Putting more state into unxml.Reader

1 files changed, 47 insertions, 24 deletions

diff --git a/unxml.go b/unxml.go
index d8d588e..f5b7c62 100644
--- a/unxml.go
+++ b/unxml.go
@@ -12,8 +12,11 @@ import (
 )
 
 type Reader struct {
-	reader io.Reader
-	tagmap map[string]bool
+	reader    io.Reader
+	tagmap    map[string]bool
+	lastread  []byte
+	count     int
+	tokenizer *html.Tokenizer
 }
 
 func NewReaderKeepTags(r io.Reader, tagstokeep []string) *Reader {
@@ -29,42 +32,62 @@ func NewReaderKeepTags(r io.Reader, tagstokeep []string) *Reader {
 	}
 
 	return &Reader{reader: r,
-		tagmap: tagmap,
+		tagmap:    tagmap,
+		tokenizer: html.NewTokenizer(r),
 	}
 }
 
-func (r *Reader) Read(out []byte) (n int, err error) {
-	hr := html.NewTokenizer(r.reader)
-	buf := bytes.NewBuffer(make([]byte, len(out)))
-	depth := 0
+func (r *Reader) Read(out []byte) (int, error) {
+	fmt.Fprintf(os.Stderr, "Read has been called.\n")
+	var err error
+
+	r.count = 0
+	n := 0
+	lenout := len(out)
+	if lenout == 0 {
+		return r.count, nil
+	}
+
+	lenlr := len(r.lastread)
+	if lenlr > 0 {
+		n = copy(out[0:], r.lastread)
+		r.count += n
+		r.lastread = make([]byte, len(out))
+		lenout -= n
+	}
+
+	buf := bytes.NewBuffer(out)
 
 	for {
-		tt := hr.Next()
+		tt := r.tokenizer.Next()
 
 		switch tt {
 		case html.ErrorToken:
-			fmt.Fprintf(os.Stderr, "There was an error when reading from the underlying os.Reader: %s", tt)
-			return 0, hr.Err()
+			fmt.Fprintf(os.Stderr, "There was an error when parsing the html: %s, %s\n", tt, r.tokenizer.Err())
+			return r.count, io.EOF
 
 		case html.TextToken:
-			if depth > 0 {
-				// emitBytes should copy the []byte it receives,
-				// if it doesn't process it immediately.
-				n, err = buf.Write(hr.Text())
+			text := r.tokenizer.Text()
+			lentext := len(text)
+			if lentext <= lenout {
+				//n, err = buf.Write(r.tokenizer.Text())
+				n = copy(out[r.count:], text)
+				r.count += n
+				lenout -= n
+				fmt.Printf("HAD SPACE: n, err: %d, %s; out: %s; buf: %s\n", n, err, out, buf)
+			} else {
+				n = copy(out[r.count:], text[:lenout-1])
+				r.count += n
+				r.lastread = text[lenout-1:]
+				fmt.Printf("HAD NO SPACE: n, err: %d, %s; out: %s; buf: %s\n", n, err, out, buf)
+				return r.count, err
 			}
 
 		case html.StartTagToken, html.EndTagToken:
-			tn, _ := hr.TagName()
-
-			if len(tn) == 1 && tn[0] == 'a' {
-				if tt == html.StartTagToken {
-					depth++
-				} else {
-					depth--
-				}
-			}
+			_, _ = r.tokenizer.TagName()
+			//fmt.Printf("TagName: %s\n", tn)
 		}
 	}
 
-	return
+	return r.count, err
 }