Improve performance and code quality across core modules

[Rashed-alothman]

Download engine (manager.py):

• Parallelize chunk downloads with asyncio.gather + semaphore (8 concurrent)
• Stream chunk data via iter_chunked with inline SHA256 (halves peak memory)
• Add connection pooling: TCPConnector(limit=30, limit_per_host=10, dns_cache)
• Replace blocking file I/O with aiofiles in async download loop

CLI startup (cli.py):

• Lazy Console proxy defers Rich import until first print
• Move Rich traceback install into main() instead of module level

Blocking calls & caching:

• Fix psutil.cpu_percent(interval=1) → interval=None (non-blocking)
• Parallelize network endpoint checks with asyncio.gather
• Add mtime-based config caching (skip disk reads when file unchanged)

Error handling:

• Fix all 4 bare except: clauses (progress.py, p2p.py, error_handler.py)
• Fix 2 swallowed exceptions (except: pass → logging.debug)
• Fix race condition: duplicate speed sample pop(0) outside lock
• Optimize cache eviction: heapq.nsmallest vs sorted() (O(n+k) vs O(n log n))
• Fix asyncio.iscoroutinefunction deprecation in error_handler.py

Tests (47 passing):

• Add test_config.py: loading, caching, corruption handling (5 tests)
• Add test_manager.py: DownloadChunk, init, connection pooling (5 tests)
• Add test_network.py: connection check, caching behavior (2 tests)
• Add test_session.py: SessionManager init, update, lookup (4 tests)
• Fix test_cli.py: replace timeout-prone tests with fast import check

[github-actions]

Dependency Review

✅ No vulnerabilities or license issues or OpenSSF Scorecard issues found.

Scanned Files

None