From c102e2489df8cfa2b8ee148c7ba91e2e98239afe Mon Sep 17 00:00:00 2001
From: Ryan Yin <xiaoyin_c@qq.com>
Date: Fri, 9 Oct 2026 18:03:17 +0800
Subject: [PATCH] =?UTF-8?q?fix(mihomo):=20name=20the=20Singapore=20nodes?=
 =?UTF-8?q?=20instead=20of=20matching=20=E6=96=B0?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The single-char keyword also matched 新西兰, so the Singapore group held a New
Zealand node and url-test flipped the exit country mid-session.
---
 modules/nixos/desktop/networking/mihomo/policy.yaml | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/modules/nixos/desktop/networking/mihomo/policy.yaml b/modules/nixos/desktop/networking/mihomo/policy.yaml
index 3e615d2a..14143c6a 100644
--- a/modules/nixos/desktop/networking/mihomo/policy.yaml
+++ b/modules/nixos/desktop/networking/mihomo/policy.yaml
@@ -14,8 +14,9 @@
 # `filter` (Go regex, case-insensitive). The generator also adds "🌍 其他地区"
 # (manual, everything not matched below) and "🌐 全部地区" (manual, all nodes);
 # AUTO is the all-node auto pick. Edit the regions freely; the single-char
-# keywords (港/台/日/新/美/韩) deliberately catch naming variants like
-# 美西, 日服 and 狮城.
+# keywords (港/台/日/美/韩) deliberately catch naming variants like 美西, 日服
+# and 狮城. Spell a country out where one character would also catch a
+# neighbour: 新 alone catches 新西兰.
 node_groups:
   - name: "🇭🇰 香港"
     filter: "(?i)(港|hk|hong ?kong|🇭🇰)"
@@ -24,7 +25,7 @@ node_groups:
   - name: "🇯🇵 日本"
     filter: "(?i)(日|jp|japan|东京|🇯🇵)"
   - name: "🇸🇬 新加坡"
-    filter: "(?i)(新|狮城|sg|singapore|🇸🇬)"
+    filter: "(?i)(新加坡|狮城|sg|singapore|🇸🇬)"
   - name: "🇰🇷 韩国"
     filter: "(?i)(韩|kr|korea|🇰🇷)"
   - name: "🇺🇸 美国"