Add support for different Maven POM encoding (#25873)

Fixes #25853

- Maven POM files aren't always UTF-8 encoded.
- Reject the upload of unparsable POM files

2 files changed, 23 insertions, 1 deletions

diff --git a/modules/packages/maven/metadata.go b/modules/packages/maven/metadata.go
index be057c8c07..42aa250718 100644
--- a/modules/packages/maven/metadata.go
+++ b/modules/packages/maven/metadata.go
@@ -8,6 +8,8 @@ import (
 	"io"
 
 	"code.gitea.io/gitea/modules/validation"
+
+	"golang.org/x/net/html/charset"
 )
 
 // Metadata represents the metadata of a Maven package
@@ -52,7 +54,10 @@ type pomStruct struct {
 // ParsePackageMetaData parses the metadata of a pom file
 func ParsePackageMetaData(r io.Reader) (*Metadata, error) {
 	var pom pomStruct
-	if err := xml.NewDecoder(r).Decode(&pom); err != nil {
+
+	dec := xml.NewDecoder(r)
+	dec.CharsetReader = charset.NewReaderLabel
+	if err := dec.Decode(&pom); err != nil {
 		return nil, err
 	}
 
diff --git a/modules/packages/maven/metadata_test.go b/modules/packages/maven/metadata_test.go
index f4bc84268e..e675467730 100644
--- a/modules/packages/maven/metadata_test.go
+++ b/modules/packages/maven/metadata_test.go
@@ -8,6 +8,7 @@ import (
 	"testing"
 
 	"github.com/stretchr/testify/assert"
+	"golang.org/x/text/encoding/charmap"
 )
 
 const (
@@ -69,4 +70,20 @@ func TestParsePackageMetaData(t *testing.T) {
 		assert.Equal(t, dependencyArtifactID, m.Dependencies[0].ArtifactID)
 		assert.Equal(t, dependencyVersion, m.Dependencies[0].Version)
 	})
+
+	t.Run("Encoding", func(t *testing.T) {
+		// UTF-8 is default but the metadata could be encoded differently
+		pomContent8859_1, err := charmap.ISO8859_1.NewEncoder().String(
+			strings.ReplaceAll(
+				pomContent,
+				`<?xml version="1.0"?>`,
+				`<?xml version="1.0" encoding="ISO-8859-1"?>`,
+			),
+		)
+		assert.NoError(t, err)
+
+		m, err := ParsePackageMetaData(strings.NewReader(pomContent8859_1))
+		assert.NoError(t, err)
+		assert.NotNil(t, m)
+	})
 }