finalize the blocklist feature

2 files changed, 95 insertions, 16 deletions

diff --git a/internal/blocklist/blocklist.go b/internal/blocklist/blocklist.go
index 34a456d..84d758c 100644
--- a/internal/blocklist/blocklist.go
+++ b/internal/blocklist/blocklist.go
@@ -24,30 +24,50 @@ func Load(path string) (*Blocklist, error) {
 	sc := bufio.NewScanner(f)
 	for sc.Scan() {
 		line := strings.TrimSpace(sc.Text())
-		if line == "" || line[0] == '#' {
+		if line == "" || line[0] == '#' || line[0] == '!' {
 			continue
 		}
-		line = strings.ToLower(line)
-		line = strings.TrimSuffix(line, ".")
-		bl.domains[line] = struct{}{}
+		domain := parseLine(line)
+		if domain != "" && !strings.HasPrefix(domain, "*.") {
+			bl.domains[strings.ToLower(domain)] = struct {}{}
+		}
 	}
 	return bl, sc.Err()
 }
 
-func (bl *Blocklist) Blocked(qname string) bool {
-	qname = strings.ToLower(qname)
-	qname = strings.TrimSuffix(qname, ".")
+func parseLine(line string) string {
+	line = strings.TrimSpace(line)
 
-	if _, ok := bl.domains[qname]; ok {
-		return true
+	// 1. Hosts: "0.0.0.0 domain" "127.0.0.1 domain # comment"
+	if strings.HasPrefix(line, "0.0.0.0 ") {
+		line = line[8:]
+	} else if strings.HasPrefix(line, "127.0.0.1 ") {
+		line = line[10:]
+	} else if strings.HasPrefix(line, "||") {
+		// 2. AdGuard / uBlock: "||domain^$attr"
+		line = line[2:]
+		// Cut at ^ or $
+		if idx := strings.IndexAny(line, "^$"); idx >= 0 {
+			line = line[:idx]
+		}
 	}
 
-	for i := 0; i < len(qname); i++ {
-		if qname[i] == '.' {
-			if _, ok := bl.domains[qname[i+1:]]; ok {
-				return true
-			}
-		}
+	if idx := strings.IndexAny(line, " \t#"); idx >= 0 {
+		line = line[:idx]
+	}
+
+	line = strings.TrimSpace(line)
+	line = strings.TrimSuffix(line, ".")
+
+	if strings.ContainsAny(line, "/:\\?") {
+		return ""
 	}
-	return false
+	return line
+}
+
+func (bl *Blocklist) Blocked(qname string) bool {
+	qname = strings.ToLower(qname)
+	qname = strings.TrimSuffix(qname, ".")
+	_, ok := bl.domains[qname]
+	return ok
 }
diff --git a/internal/blocklist/blocklist_test.go b/internal/blocklist/blocklist_test.go
new file mode 100644
index 0000000..0104bc9
--- /dev/null
+++ b/internal/blocklist/blocklist_test.go
@@ -0,0 +1,59 @@
+package blocklist
+
+import (
+	"os"
+	"testing"
+)
+
+func TestParseLine(t *testing.T) {
+	tests := []struct{in,out string}{
+		{"0.0.0.0 domain.com", "domain.com"},
+		{"0.0.0.0 domain.com # foo", "domain.com"},	
+		{"0.0.0.0 *.domain.com", "*.domain.com"},
+		{"||domain.com^", "domain.com"},
+		{"||domain.com^$script", "domain.com"},
+		{"",""},
+		{"# comment", ""},
+		{"||*.domain.com^", "*.domain.com"},
+	}
+	for _, tt := range tests {
+		if got := parseLine(tt.in); got != tt.out {
+			t.Errorf("parseLine(%q) = %q, want %q", tt.in, got, tt.out)
+		}
+	}
+}
+
+func TestBlocked(t *testing.T) {
+	tmp, _ := os.CreateTemp("", "blocklist-*.txt")
+	tmp.WriteString("radhitya.org\n")
+	tmp.WriteString("0.0.0.0 ads.test.com\n")
+	tmp.WriteString("||example.net^\n")
+	tmp.WriteString("||*.domain.com^\n")
+	tmp.Close()
+	defer os.Remove(tmp.Name())
+
+	bl, err := Load(tmp.Name())
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	if !bl.Blocked("radhitya.org") {
+		t.Error("radhitya.org shall be blocked")
+	}
+	if bl.Blocked("alif.radhitya.org") {
+		t.Error("alif.radhitya.org shall NOT be blocked")
+	}
+	if !bl.Blocked("ads.test.com") {
+		t.Error("ads.test.com shall be blocked (hosts format)")
+	}
+	if !bl.Blocked("example.net") {
+		t.Error("example.net shall be blocked (adguard format)")
+	}
+
+	if bl.Blocked("domain.com") {
+		t.Error("domain.com shall NOT be blocked (wildcard skipped)")
+	}
+	if bl.Blocked("sub.domain.com") {
+		t.Error("sub.domain.com shall NOT be blocked (wildcard skipped)")
+	}
+}