- Add view_autodiscover.sh helper script for testing autodiscover responses
- Support -h/--help flag for usage information
- Support -d/--domain flag to override autodiscover target (useful for testing)
- Auto-detect xmllint availability for formatted output
- Email validation with regex
- Interactive mode if no email provided
- Display response length for debugging
- Replace hardcoded error IDs with random values (1-10 billion range) for better debugging
- Cast SimpleXMLElement email to string before SQL query to prevent type errors
- Qualify ambiguous 'active' column with table names in JOIN query
- Add proper error XML response for database errors instead of die()
- Ensure all error paths return complete XML documents
- Remove HTTP Basic Authentication requirement from autodiscover.php
- Extract email address from XML request body instead of AUTH headers
- Validate mailbox existence and active status before returning config
- Improve security by eliminating password transmission
- Add comprehensive error handling for invalid/inactive mailboxes
- Follow industry standards (Microsoft, Google, Apple)
- Maintain backward compatibility with existing email clients
- Keep full logging functionality in Redis AUTODISCOVER_LOG
This change enhances security while improving user experience and
follows modern email client configuration best practices.
* Allow making spam aliases permanent
* added german translation
* updated Spamalias Twig + Rename in Spam Alias
* compose: update image tags to align to vendor version
---------
Co-authored-by: DerLinkman <niklas.meyer@servercow.de>
* Replace pigz with zstd for backup compression
This change replaces pigz (parallel gzip) with zstd (Zstandard) as the
compression algorithm for mailcow backups while maintaining full backward
compatibility with existing .tar.gz backups.
Benefits:
- Better compression ratios (12-37% improvement in tests)
- Improved compression speed with modern algorithm
- Maintains rsyncable functionality for incremental backups
- Full backward compatibility for restoring old .tar.gz backups
- Wide industry adoption and active development
Changes:
- Backup compression: pigz --rsyncable -p → zstd --rsyncable -T
- Backup decompression: pigz -d -p → zstd -d -T
- File extensions: .tar.gz → .tar.zst
- Added get_archive_info() function for intelligent format detection
- Updated backup Dockerfile to install zstd alongside pigz
- Restore function now auto-detects and handles both formats
- Updated FILE_SELECTION regex to recognize both .tar.zst and .tar.gz
- Updated comments to reflect new file extension
Backward Compatibility:
- Restore automatically detects .tar.zst (preferred) or .tar.gz (legacy)
- Existing .tar.gz backups can still be restored without issues
- pigz remains installed in backup image for legacy support
- Graceful fallback if backup file format not found
Testing:
- Added comprehensive test suite (test_backup_and_restore.sh)
- 12 automated tests covering all scenarios:
* Backup creation (both formats)
* Restore (both formats)
* Format detection and priority
* Error handling (missing files, empty dirs)
* Content integrity verification
* Multi-threading configuration
* Large file compression (8.59 MB realistic data)
Test Results:
✓ zstd compression working
✓ pigz compression working (legacy)
✓ zstd decompression working
✓ pigz decompression working (backward compatible)
✓ Archive detection working
✓ Content integrity verified
✓ Format priority correct (.tar.zst preferred)
✓ Error handling for missing files
✓ Error handling for empty directories
✓ Multi-threading configuration verified
✓ Large file compression: 37.05% improvement
✓ Small file compression: 12.18% improvement
* move testing script into development folder
---------
Co-authored-by: DerLinkman <niklas.meyer@servercow.de>
