LitematicaFolia

Fixed — PasteOperation semaphore deadlock (Folia region self-deadlock)

Completes the v0.4.1 fix, which only moved the pass-1 dispatch loop off the region tick thread.

Root cause. The pass-2 and deferred-physics-sweep loops were chained with a plain (non-async) thenCompose, which runs on the thread that completed the upstream future — a Folia region tick thread (FoliaCompat.runOnRegion completes it there). Those loops then called the blocking acquireSlot() on that region thread; when the region still had more pending chunk tasks than free CHUNK_THROTTLE permits, it parked in Semaphore.acquire() waiting for permits only it could release → self-deadlock. Regions froze indefinitely (observed ~116h on a production server); the Folia Watchdog logged Tick region ... has not responded without ever crashing.

Fix.

1. Both continuation loops now run via thenComposeAsync(..., offRegionExecutor) (Folia async pool / Paper async worker), so every acquireSlot() runs off-region — a blocked dispatch thread can no longer park a region.
2. acquireSlot() now uses tryAcquire(60s): a timeout dispatches the chunk unthrottled and logs a warning instead of blocking forever. The release guard is seeded so an untaken permit is never over-released.

No config/schema change vs 0.4.1 — in-place drop-in. Normal-load behaviour (32 in-flight chunk-task throttle) unchanged.