// NextStartTag skips everything until we find a start tag of the given names.
func NextStartTag(t *html.Tokenizer, tagNames ...string) (html.Token, error) {
for {
switch t.Next() {
case html.ErrorToken:
return html.Token{}, t.Err()
case html.SelfClosingTagToken, html.StartTagToken:
tt := t.Token()
for _, v := range tagNames {
if v == tt.Data {
return tt, nil
}
}
}
}
panic("unreachable")
}
// TextInTag returns nodes containing all text until the given startTag is
// closed, including the tags of the given names.
//
// It expects well-formed HTML: mismatching closing tags will result in error.
func TextInTag(t *html.Tokenizer, startTag string, tagNames ...string) ([]Node, error) {
stack := []string{startTag}
var c []Node
for len(stack) > 0 {
// A token can be of the following types:
//
// html.ErrorToken
// html.TextToken
// html.StartTagToken
// html.EndTagToken
// html.SelfClosingTagToken
// html.CommentToken
// html.DoctypeToken
switch t.Next() {
case html.ErrorToken:
return nil, fmt.Errorf("unclosed tags: %v", stack)
case html.SelfClosingTagToken:
tt := t.Token()
tag := tt.Data
SelfClosingTagLoop:
for _, v := range tagNames {
if v == tag {
c = append(c, Node{Token: tt})
break SelfClosingTagLoop
}
}
case html.StartTagToken:
tt := t.Token()
tag := tt.Data
found := false
StartTagLoop:
for _, v := range tagNames {
if v == tag {
found = true
list, err := TextInTag(t, tag, tagNames...)
if err != nil {
return nil, err
}
c = append(c, Node{Token: tt, List: list})
break StartTagLoop
}
}
if !found {
stack = append(stack, tag)
}
case html.EndTagToken:
var err error
tag, _ := t.TagName()
if stack, err = popTag(stack, string(tag)); err != nil {
return nil, err
}
case html.TextToken:
tt := t.Token()
tt.Data = strings.TrimSpace(tt.Data)
if len(tt.Data) > 0 {
c = append(c, Node{Token: tt})
}
}
}
return c, nil
}