Class: HTML::Proofer::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

create_nokogiri, swap

Constructor Details

#initialize(logger, external_urls, options, typhoeus_opts, hydra_opts) ⇒ 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, typhoeus_opts, hydra_opts)
  @logger = logger
  @external_urls = external_urls
  @failed_tests = []
  @options = options
  @hydra = Typhoeus::Hydra.new(hydra_opts)
  @typhoeus_opts = typhoeus_opts
end

Instance Attribute Details

#external_urlsObject

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

#hydraObject

Returns the value of attribute hydra


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

def hydra
  @hydra
end

#loggerObject

Returns the value of attribute logger


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

def logger
  @logger
end

Instance Method Details

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


120
121
122
123
124
125
126
# File 'lib/html/proofer/url_validator.rb', line 120

def add_external_issue(filenames, desc, status = nil)
  if filenames.nil?
    @failed_tests << CheckRunner::Issue.new('', desc, nil, status)
  else
    filenames.each { |f| @failed_tests << CheckRunner::Issue.new(f, desc, nil, 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


97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
# File 'lib/html/proofer/url_validator.rb', line 97

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

  body_doc = create_nokogiri(response.body)

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

  return unless body_doc.xpath(xpath).empty?

  add_external_issue filenames, "External link #{href} failed: #{effective_url} exists, but the hash '#{hash}' does not", response.code
end

#clean_url(href) ⇒ Object


62
63
64
# File 'lib/html/proofer/url_validator.rb', line 62

def clean_url(href)
  Addressable::URI.parse(href).normalize
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 the check_directory_of_files process.

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 an option.


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

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

  count = external_urls.length
  check_text = "#{count} " << (count == 1 ? 'external link' : 'external links')
  logger.log :info, :blue, "Checking #{check_text}..."

  Ethon.logger = logger # log from Typhoeus/Ethon

  url_processor(external_urls)

  logger.log :debug, :yellow, "Running requests for all #{hydra.queued_requests.size} external URLs..."
  hydra.run
end

#handle_timeout(href, filenames, response_code) ⇒ Object


115
116
117
118
# File 'lib/html/proofer/url_validator.rb', line 115

def handle_timeout(href, filenames, response_code)
  return if @options[:only_4xx]
  add_external_issue filenames, "External link #{href} failed: got a time out", response_code
end

#hash?(url) ⇒ Boolean

Returns:

  • (Boolean)

128
129
130
131
132
# File 'lib/html/proofer/url_validator.rb', line 128

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

#queue_request(method, href, filenames) ⇒ Object


66
67
68
69
70
# File 'lib/html/proofer/url_validator.rb', line 66

def queue_request(method, href, filenames)
  request = Typhoeus::Request.new(href, @typhoeus_opts.merge({ :method => method }))
  request.on_complete { |response| response_handler(response, filenames) }
  hydra.queue request
end

#response_handler(response, filenames) ⇒ Object


72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
# File 'lib/html/proofer/url_validator.rb', line 72

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

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

  if response_code.between?(200, 299)
    check_hash_in_2xx_response(href, effective_url, response, filenames)
  elsif response.timed_out?
    handle_timeout(href, filenames, response_code)
  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.
    add_external_issue(filenames, "External link #{href} failed: #{response_code} #{response.return_message}", response_code)
  end
end

#runObject


21
22
23
24
# File 'lib/html/proofer/url_validator.rb', line 21

def run
  external_link_checker(external_urls)
  @failed_tests
end

#url_processor(external_urls) ⇒ Object


51
52
53
54
55
56
57
58
59
60
# File 'lib/html/proofer/url_validator.rb', line 51

def url_processor(external_urls)
  external_urls.each_pair do |href, filenames|
    href = clean_url(href)
    if hash?(href) && @options[:check_external_hash]
      queue_request(:get, href, filenames)
    else
      queue_request(:head, href, filenames)
    end
  end
end