add punycode normalization for URL

readium · Oct 15, 2024 · 41e1716 · 41e1716
1 parent e8791be
commit 41e1716
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 0 deletions.
diff --git a/pkg/util/url/url.go b/pkg/util/url/url.go
@@ -9,6 +9,7 @@ import (
 	"strings"
 
 	"github.com/readium/go-toolkit/pkg/internal/extensions"
+	"golang.org/x/net/idna"
 )
 
 /*
@@ -302,6 +303,11 @@ func (u AbsoluteURL) Normalize() URL {
 	}
 
 	u.url.Scheme = SchemeFromString(u.url.Scheme).String()
+	asciiHost, err := idna.ToASCII(u.url.Host)
+	if err == nil {
+		u.url.Host = asciiHost
+	}
+
 	return AbsoluteURL{url: u.url, scheme: Scheme(u.url.Scheme)}
 }
 

diff --git a/pkg/util/url/url_test.go b/pkg/util/url/url_test.go
@@ -420,6 +420,10 @@ func TestNormalize(t *testing.T) {
 	u, _ := URLFromString("HTTP://example.com/foo")
 	assert.Equal(t, "http://example.com/foo", u.Normalize().String())
 
+	// Host becomes punycode equivalent.
+	u, _ = URLFromString("http://도메인.com/foo")
+	assert.Equal(t, "http://xn--hq1bm8jm9l.com/foo", u.Normalize().String())
+
 	// Percent encoding of path is normalized.
 	u, _ = URLFromString("HTTP://example.com/c'est%20valide")
 	assert.Equal(t, "http://example.com/c'est%20valide", u.Normalize().String())