From: Eleanor McHugh <eleanor@...>
Date: 2009-04-07T00:15:28+09:00
Subject: Re: Code duplication

On 6 Apr 2009, at 14:50, Arun Kumar wrote:
> My tutor is saying that there is a duplication in the above code. ie.
> code for html reading is specified twice without any purpose and it
> should be removed. I've no idea where there is a mistake. I'm a newbee
> to ruby and i don't understand the problem correctly or where things
> went wrong. Can anyone please help me to find the mistake.


There are probably better solutions, but the following illustrates the  
point your tutor is making:

MOZILLA_HEADER = { "User-Agent"=>"Mozilla/4.0 (compatible; MSIE 5.5;  
Windows NT 5.0)" }

def get_http_response uri, max_redirects = 0
   Net::HTTP.start(uri) do |connection|
     response = connection.get(uri.path, MOZILLA_HEADER)
     response &&= case response
     when Net::HTTPRedirection
       if max_redirects > 0 then
         get_http_response URI.parse(response['Location']),  
(max_redirects - 1)
       else
         raise "Too many redirects"
       end
     when Net::HTTPRedirection
       get_http_response URI.parse("http://#{uri.host}:#{uri.port}/"),  
max_redirects
     end
   end
end

data = get_http_response(my_uri, 3).body

See how get_http_response is recursive in the case of an erroneous  
response? This minimises the actual HTTP interaction code as well as  
elegantly handling redirects. Whilst this could result in many more  
http connections being used, it also makes them clear up after  
themselves which is always good.


Ellie

Eleanor McHugh
Games With Brains
http://slides.games-with-brains.net
----
raise ArgumentError unless @reality.responds_to? :reason