Reconstruct binary data saved with unpack("H*", $b)

[andreas1234567]

Honorable Monks,

An application reads binary data from an external system, and saves it using $h = unpack("H*", $b). If I try to reconstruct the binary data with $b = pack("H*", $h), the binary data will sometimes have a trailing null byte. Is it possible to make sure the binary data as it were (i.e. never add the null byte)?

A little code sample to illustrate the problem indicates that is works when input length is even only:

$ cat pack.t 
use strict;
use warnings;
use Test::More qw /no_plan/;

my @arr = (0 .. 9, 'a' .. 'f');
my $hexdata = q{};

while (@arr) {
  $hexdata .= shift @arr;
  my $binary_data = pack("H*", $hexdata);
  cmp_ok(unpack('H*', $binary_data), 'eq', $hexdata);
}

__END__

$ perl pack.t 
not ok 1
#   Failed test at pack.t line 11.
#          got: '00'
#     expected: '0'
ok 2
not ok 3
#   Failed test at pack.t line 11.
#          got: '0120'
#     expected: '012'
ok 4
not ok 5
#   Failed test at pack.t line 11.
#          got: '012340'
#     expected: '01234'
ok 6
not ok 7
#   Failed test at pack.t line 11.
#          got: '01234560'
#     expected: '0123456'
ok 8
not ok 9
#   Failed test at pack.t line 11.
#          got: '0123456780'
#     expected: '012345678'
ok 10
not ok 11
#   Failed test at pack.t line 11.
#          got: '0123456789a0'
#     expected: '0123456789a'
ok 12
not ok 13
#   Failed test at pack.t line 11.
#          got: '0123456789abc0'
#     expected: '0123456789abc'
ok 14
not ok 15
#   Failed test at pack.t line 11.
#          got: '0123456789abcde0'
#     expected: '0123456789abcde'
ok 16
1..16
# Looks like you failed 8 tests of 16.

[Anonymous Monk]

I think the solution is to use u template (uuencoding)

The problem with your sample data is it doesn't come in pairs -- this is where the null is coming from, its padding -- H* demands pairs

[ikegami]

Unless your ok with just using chop on the result — at which point it will only work if the input length is odd — it is impossible to recreate your original data since multiple inputs encode identically.

$ perl -e'print pack "H*", "abc";' | od -t x1
0000000 ab c0
0000002

$ perl -e'print pack "H*", "abc0";' | od -t x1
0000000 ab c0
0000002

If you don't care about preserving leading zeroes, you can use

my $bin = pack("H*", ( length($hex) % 2 ? '0' : '' ).$hex);
...
( my $hex = unpack('H*', $bin) ) =~ s/^0+(?!\z)//;

If you don't care about preserving leading zeroes, you can use

my $bin = pack("H*", "".reverse $hex);
...
( my $hex = reverse unpack('H*', $bin) ) =~ s/^0+(?!\z)//;

If you do care about preserving leading zeroes, you can use

my $bin = pack("NH*", length($hex), $hex)
...
my ($length, $hex) = unpack('NH*', $bin);
substr($hex, $length, length($hex), '');