search plugin: fix gcalc regexp

[user/henk/code/ruby/rbot.git] / data / rbot / plugins / url.rb

diff --git a/data/rbot/plugins/url.rb b/data/rbot/plugins/url.rb
index c4fce2ae6a7f3d03d36385e51c08accb2ac89422..2cee5f4f5d37b2d1443b16355fe9ae60c086fc20 100644 (file)
--- a/data/rbot/plugins/url.rb
+++ b/data/rbot/plugins/url.rb
@@ -57,8 +57,19 @@ class UrlPlugin < Plugin
     url = uri_str.kind_of?(URI) ? uri_str : URI.parse(uri_str)
     return if url.scheme !~ /https?/
 
-    if url.host =~ @no_info_hosts
-      return "Sorry, info retrieval for #{url.host} is disabled"
+    # also check the ip, the canonical name and the aliases
+    begin
+      checks = TCPSocket.gethostbyname(url.host)
+      checks.delete_at(-2)
+    rescue => e
+      return "Unable to retrieve info for #{url.host}: #{e.message}"
+    end
+
+    checks << url.host
+    checks.flatten!
+
+    unless checks.grep(@no_info_hosts).empty?
+      return "Sorry, info retrieval for #{url.host} (#{checks.first}) is disabled"
     end
 
     logopts = opts.dup
@@ -68,7 +79,7 @@ class UrlPlugin < Plugin
 
     begin
       debug "+ getting info for #{url.request_uri}"
-      info = Utils.get_html_info(url)
+      info = @bot.filter(:htmlinfo, url)
       debug info
       resp = info[:headers]
 
@@ -120,19 +131,30 @@ class UrlPlugin < Plugin
       next unless urlstr =~ /^https?:/
       title = nil
       debug "Getting title for #{urlstr}..."
+      reply = nil
       begin
         title = get_title_for_url(urlstr,
                                   :nick => m.source.nick,
                                   :channel => m.channel,
                                   :ircline => m.message)
         debug "Title #{title ? '' : 'not '} found"
+        reply = "#{LINK_INFO} #{title}" if title
       rescue => e
-        m.reply "Error #{e.message}"
+        debug e
+        # we might get a 404 because of trailing punctuation, so we try again
+        # with the last character stripped. this might generate invalid URIs
+        # (e.g. because "some.url" gets chopped to some.url%2, so catch that too
+        if e.message =~ /\(404 - Not Found\)/i or e.kind_of?(URI::InvalidURIError)
+          # chop off last character, and retry if we still have enough string to
+          # look like a minimal URL
+          retry if urlstr.chop! and urlstr =~ /^https?:\/\/./
+        end
+        reply = "Error #{e.message}"
       end
 
       if display_info > urls_displayed
-        if title
-          m.reply("#{LINK_INFO} #{title}", :overlong => :truncate)
+        if reply
+          m.plainreply(reply, :overlong => :truncate)
           urls_displayed += 1
         end
       end
@@ -158,12 +180,12 @@ class UrlPlugin < Plugin
     Thread.new { handle_urls(m, urls, params[:urls].length) }
   end
 
-  def listen(m)
-    return unless m.kind_of?(PrivMessage)
+  def message(m)
     return if m.address?
 
     escaped = URI.escape(m.message, OUR_UNSAFE)
-    urls = URI.extract(escaped)
+    urls = URI.extract(escaped, ['http', 'https'])
+    return if urls.empty?
     Thread.new { handle_urls(m, urls) }
   end