rdoc_link_checker 0.2.0 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: dc87ba9f414e2d9c949f1eb2f85c20ff6ad3a165f0bed638211ea7a84ac16d69
-  data.tar.gz: e801fe955f57f847e7d172540a23a6acfc94597e878ee8dad15eb772da155d32
+  metadata.gz: 3a9ffa7ca1ef044fdc73a544cd7fb31dce91b0fc0a2f0937d071a8ce1d033ac6
+  data.tar.gz: 7e93a01cbb75c2db88050e1ab10aee8e549316d4ae7e06ab80dec38e3af5f9e6
 SHA512:
-  metadata.gz: 7487c3faaed27b9e8a3d3c9922dd650891eb7aeb1e9bd50704006037c9c4dd234bacc95482dcd2247ef9222e4288f2b2e789485bcd044d113a565d0fac733f96
-  data.tar.gz: ebcb50284bb5f96f184d1f0047ed2a21994474c6c3e135dc09532a72ee88fe1fb25fa50e6253d7599966d73287543f7543bde72a34ab536d73b0b165c88a1670
+  metadata.gz: ffd501ff86f9bda291348d6a9aa0a982a4ad1b7b5020d963c69a388343a34427b2e9f5282732da8249b1fc2604482938dd96a610db236cd537f37f9abee7d075
+  data.tar.gz: 5ee92b068a8cb0feea303e4bf8f669bac3f587086a213cd9324cdd6d66e07c22bf6639c42d5119e25d63ad312bac3211bc2675424fec7c4ad76b0ca1f4770c32

data/README.md

@@ -1,6 +1,35 @@
 # RDoc Link Checker
 
-Not ready for prime time.  Just wanted to reserve the name.
+A gem to find broken links in HTML files generated by Ruby RDoc.
 
-Development is active,
-so should be up in a few days (say, by the end of May 2023).
+Reports a link as broken if:
+
+- The target page given by +href+ is not found.
+- The target page is found, but the fragment given by +href+
+  is not a link target on that page;
+  this usually causes a browser to open at the top of the page
+  instead of at the given fragment.
+
+Note that some browsers are forgiving, and will open the target
+page at a link target similar to the given fragment;
+for example, fragment ```bar``` may be opened at an element
+with id ```foobar```.
+
+```
+Usage:
+rdoc_link_checker html_dirpath options
+
+The argument is the path to a directory containing a tree
+of RDoc-generated HTML files, such as those generated by command
+
+rdoc --visibility=private --op html . # Note the trailing dot.
+
+Options:
+--onsite_only   Check link targets only on pages in the file tree at <html_dirpath>,
+                and not those on other local pages or those on the web.
+--no_toc        Do not check links on the TOC page (table_of_contents.html).
+--version       Print the version and exit.
+--help          Print this help and exit.
+
+The output is file <html_dirpath>/Report.htm, which reports broken links.
+```

data/bin/rdoc_link_checker

@@ -5,31 +5,33 @@ require_relative '../lib/rdoc_link_checker'
 
 options = GetoptLong.new(
   ['--html_dirpath', '-d', GetoptLong::REQUIRED_ARGUMENT],
-  ['--version', '-v', GetoptLong::NO_ARGUMENT],
-  ['--help', '-h', GetoptLong::NO_ARGUMENT]
+  ['--onsite_only', '-l', GetoptLong::NO_ARGUMENT],
+  ['--no_toc',      '-n', GetoptLong::NO_ARGUMENT],
+  ['--version',     '-v', GetoptLong::NO_ARGUMENT],
+  ['--help',        '-h', GetoptLong::NO_ARGUMENT]
 )
 
-message = nil
-case ARGV.size
-when 0
-  message = "Expected one argument; got none."
-when 1
-  # Okay.
-else
-  message = "Expected one argument, not #{ARGV.inspect}."
-end
-raise ArgumentError.new(message) if message
-
 def help
-  puts 'Boo!'
+  path = File.absolute_path(__FILE__)
+  dirname = File.dirname(File.dirname(path))
+  filepath = File.join(dirname, 'doc', 'help.txt')
+  puts File.read(filepath)
+  exit
 end
 
 def version
   puts RDocLinkChecker::VERSION
+  exit
 end
 
+onsite_only = false
+no_toc = false
 options.each do |option, argument|
   case option
+  when '--onsite_only'
+    onsite_only = true
+  when '--no_toc'
+    no_toc = true
   when '--help'
     help
   when '--version'
@@ -37,5 +39,20 @@ options.each do |option, argument|
   end
 end
 
+message = nil
+case ARGV.size
+when 0
+  message = "Expected one argument; got none."
+when 1
+  # Okay.
+else
+  message = "Expected one argument, not #{ARGV.inspect}."
+end
+raise ArgumentError.new(message) if message
+
 html_dirpath = ARGV[0]
-RDocLinkChecker.new(html_dirpath)
+RDocLinkChecker.new(
+  html_dirpath,
+  onsite_only: onsite_only,
+  no_toc: no_toc
+).check

data/doc/help.txt

@@ -0,0 +1,16 @@
+Usage:
+  rdoc_link_checker html_dirpath options
+
+The argument is the path to a directory containing a tree
+of RDoc-generated HTML files, such as those generated by command
+
+  rdoc --visibility=private --op html . # Note the trailing dot.
+
+Options:
+  --onsite_only   Check link targets only on pages in the file tree at <html_dirpath>,
+                  and not those on other local pages or those on the web.
+  --no_toc        Do not check links on the TOC page (table_of_contents.html).
+  --version       Print the version and exit.
+  --help          Print this help and exit.
+
+The output is file <html_dirpath>/Report.htm, which reports broken links.

data/lib/rdoc_link_checker/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 class RDocLinkChecker
-  VERSION = "0.2.0"
+  VERSION = "0.3.0"
 end

data/lib/rdoc_link_checker.rb

@@ -1,13 +1,686 @@
 # frozen_string_literal: true
 
+require 'nokogiri'
+require 'rexml/document'
+require 'find'
+require 'net/http'
+
 require_relative 'rdoc_link_checker/version'
 
 class RDocLinkChecker
 
-  def initialize(html_dirpath)
-    puts html_dirpath
+  include REXML
+
+  attr_accessor :html_dirpath, :onsite_only, :no_toc
+
+  def initialize(
+    html_dirpath,
+    onsite_only: false,
+    no_toc: false
+  )
+    self.html_dirpath = html_dirpath
+    self.onsite_only = onsite_only
+    self.no_toc = no_toc
+    @pages = {}
+    @counts = {
+      source_pages: 0,
+      target_pages: 0,
+      links_checked: 0,
+      links_broken: 0,
+    }
+    @verbose = false
+  end
+
+  def check
+    # All work is done in the HTML directory,
+    # and that is where Report.htm will be put.
+    Dir.chdir(html_dirpath) do |dir|
+      @counts[:start_time] = Time.now
+      gather_source_paths
+      create_source_pages
+      create_target_pages
+      verify_links
+      @counts[:end_time] = Time.now
+      report
+    end
+  end
+
+  # Gather paths to source HTML pages.
+  def gather_source_paths
+    paths = []
+    puts 'Gathering source paths' if @verbose
+    paths = Find.find('.').select {|path| path.end_with?('.html') }
+    # Remove leading './'.
+    @source_paths = paths.map{|path| path.sub(%r[^\./], '')}
+    @source_paths.delete('table_of_contents.html') if no_toc
+    if @verbose
+      @source_paths.each_with_index do |source_path, i|
+        puts '- %4d %s' % [i, source_path]
+      end
+    end
+    @counts[:source_pages] = @source_paths.size
+    puts "Gathered #{@source_paths.size} source paths" if @verbose
+  end
+
+  # Create a source \Page object for each source path.
+  # Gather its links and ids.
+  def create_source_pages
+    puts "Creating #{@source_paths.size} source pages" if @verbose
+    @source_paths.sort.each_with_index do |source_path, i|
+      progress_s = RDocLinkChecker.progress_s(i + 1, @source_paths.size)
+      puts "Creating source page #{source_path} #{progress_s}" if @verbose
+      source_page = Page.new(source_path, @verbose, @pages, @counts, onsite_only)
+      @pages[source_path] = source_page
+      source_text = File.read(source_path)
+      doc = Nokogiri::HTML(source_text)
+      source_page.gather_links(doc)
+      source_page.gather_ids(doc)
+      puts "Created source page #{progress_s}" if @verbose
+    end
+    puts "Created #{@pages.size} source pages" if @verbose
+  end
+
+  # Create a target \Page object for each link
+  # (unless already created as a source page).
+  def create_target_pages
+    doc = nil
+    target_page_count = 0
+    @source_paths = @pages.keys
+    @source_paths.each do |source_path|
+      # Need for relative links to work.
+      dirname = File.dirname(source_path)
+      Dir.chdir(dirname) do
+        source_page = @pages[source_path]
+        puts "Creating target pages for #{source_page.links.size} links in #{source_path}" if @verbose
+        source_page.links.each_with_index do |link, i|
+          next if link.path.nil?
+          link.puts(i) if @verbose
+          target_path = link.real_path
+          if @pages[target_path]
+            puts "Page #{target_path} already created" if @verbose
+            target_page = @pages[target_path]
+          else
+            if File.readable?(link.path)
+              puts "Creating target page #{target_path}" if @verbose
+              target_page_count += 1
+              target_page = Page.new(target_path, @verbose, @pages, @counts, onsite_only)
+              @pages[target_path] = target_page
+              target_text = File.read(link.path)
+              doc = Nokogiri::HTML(target_text)
+              target_page.gather_ids(doc)
+              puts "Created target page #{target_path}" if @verbose
+            elsif RDocLinkChecker.checkable?(link.path)
+              puts "Creating target page #{target_path}" if @verbose
+              target_page_count += 1
+              target_page = Page.new(target_path, @verbose, @pages, @counts, onsite_only)
+              @pages[target_path] = target_page
+              puts "Created target page #{target_path}" if @verbose
+              link.exception = fetch(link.path, target_page)
+              link.valid_p = false if link.exception
+            else
+              puts "File not readable or checkable: #{target_path}" if @verbose
+            end
+          end
+          next if target_page.nil?
+          if link.has_fragment? && target_page.ids.empty?
+            doc || doc = Nokogiri::HTML(target_text)
+            target_page.gather_ids(doc)
+          end
+        end
+        puts "Created target pages for #{source_page.links.size} links in #{source_path}" if @verbose
+      end
+    end
+    puts "Created #{target_page_count} target pages" if @verbose
+    @counts[:target_pages] = target_page_count
+  end
+
+  # Verify that each link target exists.
+  def verify_links
+    linking_pages = @pages.select do |path, page|
+      !page.links.empty?
+    end
+    puts "Checking links on #{linking_pages.size} pages" if @verbose
+    link_count = 0
+    broken_count = 0
+    linking_pages.each_pair do |path, page|
+      puts "Checking #{page.links.size} links on page #{path}" if @verbose
+      link_count += page.links.size
+      page.links.each_with_index do |link, i|
+        if link.valid_p.nil? # Don't disturb if already set to false.
+          target_page = @pages[link.real_path]
+          if target_page
+            target_id = link.fragment
+            link.valid_p = target_id.nil? || target_page.ids.include?(target_id)
+          else
+            link_valid_p = false
+          end
+        end
+        link.puts(i) if @verbose
+        broken_count += 1 unless link.valid_p
+      end
+      puts "Checked #{page.links.size} links on page #{path}" if @verbose
+    end
+    puts "Checked #{link_count} links on #{linking_pages.size} pages" if @verbose
+    @counts[:links_checked] = link_count
+    @counts[:links_broken] = broken_count
+  end
+
+  # Fetch the page from the web and gather its ids into the target page.
+  # Returns exception or nil.
+  def fetch(url, target_page)
+    puts "Begin fetch target page #{url}" if @verbose
+    puts "Getting return code for #{url}" if @verbose
+    code = 0
+    exception = nil
+    begin
+      response =  Net::HTTP.get_response(URI(url))
+      code = response.code.to_i
+      target_page.code = code
+      puts "Returned #{code} (#{response.class})" if @verbose
+    rescue => x
+      puts "Raised #{x.class} #{x.message}" if @verbose
+      raise unless x.class.name.match(/^(Net|SocketError|IO::TimeoutError|Errno::)/)
+      exception = RDocLinkChecker::HttpResponseError.new(url, x)
+    end
+    puts "Got return code #{code} for #{url} " if @verbose
+    # Don't load if bad code, or no response, or if not html.
+    if !code_bad?(code)
+      if content_type_html?(response)
+        doc = Nokogiri::HTML(response.body)
+        target_page.gather_ids(doc)
+      end
+    end
+    puts "End fetch target page #{url}" if @verbose
+    exception
+  end
+
+  # Returns whether the code is bad (zero or >= 400).
+  def code_bad?(code)
+    return false if code.nil?
+    (code == 0) || (code >= 400)
+  end
+
+  # Returns whether the response body should be HTML.
+  def content_type_html?(response)
+    return false unless response
+    return false unless response['Content-Type']
+    response['Content-Type'].match('html')
+  end
+
+  # Returns whether the path is offsite.
+  def self.offsite?(path)
+    path.start_with?('http')
+  end
+
+  # Returns the string fragment for the given path or ULR, or +nil+
+  def self.get_fragment(s)
+    a = s.split('#', 2)
+    a.size == 2 ? a[1] : nil
+  end
+
+  # Returns a progress string giving a fraction and percentage.
+  def self.progress_s(i, total)
+    fraction_s = "#{i}/#{total}"
+    percent_i = (i*100.0/total).round
+    "(#{fraction_s}, #{percent_i}%)"
+  end
+
+  # Returns whether the path is checkable.
+  def self.checkable?(path)
+    return false unless path
+    begin
+      uri = URI(path)
+      return ['http', 'https', nil].include?(uri.scheme)
+    rescue
+      return false
+    end
+  end
+
+  # Generate the report; +checker+ is the \RDocLinkChecker object.
+  def report
+
+    doc = Document.new('')
+    root = doc.add_element(Element.new('root'))
+
+    head = root.add_element(Element.new('head'))
+    title = head.add_element(Element.new('title'))
+    title.text = 'RDocLinkChecker Report'
+    style = head.add_element(Element.new('style'))
+    style.text = <<EOT
+*        { font-family: sans-serif }
+.data    { font-family: courier }
+.center  { text-align: center }
+.good    { color: rgb(  0,  97,   0); background-color: rgb(198, 239, 206) } /* Greenish */
+.iffy    { color: rgb(156, 101,   0); background-color: rgb(255, 235, 156) } /* Yellowish */
+.bad     { color: rgb(156,   0,   6); background-color: rgb(255, 199, 206) } /* Reddish */
+.neutral { color: rgb(  0,   0,   0); background-color: rgb(217, 217, 214) } /* Grayish */
+EOT
+
+    body = root.add_element(Element.new('body'))
+    h1 = body.add_element(Element.new('h1'))
+    h1.text = 'RDocLinkChecker Report'
+
+    add_summary(body)
+    add_broken_links(body)
+    add_offsite_links(body) unless onsite_only
+    report_file_path = 'Report.htm' # _Not_ .html.
+    doc.write(File.new(report_file_path, 'w'), 2)
+  end
+
+  def add_summary(body)
+    h2 = body.add_element(Element.new('h2'))
+    h2.text = 'Summary'
+
+    # Parameters table.
+    data = []
+    [
+      :html_dirpath,
+      :onsite_only,
+      :no_toc
+    ].each do |sym|
+      value = send(sym).inspect
+      row = {sym => :label, value => :good}
+      data.push(row)
+    end
+    table2(body, data, 'Parameters')
+    body.add_element(Element.new('p'))
+
+    # Times table.
+    elapsed_time = @counts[:end_time] - @counts[:start_time]
+    seconds = elapsed_time % 60
+    minutes = (elapsed_time / 60) % 60
+    hours = (elapsed_time/3600)
+    elapsed_time_s = "%2.2d:%2.2d:%2.2d" % [hours, minutes, seconds]
+    format = "%Y-%m-%d-%a-%H:%M:%S"
+    start_time_s = @counts[:start_time].strftime(format)
+    end_time_s = @counts[:end_time].strftime(format)
+    data = [
+      {'Start Time' => :label, start_time_s => :good},
+      {'End Time' => :label, end_time_s => :good},
+      {'Elapsed Time' => :label, elapsed_time_s => :good},
+    ]
+    table2(body, data, 'Times')
+    body.add_element(Element.new('p'))
+
+    # Counts.
+    data = [
+      {'Source Pages' => :label, @counts[:source_pages] => :good},
+      {'Target Pages' => :label, @counts[:target_pages] => :good},
+      {'Links Checked' => :label, @counts[:links_checked] => :good},
+      {'Links Broken' => :label, @counts[:links_broken] => :bad},
+    ]
+    table2(body, data, 'Counts')
+    body.add_element(Element.new('p'))
+
+  end
+
+  def add_broken_links(body)
+    h2 = body.add_element(Element.new('h2'))
+    h2.text = 'Broken Links by Source Page'
+
+    if @counts[:links_broken] == 0
+      p = body.add_element('p')
+      p.text = 'None.'
+      return
+    end
+
+    ul = body.add_element(Element.new('ul'))
+    li = ul.add_element(Element.new('li'))
+    li.text = 'Href: the href of the anchor element.'
+    li = ul.add_element(Element.new('li'))
+    li.text = 'Text: the text of the anchor element.'
+    li = ul.add_element(Element.new('li'))
+    li.text = 'Path: the URL or path of the link (not including the fragment):'
+    ul2 = li.add_element(Element.new('ul'))
+    li2 = ul2.add_element(Element.new('li'))
+    li2.text = 'For an on-site link, an abbreviated path is given.'
+    li2 = ul2.add_element(Element.new('li'))
+    li2.text = <<EOT
+For an off-site link, the full URL is given.
+If the path is reddish, the page was not found.
+EOT
+    li = ul.add_element(Element.new('li'))
+    li.text = <<EOT
+Fragment: the fragment of the link.
+If the fragment is reddish, fragment was not found.
+EOT
+
+    @pages.each_pair do |path, page|
+      broken_links = page.links.select {|link| !link.valid_p }
+      next if broken_links.empty?
+
+      h3 = body.add_element(Element.new('h3'))
+      a = Element.new('a')
+      a.text = path
+      a.add_attribute('href', path)
+      h3.add_element(a)
+
+      broken_links.each do |link|
+        data = []
+        # Text, URL, fragment
+        a = Element.new('a')
+        a.text = link.href
+        a.add_attribute('href', link.href)
+        data.push({'Href' => :label, a => :bad})
+        data.push({'Text' => :label, link.text => :good})
+        fragment_p = !link.fragment.nil?
+        class_ = fragment_p ? :good : :bad
+        data.push({'Path' => :label, link.real_path => class_})
+        class_ = fragment_p ? :bad : :good
+        data.push({'Fragment' => :label, link.fragment => class_})
+        if link.exception
+          data.push({'Exception' => :label, link.exception.class => :bad})
+          data.push({'Message' => :label, link.exception.message => :bad})
+        end
+        table2(body, data)
+        body.add_element(Element.new('p'))
+      end
+    end
+
+  end
+
+  def add_offsite_links(body)
+    h2 = body.add_element(Element.new('h2'))
+    h2.text = 'Off-Site Links by Source Page'
+    @pages.each_pair do |path, page|
+      offsite_links = page.links.select do |link|
+        RDocLinkChecker.offsite?(link.href)
+      end
+      next if offsite_links.empty?
+
+      h3 = body.add_element(Element.new('h3'))
+      a = Element.new('a')
+      a.text = path
+      a.add_attribute('href', path)
+      h3.add_element(a)
+
+      offsite_links.each do |link|
+        data = []
+        # Text, URL, fragment
+        a = Element.new('a')
+        a.text = link.href
+        a.add_attribute('href', link.href)
+        class_ = link.valid_p ? :good : :bad
+        data.push({'Href' => :label, a => class_})
+        data.push({'Text' => :label, link.text => :good})
+        table2(body, data)
+        body.add_element(Element.new('p'))
+      end
+    end
+  end
+
+  Classes = {
+    label: 'label center neutral',
+    good: 'data center good',
+    iffy: 'data center iffy',
+    bad: 'data center bad',
+  }
+
+  def table2(parent, data, title = nil)
+    data = data.dup
+    table = parent.add_element(Element.new('table'))
+    if title
+      tr = table.add_element(Element.new('tr)'))
+      th = tr.add_element(Element.new('th'))
+      th.add_attribute('colspan', 2)
+      if title.kind_of?(REXML::Element)
+        th.add_element(title)
+      else
+        th.text = title
+      end
+    end
+    data.each do |row_h|
+      label, label_class, value, value_class = row_h.flatten
+      tr = table.add_element(Element.new('tr'))
+      td = tr.add_element(Element.new('td'))
+      td.text = label
+      td.add_attribute('class', Classes[label_class])
+      td = tr.add_element(Element.new('td'))
+      if value.kind_of?(REXML::Element)
+        td.add_element(value)
+      else
+        td.text = value
+      end
+      td.add_attribute('class', Classes[value_class])
+    end
+  end
+
+  class Error; end
+
+  class HttpResponseError < Error
+
+    attr_accessor :url, :x
+
+    def initialize(url, x)
+      self.url = url
+      self.x = x
+    end
+
+    def message
+      <<EOT
+#{self.class.name}:
+An exception was raised when checking page availability with Net::HTTP:
+  Url: #{url}
+  Class: #{x.class}
+  Message: #{x.message}
+EOT
+    end
+
+  end
+
+  class HttpStatusCodeError < Error
+
+    attr_accessor :url, :code
+
+    def initialize(url, code)
+      self.url = url
+      self.code = code
+    end
+
+    def message
+      <<EOT
+#{self.class.name}:
+  The return code for the page was not 200:
+    Url: #{url}
+    Return code: #{code}
+EOT
+    end
+
+  end
+
+  # Class to represent a page.
+  class Page
+
+    attr_accessor :path, :type, :verbose, :pages, :counts, :code, :links, :ids, :dirname, :onsite_only
+
+    # Returns a new \Page object:
+    #
+    # - +path+: a path relative to the HTML directory (if on-site)
+    #   or a URL (if off-site).
+    # - +verbose+: whether to put progress message to $stdout.
+    # - +pages+: hash of path/page pairs.
+    # - +counts+: hash of counts.
+    #
+    def initialize(path, verbose, pages, counts, onsite_only)
+      self.path = path
+      self.verbose = verbose
+      self.pages = pages
+      self.counts = counts
+      self.onsite_only = onsite_only
+      self.code = nil
+      self.links = []
+      self.ids = []
+      self.dirname = File.dirname(path)
+      self.dirname = self.dirname == '.' ? '' : dirname
+    end
+
+    # Gather links for the page:
+    #
+    # - +doc+: Nokogiri document to be parsed for links.
+    #
+    def gather_links(doc)
+      puts 'Gathering links' if @verbose
+      i = 0
+      # The links are in the anchors.
+      doc.search('a').each do |a|
+        # Ignore pilcrow (paragraph character) and up-arrow.
+        next if a.text == "\u00B6"
+        next if a.text == "\u2191"
+
+        href = a.attr('href')
+        next if href.nil? or href.empty?
+        next if RDocLinkChecker.offsite?(href) && onsite_only
+        next unless RDocLinkChecker.checkable?(href)
+
+        link = Link.new(href, a.text, dirname)
+        next if link.path.nil? || link.path.empty?
+
+        links.push(link)
+        link.puts(i) if @verbose
+        i += 1
+      end
+      puts "Gathered #{i} links" if @verbose
+    end
+
+    # Gather ids for the page.
+    # +doc+ is the Nokogiri document to be parsed.
+    def gather_ids(doc)
+      # Don't do twice (some pages are both source and target).
+      return unless ids.empty?
+
+      # For off-site, gather all ids, regardless of element.
+      if RDocLinkChecker.offsite?(path)
+        doc.xpath("//*[@id]").each do |element|
+          id = element.attr('id')
+          ids.push(id)
+        end
+        return
+      end
+
+      # We're on-site, which means that the page is RDoc-generated
+      # and we know what to expect.
+      # In theory, an author can link to any element that has an attribute :id.
+      # In practice, gathering all such elements is very time-consuming.
+      # These are the elements currently linked to:
+      #
+      # - body
+      # - a
+      # - div
+      # - dt
+      # - h*
+      #
+      # We can add more as needed (i.e., if/when we have actual broken links).
+      puts 'Gathering potential link targets' if @verbose
+
+      # body element has 'top', which is a link target.
+      body = doc.at('//body')
+      id = body.attribute('id')
+      ids.push(id) if id
+
+      # Some ids are in the as (anchors).
+      body.search('a').each do |a|
+        id = a.attr(id)
+        ids.push(id) if id
+      end
+
+      # Method ids are in divs, but gather only method-detail divs.
+      body.search('div').each do |div|
+        class_ = div.attr('class')
+        next if class_.nil?
+        next unless class_.match('method-')
+        id = div.attr('id')
+        ids.push(id) if id
+      end
+
+      # Constant ids are in dts.
+      body.search('dt').each do |dt|
+        id = dt.attr('id')
+        ids.push(id) if id
+      end
+
+      # Label ids are in headings.
+      %w[h1 h2 h3 h4 h5 h6].each do |tag|
+        body.search(tag).each do |h|
+          id = h.attr('id')
+          ids.push(id) if id
+        end
+      end
+      if @verbose
+        ids.each_with_index do |id, i|
+          puts '%4d %s' % [i, id]
+        end
+      end
+      puts "Gathered #{ids.size} potential link targets" if @verbose
+
+    end
+
   end
 
-  class Error < StandardError; end
+  # Class to represent a link.
+  class Link
+
+    attr_accessor :href, :text, :dirname, :path, :fragment, :valid_p, :real_path, :exception
+
+    # Returns a new \Link object:
+    #
+    # - +href+: attribute href from anchor element.
+    # - +text+: attribute text from anchor element.
+    # - +dirname+: directory path of the linking page.
+    #
+    # TODO: accept the anchor element, instead of its href and text.
+    def initialize(href, text, dirname)
+      self.href = href
+      self.text = text
+      self.dirname = dirname
+      path, fragment = href.split('#', 2)
+      self.path = path
+      self.fragment = fragment
+      self.valid_p = nil
+      self.real_path = make_real_path(dirname, path)
+      self.exception = nil
+    end
+
+    # Return the real (not relative) path of the link.
+    def make_real_path(dirname, path)
+      # Trim single dot.
+      return path.sub('./', '') if path.start_with?('./')
+      return path if dirname.nil? || dirname.empty?
+
+      # May have one or more leading '../'.
+      up_dir = '../'
+      levels = path.scan(/(?=#{up_dir})/).count
+      dirs = dirname.split('/')
+      if levels == 0
+        dirs.empty? ? path : File.join(dirname, path)
+      else
+        # Remove leading '../' elements.
+        path = path.gsub(%r[\.\./], '')
+        # Remove the corresponding parts of dirname.
+        dirs.pop(levels)
+        return path if dirs.empty?
+        dirname = dirs.join('/')
+        File.join(dirname, path)
+      end
+    end
+
+    # Returns whether the link has a fragment.
+    def has_fragment?
+      fragment ? true : false
+    end
+
+    # Puts link info onto $stdout.
+    def puts(i)
+      $stdout.puts <<EOT
+Link #{i}:
+  Href:      #{href}
+  Text:      #{text}
+  Path:      #{path}
+  Fragment:  #{fragment}
+  Valid:     #{valid_p}
+  Real path: #{real_path}
+  Dirname:   #{dirname}
+EOT
+    end
+  end
 
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rdoc_link_checker
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.3.0
 platform: ruby
 authors:
 - burdettelamar
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-05-19 00:00:00.000000000 Z
+date: 2023-05-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -66,6 +66,7 @@ files:
 - README.md
 - Rakefile
 - bin/rdoc_link_checker
+- doc/help.txt
 - lib/rdoc_link_checker.rb
 - lib/rdoc_link_checker/version.rb
 - rdoc_link_checker.gemspec