In one of the cases you forgot the *

Interestingly, I don't get any difference in SHA1 (with or without the * — though the latter doesn't work with strictures of course, i.e. it gives Bareword "IN" not allowed while "strict subs" in use).

#!/usr/bin/perl -l

#use strict;
use warnings;
use Digest::SHA1;

my $fname = $ARGV[0];
my $sha1;

open(IN, $fname) || die "can't open '$fname' for input: $!";
binmode IN;
$sha1 = Digest::SHA1->new;
$sha1->addfile(IN);
print $sha1->hexdigest;

open(IN, $fname) || die "can't open '$fname' for input: $!";
binmode IN;
$sha1 = Digest::SHA1->new;
$sha1->addfile(*IN);
print $sha1->hexdigest;

open(my $fhIN, $fname) || die "can't open '$fname' for input: $!";
binmode $fhIN;
$sha1 = Digest::SHA1->new;
$sha1->addfile($fhIN);
print $sha1->hexdigest;

system 'sha1sum', $fname;
[download]

Output for a file "foo" containing "bar\n":

$ ./840698.pl foo
e242ed3bffccdf271b7fbaf34ed72d089537b42f
e242ed3bffccdf271b7fbaf34ed72d089537b42f
e242ed3bffccdf271b7fbaf34ed72d089537b42f
e242ed3bffccdf271b7fbaf34ed72d089537b42f  foo
[download]

Anyhow, to avoid the confusion, the OP could've used a lexical file handle, as shown in the third variant above.

So you see no difference? Weird! I'm using a zip file of 60mb in my example. I'm on perl 5.8 on a sun solaris server. COuld there be some sort of problem relating to the size of the source file I want to make a digest of? Thanks

If it reads from the handle as long as it can, it hashes the empty string.

But you'd have to look into the source to find out for sure.