feat: parallel multi-drive search with streaming results

- Rewrite search_multi_drive_filtered() to use tokio::spawn for parallel I/O
- All drives read concurrently instead of sequentially
- Results stream via mpsc channel as each drive completes
- Add OwnedQueryFilters for thread-safe filter sharing
- Add DriveResult struct for per-drive result tracking
- Merge filtered results incrementally as they arrive
- Log shows progress as 'X/Y drives completed'

Performance: N drives now complete in ~max(drive_time) instead of sum(drive_times)

Author: githubrobbi

Cargo.lock

@@ -99,7 +99,7 @@ ratatui = "0.30.0"
 crossterm = "0.29.0"
 
 # ───── Error Handling ─────
-thiserror = "2.0.17"
+thiserror = "2.0.18"
 anyhow = "1.0.100"
 miette = { version = "7.6.0", features = ["fancy"] }