- Derive Default trait instead of manual implementation for McpRoot and ProviderManager
- Remove redundant closures in codex_config.rs and config.rs
- Simplify match statements to if let patterns in migration.rs and lib.rs
- Remove unnecessary type conversions and borrows in lib.rs
- Fix i18n key inconsistency: sequentialThinking → sequential-thinking
- Format TypeScript files to match Prettier style
All clippy warnings resolved, code passes all quality checks.
- Add normalize_server_keys() to ensure MCP server map keys match internal id fields
- Auto-normalize on all read/write operations (get, upsert, delete, import, sync)
- Handle edge cases: empty/whitespace ids, key renaming, conflict resolution
- Auto-save config when normalization detects changes
- Apply cargo fmt for code formatting consistency
This enhancement improves data integrity by automatically fixing inconsistencies
between server entry keys and their id fields, especially after manual config edits.
- Add persistent app settings with custom Claude Code and Codex config directories
- Add config directory override UI in settings modal with manual input, browse, and reset options
- Integrate tauri-plugin-dialog for native directory picker
- Support WSL and other special environments where config paths need manual specification
Changes:
- settings.rs: Implement settings load/save and directory override logic
- SettingsModal: Add config directory override UI components
- API: Add get_config_dir and pick_directory commands
- Add settings module for managing common configuration snippets
- Implement UI for creating, editing, and deleting snippets
- Add tauri-plugin-fs for file operations
- Replace co-authored setting with flexible snippet system
- Enable users to define custom config snippets for frequently used settings
- Extract common TOML read/validation logic into dedicated helper functions
- Add graceful degradation for corrupted config files during migration
- Centralize Codex config.toml processing to ensure consistency across operations
- Improve error handling: log warnings for invalid files but continue processing
- Eliminate code duplication between migration, import, and runtime operations
This makes the system more resilient to user configuration issues while maintaining data integrity through unified validation logic.
- Fix operation order: write live files first, then save config.json for consistency
- Implement short-lock pattern: read state quickly, release lock before I/O operations
- Add atomic Codex dual-file writes with rollback on failure
- Simplify add_provider and update_provider logic with consistent structure
- Remove unnecessary duplicate code and improve error handling reliability
This ensures data consistency when operations fail and significantly improves concurrency by minimizing lock holding time during file I/O.
