fix: avoid caching cn rules when verifying geodata files

component/geodata/utils.go

@@ -1,7 +1,11 @@
 package geodata
 
 import (
+	"bufio"
+	"bytes"
 	"fmt"
+	"io"
+	"os"
 	"strings"
 
 	"github.com/metacubex/mihomo/common/singleflight"
@@ -50,14 +54,84 @@ func SetSiteMatcher(newMatcher string) {
 	}
 }
 
+func verifyGeodataReader(r io.Reader, fileSize int64) error {
+	if fileSize == 0 {
+		return fmt.Errorf("invalid geodata file: empty file")
+	}
+
+	br := bufio.NewReader(r)
+	var pos int64
+	var b [1]byte
+
+	for pos < fileSize {
+		// Each top-level entry starts with 0x0A (protobuf field 1, wire type 2).
+		if _, err := io.ReadFull(br, b[:]); err != nil {
+			return fmt.Errorf("invalid geodata file: %w", err)
+		}
+		pos++
+		if b[0] != 0x0A {
+			return fmt.Errorf("invalid geodata file: unexpected byte 0x%02X at offset %d", b[0], pos-1)
+		}
+
+		// Decode the entry length varint.
+		var entryLen uint64
+		var shift uint
+		for {
+			if _, err := io.ReadFull(br, b[:]); err != nil {
+				return fmt.Errorf("invalid geodata file: truncated varint at offset %d: %w", pos, err)
+			}
+			pos++
+			entryLen |= uint64(b[0]&0x7F) << shift
+			if b[0] < 0x80 {
+				break
+			}
+			shift += 7
+			if shift >= 64 {
+				return fmt.Errorf("invalid geodata file: varint overflow at offset %d", pos)
+			}
+		}
+
+		if entryLen == 0 {
+			return fmt.Errorf("invalid geodata file: zero-length entry at offset %d", pos)
+		}
+
+		if _, err := io.CopyN(io.Discard, br, int64(entryLen)); err != nil {
+			return fmt.Errorf("invalid geodata file: truncated entry at offset %d: %w", pos, err)
+		}
+		pos += int64(entryLen)
+	}
+
+	if pos != fileSize {
+		return fmt.Errorf("invalid geodata file: truncated (last entry ends at %d, file size %d)", pos, fileSize)
+	}
+
+	return nil
+}
+
+func verifyGeodataFile(path string) error {
+	f, err := os.Open(path)
+	if err != nil {
+		return err
+	}
+	defer f.Close()
+
+	info, err := f.Stat()
+	if err != nil {
+		return err
+	}
+	return verifyGeodataReader(f, info.Size())
+}
+
+func VerifyGeodataBytes(data []byte) error {
+	return verifyGeodataReader(bytes.NewReader(data), int64(len(data)))
+}
+
 func Verify(name string) error {
 	switch name {
 	case C.GeositeName:
-		_, err := LoadGeoSiteMatcher("CN")
-		return err
+		return verifyGeodataFile(C.Path.GeoSite())
 	case C.GeoipName:
-		_, err := LoadGeoIPMatcher("CN")
-		return err
+		return verifyGeodataFile(C.Path.GeoIP())
 	default:
 		return fmt.Errorf("not support name")
 	}

component/updater/update_geo.go

@@ -11,7 +11,6 @@ import (
 	"github.com/metacubex/mihomo/common/atomic"
 	"github.com/metacubex/mihomo/common/utils"
 	"github.com/metacubex/mihomo/component/geodata"
-	_ "github.com/metacubex/mihomo/component/geodata/standard"
 	"github.com/metacubex/mihomo/component/mmdb"
 	"github.com/metacubex/mihomo/component/resource"
 	C "github.com/metacubex/mihomo/constant"
@@ -107,8 +106,6 @@ func UpdateASN() (err error) {
 }
 
 func UpdateGeoIp() (err error) {
-	geoLoader, err := geodata.GetGeoDataLoader("standard")
-
 	vehicle := resource.NewHTTPVehicle(geodata.GeoIpUrl(), C.Path.GeoIP(), "", nil, defaultHttpTimeout, 0)
 	var oldHash utils.HashType
 	if buf, err := os.ReadFile(vehicle.Path()); err == nil {
@@ -125,8 +122,8 @@ func UpdateGeoIp() (err error) {
 		return fmt.Errorf("can't download GeoIP database file: no data")
 	}
 
-	if _, err = geoLoader.LoadIPByBytes(data, "cn"); err != nil {
-		return fmt.Errorf("invalid GeoIP database file: %s", err)
+	if err = geodata.VerifyGeodataBytes(data); err != nil {
+		return fmt.Errorf("invalid GeoIP database file: %w", err)
 	}
 
 	defer geodata.ClearGeoIPCache()
@@ -137,8 +134,6 @@ func UpdateGeoIp() (err error) {
 }
 
 func UpdateGeoSite() (err error) {
-	geoLoader, err := geodata.GetGeoDataLoader("standard")
-
 	vehicle := resource.NewHTTPVehicle(geodata.GeoSiteUrl(), C.Path.GeoSite(), "", nil, defaultHttpTimeout, 0)
 	var oldHash utils.HashType
 	if buf, err := os.ReadFile(vehicle.Path()); err == nil {
@@ -155,8 +150,8 @@ func UpdateGeoSite() (err error) {
 		return fmt.Errorf("can't download GeoSite database file: no data")
 	}
 
-	if _, err = geoLoader.LoadSiteByBytes(data, "cn"); err != nil {
-		return fmt.Errorf("invalid GeoSite database file: %s", err)
+	if err = geodata.VerifyGeodataBytes(data); err != nil {
+		return fmt.Errorf("invalid GeoSite database file: %w", err)
 	}
 
 	defer geodata.ClearGeoSiteCache()