hpricot-scrub 0.2.0

data/test/hpricot_scrub_test.rb

@@ -0,0 +1,45 @@
+require File.dirname(__FILE__) + '/test_helper.rb'
+require File.dirname(__FILE__) + '/scrubber_data.rb'
+
+class HpricotScrubTest < Test::Unit::TestCase
+
+  def setup
+    @clean = Hpricot(MARKUP).scrub.inner_html
+    @config = YAML.load_file('examples/config.yml')
+  end
+
+  def test_full_scrub
+    doc = Hpricot(MARKUP).scrub
+    # using the divisor search throws warnings in test
+    assert_tag_count(doc, 'a', 0)
+    assert_tag_count(doc, 'p', 0)
+    assert_tag_count(doc, 'img', 0)
+    assert_tag_count(doc, 'br', 0)
+    assert_tag_count(doc, 'div', 0)
+    assert_tag_count(doc, 'script', 0)
+  end
+
+  def test_partial_scrub
+    full = Hpricot(MARKUP)
+    doc = Hpricot(MARKUP).scrub(@config)
+    # using the divisor search throws warnings in test
+    assert_tag_count(doc, 'a', 0)
+    assert_tag_count(doc, 'p', full.search('//p').size)
+    assert_tag_count(doc, 'div', full.search('//div').size)
+    assert_tag_count(doc, 'img', full.search('//img').size)
+    assert_tag_count(doc, 'br', full.search('//br').size)
+    assert_tag_count(doc, 'script', 0)
+  end
+
+  def test_string_scrub
+    formatted = MARKUP
+    assert formatted.scrub == @clean
+    assert formatted == MARKUP
+  end
+  
+  def test_string_scrub!
+    formatted = MARKUP
+    assert formatted.scrub! == @clean
+    assert formatted == @clean
+  end
+end

data/test/scrubber_data.rb

@@ -0,0 +1,14 @@
+MARKUP = <<-EOS
+<p>Some <b>bold junk</b> here...</p>
+<p>Some more junk <i>here...</i><br></p>
+<p>Still more junk <u>here</u>... </p>
+<p><img title="nothing to see here" alt="nothing to see here" mce_src="http://example.com/imgtest.png" src="http://example.com/imgtest.png" align="middle" border="1" height="240" hspace="5" vspace="5" width="320">&nbsp;</p>
+<p>&nbsp;And a <a title="Just a link" target="_blank" mce_href="http://example.com/nothing.html" href="http://example.com/nothing.html">link</a> just because</p>
+<p><div>some stuff in here</div><img title="nothing to see here" alt="nothing to see here" mce_src="http://example.com/imgtest.png" src="http://example.com/imgtest.png" align="middle" border="1" height="240" hspace="5" vspace="5" width="320"></p>
+<a name="junk"></a>
+<p><script type="text/javascript">//nasty bits go here
+alert("gotcha");</script><img src="http://content.example.com/content/3587a2f6ee641074fec4e7534c01655326c218ec">how about an <a href="javascript:alert('gotcha')">inline script</a>
+</p>
+<span>some random unclosed span
+<style type="text/css">.foo {color:blue}</style>
+EOS

data/test/test_helper.rb

@@ -0,0 +1,8 @@
+require 'test/unit'
+require File.dirname(__FILE__) + '/../lib/hpricot_scrub'
+
+def assert_tag_count(doc, tag, expected)
+  found = doc.search("//#{tag}").size
+  assert found == expected,
+    "Expected to find #{expected} '#{tag}' tag(s), found #{found}"
+end

metadata

@@ -0,0 +1,64 @@
+--- !ruby/object:Gem::Specification 
+rubygems_version: 0.9.1
+specification_version: 1
+name: hpricot-scrub
+version: !ruby/object:Gem::Version 
+  version: 0.2.0
+date: 2007-03-04 00:00:00 -08:00
+summary: Scrub HTML with Hpricot
+require_paths: 
+- lib
+email: michael@underpantsgnome.com
+homepage: http://trac.underpantsgnome.com/hpricot_scrub/
+rubyforge_project: hpricot-scrub
+description: Scrub HTML with Hpricot
+autorequire: 
+default_executable: 
+bindir: bin
+has_rdoc: true
+required_ruby_version: !ruby/object:Gem::Version::Requirement 
+  requirements: 
+  - - ">"
+    - !ruby/object:Gem::Version 
+      version: 0.0.0
+  version: 
+platform: ruby
+signing_key: 
+cert_chain: 
+post_install_message: 
+authors: 
+- UnderpantsGnome
+files: 
+- Rakefile
+- README.txt
+- CHANGELOG.txt
+- Manifest.txt
+- setup.rb
+- lib/hpricot_scrub/version.rb
+- lib/hpricot_scrub.rb
+- test/test_helper.rb
+- test/scrubber_data.rb
+- test/hpricot_scrub_test.rb
+- examples/config.yml
+test_files: 
+- test/hpricot_scrub_test.rb
+rdoc_options: []
+
+extra_rdoc_files: []
+
+executables: []
+
+extensions: []
+
+requirements: []
+
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: hpricot
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Version::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0.5"
+    version: