searchelemattr

README

@@ -1,3 +1,5 @@
 Basic HTML scraper
 
-Was done long-long time ago. I'm surprised how I managed to force myself to make it work.
+Was done long-long time ago. I'm surprised how I managed to force myself to make it work.
+
+TODO: add context, to prevent endless crawling

match.go

@@ -17,13 +17,19 @@ var body = &html.Node{
 
 var ErrNotAnElementNode = errors.New("not an ElementNode")
 
+func parseFragment(s string) (*html.Node, error) {
+	n, err := html.ParseFragment(strings.NewReader(s), body)
+	if err != nil {
+		return n, err
+	}
+	return n[0], nil
+}
+
 func MatchElemAttr(s string, n2 *html.Node) (bool, error) {
-	r := strings.NewReader(s)
+	n1, err := parseFragemnt(s)
-	n, err := html.ParseFragment(r, body)
 	if err != nil {
 		return false, err
 	}
-	n1 := n[0]
 
 	if !(n1.Type == html.ElementNode &&
 		n2.Type == html.ElementNode) {
@@ -36,3 +42,32 @@ func MatchElemAttr(s string, n2 *html.Node) (bool, error) {
 	}
 	return false, nil
 }
+
+
+func SearchElemAttr2(s string, n2 *html.Node) (chan *html.Node, error) {
+	ch := make(chan *html.Node)
+	var crawl func(*html.Node)
+	crawl = func(n *html.Node) {
+		matches, err := MatchElemAttr(s, n)
+		if err != nil {
+			panic(err)
+		}
+		if matches {
+			ch <- n
+		}
+		c := n.FirstChild
+		for c != nil {
+			crawl(c)
+			c = c.NextSibling
+		}
+	}
+	go func() {
+		defer close(ch)
+		crawl(n2)
+	}()
+
+	if r := recover(); r != nil {
+		return nil, r
+	}
+	return <-ch, nil
+}