Replace expect() and unwrap() calls with proper error handling

Problem

The codebase contains 233+ expect(), unwrap(), and panic!() calls that could cause the application to crash. Many of these are in production code paths where errors should be handled gracefully.

Crates Affected

| Crate | Severity | Examples | |-------|----------|----------| | chicken_network | High | Certificate generation, socket binding | | chicken_settings | Medium | File operations, serialization | | chicken_steam | Medium | Steam API initialization | | chicken_polygon | Low | Geometry operations |

High-Risk Locations

// Network - could crash on startup
.expect("all given SANs should be valid DNS names");
.expect("should be a valid certificate");
.expect("failed to bind discovery socket");

// Settings - could crash on file operations
.unwrap();  // 15+ occurrences in settings_edit

// Steam - could crash if Steam not running
.expect("Some error occurred");

Recommended Approach

1. Network/Server: Return Result and propagate errors to caller
2. Settings: Use thiserror or anyhow for context-rich errors
3. Plugin Init: Log error and disable plugin rather than panic
4. Tests: OK to keep unwrap in test code

Acceptance Criteria

• [ ] Audit all expect()/unwrap() calls in production code
• [ ] Replace with proper error types (define custom errors where needed)
• [ ] Add error context using anyhow or thiserror
• [ ] Ensure graceful degradation (e.g., disable features that fail)
• [ ] Update documentation with error handling patterns
• [ ] Keep #![warn(clippy::unwrap_used)] enabled

Related

• Several TODOs mention error handling improvements
• Should align with Rust error handling best practices