PuerkitoBio/goquery

doc.Find("body").Contents().Length() is greater than zero even though <body> is empty

Closed this issue · 2 comments

sognokdev commented

Hello,

The following program:

package main

import (
        "bytes"
        "fmt"
        "github.com/PuerkitoBio/goquery"
)

func numNodes(code string) {
        r := bytes.NewReader([]byte(code))

        doc, err := goquery.NewDocumentFromReader(r)
        if err != nil {
                panic(err)
        }

        fmt.Println("Length of selection =", doc.Find("body").Contents().Length())
}

func main() {
        code1 := "<html><head></head><body></body></html>"
        numNodes(code1)

        code2 := "<html><head></head><body></body>\n</html>"
        numNodes(code2)
}

produces this output:

Length of selection = 0
Length of selection = 1

Is it normal to get "Length of selection = 1" for the HTML code that is contained in the code2 variable? I expect to get "Length of selection = 0" because the <body> tag is empty.

mna commented

Hello,

It's normal due to how the html5 parser interprets this html - it moves the newline inside the <body> so it is no longer empty. When you see weird things with the html, it's always a good idea to print the document after it has been parsed into goquery to see what goquery sees, e.g.:

	doc, err := goquery.NewDocumentFromReader(r)
	if err != nil {
		panic(err)
	}
	fmt.Println(goquery.OuterHtml(doc.Selection))

You can see that in your second call, the actual document looks like this:

<html><head></head><body>
</body></html>

And since the Contents method selects not only elements, but also comments and text nodes, it has selected the text node containing the newline.

Hope this helps,
Martin

  • 👍1
sognokdev commented

Thank you very much for your explanation, mna!