webtranslateit-hpricot 0.9.0

data/test/files/why.xml

@@ -0,0 +1,19 @@
+<?xml version='1.0'?><rss xmlns:admin='http://webns.net/mvcb/' version='2.0' xmlns:sy='http://purl.org/rss/1.0/modules/syndication/' xmlns:dc='http://purl.org/dc/elements/1.1/' xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'>
+    <channel>
+    <title>why the lucky stiff</title>
+    <link>http://whytheluckystiff.net</link>
+    <description>hex-editing reality to give us infinite grenades!!</description>
+    <dc:language>en-us</dc:language>
+    <dc:creator/>
+    <dc:date>2007-01-16T22:39:04+00:00</dc:date>
+    <admin:generatorAgent rdf:resource='http://hobix.com/?v=0.4'/>
+    <sy:updatePeriod>hourly</sy:updatePeriod>
+    <sy:updateFrequency>1</sy:updateFrequency>
+    <sy:updateBase>2000-01-01T12:00+00:00</sy:updateBase>
+    <item><title>1.3</title><link>http://whytheluckystiff.net/quatrains/1.3.html</link><guid isPermaLink='false'>quatrains/1.3@http://whytheluckystiff.net</guid><dc:subject>quatrains</dc:subject><dc:subject>quatrains</dc:subject><dc:creator>why the lucky stiff</dc:creator><dc:date>2007-01-14T08:47:05+00:00</dc:date><description>&lt;blockquote&gt;
+	&lt;p&gt;That cadillac of yours and that driver of yours!&lt;br /&gt;You and your teacups rattling away in the back seat!&lt;br /&gt;You always took the mike, oh, and all those cowboys you shot!&lt;br /&gt;I held your hand! And I&amp;#8217;ll shoot a cowboy one day!&lt;/p&gt;
+	&lt;/blockquote&gt;
+	&lt;blockquote&gt;
+	&lt;p&gt;You said, &amp;#8220;Let&amp;#8217;s run into the woods like kids!&amp;#8221; &lt;br /&gt;You said, &amp;#8220;Let&amp;#8217;s rub our hands together super-hot!&amp;#8221; &lt;br /&gt;And we scalded the trees and left octagons, I think that was you and&lt;br /&gt;You threw parties on the roof!&lt;/p&gt;
+	&lt;/blockquote&gt;</description></item></channel>
+</rss>

data/test/load_files.rb

@@ -0,0 +1,7 @@
+module TestFiles
+    Dir.chdir(File.dirname(__FILE__)) do
+        Dir['files/*.{html,xhtml,xml}'].each do |fname|
+            const_set fname[%r!/(\w+)\.\w+$!, 1].upcase, IO.read(fname)
+        end
+    end
+end

data/test/nokogiri-bench.rb

@@ -0,0 +1,64 @@
+#!/usr/bin/env ruby
+require 'rubygems'
+require 'open-uri'
+require 'hpricot'
+require 'nokogiri'
+require 'benchmark'
+
+content = File.read("test/files/boingboing.html")
+
+N = 100
+
+unless Gem.loaded_specs['hpricot'].version > Gem::Version.new('0.6.161')
+  abort "** Use higher than Hpricot 0.6.161!"
+end
+
+puts "Hpricot #{Gem.loaded_specs['hpricot'].version} vs. Nokogiri #{Gem.loaded_specs['nokogiri'].version}"
+hdoc = Hpricot(content)
+ndoc = Nokogiri.Hpricot(content)
+
+Benchmark.bm do |x|
+  x.report('hpricot:doc') do
+    N.times do
+      Hpricot(content)
+    end
+  end
+
+  x.report('nokogiri:doc') do
+    N.times do
+      Nokogiri.Hpricot(content)
+    end
+  end
+end
+
+Benchmark.bm do |x|
+  x.report('hpricot:xpath') do
+    N.times do
+      info = hdoc.search("//a[@name='027906']").first.inner_text
+      url = hdoc.search("h3[text()='College kids reportedly taking more smart drugs']").first.inner_text
+    end
+  end
+
+  x.report('nokogiri:xpath') do
+    N.times do
+      info = ndoc.search("//a[@name='027906']").first.inner_text
+      url = ndoc.search("h3[text()='College kids reportedly taking more smart drugs']").first.inner_text
+    end
+  end
+end
+
+Benchmark.bm do |x|
+  x.report('hpricot:css') do
+    N.times do
+      info = hdoc.search('form input[@checked]').first
+      url = hdoc.search('td spacer').first.inner_text
+    end
+  end
+
+  x.report('nokogiri:css') do
+    N.times do
+      info = ndoc.search('form input[@checked]').first
+      url = ndoc.search('td spacer').first.inner_text
+    end
+  end
+end

data/test/test_alter.rb

@@ -0,0 +1,96 @@
+# -*- coding: utf-8 -*-
+#!/usr/bin/env ruby
+
+require 'test/unit'
+require 'hpricot'
+require 'load_files'
+
+class TestAlter < Test::Unit::TestCase
+  def setup
+    @basic = Hpricot.parse(TestFiles::BASIC)
+  end
+
+  def test_before
+    test0 = "<link rel='stylesheet' href='test0.css' />"
+    @basic.at("link").before(test0)
+    assert_equal 'test0.css', @basic.at("link").attributes['href']
+  end
+
+  def test_after
+    test_inf = "<link rel='stylesheet' href='test_inf.css' />"
+    @basic.search("link")[-1].after(test_inf)
+    assert_equal 'test_inf.css', @basic.search("link")[-1].attributes['href']
+  end
+
+  def test_wrap
+    ohmy = (@basic/"p.ohmy").wrap("<div id='wrapper'></div>")
+    assert_equal 'wrapper', ohmy[0].parent['id']
+    assert_equal 'ohmy', Hpricot(@basic.to_html).at("#wrapper").children[0]['class']
+  end
+
+  def test_add_class
+    first_p = (@basic/"p:first").add_class("testing123")
+    assert first_p[0].get_attribute("class").split(" ").include?("testing123")
+    assert (Hpricot(@basic.to_html)/"p:first")[0].attributes["class"].split(" ").include?("testing123")
+    assert !(Hpricot(@basic.to_html)/"p:gt(0)")[0].attributes["class"].split(" ").include?("testing123")
+  end
+
+  def test_change_attributes
+    all_ps = (@basic/"p").attr("title", "Some Title & Etc…")
+    all_as = (@basic/"a").attr("href", "http://my_new_href.com")
+    all_lb = (@basic/"link").attr("href") { |e| e.name }
+    assert_changed(@basic, "p", all_ps) {|p| p.raw_attributes["title"] == "Some Title &amp; Etc&#8230;"}
+    assert_changed(@basic, "a", all_as) {|a| a.attributes["href"] == "http://my_new_href.com"}
+    assert_changed(@basic, "link", all_lb) {|a| a.attributes["href"] == "link" }
+  end
+
+  def test_change_attributes2
+    all_as = (@basic%"a").attributes["href"] = "http://my_new_href.com"
+    all_ps = (@basic%"p").attributes["title"] = "Some Title & Etc…"
+    assert_equal (@basic%"a").raw_attributes["href"], "http://my_new_href.com"
+    assert_equal (@basic%"p").raw_attributes["title"], "Some Title &amp; Etc&#8230;"
+    assert_equal (@basic%"p").attributes["title"], "Some Title & Etc…"
+  end
+
+  def test_remove_attr
+    all_rl = (@basic/"link").remove_attr("href")
+    assert_changed(@basic, "link", all_rl) { |link| link['href'].nil? }
+  end
+
+  def test_remove_class
+    all_c1 = (@basic/"p[@class*='last']").remove_class("last")
+    assert_changed(@basic, "p[@class*='last']", all_c1) { |p| p['class'] == 'final' }
+  end
+
+  def test_remove_all_classes
+    all_c2 = (@basic/"p[@class]").remove_class
+    assert_changed(@basic, "p[@class]", all_c2) { |p| p['class'].nil? }
+  end
+
+  def test_xml_casing
+    doc = Hpricot.XML("<root><wildCat>text</wildCat></root>")
+    (doc/:root/:wildCat).after("<beanPole>gravity</beanPole>")
+    assert_equal doc.to_s, "<root><wildCat>text</wildCat><beanPole>gravity</beanPole></root>"
+
+    frag = Hpricot.XML do
+      b { i "A bit of HTML" }
+    end
+    (frag/:b).after("<beanPole>gravity</beanPole>")
+    assert_equal frag.to_s, "<b><i>A bit of HTML</i></b><beanPole>gravity</beanPole>"
+  end
+
+  def test_reparent_empty_nodes
+    doc = Hpricot("<div/>")
+    doc.root.inner_html = "foo"
+    assert_equal doc.root.inner_html, "foo"
+    doc.root.inner_html = ""
+    assert_equal doc.root.inner_html, ""
+    doc.root.swap { b "test" }
+    assert_equal doc.root.inner_html, "test"
+  end
+
+  def assert_changed original, selector, set, &block
+    assert set.all?(&block)
+    assert Hpricot(original.to_html).search(selector).all?(&block)
+  end
+end

data/test/test_builder.rb

@@ -0,0 +1,37 @@
+# -*- coding: utf-8 -*-
+#!/usr/bin/env ruby
+
+require 'test/unit'
+require 'hpricot'
+
+class TestBuilder < Test::Unit::TestCase
+  def test_escaping_text
+    doc = Hpricot() { b "<a\"b>" }
+    assert_equal "<b>&lt;a&quot;b&gt;</b>", doc.to_html
+    assert_equal %{<a"b>}, doc.at("text()").to_s
+  end
+
+  def test_no_escaping_text
+    doc = Hpricot() { div.test.me! { text "<a\"b>" } }
+    assert_equal %{<div class="test" id="me"><a"b></div>}, doc.to_html
+    assert_equal %{<a"b>}, doc.at("text()").to_s
+  end
+
+  def test_latin1_entities
+    doc = Hpricot() { b "ۥ" }
+    assert_equal "<b>&#8364;&#8226;</b>", doc.to_html
+    assert_equal "ۥ", doc.at("text()").to_s
+  end
+
+  def test_escaping_attrs
+    text = "<span style='font-family:\"MS Mincho\"'>Some text</span>"
+    assert_equal "<span style=\"font-family:\\\"MS Mincho\\\"\">Some text</span>",
+      Hpricot(text).to_html
+  end
+
+  def test_korean_utf8_entities
+    a = '한글'
+    doc = Hpricot() { b a }
+    assert_equal "<b>&#54620;&#44544;</b>", doc.to_html
+  end
+end