Should Nokogiri::XML.parse be creating separate Text nodes for linefeeds? - ruby

Question

I have an XML document created by an outside tool:
<?xml version="1.0" encoding="UTF-8"?>
<suite>
<id>S1</id>
<name>First Suite</name>
<description></description>
<sections>
<section>
<name>section 1</name>
<cases>
<case>
<id>C1</id>
<title>Test 1.1</title>
<type>Other</type>
<priority>4 - Must Test</priority>
<estimate></estimate>
<milestone></milestone>
<references></references>
</case>
<case>
<id>C2</id>
<title>Test 1.2</title>
<type>Other</type>
<priority>4 - Must Test</priority>
<estimate></estimate>
<milestone></milestone>
<references></references>
</case>
</cases>
</section>
</sections>
</suite>
From irb, I do the following: (Output suppressed until final command)
> require('nokogiri')
> doc = Nokogiri::XML.parse(open('./test.xml'))
> test_case = doc.search('case').first
=> #<Nokogiri::XML::Element:0x3ff75851bc44 name="case" children=[#<Nokogiri::XML::Text:0x3ff75851b8fc "\n ">, #<Nokogiri::XML::Element:0x3ff75851b7bc name="id" children=[#<Nokogiri::XML::Text:0x3ff75851b474 "C1">]>, #<Nokogiri::XML::Text:0x3ff75851b1cc "\n ">, #<Nokogiri::XML::Element:0x3ff75851b078 name="title" children=[#<Nokogiri::XML::Text:0x3ff75851ad58 "Test 1.1">]>, #<Nokogiri::XML::Text:0x3ff75851aa9c "\n ">, #<Nokogiri::XML::Element:0x3ff75851a970 name="type" children=[#<Nokogiri::XML::Text:0x3ff75851a6c8 "Other">]>, #<Nokogiri::XML::Text:0x3ff7585191d8 "\n ">, #<Nokogiri::XML::Element:0x3ff7585190d4 name="priority" children=[#<Nokogiri::XML::Text:0x3ff758518d64 "4 - Must Test">]>, #<Nokogiri::XML::Text:0x3ff758518ad0 "\n ">, #<Nokogiri::XML::Element:0x3ff7585189a4 name="estimate">, #<Nokogiri::XML::Text:0x3ff758518670 "\n ">, #<Nokogiri::XML::Element:0x3ff758518558 name="milestone">, #<Nokogiri::XML::Text:0x3ff7585182b0 "\n ">, #<Nokogiri::XML::Element:0x3ff758518184 name="references">, #<Nokogiri::XML::Text:0x3ff758517ef0 "\n ">]>
This results in a number of children that look like the following:
#<Nokogiri::XML::Text:0x3ff758517ef0 "\n ">
I want to iterate through these XML nodes without having to do something like:
> real_nodes = test_case.children.reject{|n| n.node_name == 'text' && n.content.strip!.empty?}
I couldn't find a parse parameter in the Nokogiri docs to suppress the treating of newlines as separate nodes. Is there a way to do this during the parse instead of after?

Answer 1

Check the documentation. You can just do this:
doc = Nokogiri::XML.parse(open('./test.xml')) do |config|
config.noblanks
end
That will load the file without any empty nodes.

Answer 2

The text nodes are the result of pretty-printing the XML. The spec doesn't require whitespace between tags, and, for efficiency, a huge XML file could be stripped of inter-tag whitespace to save space and reduce transfer time, without sacrificing the data content.
This might show what's happening:
require 'nokogiri'
xml = '<foo></foo>'
Nokogiri::XML(xml).at('foo').child
=> nil
With no whitespace between the tags there's no text node either.
xml = '<foo>
</foo>'
Nokogiri::XML(xml).at('foo').child
=> #<Nokogiri::XML::Text:0x3fcee9436ff0 "\n">
doc.at('foo').child.class
=> Nokogiri::XML::Text
With whitespace for pretty-printing, the XML has a text node following the foo tag.