etreeutils: sort attrs with matching namespaces by their NS URIs

In several of the XML Canonicalization specifications it states that
attributes with matching namespaces need to be sorted by their
Namespace URIs and not the prefixes.

> However, the namespace URI was selected as the primary key because
> this is closer to the intent of the Namespaces in XML specification,
> which is to identify namespaces by URI and local name, not by a prefix
> and local name.

The specifications call out that sorting by prefix is easier, but incorrect.

Further, the specifications state:

> An element's attribute nodes are sorted lexicographically with
> namespace URI as the primary key and local name as the secondary
> key (an empty namespace URI is lexicographically least).

Sources:

- https://www.w3.org/TR/xml-c14n11/#SortByNSURI
- https://www.w3.org/TR/2007/CR-xml-c14n11-20070621/#SortByNSURI
- https://www.w3.org/TR/2001/REC-xml-c14n-20010315#SortByNSURI

etreeutils/sort.go

@@ -55,12 +55,25 @@ func (a SortedAttrs) Less(i, j int) bool {
 		return false
 	}
 
-	// Wow. We're still going. Finally, attributes in the same namespace should be
-	// sorted by key. Attributes in different namespaces should be sorted by the
-	// actual namespace (_not_ the prefix). For now just use the prefix.
+	// Attributes with the same prefix should be sorted by their keys.
 	if a[i].Space == a[j].Space {
 		return a[i].Key < a[j].Key
 	}
 
-	return a[i].Space < a[j].Space
+	// Attributes in the same namespace are sorted by their Namespace URI, not the prefix.
+	if a[i].Key == a[j].Key {
+		var leftNS, rightNS etree.Attr
+		for n := range a {
+			if a[i].Space == a[n].Key {
+				leftNS = a[n]
+			}
+			if a[j].Space == a[n].Key {
+				rightNS = a[n]
+			}
+		}
+		// Sort based on the NS URIs
+		return leftNS.Value < rightNS.Value
+	}
+
+	return a[i].Key < a[j].Key
 }

etreeutils/sort_test.go

@@ -0,0 +1,30 @@
+package etreeutils
+
+import (
+	"sort"
+	"testing"
+
+	"github.com/beevik/etree"
+	"github.com/stretchr/testify/require"
+)
+
+func TestSortedAttrs(t *testing.T) {
+	// Adapted from https://www.w3.org/TR/2001/REC-xml-c14n-20010315#Example-SETags
+	input := `<e5 a:attr="out" b:attr="sorted" attr2="all" attr="I m" xmlns:b="http://www.ietf.org" xmlns:a="http://www.w3.org" xmlns="http://example.org"></e5>`
+	expected := `<e5 xmlns="http://example.org" xmlns:a="http://www.w3.org" xmlns:b="http://www.ietf.org" attr="I m" attr2="all" b:attr="sorted" a:attr="out"></e5>`
+
+	inDoc := etree.NewDocument()
+	inDoc.ReadFromString(input)
+
+	outElm := inDoc.Root().Copy()
+	sort.Sort(SortedAttrs(outElm.Attr))
+	outDoc := etree.NewDocument()
+	outDoc.SetRoot(outElm)
+	outDoc.WriteSettings = etree.WriteSettings{
+		CanonicalEndTags: true,
+	}
+
+	outStr, err := outDoc.WriteToString()
+	require.NoError(t, err)
+	require.Equal(t, expected, outStr)
+}