Adding non-escaped Ampersands to HTML with Nokogiri::XML::Builder

Question

I would like to add things like bullet points "•" and such to html using the XML Builder in Nokogiri, but everything is being escaped. How do I prevent it from being escaped?

I would like the result to be:

<span>&#8226;</span> 

rather than

<span>&amp;#8226;</span> 

What am I missing?

I'm just doing this:

xml.span { 
  xml.text "•\ " 
}

Thanks!

Answer 1

When you're setting the text of an element, you really are setting text, not HTML source. < and & don't have any special meaning in plain text.

So just type a bullet: '•'. Of course your source code and your XML file will have to be using the same encoding for that to come out right. If your XML file is UTF-8 but your source code isn't, you'd probably have to say '\xe2\x80\xa2' which is the UTF-8 byte sequence for the bullet character as a string literal.

(In general non-ASCII characters in Ruby 1.8 are tricky. The byte-based interfaces don't mesh too well with XML's world of all-text-is-Unicode.)

Answer 2

If you define

  class Nokogiri::XML::Builder
    def entity(code)
      doc = Nokogiri::XML("<?xml version='1.0'?><root>&##{code};</root>")
      insert(doc.root.children.first)
    end
  end

then this

  builder = Nokogiri::XML::Builder.new do |xml|
    xml.span {
      xml.text "I can has "
      xml.entity 8665
      xml.text " entity?"
    }
  end
  puts builder.to_xml

yields

<?xml version="1.0"?>
<span>I can has &#x2022; entity?</span>

 

PS this a workaround only, for a clean solution please refer to the libxml2 documentation (Nokogiri is built on libxml2) for more help. However, even these folks admit that handling entities can be quite ..err, cumbersome sometimes.