Tests for failing Whitehall Govspeak

Test cases for content in Whitehall which currently fails Govspeak
6+ validation.

Author: erino

test/html_validator_test.rb

@@ -101,4 +101,59 @@ class HtmlValidatorTest < Minitest::Test
     assert Govspeak::HtmlValidator.new("{button start}[Start now](https://gov.uk){/button}").valid?
     assert Govspeak::HtmlValidator.new("{button start cross-domain-tracking:UA-XXXXXX-Y}[Start now](https://gov.uk){/button}").valid?
   end
+
+  test "allow a table with no tbody" do
+    # The publication https://whitehall-admin.integration.publishing.service.gov.uk/government/admin/publications/889052
+    # with the HTML attachment "What Works Network membership requirements (HTML)"
+    # has a table with no tbody
+    assert Govspeak::HtmlValidator.new("<table><tr><td>Hello</td></tr></table>").valid?, "No <tbody> is valid"
+  end
+
+  test "allow a table with a tbody" do
+    # If the above test is made to pass, there is a chance this will break
+    assert Govspeak::HtmlValidator.new("<table><tbody><tr><td>Hello</td></tr></tbody></table>").valid?, "<tbody> is valid"
+  end
+
+  test "allow a table with a quotes round an element" do
+    # The publication https://whitehall-admin.integration.publishing.service.gov.uk/government/admin/publications/964161
+    # with the HTML attachment "Upper Tribunal (Tax and Chancery) financial services hearings and register 2014 to date"
+    # has quotes spanning table elements
+    skip "This should probabably fail"
+    assert Govspeak::HtmlValidator.new("<table><tbody><tr><td>\"Hello</td><td>World\"</td></tr></tbody></table>").valid?
+  end
+
+  test "allow a govspeak table" do
+    # This ensures that tables created with Markdown are still valid
+    # https://whitehall-admin.integration.publishing.service.gov.uk/government/admin/detailed-guides/968711/edit
+    govspeak = "
+      Table | Header
+      - | -
+      Build | cells"
+    assert Govspeak::HtmlValidator.new(govspeak).valid?, "Markdown tables are OK"
+  end
+
+  test "allow a document with 2 tables: one with a tbody and one without" do
+    # The guidance https://whitehall-admin.integration.publishing.service.gov.uk/government/admin/publications/938674
+    # with the HTML attachment "2018 to 2019: Student Loan deduction tables" has tables with and without tbody tags
+    html = "
+    <table><tr><td>Hello</td></tr></table>
+
+    <table><tbody><tr><td>Hello</td></tr></tbody></table>
+    "
+    assert Govspeak::HtmlValidator.new(html).valid?, "All tables are valid"
+  end
+
+  test "allow unsuitable chars" do
+    # The Sanatize gem remove Unsuitable Chars https://github.com/rgrove/sanitize/blob/master/lib/sanitize.rb#L170
+    values = [
+      "\u2028", # https://whitehall-admin.integration.publishing.service.gov.uk/government/admin/news/968745/edit
+      "\u202a", # https://whitehall-admin.integration.publishing.service.gov.uk/government/admin/news/947317/edit
+      "\u202c", # https://whitehall-admin.integration.publishing.service.gov.uk/government/admin/news/947317/edit
+      "\ufffc"  # https://whitehall-admin.integration.publishing.service.gov.uk/government/admin/news/951534
+    ]
+    values.each do |value|
+      assert Govspeak::HtmlValidator.new(value).valid?, "value: #{value} (#{value.unpack("U*")})"
+    end
+  end
+
 end