-require 'open-uri'
-
+require 'net/http'
+require 'uri'
+require 'cgi'
+\r
Url = Struct.new("Url", :channel, :nick, :time, :url)
-TITLE_RE = /<\s*title\s*>(.+)<\s*\/title\s*>/im
+TITLE_RE = /<\s*?title\s*?>(.+?)<\s*?\/title\s*?>/im
class UrlPlugin < Plugin
BotConfig.register BotConfigIntegerValue.new('url.max_urls',
:default => 100, :validate => Proc.new{|v| v > 0},
:desc => "Maximum number of urls to store. New urls replace oldest ones.")
- BotConfig.register BotConfigBooleanValue.new('url.say_titles',
+ BotConfig.register BotConfigBooleanValue.new('url.display_link_info',
:default => true,
- :desc => "Get the title of any links pasted to the channel and display it (Also, tells if the link is broken)")
+ :desc => "Get the title of any links pasted to the channel and display it (also tells if the link is broken or the site is down)")
def initialize
super
"urls [<max>=4] => list <max> last urls mentioned in current channel, urls search [<max>=4] <regexp> => search for matching urls. In a private message, you must specify the channel to query, eg. urls <channel> [max], urls search <channel> [max] <regexp>"
end
- def get_title_for_url(url)
- begin
- pagedata = open(url).read
- return unless TITLE_RE.match(pagedata)
- title = $1.strip.gsub(/\s*\n\s*/, " ")
- title = title[0..255] if title.length > 255
- "Title: #{title}"
+ def get_title_from_html(pagedata)
+ return unless TITLE_RE.match(pagedata)
+ title = $1.strip.gsub(/\s*\n+\s*/, " ")
+ title = CGI::unescapeHTML title
+ title = title[0..255] if title.length > 255
+ "[Link Info] title: #{title}"
+ end
- rescue SocketError => e
- "Title: ^- Error connecting to site (#{e.message})"
- rescue OpenURI::HTTPError => e
- "Title: ^- Error getting page (#{e.message})"
- rescue Exception => e
- "Title: ^- Error: #{e.inspect}"
+ def get_title_for_url(uri_str)
+ # This god-awful mess is what the ruby http library has reduced me to.
+ # Python's HTTP lib is so much nicer. :~(
+
+ puts "+ Getting #{uri_str}"
+ url = URI.parse(uri_str)
+ return if url.scheme !~ /https?/
+
+ puts "+ connecting to #{url.host}:#{url.port}"\r
+ http = @bot.httputil.get_proxy(url)
+ title = http.start do |http|
+ url.path = '/' if url.path == ''
+ head = http.request_head(url.path)
+ case head
+ when Net::HTTPRedirection then
+ # call self recursively if this is a redirect
+ redirect_to = head['location']
+ puts "+ redirect location: #{redirect_to}"
+ url = URI.join url.to_s, redirect_to
+ puts "+ whee, redirecting to #{url.to_s}!"
+ title = get_title_for_url(url.to_s)
+ when Net::HTTPSuccess then
+ if head['content-type'] =~ /^text\//
+ # content is 'text/*'
+ # retrieve the title from the page
+ puts "+ getting #{url.path}"
+ response = http.request_get(url.request_uri)
+ return get_title_from_html(response.body)
+ else
+ # content isn't 'text/*'... display info about the file.
+ size = head['content-length'].gsub(/(\d)(?=\d{3}+(?:\.|$))(\d{3}\..*)?/,'\1,\2')
+ #lastmod = head['last-modified']
+ return "[Link Info] type: #{head['content-type']}#{size ? ", size: #{size} bytes" : ""}"
+ end
+ when Net::HTTPClientError then
+ return "[Link Info] Error getting link (#{head.code} - #{head.message})"
+ when Net::HTTPServerError then
+ return "[Link Info] Error getting link (#{head.code} - #{head.message})"\r
+ end
end
+ rescue SocketError => e
+ return "[Link Info] Error connecting to site (#{e.message})"
end
def listen(m)