Create and use new rspec matcher 'match_html'

This introduces match_html, which converts actual and expected HTML strings into "canonical" HTML using Nokogiri with NOBLANKS and COMPACT, and then does a simple equality comparison.

This eliminates whitespace differences introduced by library changes (e.g. the libxml2 2.9.0 change).

pretty_text_spec.rb has been changed to use match_html where appropriate. and all tests pass under libxml2 2.7.6, 2.8.0 or 2.9.0

spec/support/match_html_matcher.rb

@@ -0,0 +1,21 @@
+require 'nokogiri/xml/parse_options'
+RSpec::Matchers.define :match_html do |expected|
+  match do |actual|
+    a = make_canonical_html expected
+    b = make_canonical_html actual
+    a.to_html == b.to_html
+  end
+
+  failure_message_for_should do |actual|
+    "after sanitizing for extra white space and compactness, expected #{actual} to match #{expected}"
+  end
+
+  failure_message_for_should_not do |actual|
+    "after sanitizing for extra white space and compactness, expected #{actual} not to match #{expected}"
+  end
+
+  def make_canonical_html(html)
+    Nokogiri::HTML(html) { |config| config.options = Nokogiri::XML::ParseOptions::NOBLANKS | Nokogiri::XML::ParseOptions::COMPACT }
+  end
+
+end