url: add benchmark

2 files changed, 15 insertions, 10 deletions

diff --git a/skate/url.go b/skate/url.go
index 92c9f61..f0edddf 100644
--- a/skate/url.go
+++ b/skate/url.go
@@ -19,16 +19,15 @@ var (
 // http://agingmind.utdallas.edu/facedb/view/neutral-faces.doi:10.3758/BF03206543.
 // http://10.3386/w20634https://doi.org/10.3386/w20634
 func SanitizeURL(s string) string {
-	if len(s) > 14 && strings.Contains(s[14:], "http") {
-		var (
-			index   = suffixarray.New([]byte(s))
-			indices = index.Lookup([]byte("http"), -1)
-		)
-		if len(indices) > 1 {
-			s = s[0:indices[1]] // only use the first
-			s = strings.TrimRight(s, ":")
-			s = strings.TrimRight(s, ";")
-		}
+	var (
+		// seems to only be about 15% of total time spent
+		index   = suffixarray.New([]byte(s))
+		indices = index.Lookup([]byte("http"), -1)
+	)
+	if len(indices) > 1 {
+		s = s[0:indices[1]] // only use the first
+		s = strings.TrimRight(s, ":")
+		s = strings.TrimRight(s, ";")
 	}
 	// http://!!!:
 	// http://!
diff --git a/skate/url_test.go b/skate/url_test.go
index 8dddb09..c138a0e 100644
--- a/skate/url_test.go
+++ b/skate/url_test.go
@@ -106,3 +106,9 @@ func TestSanitizeURL(t *testing.T) {
 		}
 	}
 }
+
+func BenchmarkSanitizeURL(b *testing.B) {
+	for n := 0; n < b.N; n++ {
+		SanitizeURL(`http://acrf.com.au/2012/world-firsthpv-vaccina-tion-plan-will-protect-young-australian-men-from-cancer/`)
+	}
+}