- Updated frontend to Node 20 to fix Vite crypto.hash error
- Removed mailhog service as not needed for development
- Updated email configuration to be disabled by default in dev
- Fixed frontend port mapping to use 3005 consistently
- Added script to show/reset admin credentials
- Removed unnecessary storage volume mount
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Added DATABASE_CLIENT=pg to docker-compose.dev.yml for PostgreSQL connection
- Fixed migration 032 to check if tables exist before creating
- Removed language-specific email template columns (use standard columns)
- Added conditional checks for app_settings and email_templates inserts
- Created helper scripts for migration state management
- Added .env.dev with PostgreSQL configuration for development
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fixed migration 030 to use standard email_templates columns (subject, body_html, body_text)
- Removed language-specific columns that don't exist in base schema
- Updated docker-compose.dev.yml for development environment
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fixed migration 029: Use base email_templates columns (subject, body_html, body_text)
instead of language-specific columns that don't exist yet
- Fixed migration 004: JSON.stringify the setting_value for app_settings table
- Removed German translations from backup email templates in core migration
The errors occurred because:
1. Migration 029 assumed language columns existed, but they're added by later migrations
2. Migration 004 passed a plain string to a JSON column in PostgreSQL
- Added check in markMigrationAsApplied to prevent duplicate inserts
- Now checks if migration is already marked before inserting
- Prevents 'duplicate key value violates unique constraint' error
The error occurred when detectExistingSchema() marked a migration
as applied, then the migration runner caught a 'schema exists' error
and tried to mark it as applied again.
- Wrapped file writing in try-catch to prevent migration failure
- Credentials are always shown in console output
- File writing is now optional - if it fails, migration continues
- Added informative message when file cannot be written
This prevents the migration from failing in environments where
the data directory has permission issues, while still ensuring
administrators can see and copy the credentials from console output.
- Changed credential file location from /app/ to /app/data/
- Added directory creation with recursive flag
- Updated console messages to show correct file location
- The data/ directory is already owned by nodejs user in Dockerfile
The error occurred because the nodejs user doesn't have write
permission to /app/ directory, but does have permission to /app/data/
which is explicitly created and chowned in the Dockerfile.
- Removed must_change_password field that doesn't exist in admin_users table
- Changed from using db to knex parameter for database operations
- Fixed require statement that was accidentally changed
- Updated security message to reflect no forced password change
- Removed debug logging after identifying the issue
The error occurred because 001_init.js was trying to insert a column
that doesn't exist in the admin_users table schema created by
initializeDatabase().
- Changed from standalone script to proper migration with exports.up/down
- Removed process.exit() calls that were terminating the migration runner
- Removed immediate execution of runMigrations()
- Now properly exports migration functions like other migrations
This was the root cause - 001_init.js was executing immediately when
required and calling process.exit(), preventing it from being run as
a migration and causing 029 to run first on an empty database.
- Changed from string sort to numeric sort for migration files
- String sort was causing '029' to run before '001'
- Now properly extracts and compares numeric prefixes
- Applied fix to both run-migrations.js and run-migrations-safe.js
This ensures 001_init.js runs first and creates all necessary tables
before other migrations try to use them.
- Check for essential tables (events, photos, admin_users, activity_logs)
to determine if it's truly a new deployment
- Only run detectExistingSchema() for actual existing deployments
- Remove obsolete init.js references (now 001_init.js)
- Fix migration filters to handle renamed init file
The issue was that detectExistingSchema() was marking migrations as
applied from previous failed runs, causing the system to incorrectly
treat new deployments as existing ones and run legacy migrations that
expect tables to already exist.
- Renamed core/init.js to core/001_init.js to ensure it runs first
- Updated detectExistingSchema() to reference 001_init.js
- This fixes the issue where backup migrations tried to access
app_settings table before it was created
- Migrations now run in correct order: init first, then numbered
The error occurred because alphabetical sorting put 029 before init,
causing migrations to fail on new deployments.
- Updated all core migrations to use ../../src/ instead of ../src/
- Updated legacy migrations with the same path fix
- This fixes MODULE_NOT_FOUND errors during deployment
The error occurred because migrations were moved one level deeper
into core/ and legacy/ subdirectories without updating the relative
paths to the source files.
- Created core/ directory for essential migrations that always run
- Created legacy/ directory for migrations only needed when upgrading
- New deployments will only run core migrations for a clean database
- Existing deployments will run all migrations in proper sequence
- Fixed duplicate migration numbers (014 and 027)
- Updated migration runners to handle new directory structure
- Added README explaining the migration organization
This change optimizes deployment for new users who will get a clean
schema without running unnecessary upgrade migrations.
- Remove all @example.com email addresses from documentation
- Replace security@example.com with GitHub security issue links
- Replace conduct@example.com with GitHub issue link
- Update CONTRIBUTING.md to use GitHub issues instead of email
- Ensure all communication happens through GitHub's issue tracking system
- Avoid direct email communication for better transparency and tracking
- Add acknowledgment section about AI generation
- Clarify human testing and security auditing
- Emphasize production testing and code review
- Remove unnecessary .gitkeep files
- Create docker-compose.dev.yml with Mailhog for development email testing
- Standardize all configurations to use PORT=3001 for backend
- Fix database service naming (postgres → db) across all files
- Add missing BACKEND_URL environment variable to all configs
- Update .env examples to match actual Docker setup requirements
- Remove orphaned postgres-init directory (Umami handles its own DB)
- Update README roadmap: mark gallery feedback as implemented, add multi-admin support
- Update deployment guide with development setup instructions
- Fix frontend Dockerfile.dev for proper hot-reload development
- Remove unused files (wedding-photos.db, frontend/README.md)
This ensures all configuration files are consistent and aligned with the deployment guide.
- Merge all deployment docs into single comprehensive DEPLOYMENT_GUIDE.md
- Add instructions for non-nginx deployment options
- Reference utility scripts in deployment guide
- Remove orphaned migrations folder at root level
- Remove redundant deployment documentation files
- Keep all utility scripts in scripts/ folder
- Update CLAUDE.md to reference new deployment guide
This provides a single source of truth for all deployment scenarios.
- Hide "Manage Feedback" button when feedback is disabled for an event
- Fix 500 error on feedback API endpoint by adding null-safe operators
- Fix TypeError on analytics page by calculating average_rating in backend
- Fix password validation for event creation by properly awaiting async validation
- Add proper null checks and fallbacks for feedback statistics
These fixes ensure:
- Date passwords like "19.07.2025" work with simple password complexity settings
- Feedback management page loads without errors
- Analytics display correctly even with no feedback data
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Added /admin/events/:id/feedback route to App.tsx
- This fixes the empty page issue when navigating to feedback management
- EventFeedbackPage component was already implemented but route was missing
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add German translations for event dropdown menu actions
- Add feedback settings to event edit form
- Hide comment button in gallery when feedback is disabled
- Add feedback moderation panel to event details page
Implements:
1. German translation for three dots menu actions (viewDetails, archiveEventAction, etc.)
2. Feedback enable option now visible when editing existing events
3. Comment button in photo lightbox only shows when feedback is enabled
4. Inline comment moderation in admin event detail view
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Replace plugins/github-release with plugins/gitea-release
- Fix API endpoint compatibility issue (was using GitHub API v3)
- Update base_url to gitea.local.nothaft.cloud
- Change secret from GITHUB_TOKEN to GITEA_TOKEN
- Update release notes to reference local Gitea URLs
This fixes the 401 authentication error when creating releases
on Gitea instances.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add convertDateFormat function to automatically fix DD->dd, YYYY->yyyy
- Handles existing database values with old format strings
- Prevents RangeError when using old formats stored in settings
- Ensures backward compatibility without requiring database updates
This fix converts formats on-the-fly so existing production data
with old formats like 'DD.MM.YYYY' will work correctly.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fix TypeError "e.match is not a function" when creating events
- Update useLocalizedDate hook to handle both string and object date formats
- Add type safety for date format configuration
- Fix date format strings to use correct date-fns format (lowercase)
- Ensure backward compatibility with existing date settings
The issue was caused by SettingsPage saving date formats as objects
while useLocalizedDate expected strings. This fix handles both formats
gracefully and prevents the error page redirect.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fix restoreTypes translation structure (was under options.types)
- Add missing restore.messages.restoreStarted translation
- Ensure all restore wizard strings use translations
- Add corresponding German translations for restore section
- Fix translation key structure to match component expectations
🤖 Generated with Claude Code
Co-Authored-By: Claude <noreply@anthropic.com>
- Fix '0 files' hardcoded string to use translation
- Fix 'local' destination type to show translated name
- Add missing field placeholders for rsync and S3 configurations
- Add missing backup.history.columns.* translations
- Add missing backup.history.filter.* translations
- Add missing backup.history.details.* translations
- Fix backup destination display to use proper translation key
- Replace TestTube icon with Wifi icon for connection testing
- Add all corresponding German translations
- Ensure Backup Health and Coverage titles use translations
🤖 Generated with Claude Code
Co-Authored-By: Claude <noreply@anthropic.com>
- Add CPU, RAM, and storage requirements
- Include OS and software dependencies
- Add Docker requirements for containerized deployment
- Keep it concise and focused on minimum requirements only
🤖 Generated with Claude Code
Co-Authored-By: Claude <noreply@anthropic.com>
- Replace manual curl approach with plugins/github-release
- Fixes shell parsing issues with multiline strings
- Properly passes GITHUB_TOKEN via api_key setting
- Uses YAML multiline string (|) for release notes
- Cleaner and more reliable approach
🤖 Generated with Claude Code
Co-Authored-By: Claude <noreply@anthropic.com>
- Use echo with single JSON string instead of heredoc
- Use > for folded scalar to avoid newline issues
- Properly escape quotes in JSON body
- Ensure GITHUB_TOKEN is properly passed
🤖 Generated with Claude Code
Co-Authored-By: Claude <noreply@anthropic.com>
- Add comprehensive backup translation keys to en.json and de.json
- Update all backup components to use i18next translations:
- BackupManagement.jsx: main page with tab navigation
- BackupDashboard.jsx: health status and statistics
- BackupConfiguration.jsx: settings and destination configuration
- BackupHistory.jsx: backup history table and details
- RestoreWizard.jsx: multi-step restore process
- Replace all hardcoded strings with translation keys
- Support dynamic values with interpolation
- Fix Drone CI/CD github-release step:
- Write release.json to /tmp to avoid permission issues
- Use quoted heredoc to prevent shell interpretation errors
- Replace placeholders with actual tag values using sed
🤖 Generated with Claude Code
Co-Authored-By: Claude <noreply@anthropic.com>
The github-release plugin was incorrectly detecting and using the
Gitea API instead of GitHub's API. Replaced with direct curl command
that explicitly calls GitHub API to create releases.
This approach:
- Uses curlimages/curl image for lightweight execution
- Directly calls GitHub API v3 with proper authentication
- Avoids any auto-detection issues from the plugin
- Creates releases with full markdown formatting
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Users were being redirected from gallery pages to admin login due to
useLocalizedDate hook trying to fetch admin settings. Fixed by:
1. Added general_date_format to public settings endpoint
2. Created publicSettingsService for unauthenticated access
3. Updated useLocalizedDate to use public settings instead of admin
4. Fixed API interceptor to not redirect on public endpoint 401s
5. Added backups/ and test-archiver/ to .gitignore
This restores gallery access for all users.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
paul
Gitea Actions Bot