Class: HTMLProofer::UrlValidator::External

Inherits:
HTMLProofer::UrlValidator show all
Includes:
HTMLProofer::Utils
Defined in:
lib/html_proofer/url_validator/external.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from HTMLProofer::Utils

#blank?, #create_nokogiri, #pluralize

Constructor Details

#initialize(runner, external_urls) ⇒ External

Returns a new instance of External.



16
17
18
19
20
21
22
23
24
 
# File 'lib/html_proofer/url_validator/external.rb', line 16

def initialize(runner, external_urls)
  super(runner)

  @external_urls = external_urls
  @hydra = Typhoeus::Hydra.new(@runner.options[:hydra])
  @before_request = []

  @paths_with_queries = {}
end

Instance Attribute Details

#before_request=(value) ⇒ Object (writeonly)

Sets the attribute before_request

Parameters:

  • value —

    the value to set the attribute before_request to.



14
15
16
 
# File 'lib/html_proofer/url_validator/external.rb', line 14

def before_request=(value)
  @before_request = value
end

#external_urls ⇒ Object (readonly)

Returns the value of attribute external_urls.



13
14
15
 
# File 'lib/html_proofer/url_validator/external.rb', line 13

def external_urls
  @external_urls
end

Instance Method Details

#add_failure(metadata, description, status = nil) ⇒ Object 



193
194
195
196
197
198
199
200
201
 
# File 'lib/html_proofer/url_validator/external.rb', line 193

def add_failure(metadata, description, status = nil)
  if blank?(metadata) # possible if we're checking an array of links
    @failed_checks << Failure.new("", "Links > External", description, status: status)
  else
    metadata.each do |m|
      @failed_checks << Failure.new(m[:filename], "Links > External", description, line: m[:line], status: status)
    end
  end
end

#check_hash_in_2xx_response(href, 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



118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
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
 
# File 'lib/html_proofer/url_validator/external.rb', line 118

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

  hash = url.hash
  headers = response.options.fetch(:headers, {})
  content_type = headers.find { |k, _| k.casecmp("content-type").zero? }

  # attempt to verify PDF hash ref; see #787 for more details
  # FIXME: this is re-reading the PDF response
  if content_type && content_type[1].include?("pdf")
    io = URI.parse(url.to_s).open
    reader = PDF::Reader.new(io)

    pages = reader.pages
    if hash =~ /\Apage=(\d+)\z/
      page = Regexp.last_match[1].to_i

      unless pages[page - 1]
        msg = "External link #{href} failed: #{url.without_hash} exists, but the hash '#{hash}' does not"
        add_failure(filenames, msg, response.code)
        @cache.add_external(href, filenames, response.code, msg, false)
      end

      return true
    end
  end

  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 url.host =~ /github\.com/i
    xpath << [%(//*[@name="user-content-#{hash}"]|//*[@id="user-content-#{hash}"])]
    # when linking to a file on GitHub, like #L12-L34, only the first "L" portion
    # will be identified as a linkable portion
    xpath << [%(//td[@id="#{Regexp.last_match[1]}"])] if hash =~ /\A(L\d)+/
  end

  return unless body_doc.xpath(xpath.join("|")).empty?

  msg = "External link #{href} failed: #{url.without_hash} exists, but the hash '#{hash}' does not"
  add_failure(filenames, msg, response.code)
  @cache.add_external(href, filenames, response.code, msg, false)
  true
end

#handle_connection_failure(href, metadata, response_code, status_message) ⇒ Object 



175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
 
# File 'lib/html_proofer/url_validator/external.rb', line 175

def handle_connection_failure(href, metadata, response_code, status_message)
  msgs = [<<~MSG,
    External link #{href} failed with something very 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.
  MSG
  ]

  msgs << "Either way, the return message from the server is: #{status_message}" unless blank?(status_message)

  msg = msgs.join("\n").chomp

  @cache.add_external(href, metadata, 0, msg, false)
  return if @runner.options[:only_4xx]

  add_failure(metadata, msg, response_code)
end

#handle_timeout(href, filenames, response_code) ⇒ Object 



167
168
169
170
171
172
173
 
# File 'lib/html_proofer/url_validator/external.rb', line 167

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

  add_failure(filenames, msg, response_code)
end

#queue_request(method, url, filenames) ⇒ Object 



72
73
74
75
76
77
78
79
80
 
# File 'lib/html_proofer/url_validator/external.rb', line 72

def queue_request(method, url, filenames)
  opts = @runner.options[:typhoeus].merge(method: method)
  request = Typhoeus::Request.new(url.url, opts)
  @before_request.each do |callback|
    callback.call(request)
  end
  request.on_complete { |response| response_handler(response, url, filenames) }
  @hydra.queue(request)
end

#response_handler(response, url, filenames) ⇒ Object 



82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
 
# File 'lib/html_proofer/url_validator/external.rb', line 82

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

  @logger.log(:debug, "Received a #{response_code} for #{href}")

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

  if response_code.between?(200, 299)
    @cache.add_external(href, filenames, response_code, "OK", true) unless check_hash_in_2xx_response(
      href,
      url,
      response,
      filenames,
    )
  elsif response.timed_out?
    handle_timeout(href, filenames, response_code)
  elsif response_code.zero?
    handle_connection_failure(href, filenames, response_code, response.status_message)
  elsif method == :head # some servers don't support HEAD
    queue_request(:get, url, filenames)
  else
    return if @runner.options[:only_4xx] && !response_code.between?(400, 499)

    # Received a non-successful http response.
    status_message = blank?(response.status_message) ? "" : ": #{response.status_message}"
    msg = "External link #{href} failed#{status_message}"
    add_failure(filenames, msg, response_code)
    @cache.add_external(href, filenames, response_code, msg, false)
  end
end

#run_external_link_checker(external_urls) ⇒ Object

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.



46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
 
# File 'lib/html_proofer/url_validator/external.rb', line 46

def run_external_link_checker(external_urls)
  # Route log from Typhoeus/Ethon to our own logger
  Ethon.logger = @logger

  external_urls.each_pair do |external_url, metadata|
    url = Attribute::Url.new(@runner, external_url, base_url: nil)

    unless url.valid?
      add_failure(metadata, "#{url} is an invalid URL", 0)
      next
    end

    next unless new_url_query_values?(url)

    method = if @runner.options[:check_external_hash] && url.hash?
      :get
    else
      :head
    end

    queue_request(method, url, metadata)
  end

  @hydra.run
end

#validate ⇒ Object 



26
27
28
29
30
31
32
33
34
 
# File 'lib/html_proofer/url_validator/external.rb', line 26

def validate
  urls_to_check = @cache.external_enabled? ? @runner.load_external_cache : @external_urls
  urls_detected = pluralize(urls_to_check.count, "external link", "external links")
  @logger.log(:info, "Checking #{urls_detected}")

  run_external_link_checker(urls_to_check)

  @failed_checks
end