regex to grab file extension 1

[LawnBoy]

I'm trying to extract a file extension (.any) and populate a scalar with it. I've got 4 possible situations and I can't come up with a regex that matches in all situations.
I've got:
1 - somefile.txt
2 - some.file.txt
3 - somefile..text (don't ask, my users will type in anything for a filename)
4 - somefile

After browsing this and other forums, I came up with (filename is in $_)

($name, $ext) = (/^(.*?)\.?([^\.]*)$/);

This works great for situations 1, 2, and 3, but in situation 4 I end up with $name = null and $ext="somefile". What I did was to throw in an extra check:

($name, $ext) = (/^(.*?)\.?([^\.]*)$/);
if ($name eq '') {
	$ext = '';
}

This works well enough, but it's ugly code.
How can I meet all requirements with a single regex?

 

[ishnid]

In the second example you've posted, what do you want to be returned as the extension? `file.txt' or `.txt'. The reason I ask is because for gzipped tarballs, the extension is `.tar.gz', in which case just grabbing the text after the last dot isn't enough.

 

[LawnBoy]

Good point. I want to end up with ".ext" or "ext". In your example I would need ".tar.gz" or "tar.gz".

So make that 5 conditions to meet.

 

[KevinADC]

use the FIle::Basename module, but this is still "messy" when dealing with incorrectly or oddly named files.

 

[KevinADC]

that should be: File::Basename

 

[LawnBoy]

Thx, I'll check out File::Basename.

 

[LawnBoy]

Another question, has anybody ever seen a file ending in ".gz" that isn't ".tar.gz"? If not, I can slop my way around that too.

 

[duncdude]

[b]#!/usr/bin/perl[/b]

while (<DATA>) {
  print "$1\n" if m/((\.tar)?\.[a-z]+)$/;
}

[blue]__DATA__
somefile.txt
some.file.txt
somefile..text
somefile.tar.gz
somefile[/blue]

Kind Regards
Duncan

 

[LawnBoy]

That works nicely Duncan, thx. I'm gonna have to get a good book on regex's...

 

[duncdude]

thank you dude!

Kind Regards
Duncan