio: 1.0
`)
changed := []byte(`telemetry:
sampler:
type: traceidratio
ratio: 1.0
`)
require.NoError(t, os.WriteFile(path, initial, 0644))
c := loadFromFile(t, path)
assert.Equal(t, "parentbased_always_on", c.GetSamplerType())
// Setup callback tracker
var mu sync.Mutex
callbackCalled := false
var recordedType string
var recordedRatio float64
c.SetSamplerReconfigureCallback(func(ctx context.Context, samplerType string, samplerRatio float64) error {
mu.Lock()
defer mu.Unlock()
callbackCalled = true
recordedType = samplerType
recordedRatio = samplerRatio
return nil
})
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
c.WatchAndApply(ctx)
// Mutate the file
require.NoError(t, os.WriteFile(path, changed, 0644))
// Poll for up to 2s waiting for callback
deadline := time.Now().Add(2 * time.Second)
for time.Now().Before(deadline) {
mu.Lock()
if callbackCalled {
mu.Unlock()
assert.Equal(t, "traceidratio", recordedType)
assert.Equal(t, 1.0, recordedRatio)
return
}
mu.Unlock()
time.Sleep(20 * time.Millisecond)
}
mu.Lock()
defer mu.Unlock()
t.Fatalf("sampler reconfigure callback was not invoked: callbackCalled=%v", callbackCalled)
}
// TestWatchAndApply_TelemetrySamplerRatio proves Phase 3 of ADR-0023:
// when telemetry.sampler.ratio changes, the callback registered via
// SetSamplerReconfigureCallback is invoked exactly once with the new value.
func TestWatchAndApply_TelemetrySamplerRatio(t *testing.T) {
dir := t.TempDir()
path := filepath.Join(dir, "config.yaml")
initial := []byte(`telemetry:
sampler:
type: parentbased_always_on
ratio: 1.0
`)
changed := []byte(`telemetry:
sampler:
type: parentbased_always_on
ratio: 0.5
`)
require.NoError(t, os.WriteFile(path, initial, 0644))
c := loadFromFile(t, path)
assert.Equal(t, 1.0, c.GetSamplerRatio())
// Setup callback tracker
var mu sync.Mutex
callbackCalled := false
var recordedType string
var recordedRatio float64
c.SetSamplerReconfigureCallback(func(ctx context.Context, samplerType string, samplerRatio float64) error {
mu.Lock()
defer mu.Unlock()
callbackCalled = true
recordedType = samplerType
recordedRatio = samplerRatio
return nil
})
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
c.WatchAndApply(ctx)
// Mutate the file
require.NoError(t, os.WriteFile(path, changed, 0644))
// Poll for up to 2s waiting for callback
deadline := time.Now().Add(2 * time.Second)
for time.Now().Before(deadline) {
mu.Lock()
if callbackCalled {
mu.Unlock()
assert.Equal(t, "parentbased_always_on", recordedType)
assert.Equal(t, 0.5, recordedRatio)
return
}
mu.Unlock()
time.Sleep(20 * time.Millisecond)
}
mu.Lock()
defer mu.Unlock()
t.Fatalf("sampler reconfigure callback was not invoked: callbackCalled=%v", callbackCalled)
}
// TestWatchAndApply_SamplerCallbackNotCalledWhenNoChange proves that
// the sampler callback is NOT invoked when the config file changes but
// sampler type and ratio remain the same.
func TestWatchAndApply_SamplerCallbackNotCalledWhenNoChange(t *testing.T) {
dir := t.TempDir()
path := filepath.Join(dir, "config.yaml")
initial := []byte(`telemetry:
sampler:
type: parentbased_always_on
ratio: 1.0
logging:
level: info
`)
changed := []byte(`telemetry:
sampler:
type: parentbased_always_on
ratio: 1.0
logging:
level: debug
`)
require.NoError(t, os.WriteFile(path, initial, 0644))
c := loadFromFile(t, path)
// Setup callback tracker
var mu sync.Mutex
callbackCalled := false
c.SetSamplerReconfigureCallback(func(ctx context.Context, samplerType string, samplerRatio float64) error {
mu.Lock()
defer mu.Unlock()
callbackCalled = true
return nil
})
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
c.WatchAndApply(ctx)
// Mutate the file (logging level changes, but sampler stays the same)
require.NoError(t, os.WriteFile(path, changed, 0644))
// Poll for up to 2s - callback should NOT be called
deadline := time.Now().Add(2 * time.Second)
for time.Now().Before(deadline) {
mu.Lock()
wasCalled := callbackCalled
mu.Unlock()
if wasCalled {
t.Fatalf("sampler reconfigure callback was invoked but sampler did not change")
}
time.Sleep(20 * time.Millisecond)
}
}
// TestWatchAndApply_SamplerCallbackErrorHandling proves that when the
// sampler reconfigure callback returns an error, the previous sampler values
// are NOT updated, allowing retry on next config change.
func TestWatchAndApply_SamplerCallbackErrorHandling(t *testing.T) {
dir := t.TempDir()
path := filepath.Join(dir, "config.yaml")
initial := []byte(`telemetry:
sampler:
type: parentbased_always_on
ratio: 1.0
`)
changed := []byte(`telemetry:
sampler:
type: traceidratio
ratio: 0.5
`)
require.NoError(t, os.WriteFile(path, initial, 0644))
c := loadFromFile(t, path)
// Setup callback that returns an error
expectedErr := errors.New("reconfigure failed")
var mu sync.Mutex
callbackCalled := false
c.SetSamplerReconfigureCallback(func(ctx context.Context, samplerType string, samplerRatio float64) error {
mu.Lock()
defer mu.Unlock()
callbackCalled = true
return expectedErr
})
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
c.WatchAndApply(ctx)
// Mutate the file
require.NoError(t, os.WriteFile(path, changed, 0644))
// Poll for up to 2s waiting for callback error
deadline := time.Now().Add(2 * time.Second)
for time.Now().Before(deadline) {
mu.Lock()
if callbackCalled {
mu.Unlock()
// Verify previous values were NOT updated (so retry can work)
c.reloadMu.RLock()
assert.Equal(t, "parentbased_always_on", c.prevSamplerType)
assert.Equal(t, 1.0, c.prevSamplerRatio)
c.reloadMu.RUnlock()
return
}
mu.Unlock()
time.Sleep(20 * time.Millisecond)
}
mu.Lock()
defer mu.Unlock()
t.Fatalf("sampler reconfigure callback was not invoked: callbackCalled=%v", callbackCalled)
}

Some files were not shown because too many files have changed in this diff Show More