Class: LinkCheck

Inherits:
HTML::Proofer::CheckRunner show all
Includes:
HTML::Utils
Defined in:
lib/html/proofer/checks/links.rb

Instance Attribute Summary

Attributes inherited from HTML::Proofer::CheckRunner

#alt_ignores, #empty_alt_ignore, #external_urls, #href_ignores, #hydra_opts, #issues, #options, #parallel_opts, #path, #src, #typhoeus_opts, #url_ignores

Instance Method Summary collapse

Methods included from HTML::Utils

create_nokogiri, swap

Methods inherited from HTML::Proofer::CheckRunner

#add_issue, #add_to_external_urls, checks, #initialize

Constructor Details

This class inherits a constructor from HTML::Proofer::CheckRunner

Instance Method Details



92
93
94
95
96
97
98
99
100
101
 
# File 'lib/html/proofer/checks/links.rb', line 92

def external_link_check(link)
  if !link.exists?
    add_issue("trying to find hash of #{link.href}, but #{link.absolute_path} does not exist", link.line)
  else
    target_html = create_nokogiri link.absolute_path
    unless hash_check target_html, link.hash
      add_issue("linking to #{link.href}, but #{link.hash} does not exist", link.line)
    end
  end
end

#hash_check(html, href_hash) ⇒ Object 



103
104
105
 
# File 'lib/html/proofer/checks/links.rb', line 103

def hash_check(html, href_hash)
  html.xpath("//*[@id='#{href_hash}']", "//*[@name='#{href_hash}']").length > 0
end

#run ⇒ Object 



29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
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
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
 
# File 'lib/html/proofer/checks/links.rb', line 29

def run
  @html.css('a, link').each do |l|
    link = LinkCheckable.new l, self

    next if link.ignore?
    next if link.href =~ /^javascript:/ # can't put this in ignore? because the URI does not parse
    next if link.placeholder?

    # is it even a valid URL?
    unless link.valid?
      add_issue("#{link.href} is an invalid URL", l.line)
      next
    end

    case link.scheme
    when 'mailto'
      if link.path.empty?
        add_issue("#{link.href} contains no email address", l.line)
      elsif !link.path.include?('@')
        add_issue("#{link.href} contains an invalid email address", l.line)
      end
    when 'tel'
      add_issue("#{link.href} contains no phone number", l.line) if link.path.empty?
    end

    # is there even a href?
    if link.missing_href?
      add_issue('anchor has no href attribute', l.line)
      next
    end

    # intentionally here because we still want valid? & missing_href? to execute
    next if link.non_http_remote?

    # does the file even exist?
    if link.remote?
      add_to_external_urls link.href
      next
    elsif !link.internal?
      add_issue("internally linking to #{link.href}, which does not exist", l.line) unless link.exists?
    end

    # does the local directory have a trailing slash?
    if link.unslashed_directory? link.absolute_path
      add_issue("internally linking to a directory #{link.absolute_path} without trailing slash", l.line)
      next
    end

    # verify the target hash
    if link.hash
      if link.internal?
        unless hash_check @html, link.hash
          add_issue("linking to internal hash ##{link.hash} that does not exist", l.line)
        end
      elsif link.external?
        external_link_check(link)
      end
    end
  end

  external_urls
end