refactor: _acquire_sync_lock should raise instead of calling sys.exit

[ergut]

Background

_acquire_sync_lock in logseq_sync.py currently calls sys.exit(1) directly when it fails to acquire the lock:

def _acquire_sync_lock(db_path: str):
    ...
    except portalocker.LockException:
        lock_file.close()
        print("Error: another sync process is already running. ...", file=sys.stderr)
        sys.exit(1)  # ← design smell

Problem

A lock acquisition function shouldn't decide to terminate the process — that conflates two concerns:

• What happened: couldn't acquire the lock
• What to do about it: that's the caller's decision

It also makes the function harder to test (requires mocking sys.exit) and prevents any caller from handling the failure gracefully (e.g., logging, cleanup, returning a structured error).

Proposed Fix

Raise a specific exception (e.g., RuntimeError or a dedicated SyncLockError) instead of calling sys.exit. Move the exit logic to the caller (_run_sync, _watch) where it belongs.

def _acquire_sync_lock(db_path: str):
    ...
    except portalocker.LockException:
        lock_file.close()
        raise RuntimeError(
            "Another sync process is already running. "
            "Wait for it to finish or check for stale sync.lock."
        )

Impact

Low risk, improves testability and separation of concerns.