# Changelog

All notable changes to StarPunk 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]

## [1.1.1] - 2025-11-25

### Added
- **Structured Logging** - Enhanced logging system for production readiness
  - RotatingFileHandler with 10MB files, keeping 10 backups
  - Correlation IDs for request tracing across the entire request lifecycle
  - Separate log files in `data/logs/starpunk.log`
  - All print statements replaced with proper logging
  - See ADR-054 for architecture details

- **Database Connection Pooling** - Improved database performance
  - Connection pool with configurable size (default: 5 connections)
  - Request-scoped connections via Flask's g object
  - Pool statistics available for monitoring via `/admin/metrics`
  - Transparent to calling code (maintains same interface)
  - See ADR-053 for implementation details

- **Enhanced Configuration Validation** - Fail-fast startup validation
  - Validates both presence and type of all required configuration values
  - Clear, detailed error messages with specific fixes
  - Validates LOG_LEVEL against allowed values
  - Type checking for strings, integers, and Path objects
  - Non-zero exit status on configuration errors
  - See ADR-052 for configuration strategy

### Changed
- **Centralized Error Handling** - Consistent error responses
  - Moved error handlers from inline decorators to `starpunk/errors.py`
  - Micropub endpoints return spec-compliant JSON errors
  - HTML error pages for browser requests
  - All errors logged with correlation IDs
  - MicropubError exception class for spec compliance
  - See ADR-055 for error handling strategy

- **Database Module Reorganization** - Better structure
  - Moved from single `database.py` to `database/` package
  - Separated concerns: `init.py`, `pool.py`, `schema.py`
  - Maintains backward compatibility with existing imports
  - Cleaner separation of initialization and connection management

- **Performance Monitoring Infrastructure** - Track system performance
  - MetricsBuffer class with circular buffer (deque-based)
  - Per-process metrics with process ID tracking
  - Configurable sampling rates per operation type
  - Database pool statistics endpoint (`/admin/metrics`)
  - See Phase 2 implementation report for details

- **Three-Tier Health Checks** - Comprehensive health monitoring
  - Basic `/health` endpoint (public, load balancer-friendly)
  - Detailed `/health?detailed=true` (authenticated, comprehensive)
  - Full `/admin/health` diagnostics (authenticated, with metrics)
  - Progressive detail levels for different use cases
  - See developer Q&A Q10 for architecture

- **Admin Metrics Dashboard** - Visual performance monitoring (Phase 3)
  - Server-side rendering with Jinja2 templates
  - Auto-refresh with htmx (10-second interval)
  - Charts powered by Chart.js from CDN
  - Progressive enhancement (works without JavaScript)
  - Database pool statistics, performance metrics, system health
  - Access at `/admin/dashboard`
  - See developer Q&A Q19 for design decisions

### Changed

- **RSS Feed Streaming Optimization** - Memory-efficient feed generation (Phase 3)
  - Generator-based streaming with `yield` (Q9)
  - Memory usage reduced from O(n) to O(1) for feed size
  - Yields XML in semantic chunks (channel metadata, items, closing tags)
  - Lower time-to-first-byte (TTFB) for large feeds
  - Note list caching still prevents repeated DB queries
  - No ETags (incompatible with streaming), but Cache-Control headers maintained
  - Recommended for feeds with 100+ items
  - Backward compatible - transparent to RSS clients

- **Search Enhancements** - Improved search robustness
  - FTS5 availability detection at startup with caching
  - Graceful fallback to LIKE queries when FTS5 unavailable
  - Search result highlighting with XSS prevention (markupsafe.escape())
  - Whitelist-only `<mark>` tags for highlighting
  - See Phase 2 implementation for details

- **Unicode Slug Generation** - International character support
  - Unicode normalization (NFKD) before slug generation
  - Timestamp-based fallback (YYYYMMDD-HHMMSS) for untranslatable text
  - Warning logs with original text for debugging
  - Never fails Micropub requests due to slug issues
  - See Phase 2 implementation for details

### Fixed

- **Migration Race Condition Tests** - Fixed flaky tests (Phase 3, Q15)
  - Corrected off-by-one error in retry count expectations
  - Fixed mock time.time() call count in timeout tests
  - 10 retries = 9 sleep calls (not 10)
  - Tests now stable and reliable

### Technical Details
- Phase 1, 2, and 3 of v1.1.1 "Polish" release completed
- Core infrastructure improvements for production readiness
- 600 tests passing (all tests stable, no flaky tests)
- No breaking changes to public API
- Complete operational documentation added

## [1.1.0] - 2025-11-25

### Added
- **Full-Text Search** - SQLite FTS5 implementation for searching note content
  - FTS5 virtual table with Porter stemming and Unicode normalization
  - Automatic index updates on note create/update/delete
  - Graceful degradation if FTS5 unavailable
  - Helper function to rebuild index from existing notes
  - See ADR-034 for architecture details
  - **Note**: Search UI (/api/search endpoint and templates) to be completed in follow-up

- **Custom Slugs** - User-specified URLs via Micropub
  - Support for `mp-slug` property in Micropub requests
  - Automatic slug sanitization (lowercase, hyphens only)
  - Reserved slug protection (api, admin, auth, feed, etc.)
  - Sequential conflict resolution with suffixes (-2, -3, etc.)
  - Hierarchical slugs (/) rejected (deferred to v1.2.0)
  - Maintains backward compatibility with auto-generation
  - See ADR-035 for implementation details

### Fixed
- **RSS Feed Ordering** - Feed now correctly displays newest posts first
  - Added `reversed()` wrapper to compensate for feedgen internal ordering
  - Regression test ensures feed matches database DESC order

- **Custom Slug Extraction** - Fixed bug where mp-slug was ignored in Micropub requests
  - Root cause: mp-slug was extracted after normalize_properties() filtered it out
  - Solution: Extract mp-slug from raw request data before normalization
  - Affects both form-encoded and JSON Micropub requests
  - See docs/reports/custom-slug-bug-diagnosis.md for detailed analysis

### Changed
- **Database Migration System** - Renamed for clarity
  - `SCHEMA_SQL` renamed to `INITIAL_SCHEMA_SQL`
  - Documentation clarifies this represents frozen v1.0.0 baseline
  - All schema changes after v1.0.0 must go in migration files
  - See ADR-033 for redesign rationale

### Technical Details
- Migration 005: FTS5 virtual table with DELETE trigger
- New modules: `starpunk/search.py`, `starpunk/slug_utils.py`
- Modified: `starpunk/notes.py` (custom_slug param, FTS integration)
- Modified: `starpunk/micropub.py` (mp-slug extraction)
- Modified: `starpunk/feed.py` (reversed() fix)
- 100% backward compatible, no breaking changes
- All tests pass (557 tests)

## [1.0.1] - 2025-11-25

### Fixed
- Micropub Location header no longer contains double slash in URL
- Microformats2 query response URLs no longer contain double slash

### Technical Details
Fixed URL construction in micropub.py to account for SITE_URL having a trailing slash (required for IndieAuth spec compliance). Changed from `f"{site_url}/notes/{slug}"` to `f"{site_url}notes/{slug}"` at two locations (lines 312 and 383). Added comments explaining the trailing slash convention.

## [1.0.0] - 2025-11-24

### Released

**First production-ready release of StarPunk!** A minimal, self-hosted IndieWeb CMS with full IndieAuth and Micropub compliance.

This milestone represents the completion of all V1 features:
- Full W3C IndieAuth specification compliance with endpoint discovery
- Complete W3C Micropub specification implementation for posting
- Robust database migrations with race condition protection
- Production-ready containerized deployment
- Comprehensive test coverage (536 tests passing)

StarPunk is now ready for production use as a personal IndieWeb publishing platform.

### Summary of V1 Features

All features from release candidates (rc.1 through rc.5) are now stable:

#### IndieAuth Implementation
- External IndieAuth provider support (delegates to IndieLogin.com or similar)
- Dynamic endpoint discovery from user profile (ADMIN_ME)
- W3C IndieAuth specification compliance
- HTTP Link header and HTML link element discovery
- Endpoint caching (1 hour TTL) with graceful fallback
- Token verification caching (5 minutes TTL)

#### Micropub Implementation
- Full Micropub endpoint for creating posts
- Support for JSON and form-encoded requests
- Bearer token authentication with scope validation
- Content validation and sanitization
- Proper HTTP status codes and error responses
- Location header with post URL

#### Database & Migrations
- Automatic database migration system
- Migration race condition protection with database locking
- Exponential backoff retry logic for multi-worker deployments
- Safe container startup with gunicorn workers

#### Production Deployment
- Production-ready containerized deployment (Podman/Docker)
- Health check endpoint for monitoring
- Gunicorn WSGI server with multi-worker support
- Secure non-root user execution
- Reverse proxy configurations (Caddy/Nginx)

### Configuration Changes from RC Releases

- `TOKEN_ENDPOINT` environment variable deprecated (endpoints discovered automatically)
- `ADMIN_ME` must be a valid profile URL with IndieAuth link elements

### Standards Compliance

- W3C IndieAuth Specification (Section 4.2: Discovery by Clients)
- W3C Micropub Specification
- OAuth 2.0 Bearer Token Authentication
- Microformats2 Semantic HTML
- RSS 2.0 Feed Syndication

### Testing

- 536 tests passing (99%+ pass rate)
- 87% overall code coverage
- Comprehensive endpoint discovery tests
- Complete Micropub integration tests
- Migration system tests

### Documentation

Complete documentation available in `/docs/`:
- Architecture overview and design documents
- 31 Architecture Decision Records (ADRs)
- API contracts and specifications
- Deployment and migration guides
- Development standards and setup

### Related Documentation
- ADR-031: IndieAuth Endpoint Discovery
- ADR-030: IndieAuth Provider Removal Strategy
- ADR-023: Micropub V1 Implementation Strategy
- ADR-022: Migration Race Condition Fix
- See `/docs/reports/` for detailed implementation reports

## [1.0.0-rc.5] - 2025-11-24

### Fixed

#### Migration Race Condition (CRITICAL)
- **CRITICAL**: Migration race condition causing container startup failures with multiple gunicorn workers
  - Implemented database-level locking using SQLite's `BEGIN IMMEDIATE` transaction mode
  - Added exponential backoff retry logic (10 attempts, up to 120s total) for lock acquisition
  - Workers now coordinate properly: one applies migrations while others wait and verify
  - Graduated logging (DEBUG → INFO → WARNING) based on retry attempts
  - New connection created for each retry attempt to prevent state issues
  - See ADR-022 and migration-race-condition-fix-implementation.md for technical details

#### IndieAuth Endpoint Discovery (CRITICAL)
- **CRITICAL**: Fixed hardcoded IndieAuth endpoint configuration (violated IndieAuth specification)
  - Endpoints now discovered dynamically from user's profile URL (ADMIN_ME)
  - Implements W3C IndieAuth specification Section 4.2 (Discovery by Clients)
  - Supports both HTTP Link headers and HTML link elements for discovery
  - Endpoint discovery cached (1 hour TTL) for performance
  - Token verifications cached (5 minutes TTL)
  - Graceful fallback to expired cache on network failures
  - See ADR-031 and docs/architecture/indieauth-endpoint-discovery.md for details

### Changed

#### IndieAuth Endpoint Discovery
- **BREAKING**: Removed `TOKEN_ENDPOINT` configuration variable
  - Endpoints are now discovered automatically from `ADMIN_ME` profile
  - Deprecation warning shown if `TOKEN_ENDPOINT` still in environment
  - See docs/migration/fix-hardcoded-endpoints.md for migration guide

- **Token Verification** (`starpunk/auth_external.py`)
  - Complete rewrite with endpoint discovery implementation
  - Always discovers endpoints from `ADMIN_ME` (single-user V1 assumption)
  - Validates discovered endpoints (HTTPS required in production, localhost allowed in debug)
  - Implements retry logic with exponential backoff for network errors
  - Token hashing (SHA-256) for secure caching
  - URL normalization for comparison (lowercase, no trailing slash)

- **Caching Strategy**
  - Simple single-user cache (V1 implementation)
  - Endpoint cache: 1 hour TTL with grace period on failures
  - Token verification cache: 5 minutes TTL
  - Cache cleared automatically on application restart

### Added

#### IndieAuth Endpoint Discovery
- New dependency: `beautifulsoup4>=4.12.0` for HTML parsing
- HTTP Link header parsing (RFC 8288 basic support)
- HTML link element extraction with BeautifulSoup4
- Relative URL resolution against profile base URL
- HTTPS enforcement in production (HTTP allowed in debug mode)
- Comprehensive error handling with clear messages
- 35 new tests covering all discovery scenarios

### Technical Details

#### Migration Race Condition Fix
- Modified `starpunk/migrations.py` to wrap migration execution in `BEGIN IMMEDIATE` transaction
- Each worker attempts to acquire RESERVED lock; only one succeeds
- Other workers retry with exponential backoff (100ms base, doubling each attempt, plus jitter)
- Workers that arrive late detect completed migrations and exit gracefully
- Timeout protection: 30s per connection attempt, 120s absolute maximum
- Comprehensive error messages guide operators to resolution steps

#### Endpoint Discovery Implementation
- Discovery priority: HTTP Link headers (highest), then HTML link elements
- Profile URL fetch timeout: 5 seconds (cached results)
- Token verification timeout: 3 seconds (per request)
- Maximum 3 retries for server errors (500-504) and network failures
- No retries for client errors (400, 401, 403, 404)
- Single-user cache structure (no profile URL mapping needed in V1)
- Grace period: Uses expired endpoint cache if fresh discovery fails
- V2-ready: Cache structure can be upgraded to dict-based for multi-user

### Breaking Changes
- `TOKEN_ENDPOINT` environment variable no longer used (will show deprecation warning)
- Micropub now requires discoverable IndieAuth endpoints in `ADMIN_ME` profile
- ADMIN_ME profile must include `<link rel="token_endpoint">` or HTTP Link header

### Migration Guide
See `docs/migration/fix-hardcoded-endpoints.md` for detailed migration steps:
1. Ensure your ADMIN_ME profile has IndieAuth link elements
2. Remove TOKEN_ENDPOINT from your .env file
3. Restart StarPunk - endpoints will be discovered automatically

### Configuration
Updated requirements:
- `ADMIN_ME`: Required, must be a valid profile URL with IndieAuth endpoints
- `TOKEN_ENDPOINT`: Deprecated, will be ignored (remove from configuration)

### Tests
- 536 tests passing (excluding timing-sensitive migration race tests)
- 35 new endpoint discovery tests:
  - Link header parsing (absolute and relative URLs)
  - HTML parsing (including malformed HTML)
  - Discovery priority (Link headers over HTML)
  - HTTPS validation (production vs debug mode)
  - Caching behavior (TTL, expiry, grace period)
  - Token verification (success, errors, retries)
  - URL normalization and scope checking

## [1.0.0-rc.4] - 2025-11-24

### Complete IndieAuth Server Removal (Phases 1-4)

StarPunk no longer acts as an IndieAuth authorization server. All IndieAuth operations are now delegated to external providers (e.g., IndieLogin.com). This simplifies the codebase and aligns with IndieWeb best practices.

### Removed
- **Phase 1**: Authorization Endpoint
  - Deleted `/auth/authorization` endpoint and `authorization_endpoint()` function
  - Removed authorization consent UI template (`templates/auth/authorize.html`)
  - Removed authorization-related imports: `create_authorization_code` and `validate_scope`
  - Deleted tests: `tests/test_routes_authorization.py`, `tests/test_auth_pkce.py`

- **Phase 2**: Token Issuance
  - Deleted `/auth/token` endpoint and `token_endpoint()` function
  - Removed all token issuance functionality
  - Deleted tests: `tests/test_routes_token.py`

- **Phase 3**: Token Storage
  - Deleted `starpunk/tokens.py` module entirely
  - Dropped `tokens` and `authorization_codes` database tables (migration 004)
  - Removed token CRUD and verification functions
  - Deleted tests: `tests/test_tokens.py`

### Added
- **Phase 4**: External Token Verification
  - New module `starpunk/auth_external.py` for external IndieAuth token verification
  - `verify_external_token()` function to verify tokens with external providers
  - `check_scope()` function moved from tokens module
  - Configuration: `TOKEN_ENDPOINT` for external token endpoint URL
  - HTTP client (httpx) for token verification requests
  - Proper error handling for unreachable auth servers
  - Timeout protection (5s) for external verification requests

### Changed
- **Micropub endpoint** now verifies tokens with external IndieAuth providers
  - Updated `routes/micropub.py` to use `verify_external_token()`
  - Updated `micropub.py` to import `check_scope` from `auth_external`
  - All Micropub tests updated to mock external verification

- **Migrations**:
  - Migration 003: Remove `code_verifier` column from `auth_state` table
  - Migration 004: Drop `tokens` and `authorization_codes` tables
  - Both migrations applied automatically on startup

- **Tests**: All 501 tests passing
  - Fixed migration tests to work with current schema (no `code_verifier`)
  - Updated Micropub tests to mock external token verification
  - Fixed test fixtures and app context usage
  - Removed 38 obsolete token-related tests

### Configuration
New required configuration for production:
- `TOKEN_ENDPOINT`: External IndieAuth token endpoint (e.g., https://tokens.indieauth.com/token)
- `ADMIN_ME`: Site owner's identity URL (already required)

### Technical Details
- External token verification follows IndieAuth specification
- Tokens verified via GET request with Authorization header
- Token response validated for required fields (me, client_id, scope)
- Only tokens matching `ADMIN_ME` are accepted
- Graceful degradation if external server unavailable

### Breaking Changes
- **Micropub clients** must obtain tokens from external IndieAuth providers
- Existing internal tokens are invalid (tables dropped in migration 004)
- `TOKEN_ENDPOINT` configuration required for Micropub to function

### Migration Guide
1. Choose external IndieAuth provider (recommended: IndieLogin.com)
2. Set `TOKEN_ENDPOINT` environment variable
3. Existing sessions unaffected - admin login still works
4. Micropub clients need new tokens from external provider

### Standards Compliance
- Fully compliant with W3C IndieAuth specification
- Follows IndieWeb principle: delegate to external services
- OAuth 2.0 Bearer token authentication maintained

### Related Documentation
- ADR-030: IndieAuth Provider Removal Strategy
- ADR-050: Remove Custom IndieAuth Server
- Implementation report: `docs/reports/2025-11-24-indieauth-removal-complete.md`

### Notes
- This completes the transition from self-hosted IndieAuth to external delegation
- Simpler codebase: -500 lines of code, -5 database tables
- More secure: External providers handle token security
- More maintainable: Less code to secure and update

## [1.0.0-rc.3] - 2025-11-24

### Fixed
- **CRITICAL: Migration detection failure for partially migrated databases**: Fixed migration 002 detection logic
  - Production database had migration 001 applied but not migration 002
  - Migration 002's tables (tokens, authorization_codes) already existed from SCHEMA_SQL in v1.0.0-rc.1
  - Previous logic only used smart detection for fresh databases (migration_count == 0)
  - For partially migrated databases (migration_count > 0), it tried to run migration 002 normally
  - This caused "table already exists" error because CREATE TABLE statements would fail
  - Fixed by checking migration 002's state regardless of migration_count
  - Migration 002 now checks if its tables exist before running, skips table creation if they do
  - Missing indexes are created even when tables exist, ensuring complete database state
  - Fixes deployment failure on production database with existing tables but missing migration record

### Technical Details
- Affected databases: Any database with migration 001 applied but not migration 002, where tables were created by SCHEMA_SQL
- Root cause: Smart detection (is_migration_needed) was only called when migration_count == 0
- Solution: Always check migration 002's state, regardless of migration_count
- Backwards compatibility: Works for fresh databases, partially migrated databases, and fully migrated databases
- Migration 002 will create only missing indexes if tables already exist

## [1.0.0-rc.2] - 2025-11-24

### Fixed
- **CRITICAL: Database migration failure on existing databases**: Removed duplicate index definitions from SCHEMA_SQL
  - Migration 002 creates indexes `idx_tokens_hash`, `idx_tokens_me`, and `idx_tokens_expires`
  - These same indexes were also in SCHEMA_SQL (database.py lines 58-60)
  - When applying migration 002 to existing databases, indexes already existed from SCHEMA_SQL, causing failure
  - Removed the three index creation statements from SCHEMA_SQL to prevent conflicts
  - Migration 002 is now the sole source of truth for token table indexes
  - Fixes "index already exists" error when running migrations on databases created before v1.0.0-rc.1

### Technical Details
- Affected databases: Any database created with v1.0.0-rc.1 or earlier that had run init_db()
- Root cause: SCHEMA_SQL ran on every init_db() call, creating indexes before migration could run
- Solution: Remove index creation from SCHEMA_SQL, delegate to migration 002 exclusively
- Backwards compatibility: Fresh databases will get indexes from migration 002 automatically

## [1.0.0-rc.1] - 2025-11-24

### Release Candidate for V1.0.0
First release candidate with complete IndieWeb support. This milestone implements the full V1 specification with IndieAuth authentication and Micropub posting capabilities.

### Added
- **Phase 1: Secure Token Management**
  - Bearer token storage with Argon2id hashing
  - Automatic token expiration (90 days default)
  - Token revocation endpoint (`POST /micropub?action=revoke`)
  - Admin interface for token management with creation, viewing, and revocation
  - Comprehensive test coverage for token operations (14 tests)

- **Phase 2: IndieAuth Token Endpoint**
  - Token endpoint (`POST /indieauth/token`) for access token issuance
  - Authorization endpoint (`POST /indieauth/authorize`) for consent flow
  - PKCE verification for authorization code exchange
  - Token verification endpoint (`GET /indieauth/token`) for clients
  - Proper OAuth 2.0/IndieAuth spec compliance
  - Client credential validation and scope enforcement
  - Test suite for token and authorization endpoints (13 tests)

- **Phase 3: Micropub Endpoint**
  - Micropub endpoint (`POST /micropub`) for creating posts
  - Support for both JSON and form-encoded requests
  - Bearer token authentication with scope validation
  - Content validation and sanitization
  - Post creation with automatic timestamps
  - Location header with post URL in responses
  - Comprehensive error handling with proper HTTP status codes
  - Integration tests for complete authentication flow (11 tests)

### Changed
- Admin interface now includes token management section
- Database schema extended with `tokens` table for secure token storage
- Authentication system now supports both admin sessions and bearer tokens
- Authorization flow integrated with existing IndieAuth authentication

### Security
- Bearer tokens hashed with Argon2id (same as passwords)
- Tokens support automatic expiration
- Scope validation enforces `create` permission for posting
- PKCE prevents authorization code interception
- Token verification validates both hash and expiration

### Standards Compliance
- IndieAuth specification (W3C) for authentication and authorization
- Micropub specification (W3C) for posting interface
- OAuth 2.0 bearer token authentication
- Proper HTTP status codes and error responses
- Location header for created resources

### Testing
- 77 total tests (all passing)
- Complete coverage of token management, IndieAuth endpoints, and Micropub
- Integration tests verify end-to-end flows
- Error case coverage for validation and authentication failures

### Documentation
- Implementation reports for all three phases
- Architecture reviews documenting design decisions
- API contracts specified in docs/design/api-contracts.md
- Test coverage documented in implementation reports

### Related Standards
- ADR-023: Micropub V1 Implementation Strategy
- W3C IndieAuth Specification
- W3C Micropub Specification

### Notes
This is a release candidate for testing. Stable 1.0.0 will be released after testing period and any necessary fixes.

## [0.9.5] - 2025-11-23

### Fixed
- **SECRET_KEY empty string handling**: Fixed config.py to properly handle empty `FLASK_SECRET_KEY` environment variable
  - `os.getenv()` returns empty string (not None) when env var is set to `""`
  - Empty string now correctly falls back to SESSION_SECRET
  - Prevents Flask session/flash failures when FLASK_SECRET_KEY="" in .env file

## [0.9.4] - 2025-11-22

### Fixed
- **IndieAuth authentication endpoint correction**: Changed code redemption from token endpoint to authorization endpoint
  - Per IndieAuth spec: authentication-only flows use `/authorize`, not `/token`
  - StarPunk only needs identity verification, not access tokens
  - Removed unnecessary `grant_type` parameter (only needed for token endpoint)
  - Updated debug logging to reflect "code verification" terminology
  - Fixes authentication with IndieLogin.com and spec-compliant providers

### Changed
- Code redemption now POSTs to `/authorize` endpoint instead of `/token`
- Log messages updated from "token exchange" to "code verification"

## [0.9.3] - 2025-11-22

### Fixed
- **IndieAuth token exchange missing grant_type**: Added required `grant_type=authorization_code` parameter to token exchange request
  - OAuth 2.0 spec requires this parameter for authorization code flow
  - Some IndieAuth providers reject token exchange without this parameter
  - Fixes authentication failures with spec-compliant IndieAuth providers

## [0.9.2] - 2025-11-22

### Fixed
- **IndieAuth callback 404 error**: Fixed auth blueprint URL prefix mismatch
  - Auth blueprint was using `/admin` prefix but redirect_uri used `/auth/callback`
  - Changed blueprint prefix from `/admin` to `/auth` as documented in ADR-022
  - Auth routes now correctly at `/auth/login`, `/auth/callback`, `/auth/logout`
  - Admin dashboard routes remain at `/admin/*` (unchanged)

### Changed
- Updated test expectations to use new `/auth/*` URL patterns

## [0.9.1] - 2025-11-19

### Fixed
- **IndieAuth client_id trailing slash**: Added automatic trailing slash normalization to SITE_URL
  - IndieLogin.com spec requires client_id URLs to have trailing slash for root domains
  - Fixes "client_id is not registered" authentication errors
  - Normalizes https://example.com to https://example.com/
- **Enhanced debug logging**: Added detailed httpx request/response logging for token exchange
  - Shows exact HTTP method, URL, headers, and body being sent to IndieLogin.com
  - Helps troubleshoot authentication issues with full visibility
  - All sensitive data (tokens, verifiers) automatically redacted

### Changed
- SITE_URL configuration now automatically adds trailing slash if missing

## [0.9.0] - 2025-11-19

### Added
- **Automatic Database Migration System**: Zero-touch database schema updates on application startup
- Migration runner module (`starpunk/migrations.py`) with automatic execution
- Fresh database detection to prevent unnecessary migration execution
- Legacy database detection to apply pending migrations automatically
- Migration tracking table (`schema_migrations`) to record applied migrations
- Helper functions for database introspection (table_exists, column_exists, index_exists)
- Comprehensive migration test suite (26 tests covering all scenarios)

### Changed
- `init_db()` now automatically runs migrations after creating schema
- Database initialization is fully automatic in containerized deployments
- Migration files in `migrations/` directory are executed in alphanumeric order

### Features
- **Fresh Database Behavior**: New installations detect current schema and mark migrations as applied without execution
- **Legacy Database Behavior**: Existing databases automatically apply pending migrations on startup
- **Migration Tracking**: All applied migrations recorded with timestamps in schema_migrations table
- **Idempotent**: Safe to run multiple times, only applies pending migrations
- **Fail-Safe**: Application fails to start if migrations fail, preventing inconsistent state

### Infrastructure
- Container deployments now self-initialize with correct schema automatically
- No manual SQL execution required for schema updates
- Clear migration history in database for audit purposes
- Migration failures logged with detailed error messages

### Standards Compliance
- Sequential migration numbering (001, 002, 003...)
- One migration per schema change for clear audit trail
- Migration files include date and ADR reference headers
- Follows standard migration patterns from Django/Rails

### Testing
- 100% test coverage for migration system (26/26 tests passing)
- Tests cover fresh DB, legacy DB, partial migrations, failures
- Integration tests with actual migration file (001_add_code_verifier_to_auth_state.sql)
- Verified both automatic detection scenarios in production

### Related Documentation
- ADR-020: Automatic Database Migration System
- Implementation guidance document with step-by-step instructions
- Quick reference card for migration system usage

## [0.8.0] - 2025-11-19

### Fixed
- **CRITICAL**: Fixed IndieAuth authentication to work with IndieLogin.com API
- Implemented required PKCE (Proof Key for Code Exchange) for security
- Corrected IndieLogin.com API endpoints (/authorize and /token instead of /auth)
- Added issuer validation for authentication callbacks

### Added
- PKCE code_verifier generation and storage
- PKCE code_challenge generation (SHA256, base64-url encoded)
- Database column: auth_state.code_verifier for PKCE support
- Database migration script: migrations/001_add_code_verifier_to_auth_state.sql
- Comprehensive PKCE unit tests (6 tests, all passing)

### Removed
- OAuth Client ID Metadata Document endpoint (/.well-known/oauth-authorization-server)
  - Added in v0.7.0 but unnecessary for IndieLogin.com
  - IndieLogin.com does not use OAuth client discovery
- h-app microformats markup from templates
  - Modified in v0.7.1 but unnecessary for IndieLogin.com
  - IndieLogin.com does not parse h-app for client identification
- indieauth-metadata link from HTML head

### Changed
- Authentication flow now follows IndieLogin.com API specification exactly
- Database schema: auth_state table includes code_verifier column
- State token validation now returns code_verifier for token exchange
- Token exchange uses /token endpoint (not /auth)
- Authorization requests use /authorize endpoint (not /auth)

### Security
- PKCE prevents authorization code interception attacks
- Issuer validation prevents token substitution attacks
- Code verifier securely stored and single-use
- Code verifier redacted in logs for security

### Breaking Changes
- Users mid-authentication when upgrading will need to restart login (state tokens expire in 5 minutes)
- Existing state tokens without code_verifier will be invalid (intentional security improvement)

### Notes
- **v0.7.0**: OAuth metadata endpoint added based on misunderstanding of requirements. This endpoint was never functional for our use case and is removed in v0.8.0.
- **v0.7.1**: h-app visibility changes attempted to fix authentication but addressed wrong issue. h-app discovery not used by IndieLogin.com. Removed in v0.8.0.
- **v0.8.0**: Correct implementation based on official IndieLogin.com API documentation.

### Related Documentation
- ADR-025: IndieAuth Correct Implementation Based on IndieLogin.com API
- Design Document: docs/designs/indieauth-pkce-authentication.md
- ADR-016: Superseded (h-app client discovery not required)
- ADR-017: Superseded (OAuth metadata not required)

### Migration Notes
- Database migration required: Add code_verifier column to auth_state table
- See migrations/001_add_code_verifier_to_auth_state.sql for SQL
- See docs/designs/indieauth-pkce-authentication.md for full implementation guide

## [0.7.1] - 2025-11-19

### Known Issues
- **IndieAuth authentication still broken**: This release attempted to fix authentication by making h-app visible, but IndieLogin.com does not parse h-app. Missing PKCE implementation is the actual issue. Fixed in v0.8.0.

### Fixed
- **IndieAuth h-app Visibility**: Removed `hidden` and `aria-hidden="true"` attributes from h-app microformat markup
  - h-app was invisible to IndieAuth parsers, preventing proper client discovery
  - Now visible in DOM for microformat parsers while remaining non-intrusive in footer
  - Provides backward compatibility for IndieAuth services that rely on h-app parsing

## [0.7.0] - 2025-11-19

### Known Issues
- **IndieAuth authentication still broken**: This release attempted to fix authentication by adding OAuth metadata endpoint, but this is not required by IndieLogin.com. Missing PKCE implementation is the actual issue. Fixed in v0.8.0.

### Added
- **IndieAuth Detailed Logging**: Comprehensive logging for authentication flows
- Logging helper functions with automatic token redaction (_redact_token, _log_http_request, _log_http_response)
- DEBUG-level HTTP request/response logging for IndieLogin.com interactions
- Configurable logging via LOG_LEVEL environment variable (DEBUG, INFO, WARNING, ERROR)
- Security-aware logging with automatic redaction of sensitive data (tokens, codes, secrets)
- Production warning when DEBUG logging is enabled in non-development environments
- Comprehensive test suite for logging functions (14 new tests)

### Changed
- Enhanced authentication flow visibility with structured logging
- initiate_login(), handle_callback(), create_session(), and verify_session() now include detailed logging
- Flask logger configuration now based on LOG_LEVEL environment variable
- Log format varies by level: detailed for DEBUG, concise for INFO/WARNING/ERROR

### Security
- All sensitive tokens automatically redacted in logs (show only first 6-8 and last 4 characters)
- Authorization codes, state tokens, and access tokens never logged in full
- Sensitive HTTP headers (Authorization, Cookie, Set-Cookie) excluded from logs
- Production warning prevents accidental DEBUG logging in production

### Features
- Token redaction shows pattern like "abc123...********...xyz9" for debugging while protecting secrets
- HTTP request logging includes method, URL, and redacted parameters
- HTTP response logging includes status code, safe headers, and redacted body
- Session verification and creation logging for audit trails
- Admin authorization logging for security monitoring

### Testing
- 51 authentication tests passing (100% pass rate)
- Tests verify token redaction at all levels
- Tests confirm no sensitive data appears in logs
- Tests verify logging behavior at different log levels (DEBUG vs INFO)

### Standards Compliance
- OWASP Logging Cheat Sheet: Sensitive data redaction
- Python logging best practices
- IndieAuth specification compatibility (logging doesn't interfere with auth flow)

### Related Documentation
- ADR-018: IndieAuth Detailed Logging Strategy
- Implementation includes complete specification from ADR-018

## [0.6.2] - 2025-11-19

### Fixed
- **CRITICAL**: Implemented OAuth Client ID Metadata Document to fix IndieAuth authentication
- Added `/.well-known/oauth-authorization-server` endpoint returning JSON metadata
- IndieLogin.com now correctly verifies StarPunk as a registered OAuth client
- Resolves "client_id is not registered" error preventing production authentication
- Fixes authentication flow with modern IndieAuth servers (2022+ specification)

### Added
- OAuth Client ID Metadata Document endpoint at `/.well-known/oauth-authorization-server`
- JSON metadata response with client_id, client_name, redirect_uris, and OAuth capabilities
- `<link rel="indieauth-metadata">` discovery hint in HTML head
- 24-hour caching for metadata endpoint (Cache-Control headers)
- Comprehensive test suite for OAuth metadata endpoint (12 new tests)
- Tests for indieauth-metadata link discovery (3 tests)

### Changed
- IndieAuth client discovery now uses modern JSON metadata (primary method)
- h-app microformats retained for backward compatibility (legacy fallback)
- Three-layer discovery: well-known URL, link rel hint, h-app markup

### Standards Compliance
- IndieAuth specification section 4.2 (Client Information Discovery)
- OAuth Client ID Metadata Document format
- IANA well-known URI registry standard
- OAuth 2.0 Dynamic Client Registration (RFC 7591)

### Technical Details
- Metadata endpoint uses configuration values (SITE_URL, SITE_NAME)
- client_id exactly matches document URL (spec requirement)
- redirect_uris properly formatted as array
- Supports PKCE (S256 code challenge method)
- Public client configuration (no client secret)

### Related Documentation
- ADR-017: OAuth Client ID Metadata Document Implementation
- IndieAuth Fix Summary report
- IndieAuth Client Discovery Root Cause Analysis

## [0.6.1] - 2025-11-19

### Fixed
- **CRITICAL**: Fixed IndieAuth client discovery to enable production authentication
- Added h-app microformats markup to base.html for IndieAuth client verification
- IndieLogin.com can now verify StarPunk as a legitimate OAuth client
- Resolves "client_id is not registered" error that blocked all production authentication

### Changed
- Added hidden h-app metadata div to footer with SITE_URL and SITE_NAME
- h-app markup uses aria-hidden="true" and hidden attribute for screen reader and visual hiding
- Implements IndieAuth legacy client discovery standard for backward compatibility

### Standards Compliance
- IndieAuth client discovery (legacy h-app microformats)
- Microformats2 h-app specification
- HTML5 hidden attribute standard
- ARIA accessibility standard

### Related Documentation
- ADR-016: IndieAuth Client Discovery Mechanism
- IndieAuth client discovery analysis report

## [0.6.0] - 2025-11-19

### Added
- **RSS Feed Generation**: Standards-compliant RSS 2.0 feed for published notes
- RSS feed module (`starpunk/feed.py`) with feed generation functions
- GET `/feed.xml` route for RSS feed access
- Server-side feed caching (5-minute default, configurable)
- ETag support for efficient feed updates
- Cache-Control headers for client-side caching
- RSS feed auto-discovery link in HTML templates
- RSS link in site navigation
- Comprehensive RSS feed test suite (44 tests)

### Production Container
- **Containerfile**: Multi-stage build for optimized image size (174MB)
- **Container Orchestration**: Podman and Docker Compose compatible
- **Health Check Endpoint**: GET `/health` for container monitoring
- **Gunicorn WSGI Server**: Production-ready with 4 workers
- **Security**: Non-root user execution (starpunk:1000)
- **Volume Mounts**: Data persistence for notes and database
- **Reverse Proxy Configs**: Caddy and Nginx examples with auto-HTTPS
- **Container Documentation**: Comprehensive deployment guide

### Configuration
- `FEED_MAX_ITEMS`: Maximum items in RSS feed (default: 50)
- `FEED_CACHE_SECONDS`: Server-side cache duration in seconds (default: 300)
- `VERSION`: Application version for health checks (default: 0.6.0)
- `ENVIRONMENT`: Deployment environment (development/production)
- `WORKERS`: Number of Gunicorn workers (default: 4)
- `WORKER_TIMEOUT`: Gunicorn worker timeout in seconds (default: 30)
- `MAX_REQUESTS`: Max requests per worker before restart (default: 1000)

### Features
- RSS 2.0 compliant XML generation using feedgen library
- RFC-822 date formatting for RSS pubDate elements
- Intelligent note title extraction (first line or timestamp fallback)
- HTML content in CDATA sections for feed readers
- Atom self-link for feed discovery
- Only published notes included in feed
- Absolute URLs for all feed item links

### Testing
- 88% overall test coverage (up from 87%)
- 96% coverage for feed module
- 449/450 tests passing (99.78% pass rate)
- Test isolation with automatic cache clearing
- Unicode and special character handling verified

### Standards Compliance
- RSS 2.0 specification compliant
- RFC-822 date format for timestamps
- IndieWeb feed discovery support
- W3C Feed Validator compatible

### Container Features
- Multi-stage build optimizes image size (Python 3.11-slim base)
- uv package manager for fast dependency installation
- Health checks verify database connectivity and filesystem access
- Resource limits prevent container resource exhaustion
- Log rotation (10MB max, 3 files) prevents disk space issues
- Automatic restart policy for reliability
- SELinux compatibility with volume mount flags

### Deployment
- Podman-compatible with `--userns=keep-id` for proper permissions
- Docker-compatible with standard volume mounts
- Reverse proxy examples for Caddy (auto-HTTPS) and Nginx
- HTTPS required for IndieAuth in production
- Complete backup and restore procedures documented
- Performance tuning guide for worker configuration

### Related Documentation
- ADR-014: RSS Feed Implementation Strategy
- ADR-015: Phase 5 Implementation Approach
- Phase 5 design documentation
- Phase 5 quick reference guide
- Container deployment guide

## [0.5.2] - 2025-11-18

### Fixed
- **Admin Routes**: Fixed delete route to return HTTP 404 when attempting to delete nonexistent notes, per ADR-012 (HTTP Error Handling Policy)
- Added existence check to delete route before attempting deletion, consistent with edit route pattern
- Fixed test for delete nonexistent note to match ADR-012 compliance (expect 404 status, not 200 with follow_redirects)

### Changed
- Delete route now checks note existence before deletion and returns 404 with "Note not found" flash message for nonexistent notes
- Test suite: 405/406 tests passing (99.75%)

## [0.5.1] - 2025-11-18

### Fixed
- **CRITICAL**: Fixed authentication redirect loop caused by cookie name collision between Flask's session and StarPunk's auth token
- Renamed authentication cookie from `session` to `starpunk_session` to avoid conflict with Flask's server-side session mechanism used by flash messages
- All authentication flows (dev login, IndieAuth, logout) now work correctly without redirect loops
- Flash messages now display properly without interfering with authentication state

### Changed
- **BREAKING CHANGE**: Authentication cookie renamed from `session` to `starpunk_session`
  - Existing authenticated users will be logged out and need to re-authenticate after upgrade
  - This is an unavoidable breaking change required to fix the critical bug

### Documentation
- Established cookie naming convention standard (starpunk_* prefix for all application cookies)
- Created implementation report documenting the root cause and fix

## [0.5.0] - 2025-11-19

### Added
- Development authentication module (`starpunk/dev_auth.py`) for local testing
- `is_dev_mode()` function to check development mode status
- `create_dev_session()` function for authentication bypass in development
- Web interface templates with Microformats2 markup
- Admin dashboard, note editor, and login pages
- Public note display and RSS feed support

### Fixed
- Phase 4 test suite now passing (400/406 tests, 98.5% pass rate)
- Template encoding issues (removed corrupted Unicode characters)
- Test database initialization using tmp_path fixtures
- Route URL patterns (trailing slash consistency)
- Template variable naming (g.user_me → g.me)
- Function name mismatches in tests (get_all_notes → list_notes)
- URL builder endpoint name (auth.login → auth.login_form)
- Session verification return type handling in tests
- Flake8 code quality issues (unused imports, f-strings)

### Security
- Development authentication includes prominent warning logging
- DEV_MODE validation ensures DEV_ADMIN_ME is set
- Production mode validation ensures ADMIN_ME is set

### Testing
- 87% overall test coverage
- All Phase 4 route and template tests functional
- Proper test isolation with temporary databases
- Fixed test context usage (test_request_context)

### Code Quality
- All code formatted with Black
- Passes Flake8 validation
- Removed unused imports and fixed f-string warnings

## [0.4.0] - 2025-11-18

### Added
- **Authentication module** (`starpunk/auth.py`) with IndieLogin support
- Core authentication functions: `initiate_login`, `handle_callback`, `create_session`, `verify_session`, `destroy_session`
- `require_auth` decorator for protecting admin routes
- Custom authentication exceptions (AuthError, InvalidStateError, UnauthorizedError, IndieLoginError)
- CSRF protection via state tokens
- Secure session management with SHA-256 token hashing
- Session metadata tracking (user agent, IP address)
- Automatic cleanup of expired sessions and state tokens
- URL validation utility function (`is_valid_url`)
- Comprehensive authentication test suite (37 tests, 96% coverage)

### Changed
- Updated sessions table schema to use `session_token_hash` instead of plaintext tokens
- Added `user_agent` and `ip_address` fields to sessions table
- Added `redirect_uri` field to auth_state table
- Added indexes for authentication performance (session_token_hash, me)

### Security
- Token hashing with SHA-256 for secure storage
- CSRF protection with single-use state tokens
- Cryptographically secure token generation (secrets module)
- SQL injection prevention with prepared statements
- Comprehensive security logging

## [0.3.0] - 2025-11-18

### Added
- Notes management module (`starpunk/notes.py`) with CRUD operations
- Custom exceptions for note operations (NoteError, NoteNotFoundError, InvalidNoteDataError, NoteSyncError)
- File and database synchronization with transaction safety
- Support for soft and hard note deletion
- Comprehensive test suite for notes module (85 tests, 86% coverage)
- Database schema support for soft deletes (deleted_at column)
- Slug uniqueness enforcement with random suffix generation
- Content hash calculation for integrity verification

### Changed
- Updated database schema to include `deleted_at` column in notes table
- Added index on `deleted_at` for query performance

## [0.1.0] - 2024-11-18

### Added
- Initial project structure
- Core architecture design
- Technology stack selection (Flask, SQLite, file-based storage)
- Architecture Decision Records (ADR-001 through ADR-007)
- Development documentation and standards
- Phase 1.1 design: Core utilities specification
- Python coding standards
- Documentation organization structure

### Documentation
- Complete architecture overview
- Technology stack documentation
- ADR-001: Python web framework (Flask)
- ADR-002: Flask extensions (minimal approach)
- ADR-003: Frontend technology (server-side rendering)
- ADR-004: File-based note storage
- ADR-005: IndieLogin authentication
- ADR-006: Python virtual environment (uv)
- ADR-007: Slug generation algorithm
- ADR-008: Versioning strategy

[Unreleased]: https://github.com/YOUR_USERNAME/starpunk/compare/v0.4.0...HEAD
[0.4.0]: https://github.com/YOUR_USERNAME/starpunk/compare/v0.3.0...v0.4.0
[0.3.0]: https://github.com/YOUR_USERNAME/starpunk/compare/v0.1.0...v0.3.0
[0.1.0]: https://github.com/YOUR_USERNAME/starpunk/releases/tag/v0.1.0