s!^ +!!;

Benchmark: timing 1000000 iterations of index, regex...
     index:  2 wallclock secs ( 1.30 usr +  0.00 sys =  1.30 CPU) @ 771010.02/s (n=1000000)
     regex:  2 wallclock secs ( 1.09 usr +  0.00 sys =  1.09 CPU) @ 913242.01/s (n=1000000)

use Benchmark;
$s=($" x 1000)."test";
sub a { $_=$s;s/^ +//; }
sub b { $_=$s;while(index($_, " ") == 0) { $_ = substr($_, 1) }}

timethese(10000, {a=>\&a,b=>\&b});

Ergebnis:
Benchmark: 
timing 10000 iterations of
 a, b
...

         a:  0 wallclock secs ( 0.06 usr +  0.00 sys =  0.06 CPU) @ 166666.67/s (n=10000)

            (warning: too few iterations for a reliable count)

         b: 10 wallclock secs ( 9.18 usr +  0.00 sys =  9.18 CPU) @ 1089.32/s (n=10000)

use Benchmark;
$s=($" x 1000)."test";
sub a { $_=$s;s/^ +//; }
sub b { $_=$s;while(index($_, " ") == 0) { $_ = substr($_, 1) }}
sub c { $_=$s;
    while (1) { $last = index($_, " ", $pos);
	$last > -1 and $pos = $last+1;
	$last == -1 and last;
    } substr($_, 0, $pos) = "";
} # c

timethese(10000, {a=>\&a,b=>\&b,c=>\&c} );

Benchmark: timing 10000 iterations of a, b, c...
         a:  0 wallclock secs ( 0.03 usr +  0.00 sys =  0.03 CPU) @ 312500.00/s (n=10000)
            (warning: too few iterations for a reliable count)
         b:  9 wallclock secs ( 8.50 usr +  0.00 sys =  8.50 CPU) @ 1176.47/s (n=10000)
         c:  0 wallclock secs ( 0.03 usr +  0.00 sys =  0.03 CPU) @ 322580.65/s (n=10000)
            (warning: too few iterations for a reliable count)