feat: cap concurrent connections with semaphore

Add max_connections config (default 256) with -m/--max-connections CLI
flag. Server wraps on_client in asyncio.Semaphore to prevent fd
exhaustion under load. Value reloads on SIGHUP; active connections
drain normally. Also adds pool_size/pool_max_idle config fields and
first_hop_pool wiring in server.py (used by next commits), and fixes
asyncio.TimeoutError -> TimeoutError lint warnings.

tests/test_config.py

@@ -2,7 +2,7 @@
 
 import pytest
 
-from s5p.config import ChainHop, Config, parse_proxy_url
+from s5p.config import ChainHop, Config, load_config, parse_proxy_url
 
 
 class TestParseProxyUrl:
@@ -74,3 +74,19 @@ class TestConfig:
         assert c.listen_port == 1080
         assert c.chain == []
         assert c.timeout == 10.0
+        assert c.max_connections == 256
+        assert c.pool_size == 0
+        assert c.pool_max_idle == 30.0
+
+    def test_max_connections_from_yaml(self, tmp_path):
+        cfg_file = tmp_path / "test.yaml"
+        cfg_file.write_text("max_connections: 512\n")
+        c = load_config(cfg_file)
+        assert c.max_connections == 512
+
+    def test_pool_size_from_yaml(self, tmp_path):
+        cfg_file = tmp_path / "test.yaml"
+        cfg_file.write_text("pool_size: 16\npool_max_idle: 45.0\n")
+        c = load_config(cfg_file)
+        assert c.pool_size == 16
+        assert c.pool_max_idle == 45.0