I would skip a regex solution and use substr and rindex:

my $url = qq|/directory/directory/some-file.html|;
print substr($url, 0, rindex($url,'/')),"\n";
[download]

UPDATE:

I just saw tye mention:

(split('/',$url))[-1]
[download]

in the chatterbox. Negative indexing++.

Sheer boredom has settled in me today, so i have decided to run some benchmarks. You did mention 'efficient' after all.

use strict;
use Benchmark;
use File::Basename;

my $url = qq|/directory/directory/some-file.html|;

timethese(500000, {
    'substr' => sub { return substr($url, 0, rindex($url,'/')) },
    'split'  => sub { return (split('/',$url))[-1] },
    'module' => sub { return basename($url) },
});

=for benchmark_results

Benchmark: timing 500000 iterations of module, split, substr...
module: 33 wallclock secs (33.03 usr +  0.10 sys = 33.13 CPU)
 split:  5 wallclock secs ( 5.44 usr + -0.01 sys =  5.43 CPU)
substr:  1 wallclock secs ( 1.74 usr +  0.00 sys =  1.74 CPU)

(linux kernal 2.2.14-5.0, dual 450MHz)

=cut
[download]

So what does this mean? Just because the substr method is faster is not a reason to pick it over File::Basename.

The main reason is portability. Since we are dealing with URL's, we are guaranteed that forward slashes will be used, but if you are dealing with directory paths, the substr method will not portable between Unix platforms and Windows platforms.

And i bet the reason you can't use modules is not because of a work restriction - this was a homework question wasn't it? Doh! Oh well.

Jeff

R-R-R--R-R-R--R-R-R--R-R-R--R-R-R--
L-L--L-L--L-L--L-L--L-L--L-L--L-L--