Merge pull request #18 from pzaino/develop

fixed a bug found during integration tests
pzaino · Jan 22, 2024 · c3ad40b · c3ad40b
2 parents 97c1dff + b3ea7fb
commit c3ad40b
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 3 deletions.
diff --git a/pkg/crawler/crawler.go b/pkg/crawler/crawler.go
@@ -342,10 +342,20 @@ func isExternalLink(sourceURL, linkURL string) bool {
 
 	// Takes the substring that correspond to the 1st and 2nd level domain (e.g., google.com)
 	// regardless the number of subdomains
+	var srcDomainName string
 	srcFqdnArr := strings.Split(sourceParsed.Hostname(), ".")
-	srcDomainName := strings.Join(srcFqdnArr[len(srcFqdnArr)-2:], ".")
+	if len(srcFqdnArr) < 3 {
+		srcDomainName = strings.Join(srcFqdnArr, ".")
+	} else {
+		srcDomainName = strings.Join(srcFqdnArr[len(srcFqdnArr)-2:], ".")
+	}
 	linkFqdnArr := strings.Split(linkParsed.Hostname(), ".")
-	linkDomainName := strings.Join(linkFqdnArr[len(linkFqdnArr)-2:], ".")
+	var linkDomainName string
+	if len(linkFqdnArr) < 3 {
+		linkDomainName = strings.Join(linkFqdnArr, ".")
+	} else {
+		linkDomainName = strings.Join(linkFqdnArr[len(linkFqdnArr)-2:], ".")
+	}
 
 	// Compare hostnames
 	return srcDomainName != linkDomainName

diff --git a/pkg/crawler/crawler_test.go b/pkg/crawler/crawler_test.go
@@ -41,7 +41,7 @@ func TestExtractLinks(t *testing.T) {
 	}
 }
 
-func Test_isExternalLink(t *testing.T) {
+func TestIsExternalLink(t *testing.T) {
 	type args struct {
 		sourceURL string
 		linkURL   string
@@ -57,6 +57,7 @@ func Test_isExternalLink(t *testing.T) {
 		{"test3", args{"https://www.google.com", "https://www.google.com/test/test"}, false},
 		{"test4", args{"https://www.example.com", "https://www.google.com/test/test/test"}, true},
 		{"test5", args{"https://data.example.com", "https://www.example.com"}, false},
+		{"test6", args{"www.apps.com", "javascript:void(0)"}, false},
 	}
 
 	for _, tt := range tests {