From 3a9d7a36a85cf8ae1098f213186ef83bbc3c56ba Mon Sep 17 00:00:00 2001
From: Max Resnick <max@ofmax.li>
Date: Wed, 26 Mar 2025 22:21:03 -0700
Subject: fix: domain normalization, output format

---
 main.go | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

(limited to 'main.go')

diff --git a/main.go b/main.go
index 051e6cd..c7e4212 100644
--- a/main.go
+++ b/main.go
@@ -59,6 +59,11 @@ func main() {
 	w := bufio.NewWriter(f)
 	defer w.Flush()
 
+	if _, err := fmt.Fprintln(w, "server:"); err != nil {
+		slog.Error("failed to write header", "error", err)
+		os.Exit(1)
+	}
+
 	domains := make(map[string]struct{})
 	for i, url := range urls {
 		slog.Info("fetching domains", "url", url, "progress", fmt.Sprintf("%d/%d", i+1, len(urls)))
@@ -104,7 +109,6 @@ func fetchDomainsAndWrite(url string, w *bufio.Writer, seen map[string]struct{})
 	}
 	defer resp.Body.Close()
 
-	fmt.Fprint(w, "server:\n")
 	var count int
 	scanner := bufio.NewScanner(resp.Body)
 	for scanner.Scan() {
@@ -127,6 +131,10 @@ func fetchDomainsAndWrite(url string, w *bufio.Writer, seen map[string]struct{})
 
 		// Normalize and validate domain format
 		domain = strings.ToLower(strings.TrimSpace(domain))
+		if idx := strings.Index(domain, "$"); idx != -1 {
+			domain = domain[:idx]
+		}
+		domain = strings.Trim(domain, "|^")
 		if domain == "" || !strings.Contains(domain, ".") || strings.HasPrefix(domain, ".") || strings.HasSuffix(domain, ".") {
 			continue
 		}
@@ -138,7 +146,7 @@ func fetchDomainsAndWrite(url string, w *bufio.Writer, seen map[string]struct{})
 		seen[domain] = struct{}{}
 		count++
 
-		if _, err := fmt.Fprintf(w, "	local-zone: %q refuse\n", domain); err != nil {
+		if _, err := fmt.Fprintf(w, "\tlocal-zone: %q refuse\n", domain); err != nil {
 			return fmt.Errorf("failed to write domain: %w", err)
 		}
 	}
@@ -147,6 +155,10 @@ func fetchDomainsAndWrite(url string, w *bufio.Writer, seen map[string]struct{})
 		return fmt.Errorf("scanner error: %w", err)
 	}
 
+	if err := w.Flush(); err != nil {
+		return fmt.Errorf("failed to flush writer: %w", err)
+	}
+
 	slog.Info("processed url", "url", url, "new_domains", count)
 	return nil
 }
-- 
cgit v1.2.3