MM-9072/MM-10185 Force-convert the encoding of OpenGraph metadata to UTF-8 (#8631)

* Force-convert non-UTF8 HTML to UTF8 before opengraph processing * Split the force-encoding function * Add benchmark Test for the forceHTMLEncodingToUTF8() ``` Running tool: /home/comet/go-v1.9.2/bin/go test -benchmem -run=^$ github.com/mattermost/mattermost-server/app -bench ^BenchmarkForceHTMLEncodingToUTF8$ [03:32:58 KST 2018/04/21] [INFO] (github.com/mattermost/mattermost-server/app.TestMain:28) -test.run used, not creating temporary containers goos: linux goarch: amd64 pkg: github.com/mattermost/mattermost-server/app BenchmarkForceHTMLEncodingToUTF8/with_converting-4 100000 11201 ns/op 18704 B/op 32 allocs/op BenchmarkForceHTMLEncodingToUTF8/without_converting-4 300000 3931 ns/op 4632 B/op 13 allocs/op PASS ok github.com/mattermost/mattermost-server/app 2.703s Success: Benchmarks passed. ``` * Remove an unnecessary constraint * Add pre-check if content-type header is already utf-8 * Move the checking for utf-8 into forceHTMLEncodingToUTF8() for testing * Revert df3f347213faa0d023c26d201fa6531f46391086..HEAD, without Gopkg.lock
2018-05-01 01:54:11 +09:00
--- a/app/post_test.go
+++ b/app/post_test.go
@@ -297,6 +297,34 @@ func TestImageProxy(t *testing.T) {
 	}
 }

+func BenchmarkForceHTMLEncodingToUTF8(b *testing.B) {
+	HTML := `
+		<html>
+			<head>
+				<meta property="og:url" content="https://example.com/apps/mattermost">
+				<meta property="og:image" content="https://images.example.com/image.png">
+			</head>
+		</html>
+	`
+	ContentType := "text/html; utf-8"
+
+	b.Run("with converting", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			r := forceHTMLEncodingToUTF8(strings.NewReader(HTML), ContentType)
+
+			og := opengraph.NewOpenGraph()
+			og.ProcessHTML(r)
+		}
+	})
+
+	b.Run("without converting", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			og := opengraph.NewOpenGraph()
+			og.ProcessHTML(strings.NewReader(HTML))
+		}
+	})
+}
+
 func TestMakeOpenGraphURLsAbsolute(t *testing.T) {
 	for name, tc := range map[string]struct {
 		HTML       string