Because a TAB is a single ASCII character represented by one octet / byte (pay no attention to the Unicode behind the curtain . . .); it's only by convention that it's displayed as four (or two, or eight, or . . .). You'd need to do something to expand it to spaces (e.g. s{\t}{    }gx) before you count if you're going to use length.

Edit: Also, ALWAYS CHECK THE RETURN VALUE FROM open.

Except when using autodie (in which case open doesn't return anything on failure I suppose?)

Except when using autodie (in which case open doesn't return anything on failure I suppose?)

Right. The open wrapper created by autodie dies instead of returning. (This happens somewhere in Fatal, from which autodie inherits almost everything.)

Alexander

--
Today I will gladly share my knowledge and experience, for there are no sweeter words than "I told you so". ;-)

Using  tr/// as davido suggests can very efficiently count characters and also exclude characters from a count with the  /c modifier (see  tr/// in the Quote-Like Operators section of perlop):

c:\@Work\Perl\monks>perl -wMstrict -le
"use Data::Dump qq(pp);
 ;;
 my $entab = 4;
 ;;
 my @lines = (
   qq{This line has no tab.\n},
   qq{\tThis line has tab at the beginning.\n}
   );
 ;;
 for my $line (@lines) {
   my $rend = ($line =~ tr/\t//c) + ($entab * $line =~ tr/\t//);
   printf qq{%s renders as %d chars long \n}, pp($line), $rend;
   }
"
"This line has no tab.\n" renders as 22 chars long
"\tThis line has tab at the beginning.\n" renders as 40 chars long
[download]

You could use something like my $l = length($string =~ s/\t/    /gr), though that's rather inefficient. Another solution:

my $tablength = 4;

my $string = "\tSample\ttext.";

my $expanded_length = render_length($string, $tablength);

sub render_length {
    my ($string, $tabl) = @_;
    my $length = length($string);
    my $tabcount = $string =~ tr/\t//;
    $length += ($tabl - 1) * $tabcount;
    return $length;
}
[download]

This isn't terribly elegant (wrote it on the train on the way to work), but it should work fine, except that the entire notion of tab length is a choice that you make, and convention is ambiguous.

You already got good functional answers so grumpy typist/typesetter time. This concept of tabs is infuriating. A tab is a single marker of *someplace* on the page, not necessarily four characters at all. Now, get off my lawn!