- Changed from public settings endpoint to admin settings endpoint
- Fixed "Unexpected token '<'" JSON parse error
- Properly transforms settings array to key-value map
- Uses correct setting keys (analytics_umami_*)
- Maintains fallback to environment variables
The analytics page now correctly fetches Umami configuration using
the authenticated admin API instead of the public endpoint, which
was returning errors and causing JSON parse failures.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
BREAKING: This completely replaces the previous approach and will DELETE
all existing history on GitHub, creating entirely new commit SHAs.
Key changes:
- Use orphan branch instead of cherry-pick to break history connection
- Create initial commit from target commit tree using git read-tree
- Apply subsequent changes as completely new commits with new SHAs
- Force push will COMPLETELY REPLACE GitHub history
- No trace of commits before 7aca927937 will remain on GitHub
This ensures GitHub shows only history from the target commit onwards
with no connection to previous commits or their metadata.
- Added pre-fetch and check before committing to ensure we're up-to-date
- Improved retry logic with clearer output and better error handling
- Added explicit fetch before each retry attempt
- Use for-loop instead of while for clearer retry counting
- Better fallback from rebase to merge on conflicts
- Added set -e to fail fast on errors
- More verbose logging for debugging
This should resolve the persistent "non-fast-forward" errors by:
1. Checking if we're behind before even committing
2. Pulling changes if needed
3. Retrying with proper synchronization
4. Providing clear debug output
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
BREAKING ISSUE FIXED:
- 403 errors were triggering redirects, preventing login page from loading
- Public endpoints returning 403 were causing redirect loops
Changes:
- Removed 403 status from automatic redirect logic
- Only 401 (Unauthorized) now triggers login redirect
- 403 (Forbidden) errors are passed through without redirect
This fixes the critical issue where users couldn't access the login page
because public API calls were returning 403 and triggering redirects.
403 errors should be handled differently than 401:
- 401 = Missing/invalid auth (redirect to login)
- 403 = Forbidden (could be rate limit, IP block, etc - don't redirect)
🚨 Emergency fix for production
Co-Authored-By: Claude <noreply@anthropic.com>
- Replace problematic 'find | head -20' commands that caused exit code 141
- Use 'ls -la | head -10 || true' for safer file listing
- Add better progress logging during sensitive file removal
- Add error handling with '|| true' to prevent pipe failures
The find command was outputting more than head could handle, causing
SIGPIPE when head closed the pipe early. This fix uses ls which is
more predictable and adds proper error handling.
- Add --no-merges flag to exclude merge commits during cherry-pick
- Improve error handling for cherry-pick conflicts with auto-resolution
- Add reporting of skipped merge commits for transparency
- Enhance logging to show detailed progress during commit application
Fixes the workflow failure caused by trying to cherry-pick merge commits
which require special handling that was causing exit code 128.
- Added pull before push to handle concurrent workflow executions
- Implemented retry logic with 3 attempts for push operations
- Added fallback from rebase to merge if conflicts occur
- Added proper error handling and logging for debugging
This fixes the "non-fast-forward" error that occurs when multiple
workflows run simultaneously and try to push version bumps.
The workflow now:
1. Pulls latest changes before pushing
2. Retries up to 3 times with 5-second delays
3. Falls back to merge if rebase fails
4. Provides clear error messages for debugging
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Start history from commit 7aca927937 instead of orphan branch
- Use cherry-pick to preserve meaningful commit history
- Automatically exclude files that only existed before target commit
- Add comprehensive error handling and logging
- Maintain clean linear history for GitHub repository
- Added proper HTTP status check before JSON parsing in AnalyticsPage
* Prevents "Unexpected token '<'" error when API returns HTML error pages
* Throws proper error for non-OK responses
- Enhanced API error handling to treat 403 as auth failure
* Both 401 and 403 now trigger redirect to login page
* Clears expired admin tokens automatically
* Prevents users from staying on admin pages with expired sessions
These fixes resolve:
1. JSON parse errors when fetching Umami config
2. 403 Forbidden errors not redirecting to login
3. Backend version display issues due to auth failures
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Added PRODUCTION_TODO_LIST.md with 9 completed production fixes
- Updated .gitea/workflows/mirror-to-github.yml
- Updated .gitignore
This commit includes all the production fixes implemented:
1. Password complexity settings
2. Gallery login security improvements
3. Analytics configuration fixes
4. Translation additions
5. UI/UX improvements
6. Date format consistency
7. Chrome compatibility fixes
All tasks have been completed and tested for production deployment.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fixed backend version endpoint by adding retry logic import
- Gallery login page improvements:
* Increased title size from text-xl to text-2xl (responsive scaling)
* Title now uses event's custom primary color (var(--color-primary))
* Removed event category badge from login page
- Fixed Umami analytics configuration check:
* Added proper enabled state tracking
* Warning now only shows when Umami is explicitly not configured
* Checks both admin settings and environment variables properly
These changes improve user experience and fix false warnings in production.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Disabled duplicate email service (emailService.js) that was creating redundant connections
- Increased connection pool size from 10 to 25 for production environment
- Extended session timeout cache from 5 to 30 minutes to reduce DB queries
- Added connection retry logic with exponential backoff for transient failures
- Fixed password validation to use retry wrapper and correct setting key
- Updated public settings and gallery middleware to handle connection failures gracefully
These changes address the "Connection terminated unexpectedly" errors in production by:
1. Reducing unnecessary database connections
2. Increasing available connection pool capacity
3. Implementing automatic retry for transient connection failures
4. Caching frequently accessed data for longer periods
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Password Complexity: Added 4-level complexity selector (Simple/Moderate/Strong/Very Strong) in admin security settings with dynamic backend validation
- Gallery Security: Removed event date from login page (security risk), replaced with event type badge
- Analytics Config: Fixed "Not Configured" detection logic to check both admin settings and env variables
- Analytics Accuracy: Aligned calculation logic between dashboard and analytics endpoints, added totals verification
- Translations: Added missing activity keys (analytics_settings_updated, cms_page_updated, security_settings_updated, password_reset, admin_logout, system_activity)
- UI Fixes: Fixed German text overflow in CMS page selector with proper CSS truncation
- Date Format: Event creation now respects admin-configured date format instead of browser locale
- Chrome Compatibility: Replaced emoji flags with SVG components for Windows Chrome support
All changes maintain backward compatibility and production stability.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Update publicSettings.js to handle missing analytics setting_type gracefully
- Add dedicated PUT /analytics endpoint for saving analytics settings
- Update frontend settings service to route to correct endpoints based on setting type
- Fix query to use WHERE clause that won't fail if analytics type doesn't exist
This fixes the "Connection terminated unexpectedly" error when fetching
public settings with analytics configuration.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fix backend analytics to include both 'download' and 'download_all' actions
- Add Analytics tab to Settings page for Umami configuration
- Update public settings endpoint to expose Umami config when enabled
- Implement dynamic Umami initialization from backend settings
- Fix frontend analytics calculations (remove hardcoded estimations)
- Add proper download counts and unique visitor tracking
- Update CLAUDE.md with production safety guidelines
The analytics dashboard now shows accurate data for all metrics, and Umami
can be configured through the admin panel instead of environment variables.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Original: feat: enhance security logging and ensure rate limit blocks are properly tracked
- Add comprehensive logging for rate limit blocks with full request details
- IP address (with proper proxy detection), user agent, headers, timestamps
- Rate limit info (current count, limit, remaining, reset time)
- Separate tracking for auth vs general endpoints
- Enhance authentication failure logging
- JWT validation failures with detailed error info
- Admin auth attempts without token
- Failed token validation with user context
- All events include IP, path, method, user agent
- Improve Winston logger configuration for production
- Add automatic log rotation (10MB errors, 50MB combined)
- Create separate security.log for auth/rate limit events
- Ensure logs directory exists automatically
- Add structured JSON format for log aggregation
- Support container logging with LOG_TO_CONSOLE env var
- Create comprehensive documentation
- Security logging guide with examples
- Monitoring recommendations
- Configuration reference
- Add test script to verify logging functionality
All rate limit settings remain configurable via admin panel:
- Window duration, max requests, auth limits
- Skip authenticated requests option
- Public endpoints only option
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
paul
Gitea Actions Bot