From b0cc8dca91e4444425ceda5325e94a93ad6ecece Mon Sep 17 00:00:00 2001
From: tr4v3ler <0xtr4v3ler@gmail.com>
Date: Sat, 11 Apr 2026 15:07:06 +0800
Subject: [PATCH] Avoid merging category-companies@cn into geolocation-cn.

Keep company-wide @cn buckets like google@cn out of the generic mainland direct list so generated geolocation-cn does not classify blocked Google service domains as direct-connect targets.

Made-with: Cursor
---
 main.go      | 10 ++++++++++
 main_test.go | 41 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 51 insertions(+)
 create mode 100644 main_test.go

diff --git a/main.go b/main.go
index 65eae47..e598910 100644
--- a/main.go
+++ b/main.go
@@ -246,6 +246,13 @@ func mergeTags(data map[string][]geosite.Item) {
 		codeList = append(codeList, code)
 	}
 	var cnCodeList []string
+	// Company-wide @cn buckets may contain domains that are still unsuitable for
+	// a generic mainland-direct list. For example, category-companies@cn pulls in
+	// google@cn entries such as ssl.gstatic.com and fonts.googleapis.com, which
+	// makes geolocation-cn too broad for common direct-routing use.
+	geolocationCNMergeExclusions := map[string]bool{
+		"category-companies@cn": true,
+	}
 	for _, code := range codeList {
 		codeParts := strings.Split(code, "@")
 		if len(codeParts) != 2 {
@@ -260,6 +267,9 @@ func mergeTags(data map[string][]geosite.Item) {
 		if strings.HasSuffix(codeParts[0], "-cn") || strings.HasSuffix(codeParts[0], "-!cn") {
 			continue
 		}
+		if geolocationCNMergeExclusions[code] {
+			continue
+		}
 		cnCodeList = append(cnCodeList, code)
 	}
 	for _, code := range codeList {
diff --git a/main_test.go b/main_test.go
new file mode 100644
index 0000000..8d67543
--- /dev/null
+++ b/main_test.go
@@ -0,0 +1,41 @@
+package main
+
+import (
+	"testing"
+
+	"github.com/sagernet/sing-box/common/geosite"
+)
+
+func hasItem(items []geosite.Item, want geosite.Item) bool {
+	for _, item := range items {
+		if item == want {
+			return true
+		}
+	}
+	return false
+}
+
+func TestMergeTagsSkipsCategoryCompaniesAtCN(t *testing.T) {
+	companyItem := geosite.Item{Type: geosite.RuleTypeDomain, Value: "ssl.gstatic.com"}
+	devItem := geosite.Item{Type: geosite.RuleTypeDomain, Value: "pkg.go.dev"}
+	baseItem := geosite.Item{Type: geosite.RuleTypeDomain, Value: "example.cn"}
+
+	data := map[string][]geosite.Item{
+		"geolocation-cn":        {baseItem},
+		"category-companies@cn": {companyItem},
+		"category-dev@cn":       {devItem},
+	}
+
+	mergeTags(data)
+
+	got := data["geolocation-cn"]
+	if !hasItem(got, baseItem) {
+		t.Fatalf("base geolocation-cn item missing after merge")
+	}
+	if !hasItem(got, devItem) {
+		t.Fatalf("expected category-dev@cn item to still merge into geolocation-cn")
+	}
+	if hasItem(got, companyItem) {
+		t.Fatalf("did not expect category-companies@cn item to merge into geolocation-cn")
+	}
+}