Shorthands and validation (built-ins vs refinements) #3588

forthrin · 2025-12-21T13:03:54Z

forthrin
Dec 21, 2025

During time with Nokogiri, making some convenient refinements have been found necessary (see below):

Shorthand string to DOM
Shorthand node creation
Shorthand pile on elements
Ensuring valid documents (must <!DOCTYPE> and <title>, should <meta charset>, nice <meta property="og:image">)
Ensuring plain text (stripped with <br> as newlines)

But maybe comparable features are built-in? Examples much appreciated if so.

#!/usr/bin/env ruby

require 'nokogiri'

# Make life in the DOM easier
module Haohmaru
  refine String do
    def html
      Nokogiri.HTML5(self)
    end

    def xml
      Nokogiri.XML(self)
    end
  end

  refine Nokogiri::XML::Element do
    def text
      css('br').each { it.replace("\n") }
      super.strip
    end

    def merge(opt)
      opt&.each { |k, v| self[k.to_s] = v }
    end
  end

  refine Nokogiri::XML::Node do
    def add(*)
      add_child(*)
      self
    end

    def tag(*opts)
      Nokogiri::XML::Node.new(opts.shift, self) do
        it.content = opts.shift if opts.first.is_a?(String)
        it.merge(opts.shift)
        add(it)
      end
    end
  end
end

using Haohmaru

def html(title)
  ''.html.tap do
    it.at('head').tap do
      it.tag('title', title)
      it.tag('meta', { charset: 'utf-8' })
      it.tag('link', { rel: 'stylesheet', href: "/#{title}.css" })
    end
  end
end

page = html('Hello')
page.at('body').tag('div').add('Inline text').add('<br/>').tag('p', 'Hello World')
puts page.to_html(indent: 2) # lacks <!DOCTYPE html> and indentation

forthrin · 2025-12-27T05:48:39Z

forthrin
Dec 27, 2025
Author

@flavorjones: The general syntax for creating a DOM from a terse function tree was kindly answered in #3587 (comment), but questions 1-5 actually still remain. Will await response as not to re-invent the wheel.

0 replies

flavorjones · 2025-12-27T15:51:12Z

flavorjones
Dec 27, 2025
Maintainer

I'm choosing not to respond, so don't wait for me. If what you have works for you, my advice would be to stick with it.

0 replies

forthrin · 2025-12-27T16:33:22Z

forthrin
Dec 27, 2025
Author

I was wondering simply, is there built-in short forms for ...

Nokogiri::HTML::Builder.new so you don't have to write that whole thing every time?
Nokogiri::XML::Node.new ... add_child ... self, so you don't ...
... adding lists, tables, whatever, to the initial body DOM, so you don't ... well, you get the picture...

4-5. I'll just roll my own, no problem.

1 reply

flavorjones Dec 27, 2025
Maintainer

Claude says:

Global Nokogiri() method for HTML building

Instead of Nokogiri::HTML::Builder.new, you can use:

doc = Nokogiri() {
  body {
    p "Hello world"
  }
}

This returns the root node directly (from lib/nokogiri.rb:108-114).

Block without arguments (skip the prefix)

Instead of xml.tag everywhere, use a block with no parameter:

# With parameter - need xml. prefix:
Nokogiri::XML::Builder.new do |xml|
  xml.root { xml.child "text" }
end

# Without parameter - no prefix needed:
Nokogiri::XML::Builder.new do
  root { child "text" }
end

Builder.with(node) for appending to existing nodes

Instead of Node.new + add_child:

doc = Nokogiri::HTML("<html><body></body></html>")
Nokogiri::XML::Builder.with(doc.at('body')) do |html|
  html.ul {
    html.li "Item 1"
    html.li "Item 2"
  }
end

Chained shortcuts for class/id

# Creates: <div class="container sidebar" id="main">
xml.div.container.sidebar.main!

No built-in list/table helpers

There aren't specific helpers for lists or tables - you'd use the builder DSL directly:

Nokogiri() {
  table {
    tr { td "Cell 1"; td "Cell 2" }
  }
}

I'm happy to give feedback on proposed improvements to Nokogiri, and very happy to accept documentation improvements. But I choose not to put more effort into answering questions than is spent in asking them. I encourage you to read the documentation or use LLM tools to answer how-to questions. Or, consider finding a Ruby forum or support group.

forthrin · 2025-12-27T17:40:41Z

forthrin
Dec 27, 2025
Author

You're right! Just sometimes one's state of mind does the unthinkable and asks a human. Thanks for the copy-paste and alluding how you're managing the gem. Which makes you one. (Just had to.)

0 replies

forthrin · 2026-02-23T07:08:03Z

forthrin
Feb 23, 2026
Author

OK, so the AI suggestions had to be wrangled to work as usual, but results are quite satisfying.

Assuming the Nokogiri shorthand is for creating inline elements, not documents, since no DOCTYPE is added.

Questions:

Do/could shorthands exist for Nokogiri::HTML5::Builder.new, eg. simply Nokogiri.HTML5? (Same for XML)
it.write_xml_to and dropping .to_xml raises RuntimeError about multiple root nodes. Bug?
p invokes Kernel.p. Bug?

require 'nokogiri'

# XML builder
Nokogiri::XML::Builder.new do
  root do
    child 'text'
  end
end.then { puts it.class, it.to_xml, '-' * 70 } # RunTimeError wo/to_xml

# Nokogiri -> Nokogiri::XML::Element
Nokogiri do
  html do
    body do
      h1 'Hello world'
    end
  end
end.then { puts it.class, it, '-' * 70 }

# HTML5 builder
Nokogiri::HTML5::Builder.new do
  html do
    body do
      h1 'Hello world'
      p 'This little p' # Invokes Kernel.p
      ul do
        li 'Item 1'
        li 'Item 2'
      end
      table do
        tr do
          td 'Cell 1'
          td 'Cell 2'
        end
      end
    end
  end
end.then do
  puts it.class, it.to_html, '-' * 70
  it.write_xml_to $stdout, indent: 2 # RunTimeError
end

0 replies

forthrin · 2026-03-23T06:45:34Z

forthrin
Mar 23, 2026
Author

@flavorjones: Little help?

0 replies

flavorjones · 2026-03-23T11:57:14Z

flavorjones
Mar 23, 2026
Maintainer

@forthrin Please stop tagging me when you have questions that can be answered with a little of your own research or by asking questions in a more general forum. I will not spend more time answering questions than you spend on asking them.

To directly address the bug claims, however:

write_xml_to and dropping .to_xml raises RuntimeError about multiple root nodes. Bug?

Not a bug. You're calling write_xml_to on the Builder object, not the Document.

p invokes Kernel.p. Bug?

Not a bug. This is documented in the Builder class under "Special Tags".

0 replies

forthrin · 2026-03-23T17:41:06Z

forthrin
Mar 23, 2026
Author

It's an unfortunate, but recurring phenomenon in popular open source projects where the ratio between developers and users grows disproportional to a point where maintainers end up lashing out at its users for asking questions.

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Shorthands and validation (built-ins vs refinements) #3588

Uh oh!

{{title}}

Uh oh!

Replies: 8 comments 1 reply

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Select a reply

Uh oh!

Uh oh!

Shorthands and validation (built-ins vs refinements) #3588

Uh oh!

forthrin Dec 21, 2025

Replies: 8 comments · 1 reply

Uh oh!

forthrin Dec 27, 2025 Author

Uh oh!

flavorjones Dec 27, 2025 Maintainer

Uh oh!

forthrin Dec 27, 2025 Author

Uh oh!

flavorjones Dec 27, 2025 Maintainer

Uh oh!

forthrin Dec 27, 2025 Author

Uh oh!

forthrin Feb 23, 2026 Author

Uh oh!

forthrin Mar 23, 2026 Author

Uh oh!

flavorjones Mar 23, 2026 Maintainer

Uh oh!

forthrin Mar 23, 2026 Author

forthrin
Dec 21, 2025

Replies: 8 comments 1 reply

forthrin
Dec 27, 2025
Author

flavorjones
Dec 27, 2025
Maintainer

forthrin
Dec 27, 2025
Author

flavorjones Dec 27, 2025
Maintainer

forthrin
Dec 27, 2025
Author

forthrin
Feb 23, 2026
Author

forthrin
Mar 23, 2026
Author

flavorjones
Mar 23, 2026
Maintainer

forthrin
Mar 23, 2026
Author