require File.expand_path(File.join(File.dirname(__FILE__), 'helper'))
class TestBasic < Test::Unit::TestCase
def test_nil
assert_nil Dryopteris.sanitize(nil)
end
def test_empty_string
assert_equal "", Dryopteris.sanitize("")
end
def test_removal_of_illegal_tag
html = <<-HTML
following this there should be no jim tag
jim
was there?
HTML
sane = Nokogiri::HTML(Dryopteris.sanitize(html))
assert sane.xpath("//jim").empty?
end
def test_removal_of_illegal_attribute
html = ""
sane = Nokogiri::HTML(Dryopteris.sanitize(html))
node = sane.xpath("//p").first
assert node.attributes['class']
assert node.attributes['abbr']
assert_nil node.attributes['foo']
end
def test_removal_of_illegal_url_in_href
html = <<-HTML
this link should have its href removed because of illegal urlthis link should be fine
HTML
sane = Nokogiri::HTML(Dryopteris.sanitize(html))
nodes = sane.xpath("//a")
assert_nil nodes.first.attributes['href']
assert nodes.last.attributes['href']
end
def test_css_sanitization
html = ""
sane = Nokogiri::HTML(Dryopteris.sanitize(html))
assert_match(/#000/, sane.inner_html)
assert_no_match(/foo\.com/, sane.inner_html)
end
def test_fragment_with_no_tags
assert_equal "This fragment has no tags.", Dryopteris.sanitize("This fragment has no tags.")
end
def test_fragment_in_p_tag
assert_equal "
This fragment is in a p.
", Dryopteris.sanitize("
This fragment is in a p.
")
end
def test_fragment_in_a_nontrivial_p_tag
assert_equal " \n
This fragment is in a p.
", Dryopteris.sanitize(" \n
This fragment is in a p.
")
end
def test_fragment_in_p_tag_plus_stuff
assert_equal "
This fragment is in a p.
foobar", Dryopteris.sanitize("
This fragment is in a p.
foobar")
end
def test_fragment_with_text_nodes_leading_and_trailing
assert_equal "text
fragment
text", Dryopteris.sanitize("text
fragment
text")
end
def test_fragment_with_body_tags
# ignore second open body tag, use first close body tag, ignore everything after that
assert_equal "textfragment", Dryopteris.sanitize("textfragmenttext")
end
def test_whitewash_on_microsofty_markup
html = <<-EOHTML