Detect iMessage-Preview

iMessage uses a bot to show a preview of links in the conversations. It
was wrongly seen by user_agent as Safari 9. This commit adds a hackish
way to detect it, as it impersonates both the Twitter and Facebook bots.

Author: nono

all_test.go

@@ -124,6 +124,11 @@ var uastrings = []struct {
 		ua:       "APIs-Google (+https://developers.google.com/webmasters/APIs-Google.html)",
 		expected: "Browser:APIs-Google Bot:true Mobile:false",
 	},
+	{
+		title:    "iMessage-preview",
+		ua:       "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_11_1) AppleWebKit/601.2.4 (KHTML, like Gecko) Version/9.0.1 Safari/601.2.4 facebookexternalhit/1.1 Facebot Twitterbot/1.0",
+		expected: "Mozilla:5.0 Platform:Macintosh Browser:iMessage-Preview-9.0.1 Bot:true Mobile:false",
+	},
 
 	// Internet Explorer
 	{

bot.go

@@ -57,6 +57,26 @@ func (p *UserAgent) googleOrBingBot() bool {
 	return p.undecided
 }
 
+// Returns true if we think that it is iMessage-Preview. This function also
+// modifies some attributes in the receiver accordingly.
+func (p *UserAgent) iMessagePreview() bool {
+	// iMessage-Preview doesn't advertise itself. We have a to rely on a hack
+	// to detect it: it impersonates both facebook and twitter bots.
+	// See https://medium.com/@siggi/apples-imessage-impersonates-twitter-facebook-bots-when-scraping-cef85b2cbb7d
+	if strings.Index(p.ua, "facebookexternalhit") == -1 {
+		return false
+	}
+	if strings.Index(p.ua, "Twitterbot") == -1 {
+		return false
+	}
+	p.bot = true
+	p.browser.Name = "iMessage-Preview"
+	p.browser.Engine = ""
+	p.browser.EngineVersion = ""
+	// We don't set the mobile flag because iMessage can be on iOS (mobile) or macOS (not mobile).
+	return true
+}
+
 // Set the attributes of the receiver as given by the parameters. All the other
 // parameters are set to empty.
 func (p *UserAgent) setSimple(name, version string, bot bool) {

operating_systems.go

@@ -4,7 +4,9 @@
 
 package user_agent
 
-import "strings"
+import (
+	"strings"
+)
 
 // Represents full information on the operating system extracted from the user agent.
 type OSInfo struct {
@@ -100,7 +102,7 @@ func webkit(p *UserAgent, comment []string) {
 		} else if len(comment) < 2 {
 			p.localization = comment[0]
 		} else if len(comment) < 3 {
-			if !p.googleOrBingBot() {
+			if !p.googleOrBingBot() && !p.iMessagePreview() {
 				p.os = normalizeOS(comment[1])
 			}
 		} else {