Normalize filenames before comparing.

Some multibyte characters can have multiple representations. Before
comparing strings, we need to normalize them. In this case we're
normalizing to normalized form C, but it shouldn't matter as long as
both strings are normalized to the same form.

Fixes src-d/go-git#495

Author: joshbetz

utils/merkletrie/noder/path.go

@@ -3,6 +3,8 @@ package noder
 import (
 	"bytes"
 	"strings"
+
+	"golang.org/x/text/unicode/norm"
 )
 
 // Path values represent a noder and its ancestors.  The root goes first
@@ -78,7 +80,11 @@ func (p Path) Compare(other Path) int {
 		case i == len(p):
 			return -1
 		default:
-			cmp := strings.Compare(p[i].Name(), other[i].Name())
+			form := norm.Form(norm.NFC)
+			this := form.String(p[i].Name())
+			that := form.String(other[i].Name())
+
+			cmp := strings.Compare(this, that)
 			if cmp != 0 {
 				return cmp
 			}