Class: HTMLProofer::UrlValidator

Inherits:
Object
  • Object
show all
Includes:
Utils
Defined in:
lib/html-proofer/url_validator.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from Utils

clean_content, create_nokogiri, #pluralize, swap

Constructor Details

#initialize(logger, external_urls, options) ⇒ UrlValidator

Returns a new instance of UrlValidator


12
13
14
15
16
17
18
19
# File 'lib/html-proofer/url_validator.rb', line 12

def initialize(logger, external_urls, options)
  @logger = logger
  @external_urls = external_urls
  @failed_tests = []
  @options = options
  @hydra = Typhoeus::Hydra.new(@options[:hydra])
  @cache = Cache.new(@logger, @options[:cache])
end

Instance Attribute Details

#external_urlsObject (readonly)

Returns the value of attribute external_urls


10
11
12
# File 'lib/html-proofer/url_validator.rb', line 10

def external_urls
  @external_urls
end

Instance Method Details

#add_external_issue(filenames, desc, status = nil) ⇒ Object


213
214
215
216
217
218
219
220
# File 'lib/html-proofer/url_validator.rb', line 213

def add_external_issue(filenames, desc, status = nil)
  # possible if we're checking an array of links
  if filenames.nil?
    @failed_tests << Issue.new('', desc, status: status)
  else
    filenames.each { |f| @failed_tests << Issue.new(f, desc, status: status) }
  end
end

#check_hash_in_2xx_response(href, effective_url, response, filenames) ⇒ Object

Even though the response was a success, we may have been asked to check if the hash on the URL exists on the page


174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
# File 'lib/html-proofer/url_validator.rb', line 174

def check_hash_in_2xx_response(href, effective_url, response, filenames)
  return false if @options[:only_4xx]
  return false unless @options[:check_external_hash]
  return false unless (hash = hash?(href))

  body_doc = create_nokogiri(response.body)

  unencoded_hash = Addressable::URI.unescape(hash)
  xpath = %(//*[@name="#{hash}"]|/*[@name="#{unencoded_hash}"]|//*[@id="#{hash}"]|//*[@id="#{unencoded_hash}"])
  # user-content is a special addition by GitHub.
  if URI.parse(href).host =~ /github\.com/i
    xpath << %(|//*[@name="user-content-#{hash}"]|//*[@id="user-content-#{hash}"])
  end

  return unless body_doc.xpath(xpath).empty?

  msg = "External link #{href} failed: #{effective_url} exists, but the hash '#{hash}' does not"
  add_external_issue(filenames, msg, response.code)
  @cache.add(href, filenames, response.code, msg)
  true
end

#clean_url(href) ⇒ Object


123
124
125
126
127
128
129
130
131
# File 'lib/html-proofer/url_validator.rb', line 123

def clean_url(href)
  # catch any obvious issues, like strings in port numbers
  parsed = Addressable::URI.parse(href)
  if href !~ /^([!#{$&}-;=?-\[\]_a-z~]|%[0-9a-fA-F]{2})+$/
    parsed.normalize
  else
    href
  end
end

#establish_queue(external_urls) ⇒ Object


105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
# File 'lib/html-proofer/url_validator.rb', line 105

def establish_queue(external_urls)
  external_urls.each_pair do |url, filenames|
    url = begin
             clean_url(url)
           rescue URI::Error, Addressable::URI::InvalidURIError
             add_external_issue(filenames, "#{url} is an invalid URL")
             next
           end

    method = if hash?(url) && @options[:check_external_hash]
               :get
             else
               :head
             end
    queue_request(method, url, filenames)
  end
end

Proofer runs faster if we pull out all the external URLs and run the checks at the end. Otherwise, we're halting the consuming process for every file during `process_files`.

In addition, sorting the list lets libcurl keep connections to the same hosts alive.

Finally, we'll first make a HEAD request, rather than GETing all the contents. If the HEAD fails, we'll fall back to GET, as some servers are not configured for HEAD. If we've decided to check for hashes, we must do a GET–HEAD is not available as an option.


90
91
92
93
94
95
96
97
98
99
100
101
102
103
# File 'lib/html-proofer/url_validator.rb', line 90

def external_link_checker(external_urls)
  external_urls = Hash[external_urls.sort]

  count = external_urls.length
  check_text = pluralize(count, 'external link', 'external links')
  @logger.log :info, "Checking #{check_text}..."

  # Route log from Typhoeus/Ethon to our own logger
  Ethon.logger = @logger

  establish_queue(external_urls)

  @hydra.run
end

#extract_domain_path(uri) ⇒ Object


67
68
69
# File 'lib/html-proofer/url_validator.rb', line 67

def extract_domain_path(uri)
  uri.host + uri.path
end

#handle_failure(href, filenames, response_code, return_message) ⇒ Object


203
204
205
206
207
208
209
210
211
# File 'lib/html-proofer/url_validator.rb', line 203

def handle_failure(href, filenames, response_code, return_message)
  msg = "External link #{href} failed: response code #{response_code} means something's wrong.
         It's possible libcurl couldn't connect to the server or perhaps the request timed out.
         Sometimes, making too many requests at once also breaks things.
         Either way, the return message (if any) from the server is: #{return_message}"
  @cache.add(href, filenames, 0, msg)
  return if @options[:only_4xx]
  add_external_issue(filenames, msg, response_code)
end

#handle_timeout(href, filenames, response_code) ⇒ Object


196
197
198
199
200
201
# File 'lib/html-proofer/url_validator.rb', line 196

def handle_timeout(href, filenames, response_code)
  msg = "External link #{href} failed: got a time out (response code #{response_code})"
  @cache.add(href, filenames, 0, msg)
  return if @options[:only_4xx]
  add_external_issue(filenames, msg, response_code)
end

#hash?(url) ⇒ Boolean

Does the URL have a hash?

Returns:

  • (Boolean)

223
224
225
226
227
# File 'lib/html-proofer/url_validator.rb', line 223

def hash?(url)
  URI.parse(url).fragment
rescue URI::InvalidURIError
  false
end

#load_cacheObject


71
72
73
74
75
76
77
78
# File 'lib/html-proofer/url_validator.rb', line 71

def load_cache
  cache_count = @cache.size
  cache_text = pluralize(cache_count, 'link', 'links')

  @logger.log :info, "Found #{cache_text} in the cache..."

  @cache.retrieve_urls(@external_urls)
end

#new_url_query_values?(uri, paths_with_queries) ⇒ Boolean

remember queries we've seen, ignore future ones

Returns:

  • (Boolean)

53
54
55
56
57
58
59
60
61
62
63
64
65
# File 'lib/html-proofer/url_validator.rb', line 53

def new_url_query_values?(uri, paths_with_queries)
  queries = uri.query_values.keys.join('-')
  domain_path = extract_domain_path(uri)
  if paths_with_queries[domain_path].nil?
    paths_with_queries[domain_path] = [queries]
    true
  elsif !paths_with_queries[domain_path].include?(queries)
    paths_with_queries[domain_path] << queries
    true
  else
    false
  end
end

#queue_request(method, href, filenames) ⇒ Object


133
134
135
136
137
138
# File 'lib/html-proofer/url_validator.rb', line 133

def queue_request(method, href, filenames)
  opts = @options[:typhoeus].merge(method: method)
  request = Typhoeus::Request.new(href, opts)
  request.on_complete { |response| response_handler(response, filenames) }
  @hydra.queue request
end

#remove_query_valuesObject


35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
# File 'lib/html-proofer/url_validator.rb', line 35

def remove_query_values
  return nil if @external_urls.nil?
  paths_with_queries = {}
  iterable_external_urls = @external_urls.dup
  @external_urls.each_key do |url|
    uri = begin
            Addressable::URI.parse(url)
          rescue URI::Error, Addressable::URI::InvalidURIError
            @logger.log :error, "#{url} is an invalid URL"
            nil
          end
    next if uri.nil? || uri.query.nil?
    iterable_external_urls.delete(url) unless new_url_query_values?(uri, paths_with_queries)
  end
  iterable_external_urls
end

#response_handler(response, filenames) ⇒ Object


140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
# File 'lib/html-proofer/url_validator.rb', line 140

def response_handler(response, filenames)
  effective_url = response.options[:effective_url]
  href = response.request.base_url.to_s
  method = response.request.options[:method]
  response_code = response.code
  response.body.gsub!("\x00", '')

  debug_msg = "Received a #{response_code} for #{href}"
  debug_msg << " in #{filenames.join(' ')}" unless filenames.nil?
  @logger.log :debug, debug_msg

  return if @options[:http_status_ignore].include?(response_code)

  if response_code.between?(200, 299)
    unless check_hash_in_2xx_response(href, effective_url, response, filenames)
      @cache.add(href, filenames, response_code)
    end
  elsif response.timed_out?
    handle_timeout(href, filenames, response_code)
  elsif response_code.zero?
    handle_failure(effective_url, filenames, response_code, response.return_message)
  elsif method == :head
    queue_request(:get, href, filenames)
  else
    return if @options[:only_4xx] && !response_code.between?(400, 499)
    # Received a non-successful http response.
    msg = "External link #{href} failed: #{response_code} #{response.return_message}"
    add_external_issue(filenames, msg, response_code)
    @cache.add(href, filenames, response_code, msg)
  end
end

#runObject


21
22
23
24
25
26
27
28
29
30
31
32
33
# File 'lib/html-proofer/url_validator.rb', line 21

def run
  @external_urls = remove_query_values

  if @cache.use_cache?
    urls_to_check = load_cache
    external_link_checker(urls_to_check)
    @cache.write
  else
    external_link_checker(@external_urls)
  end

  @failed_tests
end