Add tests demonstrating parse failure

README.md

@@ -321,7 +321,9 @@ You can pass in additional options to configure this validation.
 
 | Option | Description | Default |
 | :----- | :---------- | :------ |
+| `report_eof_tags` | When `check_html` is enabled, HTML markup with mismatched tags are reported as errors | `false`
 | `report_invalid_tags` | When `check_html` is enabled, HTML markup that is unknown to Nokogumbo are reported as errors. | `false`
+| `report_mismatched_tags` | When `check_html` is enabled, HTML markup with tags that are malformed are reported as errors | `false`
 | `report_missing_doctype` | When `check_html` is enabled, HTML markup with missing or out-of-order `DOCTYPE` are reported as errors. | `false`
 | `report_missing_names` | When `check_html` is enabled, HTML markup that are missing entity names are reported as errors. | `false`
 | `report_script_embeds` | When `check_html` is enabled, `script` tags containing markup are reported as errors. | `false`

bin/htmlproofer

@@ -41,6 +41,8 @@ Mercenary.program(:htmlproofer) do |p|
   p.option 'report_missing_names', '--report-missing-names', 'When `check_html` is enabled, HTML markup that are missing entity names are reported as errors (default: `false`)'
   p.option 'report_script_embeds', '--report-script-embeds', 'When `check_html` is enabled, `script` tags containing markup are reported as errors (default: `false`)'
   p.option 'report_missing_doctype', '--report-missing-doctype', 'When `check_html` is enabled, HTML markup with missing or out-of-order `DOCTYPE` are reported as errors (default: `false`)'
+  p.option 'report_eof_tags', '--report-eof-tags', 'When `check_html` is enabled, HTML markup with tags that are malformed are reported as errors (default: `false`)'
+  p.option 'report_mismatched_tags', '--report-mismatched-tags', 'When `check_html` is enabled, HTML markup with mismatched tags are reported as errors (default: `false`)'
   p.option 'log_level', '--log-level <level>', String, 'Sets the logging level, as determined by Yell. One of `:debug`, `:info`, `:warn`, `:error`, or `:fatal`. (default: `:info`)'
   p.option 'only_4xx', '--only-4xx', 'Only reports errors for links that fall within the 4xx status code range'
   p.option 'storage_dir', '--storage-dir PATH', String, 'Directory where to store the cache log (default: "tmp/.htmlproofer")'
@@ -82,6 +84,8 @@ Mercenary.program(:htmlproofer) do |p|
     options[:validation][:report_missing_names] = opts['report_missing_names'] unless opts['report_missing_names'].nil?
     options[:validation][:report_invalid_tags] = opts['report_invalid_tags'] unless opts['report_invalid_tags'].nil?
     options[:validation][:report_missing_doctype] = opts['report_missing_doctype'] unless opts['report_missing_doctype'].nil?
+    options[:validation][:report_eof_tags] = opts['report_eof_tags'] unless opts['report_eof_tags'].nil?
+    options[:validation][:report_mismatched_tags] = opts['report_mismatched_tags'] unless opts['report_mismatched_tags'].nil?
 
     options[:typhoeus] = HTMLProofer::Configuration.parse_json_option('typhoeus_config', opts['typhoeus_config']) unless opts['typhoeus_config'].nil?
 

lib/html-proofer/check/html.rb

@@ -6,7 +6,9 @@ class HtmlCheck < ::HTMLProofer::Check
   INVALID_TAG_MSG = /Tag ([\w\-:]+) invalid/.freeze
   INVALID_PREFIX = /Namespace prefix/.freeze
   PARSE_ENTITY_REF = /htmlParseEntityRef: no name/.freeze
-  DOCTYPE_MSG = /The doctype must be the first token in the document/.freeze
+  DOCTYPE_MSG = /Expected a doctype token/.freeze
+  EOF_IN_TAG = /End of input in tag/.freeze
+  MISMATCHED_TAGS = /That tag isn't allowed here/.freeze
 
   def run
     @html.errors.each do |error|
@@ -24,6 +26,10 @@ def report?(message)
       options[:validation][:report_missing_names]
     when DOCTYPE_MSG
       options[:validation][:report_missing_doctype]
+    when EOF_IN_TAG
+      options[:validation][:report_eof_tags]
+    when MISMATCHED_TAGS
+      options[:validation][:report_mismatched_tags]
     else
       true
     end

lib/html-proofer/configuration.rb

@@ -52,7 +52,9 @@ module Configuration
       report_script_embeds: false,
       report_missing_names: false,
       report_invalid_tags: false,
-      report_missing_doctype: false
+      report_missing_doctype: false,
+      report_eof_tags: false,
+      report_mismatched_tags: false
     }.freeze
 
     CACHE_DEFAULTS = {}.freeze

lib/html-proofer/middleware.rb

@@ -21,7 +21,8 @@ def self.options
         allow_hash_href: true,
         check_external_hash: true,
         check_html: true,
-        url_ignore: [/.*/] # Don't try to check local files exist
+        url_ignore: [/.*/], # Don't try to check if local files exist
+        validation: { report_eof_tags: true }
       }
     end
 

lib/html-proofer/utils.rb

@@ -15,7 +15,7 @@ def create_nokogiri(path)
                   path
                 end
 
-      Nokogiri::HTML5(content)
+      Nokogiri::HTML5(content, max_errors: -1)
     end
 
     def swap(href, replacement)

spec/html-proofer/command_spec.rb

@@ -89,9 +89,9 @@
   end
 
   it 'works with check-html' do
-    broken = "#{FIXTURES_DIR}/html/unmatched_end_tag.html"
-    output = make_bin('--check-html --report-invalid-tags', broken)
-    expect(output).to match('HTML-Proofer finished successfully')
+    broken = "#{FIXTURES_DIR}/html/missing_closing_quotes.html"
+    output = make_bin('--check-html --report-eof-tags', broken)
+    expect(output).to match('1 failure')
   end
 
   it 'works with empty-alt-ignore' do

spec/html-proofer/fixtures/html/parse_failure.html

@@ -0,0 +1 @@
+<!DOCTYPE html><html><!-- -- --></a>

spec/html-proofer/html_spec.rb

@@ -124,4 +124,11 @@
     proofer = run_proofer(file, :file, opts)
     expect(proofer.failed_tests).to eq []
   end
+
+  it 'reports failures' do
+    opts = { check_html: true, validation: { report_mismatched_tags: true } }
+    file = "#{FIXTURES_DIR}/html/parse_failure.html"
+    proofer = run_proofer(file, :file, opts)
+    expect(proofer.failed_tests.first).to match(/ERROR: That tag isn't allowed here/)
+  end
 end