fix: 修复暗色主题UI对比度、端口占用死循环及多项Web表单bug

后端修复:
- engine/instances.py: Socks5Server.start() 端口占用时死循环, 增加超时和异常透传
- engine/instances.py: import time 移至顶层
- engine/instances.py: sync_instances/start_instance 容忍单实例启动失败
- run.py: 启动时 sync_instances 失败不退出进程
- web/routes.py: instances start/restart/edit 路由捕获 OSError, flash 友好提示
- api/v1.py: instances start/restart/update API 返回 JSON 错误而非 500 崩溃
- web/routes.py: toggle_user/ban_user 支持 hidden 字段语义, 缺字段时切换状态
- web/routes.py: inject_nav 增加独立 nav.dashboard 和 page_name, 修复双 active bug
- auth.py: login 路由添加 logging import, 修 NameError

前端重构:
- templates/base.html: 全局对比度提升(--text #f1f5f9, --accent #a5b4fc), 标题色显式声明, alert/modal/progress/table 完整暗色样式
- templates/base.html: main-content padding-top 修复 navbar 遮挡内容
- templates/base.html: 侧边栏 active 高亮 (白字 + elevated bg + 左边框)
- templates/base.html: 顶栏动态显示当前页面名
- templates/*.html: 统一 page-header 类, 所有图标按钮添加 title tooltip
- templates/users.html: toggle/ban 表单添加 hidden enabled/ban 字段
- templates/connections.html: fmtBytes 改为 Jinja macro, 时间格式化
- templates/dashboard.html: 时间格式化为 YYYY-MM-DD HH:MM:SS, chart canvas 高度调整
- templates/logs.html: 时间格式化为可读格式
- templates/stats.html: 提升 chart 颜色对比度
- templates/system.html: 使用 time.localtime() 显示本地时间

engine/instances.py

@@ -2,6 +2,7 @@
 import asyncio
 import logging
 import threading
+import time
 from dataclasses import dataclass, field
 
 from engine.server import ConnectionHandler
@@ -51,20 +52,41 @@ class Socks5Server:
 
         self._active_connections = 0
         self.running = True
+        self.server = None
+        self._start_error = None  # 用于在主线程透传子线程异常
 
         def run_loop():
             asyncio.set_event_loop(self.loop)
-            self.server = self.loop.run_until_complete(
-                self._create_server()
-            )
-            self.loop.run_forever()
+            try:
+                self.server = self.loop.run_until_complete(
+                    self._create_server()
+                )
+                self.loop.run_forever()
+            except Exception as e:
+                # 端口占用等异常：保存异常并停止循环，避免主线程死等
+                self._start_error = e
+                log.error("[%s] SOCKS5 服务器启动失败: %s",
+                          self.config.name, e)
+                try:
+                    self.loop.stop()
+                except Exception:
+                    pass
 
         self.thread = threading.Thread(target=run_loop, daemon=True,
                                         name=f"socks5-{self.config.name}")
         self.thread.start()
-        # 等 server 创建完成
-        while self.server is None and self.running:
+        # 等 server 创建完成 (或失败) — 设了上限避免永久死等
+        deadline = time.time() + 5.0
+        while self.server is None and self.running and self._start_error is None:
             time.sleep(0.05)
+            if time.time() > deadline:
+                log.error("[%s] SOCKS5 服务器启动超时", self.config.name)
+                self.running = False
+                break
+        # 如果子线程失败, 把异常抛回调用方
+        if self._start_error is not None:
+            self.running = False
+            raise self._start_error
         if self.server:
             log.info("[%s] SOCKS5 服务器已启动: %s:%d",
                      self.config.name, self.config.listen_host, self.config.listen_port)
@@ -157,7 +179,11 @@ class InstanceManager:
         for name in current_names - running_names:
             cfg = self._config_cache[name]
             srv = Socks5Server(cfg, self.user_service, self.flask_app)
-            srv.start()
+            try:
+                srv.start()
+            except OSError as e:
+                log.error("[%s] 跳过启动（端口冲突?）: %s", name, e)
+                continue
             with self._lock:
                 self._instances[name] = srv
 
@@ -198,7 +224,15 @@ class InstanceManager:
             )
             self._config_cache[inst.name] = cfg
             srv = Socks5Server(cfg, self.user_service, self.flask_app)
-            srv.start()
+            try:
+                srv.start()
+            except OSError as e:
+                # 端口占用等启动失败: 回滚状态, 清理缓存, 返回 False
+                log.error("[%s] 启动实例失败: %s", inst.name, e)
+                self._config_cache.pop(inst.name, None)
+                inst.running = False
+                db.session.commit()
+                raise
             with self._lock:
                 self._instances[inst.name] = srv
             inst.running = True