Using the CGI perl module, you can use function

$cgi->virtual_host ();

i

--

thanks - I can get the hostname, it's the domain name I'm after.

--
William Tasso

** Business as usual

you mean, say the hostname is www1.domain.com, you want to extract
domain.com?

I would do a "split", and then "pop" the last two items. It is good
except for countries where domains are like domain.co.uk. perhaps if
the second pop is a 'co' or some such, you could pop one more item.

i
--

And pop-pickers?

Everyone needs a hobby....

Can you be a bit more explicit.

That's okay. The invoice is in the post.

Matt Probert

Same as always

Oh you mean you just want to extract the part that would be used for
emailing? Like setting a spider to gather possible email addresses?

I think thats been done quite abit but not openly released for obvious
reasons

Gandalf Parker

William is explicit enough without you asking him to be more so! :P Next
thing you know you'll be asking him to take off that Speedo of his...

Grey

--
The technical axiom that nothing is impossible sinisterly implies the
pitfall corollary that nothing is ridiculous.
- http://www.greywyvern.com/ringmaker - Orca Ringmaker: Host a web ring
from your website!

Yes - that is exactly the issue, consider these hosts ...

www.example.com
example.com
office.example.com
admin.office.example.com

www.example.co.uk
etc...

www.example.uk.net
etc...

--
William Tasso

** Business as usual

Ugh - as always, the invitation will be politely refused.

--
William Tasso

** Business as usual

well not really

[Email Removed]
and
[Email Removed]

may both be valid mail addresses. I'm only interested in the domain.

Heh he - no need these hosts/domains already sit nicely on my servers.
I'm writing a generic component that incidentally needs to know the domain
name regadless of hostname.

except that in the case of mail spiders there is no incentive to get it
right.

--
William Tasso

** Business as usual

Asked this some time ago in the Perl group, and uhm.. didn't get a nice
reply back :-) There is no simple rule to do this. There are also things
like example.uk.com iirc.


--
John Perl SEO tools: http://johnbokma.com/perl/
Experienced (web) developer: http://castleamber.com/
Get a SEO report of your site for just 100 USD:
http://johnbokma.com/websitedesign/seo-expert-help.html

Nice reply from a perl group? That's unrealistic... :)

i

Nah, often the not so nice replies are justified, but in this case no. (And
I am not saying that because I asked :-D )

--
John Perl SEO tools: http://johnbokma.com/perl/
Experienced (web) developer: http://castleamber.com/
Get a SEO report of your site for just 100 USD:
http://johnbokma.com/websitedesign/seo-expert-help.html

Comment 32 on this page looks promising:

http://tinyurl.com/c76d6

Can't say that I understand it, but would like to.

Ed

Yes, but so far as I can tell (I still don't parse regular expressions in
my head) it still fails when there is no hostname preceeding the domain
name. e.g. http://example.com would return ".com" - I think

It's a long road, but regular expressions can save a heap of scripting.


--
William Tasso

** Business as usual

for ( $domain ) {

# remove sub domain
s/www?d?.([^.]+.)/$1/g; # common ww(w)(d) prefix
s/.+.([^.]+.[a-z]{3})$/$1/g;
s/.+.([^.]+.[^.]+.[a-z]{2})$/$1/g;
}

is what I used for a project. Note: it's far from perfect :-D.

--
John Perl SEO tools: http://johnbokma.com/perl/
Experienced (web) developer: http://castleamber.com/
Get a SEO report of your site for just 100 USD:
http://johnbokma.com/websitedesign/seo-expert-help.html

thanks, but ...

yes, having thought on it some more I'm coming to the conclusion it cannot
be done purely with logic.

As these domains are all known to me I think I'll have to do a match
against a data store - splitting the hostname and starting at the back,
adding elements till I find a match.

Really can't think of anything else that is guaranteed to work (for an
acceptable value of 'work')

--
William Tasso

** Business as usual

What would it do with this?
http://www.www.co.uk/

What about sites that use "www." and "www2." and "www3." and ... ?

--
Can you Change: MINDWORKS to MINDWORKS (* == Book)
*HALIFAX HALIFAX*
in 76 moves? Try http://www.chebucto.ns.ca/~af380/MHPuzzle.html
(Requires a browser supporting the W3C DOM such as Firefox or IE ver 6)

Give you www.co.uk (as far as I can see).

It has a flaw though: s/www?d?.([^.]+.)/$1/g; should be written as:

s/^www?d?.([^.]+.)/$1/g

otherwise foo.blawww.com gives odd results.

stripped off. If you use the above fix, only when it starts with those.

It also handles ww (some people use ww.example.com). If you also want to
handle wwww, you might want to use:

s/^w{2,4}d?.([^.]+.)/$1/g;

But again, this is far from perfect, and there are probably a lot of
examples which make this fail.

--
John Perl SEO tools: http://johnbokma.com/perl/
Experienced (web) developer: http://castleamber.com/
Get a SEO report of your site for just 100 USD:
http://johnbokma.com/websitedesign/seo-expert-help.html