Indeed. My take on the matter is ... what this person wants to get is:   “this-or-that part of ‘a domain name.’” And it so happens that the “obvious way” to get there is... “well, (obviously) I have to (figure out how to) write a regular expression” before I'll be able to “get that.”

And I'm reaching over and hitting the Pause button and saying... “oh, really?” I'll bet that there's a very good CPAN-module already out there which will give you the answer that you want. I'll wager that in fact you don't have to tackle regular-expressions in order to get the result you want, because somebody out there has already done it and has done it very well.

Dictum Ne Agas — “Do Not Do A Thing Already Done.”

For the "this or that part", they need a different module which can extract the parts like host, path, query string, port etc. etc.

---




Nobody says perl looks like line-noise any more
kids today don't know what line-noise IS ...

Thank you for pointing me to the modules, I think those will work great. Between URI::Find and URI::Find::Schemeless I 'should' be able to get done what I need. I'll let you know my results.

Thanks,

jnbek

Is there a CPAN module that will give you the result that you want, without you having to create a regular-expression at all?

Focus on the result that you want to obtain...

I don't have a ready-made solution for you. For this kind of problem, I usually can recommend some part of Regexp::Common, but it doesn't seem to have a module for this.

It sounds as if you want sub.company.domx.domy as well as just company.domx.domy where domx is optional, depending on what domy is.

I'm not sure what the allowable alphabet for domain parts is, but for now I'll say

$dom = qr{
          [a-z]       # starts with a letter
          [a-z0-9-]*  # zero or more letters, numbers, hyphens
         }xi;
[download]

From there, get a list of Internet top-level domains. Look through and see which of those you want an "extra" domain for. For au, for example, you want

$au = qr{ $dom \.   # company
          $dom \.   # subdomain
          au        # ccTLD
          \. ?      # optional trailing dot
          \b        # word break
         }xi;
[download]

But for the mighty .com, it's just

$com = qr{ $dom \.  # company
           com      # gTLD
           \. ?     # optional trailing dot
           \b       # word break
         }xi;
[download]

The word break at the end keeps us from matching i.am.a.silly.com.administrator.at.example.com as silly.com instead of example.com. This might not be the right thing for your input, though. It might be better to say (?! $dom) there instead (negative look ahead to confirm there's no more $dom stuff).

The optional trailing dot allows "example.com." as well as "example.com" (both are "legal").

Given these, you can make

my $domain = qr{ (?:        # start of group
                   $dom \.  # company subdomain
                 ) ?        # end of group, make it optional
                 (?:        # start of group
                   $com     # .com pattern
                   |        # or
                   $au      # .au pattern
                 )          # end of group
               }xi;
[download]

Having written all this and then done some testing, I find that this is harder than I thought (I should have known!). I'll stop now and provide what I got so far along with the tests that show it doesn't work. Expanding the testing to more interesting cases once you have a solution should be easy.

One problem revealed by the tests is that the pattern will be happy to match merely example.com when faced with example.com.au. Also, it will match domains with illegal characters in them, basically by pretending that the name ends at the illegal character. For example illegal_underscore.com matches as underscore.com. Depending on your application, that may be acceptable, but I don't like it much.

Then count the components with $dots = ($host =~ y/\.//); and make decisions on how much to print based on the number of dots you find.

PS: Yes, I know that regex was weak but it's just a simple example to get started. A proper module is of course suggested to have any sort of reliability for this complicated of a problem.

#!/usr/bin/perl -w
use strict;
use Data::Dumper;

my $url = shift @ARGV || 'http://www.ietf.org/rfc/rfc2396.txt';
# [Page 28] "Parsing a URI Reference with a Regular Expression"
if (my @parts = ($url =~ m{
      ^(([^:/?#]+):)? #1-2 scheme
      (//([^/?#]*))?  #3-4 authority
      ([^?#]*)        #5 path
      (\?([^#]*))?    #6-7 query
      (\#(.*))?       #8-9 fragment
      }x)) {
   print Dumper(\@parts);
}
[download]