ulink-checker 0.0.1

data/Gemfile

@@ -0,0 +1,7 @@
+source "http://rubygems.org"
+# Add dependencies required to use your gem here.
+# Example:
+#   gem "activesupport", ">= 2.3.5"
+gem "nokogiri"
+gem "rainbow"
+gem "link-checker"

data/bin/ulink-checker

@@ -0,0 +1,161 @@
+#!/usr/bin/env ruby -W0
+ 
+ $LOAD_PATH.unshift File.dirname(__FILE__)
+
+require 'rubygems'
+require 'nokogiri'
+require 'net/http'
+require 'rainbow'
+require 'link_checker'
+require 'optparse'
+
+require 'link_checker_override.rb'
+require 'duration.rb'
+
+errors=0
+fails=0
+oks=0
+warnings=0
+redirects=0
+starttime = Time.now
+
+doc = nil
+
+
+# ======================================== read commandline parameters
+def processCommandLine
+  options = {}
+
+  optparse = OptionParser.new do |opts|
+    opts.banner = "ulink-checker v0.0.1"
+
+    #options[:followXIInclude] = false
+    #opts.on( '-x', '--xiinclude', 'Follow' ) do
+    #  options[:getSecurityLevels] = true
+    #end
+
+    options[:url] = ""
+    opts.on( '-u URL', '--url URL', 'check XML at URL (not implemented yet)' ) do | url |
+      options[:url] = url
+    end
+    
+    options[:file] = ""
+    opts.on( '-f FILE', '--file FILE', 'check XML from file' ) do | file | 
+      options[:file] = file
+    end
+
+    opts.on( '-h', '--help', 'Display this screen' ) do
+      puts opts
+      exit
+    end
+
+    if (ARGV.size == 0) then
+      puts opts
+      exit
+    end
+   
+  end
+
+  begin
+    optparse.parse!
+  rescue => error
+    puts "ERROR PARSING COMMANDLINE:"
+    puts "\t"+error.to_s
+    puts
+    puts optparse
+    exit
+  end
+
+  # if they specify URL then too bad!
+  if (options[:url] != "") then
+    puts "SORRY: URLs are not supported yet"
+    puts
+    puts optparse
+    exit
+  end
+
+  #can't specify both url and file, only one
+  if (options[:url] != "" and option[:file] != "") then
+    puts "ERROR: can only specify one of file or URL."
+    puts
+    puts optparse
+    exit
+  end
+
+  return options
+end
+
+options = processCommandLine
+
+# ======================================== 
+
+if (options[:file] != "") then
+   puts "Checking URLs in #{options[:file]} ...\n\n"
+
+   File.open(options[:file]) do |fp|
+      doc = Nokogiri::XML(fp) do |conf|
+         conf.strict.dtdload.noent.nocdata.xinclude
+      end
+   end
+end
+
+# xpath the urls out
+urls = doc.xpath("//ulink")
+
+urls.each do | url |
+
+   url_string = url.xpath("./@url").to_s
+
+   begin
+      result = LinkChecker.check_uri(URI.parse(url_string.to_s))
+   rescue => error
+      #puts "\tERROR: #{error.to_s}"
+      if url_string.to_s.start_with?("mailto:")
+         print "WARNING".color("#800040")+"\t\t"
+         warnings = warnings + 1
+      else
+         #result = LinkChecker::Error.new(:uri_string => url_string.to_s)
+         print "ERROR".color("#800040")+"\t\t"
+         errors = errors + 1
+      end
+   end
+
+   if result.class.to_s == "LinkChecker::Error"
+      print "FAIL".foreground(:red).inverse.blink+"\t\t"
+      fails = fails + 1
+   end
+   if result.class.to_s == "LinkChecker::Good"
+      print "OK".foreground(:green)+"\t\t"
+      oks = oks + 1
+   end
+   if result.class.to_s == "LinkChecker::Redirect"
+      print "REDIRECT".color("#ff8000")+"\t"
+      redirects = redirects + 1
+   end
+
+   print " #{url_string.to_s.foreground(:blue).underline}\n"
+
+   if result.class.to_s == "LinkChecker::Redirect"
+      print "\t\t -> "+result.final_destination_uri_string.foreground(:blue).underline+"\n"
+   end
+   if result.class.to_s != "LinkChecker::Good" 
+      print "\t\t at "+url.path.to_s+"\n"
+   end
+
+end
+
+endtime = Time.now
+
+puts "\n"
+puts "SUMMARY".underline
+puts "Source:\t\t\t#{ARGV[0]}"
+puts "Total URLs:\t\t#{errors+fails+oks+warnings+redirects}"
+puts "Time taken:\t\t#{(endtime-starttime).duration}"
+puts "URLs OK:\t\t#{oks}"
+puts "URL Errors:\t\t#{errors}"
+puts "URLs not found:\t\t#{fails}"
+puts "URLs with warnings:\t#{warnings}"
+puts "URLs that redirect:\t#{redirects}"
+
+print "\n"
+

data/lib/duration.rb

@@ -0,0 +1,18 @@
+class Numeric
+  def duration
+    secs  = self.to_int
+    mins  = secs / 60
+    hours = mins / 60
+    days  = hours / 24
+
+    if days > 0
+      "#{days} days and #{hours % 24} hours"
+    elsif hours > 0
+      "#{hours} hours and #{mins % 60} minutes"
+    elsif mins > 0
+      "#{mins} minutes and #{secs % 60} seconds"
+    elsif secs >= 0
+      "#{secs} seconds"
+    end
+  end
+end

data/lib/link_checker_override.rb

@@ -0,0 +1,49 @@
+require 'net/http'
+require 'net/https'
+require 'uri'
+
+class LinkChecker
+
+   def self.check_uri(uri, redirected=false)
+     http = Net::HTTP.new(uri.host, uri.port)
+     http.use_ssl = true if uri.scheme == "https"
+     http.verify_mode = OpenSSL::SSL::VERIFY_NONE if http.use_ssl?
+
+     http.start do
+       path = (uri.path.empty?) ? '/' : uri.path
+       http.request_get(path) do |response|
+         #print "\n#{response.class}(#{response.code}) "
+         case response
+
+         when Net::HTTPSuccess
+           if redirected
+             return Redirect.new(:final_destination_uri_string => uri.to_s)
+           else
+             return Good.new(:uri_string => uri.to_s)
+           end
+
+         when Net::HTTPForbidden
+           if redirected
+             return Redirect.new(:final_destination_uri_string => uri.to_s)
+           else
+             return Good.new(:uri_string => uri.to_s)
+           end
+
+         when Net::HTTPUnauthorized 
+           if redirected
+             return Redirect.new(:final_destination_uri_string => uri.to_s)
+           else
+             return Good.new(:uri_string => uri.to_s)
+           end
+
+
+         when Net::HTTPRedirection then
+           return self.check_uri(URI(response['location']), true)
+         else
+           return Error.new(:uri_string => uri.to_s, :error => response)
+         end
+       end
+     end
+   end
+
+end

data/readme.md

@@ -0,0 +1,22 @@
+Scans a specified Docbook XML file and checks the URLs for each `<ulink>` tag.  It supports `<xi:include>`
+
+Reports:
+
+   * ERROR: poorly formed URLs
+   * FAIL: URLs that fail to resolve (404 etc)
+   * WARNING: mailto links
+   * REDIRECT: when the URL gets redirected
+   * OK: if the URL works fine
+
+## setup
+
+1. `$ bundle update`
+
+## run
+
+1. `./ulink-checker.rb somedocbook.xml`
+2. See the amazing technicolour output
+
+## why isn't it a gem
+
+Because I haven't done it yet.

data/ulink-checker.gemspec

@@ -0,0 +1,19 @@
+Gem::Specification.new do |s|
+  s.name        = 'ulink-checker'
+  s.version     = '0.0.1'
+  s.date        = '2013-03-12'
+  s.summary     = "checks urls in docbook ulink tags"
+  s.description = "checks urls in docbook ulink tags"
+  s.authors     = ["darrin Mison"]
+  s.email       = 'dmison@me.com'
+  s.executables << 'ulink-checker'
+  s.files       = [
+    "Gemfile", 
+    "readme.md", 
+    "ulink-checker.gemspec", 
+    "bin/ulink-checker", 
+    "lib/duration.rb", 
+    "lib/link_checker_override.rb"]
+  s.homepage    =
+    'https://github.com/dmison/ulink-checker'
+end

metadata

@@ -0,0 +1,67 @@
+--- !ruby/object:Gem::Specification 
+name: ulink-checker
+version: !ruby/object:Gem::Version 
+  prerelease: false
+  segments: 
+  - 0
+  - 0
+  - 1
+  version: 0.0.1
+platform: ruby
+authors: 
+- darrin Mison
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2013-03-12 00:00:00 +10:00
+default_executable: 
+dependencies: []
+
+description: checks urls in docbook ulink tags
+email: dmison@me.com
+executables: 
+- ulink-checker
+extensions: []
+
+extra_rdoc_files: []
+
+files: 
+- Gemfile
+- readme.md
+- ulink-checker.gemspec
+- bin/ulink-checker
+- lib/duration.rb
+- lib/link_checker_override.rb
+has_rdoc: true
+homepage: https://github.com/dmison/ulink-checker
+licenses: []
+
+post_install_message: 
+rdoc_options: []
+
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      segments: 
+      - 0
+      version: "0"
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      segments: 
+      - 0
+      version: "0"
+requirements: []
+
+rubyforge_project: 
+rubygems_version: 1.3.6
+signing_key: 
+specification_version: 3
+summary: checks urls in docbook ulink tags
+test_files: []
+