From: Rob Biedenharn <Rob@...>
Date: 2009-09-28T14:01:56+09:00
Subject: Re: Getting a valid URL from a command line

On Sep 28, 2009, at 12:45 AM, Hunt Jon wrote:
> On Mon, Sep 28, 2009 at 1:11 PM, Rob Biedenharn
> <Rob@agileconsultingllc.com> wrote:
>> On Sep 27, 2009, at 11:48 PM, Hunt Jon wrote:
>>>
>>> I expect that if I run "URI.parse()" it should raise an error, but
>>> it doesn't happen.
>>>
>>> Can anybody help me on this one?
>>>
>>> Jon
>>
>> require 'uri'
>> print "Type a URL: "
>> begin
>>  url = gets.chomp
>>  puts "You said: #{url.inspect}"
>>  uri = URI.parse(url) # should raise if a variable 'url' is  
>> malformed.
>>  puts uri.inspect
>> rescue URI::InvalidURIError
>>  puts "That is not a valid URL. Try again."
>>  retry
>> end
>>
>> Try getting a little bit more information out (and post what input  
>> you are
>> trying that you expect to be malformed).
>>
>> Note that some URI's are HTTP and some might be Generic.  There are  
>> a lot
>> more types of URI that just those that start with http://.  Have  
>> you ever
>> seen a jdbc resource string?
>>
>> -Rob
>
> I expect a user to input a HTTP or HTTPS URL. e.g., http://abcdef.gov
> Maybe using URI seems *too* generic after the research as 'uri' means
> different protocols, not just http/https.
>
> I'll look into it. Perhaps using Regexp match would be better.
>
> Jon

You can see what the scheme is determined to be:

irb> require 'uri'
=> true
irb> u=URI.parse('http://example.com/')
=> #<URI::HTTP:0x395b34 URL:http://example.com/>
irb> u.scheme
=> "http"
irb> x=URI.parse('example.com')
=> #<URI::Generic:0x392f24 URL:example.com>
irb> x.scheme
=> nil

You probably don't want to jump down the Regexp rabbit-hole if you  
know that you want a valid URI. Let the library do the heavy lifting.

-Rob

Rob Biedenharn		http://agileconsultingllc.com
Rob@AgileConsultingLLC.com