Trimming Paths with regexes

[TGI]

As part of a project I'm working on, I need to play with a path-like data structure. It comes in the form '/foo/bar/baz'. I need to check the full path, then '/foo/bar', then '/foo' and finally '/'.

So I've put together a do {} while loop that terminates when the path variable is length 0. I'm using a regex to trim the path:

do
{   if (exists $o->{$path})
    {   # Do stuff
    }
    # trim path
    print "\t\t$path\n";
    $path =~ s:(^/$)|((?<=^/)[^/]+$)|(/[^/]+$)::x;

} while ( length $path > 0 );

I keep futzing with the regex and can't come up with something that works. I keep failing to get '/' processed.

Once again, here's the transformation I'm looking for:

 '/foo/bar'-> '/foo'
 '/foo' ----> '/'
 '/' -------> ''

TGI says moo

[damian1301]

I personally would find it much more easier to just split the string and then work with it from there.



$_.=($=+(6<<1));print(chr(my$a=$_));$^H=$_+$_;$_=$^H;
print chr($_-39); # Easy but its ok.

[jryan]

I had to deal with a similar situation one time. Instead of fiddling with a regex all day, I used split to split the path up, and then put it back together (minus the last part). It was similar to this:

my $oldpath = "/foo/bar/baz";
my @path = split(/\//, $oldpath);
my $newpath = "";
for ($i=0; $i<(@path-1); $i++)
{
     $newpath .= "/".$path[$i]; 
}
print $newpath;

It might not be as cleany coded as a regex (or as cool looking!), but it gets the job done. Sometimes the simple solutions are the best :)

[demerphq]

my $path='/foo/bar';
my @dirs=split(/\//,$path);
$"='/'; $\="\n";
while (@dirs) {
    my $str="@dirs" || "/";
    print $str;
    pop @dirs;
}

outputs

/foo/bar
/foo
/

Yves

[dga]

Returns the path shortened and handles paths with //'s together like ////usr/bin correctly. Returns undef in the special case of the root path. Does not handle relative paths

sub pathtrim
{
    my($path)=@_;
    return undef if($path eq '/');
    $path =~ s#/[^/]+$#/#;
    $path =~ s#(/[^/]*)/+$#$1#;
    return $path;
}

[Hofmator]

Just to show that it's not too complicated with a regex:

do
{   
    # your stuff here
    print "\t\t$path\n";
} while $path =~ s#(?:(^/)|/)[^/]+$#$1?$1:''#e;

You can use a subroutine as well, then the regex is pretty simple if you handle the special case beforehand:

sub pathtrim {
    my ($path) = @_;
    return '/' if ($path =~ m#^/[^/]+$/);
    $path =~ s#/[^/]+$##;
    return $path;
}

Update: Fixed small bug in the subroutine.

-- Hofmator