# encoding: utf-8 # frozen_string_literal: true require "helper" class TestHtml5Nokogumbo < Nokogiri::TestCase def test_element_text doc = Nokogiri::HTML5(buffer) assert_equal("content", doc.at("span").text) end def test_element_cdata_textarea doc = Nokogiri::HTML5(buffer) assert_equal("foobar", doc.at("textarea").text.strip) end def test_element_cdata_script doc = Nokogiri::HTML5.fragment(buffer) assert_predicate(doc.document, :html?) assert_equal("", doc.at("script").to_s) end def test_attr_value doc = Nokogiri::HTML5(buffer) assert_equal("utf-8", doc.at("meta")["charset"]) end def test_comment doc = Nokogiri::HTML5(buffer) assert_equal(" test comment ", doc.xpath("//comment()").text) end def test_unknown_element doc = Nokogiri::HTML5(buffer) assert_equal("main", doc.at("main").name) end def test_IO require "stringio" doc = Nokogiri::HTML5(StringIO.new(buffer)) assert_equal("textarea", doc.at("form").element_children.first.name) end def test_nil doc = Nokogiri::HTML5(nil) assert_equal(1, doc.search("body").count) fragment = Nokogiri::HTML5.fragment(nil) assert_equal(0, fragment.errors.length) end def test_html5_doctype doc = Nokogiri::HTML5.parse("") assert_match(//, doc.to_html) end def test_fragment_no_errors doc = Nokogiri::HTML5.fragment("no missing DOCTYPE errors", max_errors: 10) assert_equal(0, doc.errors.length) end # This should be deleted when `:max_parse_errors` is removed. def test_fragment_max_parse_errors doc = Nokogiri::HTML5.fragment("testing deprecated :max_parse_errors", max_parse_errors: 10) assert_equal(0, doc.errors.length) end def test_fragment_head doc = Nokogiri::HTML5.fragment(buffer[%r{(.*?)}m, 1]) assert_equal("hello world", doc.xpath("title").text) assert_equal("utf-8", doc.xpath("meta").first["charset"]) end def test_fragment_body doc = Nokogiri::HTML5.fragment(buffer[%r{(.*?)}m, 1]) assert_equal("content", doc.xpath("main/span").to_xml) assert_equal(" test comment ", doc.xpath("comment()").text) end def test_xlink_attribute source = <<-EOF.gsub(/^ {6}/, "") EOF doc = Nokogiri::HTML5.parse(source) a = doc.at_xpath("/html/body/svg:svg/svg:a") refute_nil(a) refute_nil(a["xlink:href"]) refute_nil(a["xmlns:xlink"]) end def test_xlink_attribute_fragment source = <<-EOF.gsub(/^ {6}/, "") EOF doc = Nokogiri::HTML5.fragment(source) a = doc.at_xpath("svg:svg/svg:a") refute_nil(a) refute_nil(a["xlink:href"]) refute_nil(a["xmlns:xlink"]) end def test_template source = <<-EOF.gsub(/^ {6}/, "") EOF doc = Nokogiri::HTML5.fragment(source) template = doc.at("template") assert_equal("productrow", template["id"]) assert_equal("record", template.at("td")["class"]) end def test_root_comments doc = Nokogiri::HTML5("") assert_equal(["html", "comment", "html", "comment"], doc.children.map(&:name)) end def test_max_attributes html = '
a
' assert_raises(ArgumentError) { Nokogiri::HTML5(html, max_attributes: 2) } assert_raises(ArgumentError) { Nokogiri::HTML5(html, max_attributes: 1) } assert_raises(ArgumentError) { Nokogiri::HTML5(html, max_attributes: 0) } # -1 disables limit doc = Nokogiri::HTML5(html, max_attributes: -1) assert_equal({ "id" => "i", "class" => "c", "title" => "t" }, attributes(doc.at_css("div"))) assert_equal({ "src" => "s", "alt" => "a" }, attributes(doc.at_css("img"))) end def test_max_attributes_boolean html = '' doc = Nokogiri::HTML5(html, max_attributes: 4) assert_equal({ "checked" => "", "type" => "checkbox", "disabled" => "", "name" => "cheese" }, attributes(doc.at_css("input"))) assert_raises(ArgumentError) { Nokogiri::HTML5(html, max_attributes: 3) } assert_raises(ArgumentError) { Nokogiri::HTML5(html, max_attributes: 2) } assert_raises(ArgumentError) { Nokogiri::HTML5(html, max_attributes: 1) } assert_raises(ArgumentError) { Nokogiri::HTML5(html, max_attributes: 0) } end def test_default_max_attributes a = +"a" attrs = 50_000.times.map do x = a.dup a.succ! x end #
contains 50,000 attributes, but default limit is 400. Parsing this would take ages if # we were not enforcing any limit on attributes. All attributes are duplicated to make sure # this doesn’t alter performance or end result. html = <<~EOF
hello
EOF assert_raises(ArgumentError) { Nokogiri::HTML5(html) } end def test_fragment_max_attributes html = '
a
' assert_raises(ArgumentError) { Nokogiri::HTML5.fragment(html, max_attributes: 2) } assert_raises(ArgumentError) { Nokogiri::HTML5.fragment(html, max_attributes: 1) } assert_raises(ArgumentError) { Nokogiri::HTML5.fragment(html, max_attributes: 0) } # -1 disables limit doc = Nokogiri::HTML5.fragment(html, max_attributes: -1) assert_equal({ "id" => "i", "class" => "c", "title" => "t" }, attributes(doc.at_css("div"))) assert_equal({ "src" => "s", "alt" => "a" }, attributes(doc.at_css("img"))) end def test_fragment_default_max_attributes a = +"a" attrs = 50_000.times.map do x = a.dup a.succ! x end #
contains 50,000 attributes, but default limit is 400. Parsing this would take ages if # we were not enforcing any limit on attributes. All attributes are duplicated to make sure # this doesn’t alter performance or end result. html = <<~EOF
hello
EOF assert_raises(ArgumentError) { Nokogiri::HTML5.fragment(html) } end def test_parse_errors doc = Nokogiri::HTML5("", max_errors: 10) assert_equal(2, doc.errors.length) doc = Nokogiri::HTML5("", max_errors: 10) assert_empty(doc.errors) end def test_max_errors # This document contains 2 parse errors, but we force limit to 1. doc = Nokogiri::HTML5("", max_errors: 1) assert_equal(1, doc.errors.length) doc = Nokogiri::HTML5("", max_errors: 1) assert_empty(doc.errors) end def test_default_max_errors # This document contains 200 parse errors, but default limit is 0. doc = Nokogiri::HTML5("" + "

" * 200) assert_equal(0, doc.errors.length) end def test_parse_fragment_errors doc = Nokogiri::HTML5.fragment("<\r\n", max_errors: 10) refute_empty(doc.errors) end def test_fragment_max_errors # This fragment contains 2 parse errors, but we force limit to 1. doc = Nokogiri::HTML5.fragment("", max_errors: 1) assert_equal(1, doc.errors.length) doc = Nokogiri::HTML5.fragment("", max_errors: 10) assert_equal(2, doc.errors.length) end def test_fragment_default_max_errors # This fragment contains 200 parse errors, but default limit is 0. doc = Nokogiri::HTML5.fragment("

" * 200) assert_equal(0, Nokogiri::Gumbo::DEFAULT_MAX_ERRORS) assert_equal(0, doc.errors.length) end def test_default_max_depth_parse depth = Nokogiri::Gumbo::DEFAULT_MAX_TREE_DEPTH + 1 assert_raises(ArgumentError) do Nokogiri::HTML5("" + "
" * (depth - 2)) end end def test_default_max_depth_fragment depth = Nokogiri::Gumbo::DEFAULT_MAX_TREE_DEPTH + 1 assert_raises(ArgumentError) do Nokogiri::HTML5.fragment("
" * depth) end end def test_max_depth_parse depth = 10 html = "" + "
" * (depth - 2) assert_raises(ArgumentError) do Nokogiri::HTML5(html, max_tree_depth: depth - 1) end assert(Nokogiri::HTML5(html, max_tree_depth: depth)) end def test_max_depth_fragment depth = 10 html = "
" * depth assert_raises(ArgumentError) do Nokogiri::HTML5.fragment(html, max_tree_depth: depth - 1) end assert(Nokogiri::HTML5.fragment(html, max_tree_depth: depth)) end def test_document_encoding html = <<-TEXT Кирилические символы TEXT doc = Nokogiri::HTML5.parse(html) assert_equal("UTF-8", doc.encoding) assert_equal("Кирилические символы", doc.at("body").text.gsub(/\n\s+/, "")) end def test_line_text doc = Nokogiri.HTML5("\ntext node") assert_equal(2, doc.at_xpath("/html/body/text()").line) end def test_line_comment doc = Nokogiri.HTML5("\n\n") assert_equal(3, doc.at_xpath("/comment()").line) end def test_line_element doc = Nokogiri.HTML5("\n

") assert_equal(2, doc.at_xpath("/html/body/p").line) end def test_line_template doc = Nokogiri.HTML5("\n\n") assert_equal(3, doc.at_xpath("/html/head/template").line) end def test_line_cdata html = "\n\n" doc = Nokogiri.HTML5(html) node = doc.at_xpath("/html/body/svg:svg/svg:script/text()") assert_predicate(node, :cdata?) assert_equal(3, node.line) end it "handles ill-formed processing instructions in a document" do html = %{ hello world

hello world

content
EOF end def attributes(element) element.attributes.to_h { |name, attribute| [name, attribute.value] } end end if Nokogiri.uses_gumbo?